Human Viral Gastroenteritis

Transcription

1 CLINICAL MICROBIOLOGY REVIEWS, Jan. 1989, p Vol. 2, No /89/ $02.00/0 Copyright A) 1989, American Society for Microbiology Human Viral Gastroenteritis MARY L. CHRISTENSEN Department of Pathology, Northwestern University Medical School, Chicago, Illinois 60611, and Virology Laboratory, Department of Pathology, The Children's Memorial Hospital, Chicago, Illinois 60614* INTRODUCTION ROTAVIRUSES Introduction Molecular Biology and Classification Epidemiology Clinical Features and Pathogenesis Clinical features Pathogenesis Laboratory Methods and Diagnosis Cultivation EM EIAs and LA tests Other detection methods Antibody detection Immunology Prevention Vaccines Chemical disinfection Treatment ADENOVIRUSES Introduction Molecular Biology and Classification Epidemiology Clinical Features Laboratory Diagnosis Immunology Prevention and Treatment NORWALK AND NORWALK-LIKE VIRUSES Introduction Physical Characteristics Epidemiology Clinical Features and Pathogenesis Clinical features...73 Pathogenesis Laboratory Diagnosis Cell and organ culture In vitro assays Immunology Prevention and treatment CALICIVIRUSES ASTROVIRUSES CORONAVIRUSES ACKNOWLEDGMENTS LITERATURE CITED INTRODUCTION and Norwalk-like viruses (6, 87, 187, 278, 401), caliciviruses (112, 231), astroviruses, (229, 230), and possibly coronavi- Until 15 years ago, the causes of acute nonbacterial ruses (322). The common enteroviruses are associated with gastroenteritis were unknown. However, during the 1970s, a relatively few cases of viral gastroenteritis (15). The various number of viruses associated with this clinical syndrome viral agents were discovered by the method of electron were discovered, and their presence in the stools of patients microscopy (EM), using EM to examine stools or intestinal with gastroenteritis were eventually correlated with the biopsies from these patients (18, 33, 110, 253). As a group, disease process. These various viruses include rotaviruses these viruses are fastidious and cannot be cultivated in (18, 109), fastidious fecal adenoviruses (77, 110), Norwalk routine cell culture (22, 53, 70, 77, 84, 110). Some can now be 51

2 52 CHRISTENSEN propagated by using special cell types or techniques or both (320, 353, 372). In addition, a variety of in vitro detection systems, including enzyme immunoassays (149, 175, 333, 408), radioimmunoassays (137, 141, 181, 264), latex agglutination (140, 169), and nucleic acid hybridization techniques (116, 175), have been developed for their rapid detection. The detection and identification of these agents are important since viral gastroenteritis is the second most common clinical entity in developed countries, second only to viral upper respiratory tract illness (188). Worldwide, acute gastroenteritis and its associated dehydration afflicts almost 500 million children annually. In underdeveloped or developing countries, acute gastroenteritis, including viral gastroenteritis, is the leading cause of death of children under the age of 4 years (360). Since 1980, a wealth of literature has developed covering many aspects of these viruses and the infections they cause, including their molecular biology, epidemiology, immunology, and clinical features. In addition, detection tests for routine use have been developed for rotaviruses (310) and fecal adenoviruses (333) and experimental rotavirus vaccines have been developed (184, 377, 379, 380). Of these viruses, rotaviruses are the most common known cause of viral gastroenteritis in infants and young children, with the fastidious fecal adenoviruses most likely being the second most common cause in this age group (301, 369). Calici-, astro-, and coronaviruses are probably responsible for a minority of illness in the young age group (253, 302). In contrast, the Norwalk and Norwalk-like viruses have caused considerable numbers of outbreaks of gastroenteritis among older children, adolescents, and adults (136, 137, 189). ROTAVIRUSES Introduction Rotaviruses are in the Reoviridae family, the members of which possess a double-layer of icosahedral shells of approximately 70 nm in diameter, with a core of double-stranded ribonucleic acid (dsrna). Rotaviruses infect a wide variety of mammals and birds as well as humans. Most rotaviruses from various species are similar and share a common group antigen, which is associated with the inner capsid layer. However, some rotaviruses do not possess this group antigen. Rotaviruses are responsible for a significant proportion of gastroenteritis in small children and infants, as well as for causing gastroenteritis in the elderly. Similarly, animal and avian rotaviruses are responsible for gastroenteritis in the young of their corresponding species. Molecular Biology and Classification Rotaviruses contain a dsrna genome consisting of 11 segments, ranging in molecular weight from 0.4 x 106 to 2.0 x 106 (275). These 11 segments can be separated by polyacrylamide gel electrophoresis (PAGE) (103, 182), using the method of Laemmli (212). Different rotavirus isolates frequently exhibit differences in the electrophoretic mobilities of their 11 segments. Rotaviruses exhibiting these different electrophoretic mobilities have been termed electropherotypes. Although frequently many of the rotaviruses isolated in the same geographical location at the same time of year may exhibit similar or identical electrophoretic mobilities, there has been considerable variation in electrophoretic mobilities of viruses isolated during different seasons or at different locales during the same season. For example, CLIN. MICROBIOL. REV. Rodger et al. (304) described 19 different electropherotypes isolated in Melbourne, Australia, from 1973 to 1979; Albert et al. (3) described 9 different electropherotypes isolated in Indonesia in 1978 and 1979; Rodriquez et al. (306) reported on 10 different electropherotypes seen in two nosocomial outbreaks for a 6-month period in 1979 and 1980 in Washington, D.C.; and Spencer et al. (340) described 32 different electropherotypes seen in Santiago, Chile, from 1979 to Attempts have been made to characterize and classify human rotaviruses based on their electrophoretic mobilities. However, the variation has been too great, except for two patterns of electrophoretic mobility that have been distinguished (103, 182). These are the "short" pattern and the "long" pattern, in which gene segments 10 and 11 migrate more slowly, creating a short pattern, or more rapidly, creating a long pattern, in the polyacrylamide gels (103, 182). To study the problem of genetic relatedness among various electropherotypes, Flores et al. (115) prepared singlestranded RNA probes for hybridization to gene segments which consist of dsrna. They found that corresponding gene segments that exhibited similar migration patterns did not necessarily exhibit RNA homology when studied by hybridization with the probes. Conversely, some corresponding gene segments that exhibited RNA homology with the probes did not have the same electrophoretic migration pattern. Thus, they concluded that similarities or differences in electrophoretic mobility did not always reflect similarities or differences in genetic relatedness between various RNA segments (115). However, it has been shown that electropherotypes are excellent markers for identifying and following the spread of viruses from one individual to another in discrete outbreaks; thus, they are good for providing epidemiological information (47, 304). In addition to segments 10 and 11, three gene segments are of particular interest since they code for three major rotaviral antigens, VP4, VP6, and VP7. The major antigens of rotavirus and the genes that code for them are shown in Table 1. Gene segment 4 codes for VP4 (179, 219, 222, 243, 335). Originally the product of gene 4 was called VP3, but Liu et al. (219) recently proposed that the gene 4 product be called VP4 while the product of gene 3 be called VP3. Gene segment 6 codes for VP6 (180). Depending on the strain of rotavirus, gene segment 8 or 9 codes for VP7 (25, 180). For example, gene segment 8 codes for VP7 in the human serotype 2 rotavirus Hu/5 isolated in Melbourne, Australia (95), and for the bovine rotavirus UK (207). Gene segment 9 codes for VP7 in the human serotype 1 rotavirus Wa and for the simian rotavirus SA11 (180, 183, 207). VP4 is an outer capsid protein which has an apparent molecular weight of 88,000 (241). VP4 is associated with two biological functions involved with virus-cell interaction: hemagglutination and protease-enhanced plaque formation (179, 222, 241, 335). VP4 is also responsible for the restriction of growth in cell culture (114, 180). VP6 is the major inner core structural protein which is present in large amounts in the virion. It has an apparent molecular weight of 42,000 (25, 93, 94, 240, 275). VP6 is the major subgroup antigen which can specify one of two rotavirus subgroups (subgroups I and II) (25, 180). These subgroups have been differentiated on the basis of a variety of tests. These include differentiation by electrophoretic migration patterns of gene segments 10 and 11 in PAGE (104), the complement fixation (CF) test (420), the immune adherence hemagglutination test, immune electron microscopy (IEM) (420), the radioimmunoassay (RIA) (181, 249), and the enzyme-linked immunosorbent assay (EIA) (357,

3 VOL. 2, 1989 VIRAL GASTROENTERITIS 53 TABLE 1. Gene code assignments for rotaviral proteins Gene Protein gene Approximate mol Location Biologic functions of VP segment segment wt of protein 1 VP1 125,000 Core 2 VP2 94,000 Core 3 VP3 88,000 Core 4 VP4 88,000 Outer capsid Hemagglutination; protease-enhanced plaque formation; restriction of growth in cell culture 5 NS53 53,000 6 VP6 41,000-42,000 Inner capsid Major inner core structural protein; major subgroup antigen 7 NS34 34,000 8 or 9 VP7 34,000-40,000 Major outer capsid Serotype specificity; major neutralization 10 NS28 28, VP9 26,000 a VP, Viral protein; NS, nonstructural protein. 418, 421), using either polyclonal sera or monoclonal antibodies (132, 355). However, some monoclonal antibodies react with viruses in both subgroups, so VP6 proteins must also share common epitopes. VP7 is the major outer capsid protein which is glycosylated and has an approximate molecular weight of 34,000 to 40,000 (93, 106, 163, 180, 183, 207, 275, 335, 350). This protein is responsible for the serotype specificity of the virus (10, 92, 94, 95, 106, 163, 180, 337), which was initially determined by virus neutralization (166, 402, 404). Although VP7 is the major determinant of neutralization, VP4 can also elicit neutralizing antibody (10, 135, 165, 272, 335). At least four human serotypes have been identified: serotypes 1, 2, 3, and 4 (106, 163, 404). Subgroup I includes human rotavirus strains of serotype 2, which is represented by prototype strain DS-1 (106, 357, 402, 404). Subgroup II includes serotype 1 (represented by prototype strain Wa) and serotypes 3 and 4 (106, 165, 357, 371, 402, 404). Although at least four human serotypes have been identified (106, 163), a possible fifth serotype from Indonesia has been described by Matsumo et al. (242). This virus strain, strain 69M, has a "super-short" RNA electrophoretic pattern of gene segments 10 and 11. By RNA-RNA hybridization, 69M was found to have a low degree of homology with the representative strains of all four known human serotypes, and it could not be classified by neutralization analysis into any of these four serotypes (242). Albert et al. (4) also detected two similar super-short strains designated B37 and B38 in Indonesia. A sixth possible serotype, strain W161 isolated in the United States, has been described by Clark et al. (61). By cross-neutralization tests, this virus was distinguished from human rotavirus serotypes 1, 2, 3, and 4, from human strain 69M, and from bovine (NCDV), porcine (OSU), and chicken (Ch2) rotaviruses (61). Until recently it was thought that all human and animal strains of rotavirus possessed a common group antigen (399). However, various rotaviruses have been isolated from humans and animals that do not possess that common group antigen originally reported by Flewett and Woode (113, 399). Rotavirus strains lacking the common group antigen have been isolated from humans, cows, lambs, pigs, rats, and birds (34, 73, 81, 105, 124, 270, 305). These more newly discovered viruses are morphologically indistinguishable from other rotaviruses in that they consist of a doubleshelled icosahedran containing 11 genome segments of protein (glycosylated) determinant dsrna; however, they have an electrophoretic pattern that differs from that of known rotaviruses. These agents have been called "rotavirus like," (170, 305) "antigenically distinct rotaviruses," (81, 101) "pararotaviruses," (24, 105, 270) "atypical" rotaviruses (285, 286), and "novel" rotaviruses (171). More recently, Pedley et al. (285), studying atypical porcine rotaviruses, introduced the designation group A, B, C, etc., analogous to influenza virus terminology. They proposed the usage of group A for the original conventional rotaviruses with the common group antigen and groups B and C for more recently discovered atypical rotaviruses that (i) possess other group antigens and (ii) are genetically different based on (a) electrophoresis of the 11 gene segments and (b) one-dimensional terminal fingerprinting analysis of the RNA segments. Later, Pedley et al. (286), after analyzing atypical porcine and chicken rotaviruses, described group D and E rotaviruses. Human rotaviruses belonging to groups B and C have now been described. Atypical strains were analyzed by antigenic analysis (IEM, immunofluorescent-antibody assay, and/or EIA) and by genome profile analysis, terminal fingerprint analysis of genome segments, and/or dot blot hybridization (34). Sporadic single observations of group C rotaviruses in humans have been made in a few laboratories (34). For example, analysis of two atypical rotaviruses from Australia and Brazil were found to be in group C (34). Eiden et al. (101), found that five of six human isolates from the United States were related to porcine and bovine group B rotaviruses. In China, group B rotaviruses have been found to cause severe epidemics of diarrheal disease (48, 170, 171). An unusual characteristic of the outbreaks in China was that a much higher attack rate was noted among adults than among children. However, Dai et al. (73) described an outbreak of diarrhea among newborns caused by the new Chinese rotavirus that was placed in group B by Chen et al. (48). This is probably the first report of neonatal infection caused by group B virus since group A rotaviruses have been the primary cause of outbreaks in neonatal nurseries in a number of countries. To summarize, rotaviruses can be classified by four main categories: group specificity, serotype specificity, subgroup specificity, and strain specificity. Epidemiology Rotaviruses were first discovered in humans 15 years ago by Bishop et al. (18, 19) by the EM examination of duodenal

4 54 CHRISTENSEN CLIN. MICROBIOL. REV. FIG. 1. Common gastroenteritis viruses from stool specimens of pediatric patients with acute gastroenteritis. For each, an approximately 20% stool suspension was prepared in 1% ammonium acetate and negatively stained with 2% phosphotungstic acid; x 140,000. (A) Rotaviruses; (B) adenoviruses in stool; (C) caliciviruses; (D) astroviruses. Electron photomicrographs courtesy of Cynthia Howard, The Children's Memorial Hospital, Chicago, Ill. mucosal epithelial cells obtained by duodenal biopsy of a group of Australian children hospitalized with nonbacterial gastroenteritis. In the same month, Flewett et al. (109) described similar reoviruslike particles seen by EM in stool specimens from patients with gastroenteritis in the United Kingdom. Shortly thereafter, the finding of these viruses in patients, but not controls, was reported from Singapore (354), Canada (252), southern Africa (65), India (164), the United States (186), and Scandinavia (277). Flewett et al. (111) suggested the name "rotavirus" based on the wheellike appearance (Fig. 1A) of these viruses. The seriousness of rotavirus infection was exemplified by the reports of Middleton et al. of seven fatal cases occurring out of 60 rotavirus-infected patients (252) and a review of 21 fatal cases (41). Others reported that this virus was also responsible for nosocomial outbreaks of gastroenteritis in nurseries (40, 56, 261). Chrystie et al. (56) also noted that in neonates the disease was mild or symptoms were lacking entirely. In addition, young adults, i.e., parents of rotavirusinfected infants, infrequently excreted large amounts of virus and infrequently developed symptoms or had only mild symptoms (56, 185, 261, 307). Often the only evidence of infection in parents was determined by fourfold or greater rises in serum antibody titers (185). 01 Rotaviruses are responsible for approximately 50% or more of the gastroenteritis in hospitalized pediatric patients during the cooler months of the year in parts of the world that have temperate climates (15, 76, 152, 185, 203). These worldwide reports of the high incidence of rotavirus infection have come from urban populations. However, the incidence of rotavirus infection appears to be somewhat lower or absent in patients from rural areas as reported from Australia and South Africa (89, ). There may be several reasons for this. One is that rotaviral colonization of neonates may lead to protection against symptomatic infection later in infancy. Another reason may be that, in some instances, a lack of overcrowding may prevent the virus from spreading easily. Also, differences in diet or lifestyle in certain rural areas may contribute to a different viral gut flora. Since many reports indicate that rotavirus infection increases during the cooler months, the role of weather in altering the rotavirus infection rate has been studied in various parts of the world by a number of investigators. In temperate climates such as in northern Japan, increases in the infection rate appeared to be related to the drop in temperature, but not to the outside relative humidity (205). In Washington, D.C., rotavirus gastroenteritis increased

5 VOL. 2, 1989 after a month of cold or dry weather rather than warm or wet weather (28). Brandt et al. (28) suggested that indoor crowding and low indoor relative humidity may increase aerosolization of rotavirus particles on surfaces, as well as contribute to dehydration of infected infants during these months. However, in another report, rotavirus prevalence in southern California during one season (October to December) did not correlate with either the coldest or the driest months of the year. In some tropical countries, including Nigeria (282), southern India (235), Bangladesh (20), Indonesia (336), Ecuador (348), and Costa Rica (152), rotavirus infections tend to peak during the dry season. However, no distinct seasonal variation occurred in other reports from tropical countries, including Ecuador (345) and Venezuela (80) and, in certain reports, South Africa (323). Thus dryness may have some effect on the spread of the virus, although in some tropical countries any variation in temperature and humidity may be too slight to be of any significance. Murphy et al. (261) found no seasonal variation in nosocomial neonatal nursery infection in Sydney, Australia. However, indoor temperature and humidity may not have varied due to environmental control, and once rotaviral colonization of a nursery occurs, it may be difficult to eradicate. Dennehy and Peter (78) found that three-fourths of the nosocomial infections they studied occurred in the winter rotavirus season. Hjelt et al. (159) found that approximately half of the nosocomial rotavirus infections they investigated occurred during an epidemic outbreak whereas the other half were scattered with regard to season. A major risk factor was the duration of patient stay. Nosocomial infections appeared to be due to failure to isolate the nosocomially infected patients, although the communityinfected patients were put in private rooms upon admission. Sharing of nursing and medical staff did not differ among control patients, although this has been a factor in some studies. In one study (156), 36% of the nursing staff in the pediatric wards were infected with rotavirus as compared with 15% of the nursing staff in other wards. Asymptomatic or mild symptoms were characteristic of staff rotavirus infection. Asymptomatic rotavirus infection and viral carriage occur frequently and have been studied by several investigators. Champsaur et al. (46) found rotavirus in 36% of hospitalized children with diarrhea and in 24% of hospitalized children without diarrhea. Of those children shedding rotavirus, 48% were not diarrheic. Virus shedding that was not associated with diarrhea was observed in 71% of neonates, 50% of 1 to 6 month olds, and 26% of 7 to 24 month olds. In a similar study (45) of patients ranging in age from newborns to 24 months admitted to the hospital, Champsaur's group found that 13% had symptomatic disease (diarrhea and serologic response), 7% had asymptomatic infection (no diarrhea, but serologic response), and 20% were viral carriers (no diarrhea and no serologic response). Asymptomatic infection as defined by serologic response but no diarrhea occurred in 2% of neonates, 20% of 1 to 6 month olds, and 37% of 7 to 24 month olds. Virus carriage as defined by no diarrhea and no serologic response occurred in 27, 19, and 14% of these children, respectively. Walther et al. (389) studied 871 children admitted to a pediatric ward for various conditions. Of 742 asymptomatic children, 38% were excreting rotavirus as determined by EIA. Of 129 symptomatic children, 50% excreted rotavirus, 26% excreted enteric pathogens, and 13% excreted both rotavirus and enteric pathogens, and in 37% no agent was found. Thus, the presence of rotavirus has not always VIRAL GASTROENTERITIS 55 correlated with disease. Rotavirus is also prevalent in daycare centers and can be spread to family contacts, thus propagating the infection in the community (191). Many day-care children are asymptomatic, indicating a large reservoir of infection. The numbers of infected children range from 10 to 30%, depending on the study (9, 191). In one study (9) 20% of the adults directly involved in child care were infected, although they were primarily asymptomatic. However, adults have also experienced more severe clinically acute cases of rotavirus gastroenteritis. von Bonsdorff et al. (385) reported acute rotavirus gastroenteritis in 45% of adult hospital employees aged 19 to 62, and Echeverria et al. (98) reported rotavirus in 5% of adult patients aged 16 to 72 admitted to the hospital for gastroenteritis. Rotavirus gastroenteritis has been more severe in the elderly, according to reports from nursing homes and other institutions for the elderly (66). Halvorsrud and Orstavik (142) reported 92 cases of mild-to-severe, rotavirus-associated gastroenteritis in 70 to 90 year olds in a nursing home. Initial symptoms of nausea and vomiting were followed by diarrhea and low-grade fever. One patient died. There appeared to be a great susceptibility to both infection and disease, since 66% of the patients in the affected wards became ill. Similarly, a serious rotavirus outbreak occurred in 56% (19 of 34) of patients in a hospital geriatric ward, as well as in staff members (238). Six patients had severe illness and two died. Forty percent of asymptomatic geriatric patients were infected. Outbreaks among the elderly can become extensive and severe, possibly because of lowered immunity. Thus, in this population, testing of all staff and patients is important to identify all infected individuals and to institute control measures. Although group A rotaviruses have been responsible for almost all of the rotavirus infections in both the young and the elderly in North America and Europe, a new group of rotaviruses has been responsible for large outbreaks of severe rotavirus diarrhea occurring in adults of all ages in China (171). These rotaviruses did not share the group antigen (group A) of known rotaviruses at the time and were found to belong to a new group, group B (48). Human infection with group B rotavirus has not been widespread outside of China, however (267, 368). Within the group B rotaviruses, subgroups and types have not been delineated as has been done with the group A rotavirus. Within the A group of rotaviruses, the prevalence of the subgroup and the type of rotavirus causing infection have been studied. White et al. (394) subgrouped 99% of 252 specimens obtained from Venezuelan children over a 45- month period. Some 14% shed subgroup I and 85% shed subgroup II rotavirus. Of the subgroup I viruses, one-half were shed during a 3-month period. There was no difference in the occurrence of fever and vomiting between the children with either subgroup, but patients with subgroup II infections had longer-lasting illness. Yolken et al. (418) studied 414 rotaviruses isolated from patients in Washington, D.C., Belgium, and Central America. They found that 23% were type 2 (subgroup I) and 77% were type 1 (subgroup II), with a similar distribution from various parts of the world. In an analysis of children who were reinfected, sequential infections usually involved different serotypes, and illness caused by one serotype did not provide resistance to illness caused by the other serotype. Studying the prevalence of various subgroups and types is an important consideration for determining which rotavirus type(s) to incorporate into rotavirus vaccines.

6 56 CHRISTENSEN Clinical Features and Pathogenesis CLIN. MICROBIOL. REV. Clinical features. Several groups have studied the clinical characteristics of rotavirus infection, in particular, in hospitalized patients (161, 367, 381). The two most prominent features were vomiting and diarrhea, usually with sudden onset. Vomiting appeared to be the most common first symptom, occurring first in 34 to 55% of patients, depending on the study. Vomiting often preceded the onset of diarrhea by a few hours to 24 h and occurred at some time during illness in 48 to 92% of patients, with a mean duration of 3 days. Diarrhea was the first symptom to occur in 23 to 29% of patients, depending on the study. Diarrhea occurred at some time during illness in 65 to 100% of patients and lasted a mean of 5.0 to 5.9 days. Vomiting and diarrhea started simultaneously as the first symptoms in 22% of patients. Characteristically, there was an absence of blood in the stools (161). Fever occurred first in 13% of patients and occurred at some time during illness in 34 to 86% of patients. Abdominal pain was somewhat infrequent, occurring in 17 to 29% of patients. Respiratory symptoms occurred first in 24% of patients and occurred at some time during the illness in 24 to 52% of patients. Overall, symptoms were more severe in hospitalized patients than in patients not requiring hospitalization, and patients in poorer general condition were more dehydrated. There have been conflicting reports on the role of rotavirus in causing respiratory infection and symptoms. Lewis et al. (216) reported respiratory illness in 66% of rotavirus gastroenteritis patients versus 26 to 38% of nonrotavirus gastroenteritis patients. Respiratory symptoms usually preceded gastrointestinal symptoms, although they also occurred concurrently, and consisted of rhinitis, pharyngitis, tonsillitis, and otitis media. Yolken and Murphy (413) identified rotavirus by EIA in tracheal aspirates from two of five infants who died of sudden infant death syndrome who had acute upper respiratory infection. Although none of these infants had gastroenteritis, rotavirus was also detected in their stools. Similarly, Santosham et al. (317) detected rotavirus in respiratory secretions from 4 of 45 pediatric patients with pneumonia. None of the four had diarrhea. However, since they had prior antibiotic treatment, and Mycoplasma sp. was also found in one child, bacteria may have been the cause of the pneumonia. Fragoso et al. (118) also detected rotavirus antigen in respiratory secretions in 2 of 30 children with both upper respiratory tract infection and vomiting. To study the role of rotavirus in respiratory infection, Prince et al. (291) used a murine rotavirus model to demonstrate that aerosol transmission can occur. The result was both respiratory and gastrointestinal infection and gastroenteritis. The authors concluded that, since rotavirus may be found in respiratory secretions of children, it may be transmitted by the respiratory route as well as the fecal-oral route. However, several groups have failed to show any significant role of rotaviruses in respiratory infection. Goldwater et al. (127), using direct EM, IEM, and EIA, failed to detect rotavirus in respiratory secretions of rotavirus-infected patients. Their study included specimens obtained from patients with respiratory symptoms prior to the onset of gastroenteritis. Similarly, Vollet et al. (384) could not detect rotavirus in 11 of 13 children with rotavirus diarrhea who also had respiratory symptoms. Maki (236) found no difference in respiratory symptoms in rotavirus gastroenteritis patients (53%) or nonrotavirus gastroenteritis patients (62%). Hjelt et al. (161) found no more upper respiratory tract symptoms in rotavirus gastroenteritis patients (36%) than in nonrotavirus gastroenteritis patients (35%). Similarly, Uhnoo et al. (367) observed respiratory symptoms in 32% of rotavirus gastroenteritis patients, but these data were not significant when compared with respiratory symptoms in patients with gastroenteritis caused by other viruses or bacteria or when no agent was detected. These data suggest that rotaviruses may sometimes cause respiratory symptoms. Differences seen in the various reports may depend on the locale, the year, and the strains of rotaviruses involved. Some patients with nonrotavirus gastroenteritis and respiratory symptoms had infection caused by adenoviruses, which are known to cause both gastrointestinal and respiratory symptoms simultaneously (see below). Thus, the respiratory symptoms occurring in the adenovirus gastroenteritis groups may have contributed to the lack of significance observed between the rotavirus and nonrotavirus gastroenteritis groups in some of the reports. Various other complications have been reported to occur with rotavirus infection. Fernbach and Lloyd-Still (107) reported three patients with severe, prolonged rotavirus colitis with bloody stools. Wong et al. (397) reported a case of aseptic meningitis associated with rotavirus gastroenteritis in which rotavirus particles were seen in the cerebral spinal fluid by IEM. Ushijima et al. (373) reported on a case of encephalitis that developed during rotavirus gastroenteritis, in which rotavirus immunoglobulin G (IgG) increased in the patient's cerebrospinal fluid. In addition, rotavirus-like particles have been detected by EM and EIA in a liver biopsy in a case of hepatic abscess (139). Pathogenesis. Rotavirus infection is spread primarily by the fecal-oral route. Although rotavirus is relatively acid labile, rotavirus can survive the ph of a stomach that is buffered, or can survive in the stomach after a meal. At ph 2.0 (that of a fasting stomach), rotavirus is rapidly inactivated in <1 min (393). However, at ph 3.0, inactivation is much slower, the viral half-life being about 10 min; at ph 4.0, inactivation is minimal. The infant gastric ph tends to be approximately 3.2, and in general, the stomach ph remains above 3.0 for at least 1 h after a meal. This probably explains the efficient transmission of rotavirus. To determine the infective dose required to produce infection with or without symptoms, adult volunteers ingested to 90,000 focus-forming units of rotavirus (390). Results showed that as little as 0.9 focus-forming unit caused infection in one of nine volunteers, as determined by viral shedding and a significant rise in antibody titer. A 9-focusforming unit dosage caused infection in 8 of 11 volunteers, with 6 of these having symptoms. Higher viral doses caused higher percentages of infection, but no increase in the number of individuals with symptoms. Rotaviruses tend to infect the small intestine, as do other gastroenteritis viruses. In particular, rotavirus replication takes place in epithelial cells on the tips of villi of the small intestine, and infection is confined primarily to these cells (405). Rotaviruses selectively infect the mature villus enterocytes of the small intestine; rotaviruses exhibit a predilection for young animals of many animal species (273). The histology of duodenal biopsies obtained from rotavirus gastroenteritis patients was first described by Bishop et al. (18). A patchy irregularity of the mucosal surface was seen in most cases. Mucosal changes ranged from mild to severe. The changes included shortening and blunting of villi and increased infiltration of the lamina propria with mononuclear cells. Epithelial cells were more cuboidal and less

7 VOL. 2, 1989 regular than usual. By EM, reovirus-like particles were seen in the epithelial cells of the duodenal mucosa. Virus particles were seen within distended cisternae of the endoplasmic reticulum in vacuolated cells. No virus-like particles were observed in any of the other types of cells in the lamina propria. Davidson and Barnes (74) found mucosal damage to be quite variable and often patchy. Mild changes in seven of their patients included broadening of villi, mild cellular infiltration of the lamina propria, and early epithelial cell damage. Moderate changes in eight patients involved considerable blunting of villi, obvious increase in inflammatory cells in the lamina propria, increased crypt depth, and flattening of epithelial cells. Severe changes in two patients showed complete villous flattening, marked inflammatory cell infiltration, crypt hypertrophy, and severe epithelial damage with cuboidal epithelium. The severe damage could be confused with the structural appearance seen in coeliac disease. Rotavirus particles were observed by EM in the cytoplasm of infected cells. Mucosal damage was rapidly repaired, as early as 3 weeks after onset. They found that children with a more severe mucosal lesion were more likely to become dehydrated and require intravenous therapy for rehydration. In bacterial gastroenteritis, abdominal pain, bloody stools, leukocytosis, and prolonged diarrhea are more likely to occur than in viral gastroenteritis (367). Viral gastroenteritis also differs from bacterial gastroenteritis in its physiology (360). In rotavirus infection, there is patchy replacement of infected mature villous tip cells by secretary crypt cells, decreased intracellular Na+,K+-adenosine triphosphatase activity, and impairment of glucose-coupled sodium transport (360). Abnormally low maltase, sucrase, and lactase levels were also found in children with rotavirus gastroenteritis (18, 74), which returned to normal after 4 to 8 weeks. Most (but not all) children with acute rotavirus gastroenteritis have lactose malabsorption and intolerance (172). An increase in diarrhea can occur after feeding lactose, so that a non-lactose-containing formula is usually given during rotavirus illness when infants are able to take fluids orally. Normal lactose tolerance reappears by at least 10 to 14 days after the start of illness, when lactose-containing products can then be introduced (172). Loss of fluids and electrolytes in rotavirus gastroenteritis can lead to severe dehydration and even death and requires fluid and electrolyte replacement therapy. In developing countries, recurrent bouts of gastroenteritis can lead to a vicious cycle of protracted diarrhea, food intolerance, and malnutrition (360). The malnutrition may be further compounded by frequent fasting, which is a measure commonly used in the management of acute gastroenteritis in some developing countries (360). For marginally nourished or malnourished children, diarrhea with associated starving can have deleterious effects. When a person has fasted for 3 to 5 days, depletion of intestinal digestive enzymes and gut mass occurs and absorption of water, salt, glucose, disaccharides, and amino acids are substantially reduced (360). In spite of these multiple physiological abnormalities, oral replacement therapy is effective in correcting dehydration in rotavirus gastroenteritis. Its effectiveness is probably due to the presence of intact glucose-coupled sodium transport in noninfected bowel (360). Laboratory Methods and Diagnosis Since the discovery of rotaviruses by EM, a number of rotavirus detection methods have been developed. Since VIRAL GASTROENTERITIS 57 rotaviruses have been difficult to propagate in cell culture, other viral and antigen detection methods have been used. These are based primarily on antigen-antibody reactions. Probably the most commonly used procedures at the present time are EIA and latex agglutination (LA) tests, since several commercial kits are available for use in many countries. EM procedures are also used. A variety of other viral and antigen detection tests have been developed, but are used primarily as research tools in the laboratories that developed them. Cultivation. Propagation of human rotaviruses is usually not carried out in diagnostic laboratories since virus is found in large quantities in stool specimens and can be rapidly detected by antigen detection tests. However, some research laboratories have cultivated rotaviruses by using various manipulations. Wyatt et al. (403) propagated a strain of human rotavirus (Wa strain, serotype 2) in primary African green monkey kidney cell cultures after virus from pooled human stools had been passed 11 times in newborn gnotobiotic piglets. Sato et al. (319) and Urasawa et al. (372) reported the propagation of a number of human rotaviruses from stool specimens by using a combination of three techniques. Using roller cultures of MA-104 cells, a line of fetal rhesus monkey kidney cells, they added low levels of trypsin to the maintenance medium and pretreated their specimens with trypsin. Each of these techniques had failed when used separately. Similarly, Kutsuzawa et al. (211) used trypsin treatment and roller cultures of MA-104 cells on two stool specimens, one of which contained a subgroup I rotavirus and the other of which contained a subgroup II human rotavirus. Distinctly recognizable cytopathic effect (CPE) was observed by passage 6 of the subgroup I isolate and by passage 3 of the subgroup II isolate. CPE consisted of obscure cell boundaries, cell fusion, cell rounding, cell detachment, and lytic foci. Supernatant fluids were trypsin treated prior to each passage. Hasegawa et al. (146) also propagated a number of human rotavirus isolates from stool samples, using trypsin pretreatment and trypsin in the maintenance medium. However, they found that rolled primary cynomolgus monkey kidney cells were more sensitive than the rolled MA-104 cells. CPE appeared at passage 2 to 7, although virus could be detected in the supernatant fluids of passage 1 by the immune adherence hemagglutination test. Passaged fluids were not trypsin treated after the initial inoculation. Birch et al. (16) studied nine strains of rotavirus in MA-104 and CV-1 cells, a line of African green monkey kidney cells. In MA-104 cells, CPE consisted of a sloughing of cells, and in CV-1 cells, CPE consisted of an increased granularity of the cells. They found that CPE was not a reliable indicator of replication; CPE occasionally disappeared for several passes, although virus was detectable in the supernatant fluids by indirect fluorescent-antibody staining, EM, or EIA. Because of inapparent rotaviral CPE, Suzuki et al. (347) described an interference test, similar to that used to detect rubella virus, to detect replicating rotavirus. Using two laboratory adapted rotavirus strains of Kutsuzawa et al. (211) and the Wa strain, they found that, by challenging with coxsackievirus B-1, interference could be detected 4 days after virus infection. Ward et al. (391) compared the growth of human rotaviruses from stool specimens by using two types of primary monkey kidney cell cultures (African green and cynomolgus) and two types of monkey cell lines (MA-104 and CV-1). Primary cells supported virus growth directly from the

8 58 CHRISTENSEN specimens much more effectively than the two continuous lines. Although viruses from the specimens could not always be grown in the cell lines, the viruses appeared to be fully adapted for growth in the cell lines after two passes in primary cells. The efficiency of viral growth also increased with cell passage. Only 1 of 46,000 virions in stool specimens infected the primary cell cultures, whereas 1 of 6,600 visions were infectious after three passes in primary cells. Agliano et al. (1) reported the isolation of eight human rotaviruses in LLC-MK2 cells (a continuous line of rhesus monkey kidney) and human embryonic fibroblasts, using no special techniques. The cells were not rolled, and trypsin was not incorporated into the maintenance medium nor used to pretreat the stool specimens. They suggested that these rotaviruses may differ from other previously isolated rotaviruses, since special techniques were not needed to isolate them. These cultivation methods, however, are not used routinely. EM. Initially, rotaviruses were detected directly in stool samples by the EM of virus particles negatively stained with phosphotungstic acid, and this method is still used as the standard (252, 253). A simple method consists of making an approximately 20% suspension of the stool sample in distilled water or 1% ammonium acetate. This suspension is then placed on a Formvar-coated EM grid, and the excess is blotted. Phosphotungstic acid solution (2%) is added and the excess is blotted. Phosphotungstic acid at ph 4.5 has been shown to be optimal for both EM and IEM (268). It is an electron-dense negative stain which does not stain the virus particles, but rather the area around them, causing the unstained virions to stand out. Alternative negative stains that have been used are ammonium molybdate and uranyl acetate (268). Since children usually have had several episodes of diarrhea before their parents seek medical attention, usually no bacteria are remaining in their stools by the time a sample is submitted to the laboratory. However, some laboratories prefer to clarify stool suspensions (20%) by low-speed centrifugation. In some instances the clarified supernatant is subjected to ultracentrifugation to pellet and concentrate any virus particles present. The resuspended pellet is then used for preparing grids, followed by negative staining. In IEM, patients' convalescent sera are mixed with a virus suspension. These suspensions can be obtained by purifying virions from stools or from infected cell culture fluids. After incubation of the virus-serum mixture, the mixture is placed on a grid and stained with phosphotungstic acid. Nicolaieff et al. (269) described an IEM technique which involved coating EM grids with Staphylococcus aureus protein A, followed by adsorption of specific rotavirus antiserum to the protein A on the grid. They found that their technique detected rotavirus particles in 3.5 times as many specimens as by routine EM. Svensson et al. (349) also described a technique almost identical to that of Nicolaieff, which they called solid-phase IEM (SPIEM). They found their SPIEM to be 30 times more sensitive than routine EM or routine IEM and 10 times more sensitive than indirect EIA. Gerna et al. (121, 124) used SPIEM to distinguish between human serotypes 1, 2, 3, and 4. S. aureus protein A was first placed on a Formvar-coated grid. Then type-specific, crossabsorbed polyclonal immune sera were adsorbed to the protein A. Viral specimens were then added and observed by EM. Various other modifications of EM have been reported. Kjeldsberg and Siebke (201) described a simple immunosorbent EM which allows one to wash specimens to remove contaminating material such as sucrose solutions from virus suspensions prior to negative staining of grids. Kjeldsberg (200) also described an EM technique for specific labeling of human rotavirus with gold-igg complexes. ETAs and LA tests. Since EM procedures are time-consuming to perform for a large number of samples, other testing procedures were developed to detect rotaviruses or rotaviral antigens. The more universally used tests today are the LA tests and the EIA. The EIA is based on principles and procedures first described by Engvall and Perlmann (102) and Voller et al. (383). The first rotaviral ETAs were described by Yolken et al. (408, 416) and others. EIAs are similar to RIAs, except that an enzyme is linked to the detector antibody, instead of a radioisotope. Most of the EIAs utilize a three-layer double-antibody sandwich technique (408). Briefly, anti-rotavirus hyperimmune serum (or monoclonal antibodies) are adsorbed to a solid phase (13, 63, 162, 287, 349). This first antibody has been termed the "capture," "coating," "catching," or "primary" antibody. Next the specimen is added, and if virus is present, it will bind to the rotavirus antibody. A second antibody, the "detector" or "secondary" antibody, is then added. In many EIAs, the second, detector antibody is conjugated to an enzyme, making it also an "indicator" antibody. In some EIAs, however, the detector antibody is not conjugated to an enzyme. In these instances, a third, "anti-antibody" or indicator antibody which is enzyme conjugated is added (144, 167, 318, 421). Often the third antibody has been used to obtain a more sensitive test or it has been used for typing. This type of test is termed a "four-layer EIA" or an "indirect EIA." Various other modifications of the EIA procedure have been described. Periera et al. (287) developed a four-layer rotavirus-adenovirus combination EIA. Either rotavirus or adenovirus antiserum as capture antiserum was added to alternate rows in a microtiter plate. After incubating the samples, a single detector antibody was added, consisting of a mixture of guinea pig (GP) antirotavirus and GP antiadenovirus sera. A third, indicator antibody, consisted of rabbit anti-gp conjugate. One major modification of the CLIN. MICROBIOL. REV. EIA has been the use of typing sera or monoclonal antibodies for typing rotaviruses that are detected in stool specimens. Although IEM can also be used to serotype rotaviruses (420), it is not practical for large numbers of specimens, and serotyping by the CF method is not as sensitive as other methods (418, 420, 421). Zissis and Lambert (421) developed a type 1 and a type 2 specific EIA, using type-specific antisera as both the capture antibody and the detector antibody, plus a third indicator antibody-conjugate. Thouless et al. (357) also used polyclonal serotyping reagents in an EIA to distinguish human rotavirus serotypes 1, 2, and 3. Singh et al. (332) and White et al. (394) developed an EIA by using monoclonal antibodies to either subgroup I or II, which reacted to the 42,000-dalton inner shell protein. Shaw et al. (326) developed an EIA with monoclonal antibodies for serotypes 1 (human strain Wa) and simian serotype 2, which reacted with the VP7 of each serotype. Coulson et al. (64) also reported on an EIA to type human rotavirus, using neutralizing mouse monoclonal antibodies specific for serotypes 1, 2, 3, and 4 as detector reagents. All of the monoclonal antisera, except one, were directed to the major outer capsid protein gp34 (VP7), and one was directed to p84 (VP4, originally called VP3). The capture antibodies were hyperimmune rabbit antiserum to human rotaviruses

9 VOL. 2, 1989 types 1, 2, 3, and 4. Previously, the rotavirus serotype had been determined by cross-neutralization assays with hyperimmune antisera (64). A number of commercial EIA rotavirus detection tests are also now available. Several of these have been extensively evaluated and reported on in the literature. These include the Rotazyme I and Rotazyme IT EIA tests (Abbott Laboratories, North Chicago, Ill.), the Enzygnost EIA test (Behring Institut, Marburg, Federal Republic of Germany), the Pathfinder EIA test (Kallestad Laboratories, Austin, Tex.), and the Bio-EnzaBead EIA test (Litton Bionetics, Charleston, S.C.). There are several reports comparing Rotazyme and Enzygnost with EM and IEM. Yolken and Leister (410) evaluated Rotazyme I, Enzygnost, and indirect EIA and compared them with EM. They found the sensitivity of the indirect EIA, Rotazyme T, and Enzygnost to be 100, 93, and 88% and the specificity to be 95, 95, and 89%, respectively. Cheung et al. (50) evaluated Rotazyme I versus EM. The overall agreement was 88.7%, and the negative results had a 91.95% agreement. The tests were read visually, and specimens with high Rotazyme readings correlated 100% with EM. Rubenstein and Miller (310) compared Rotazyme I with EM and IEM. The levels of sensitivity were 106 particles per ml for simian rotavirus SA11 and 107 particles per ml for human rotavirus. The sensitivity and specificity of Rotazyme I compared with those of IEM were 98 and 92%, respectively. Rotazyme-positive specimens included those specimens that were EM negative but EIA positive that could be blocked in a blocking assay. Keswick et al. (190) also compared Rotazyme I and EM. They found EIA to be more sensitive than EM and that the Rotazyme test detected SA11 with a titer of 2 x 103 plaque-forming units (PFU)/ml, which was a level of sensitivity greater than that reported by Rubenstein and Miller. They also carried out blocking assays on EM-negative and EIA-positive specimens and found the EIA-positive specimens to be true positives. Chernesky et al. (49) evaluated Rotazyme TT, a version of Rotazyme with a shorter incubation time. They found that it was 99.4% sensitive and 97.3% specific with an overall agreement of 98.7% when compared with EM on 229 samples from patients aged 6 months to 6 years. In addition to ETAs, LA tests have been developed by Haikala et al. (140) and Hughes et al. (169). Like EIA, a number of LA tests are also commercially available and are sold under various brand names. One, Rotalex (Orion Diagnostica, Helsinki, Finland; sold in the United States by Medical Technology Corp., Somerset, N.J.) has been evaluated and reported on by several groups of investigators. It consists of latex beads coated with anti-human strain Wa rabbit serum (330). Others include the Slidex Rota-Kit (Biomerieux, Marcy-l'Etoile, France). Doern et al. (83) compared 176 specimens in Rotazyme I, Rotazyme TT, and Rotalex LA with a highly sensitive and specific monoclonal antibody EIA, described by Herrmann et al. (151). They found the sensitivities of the Rotazyme I and TT and LA to be 97.4, 100, and 81.6%, respectively, and the specificities to be 88.8, 83.9, and 100%, respectively. Thus, the Rotazyme TT was more sensitive but less specific than Rotazyme I or LA. Overall, Rotazyme I and TT were highly sensitive, but both lacked specificity. They also noted a problem with a large number of specimens having equivocal results with Rotazyme I and TT. Knisley et al. (202) used 100 specimens to evaluate four VIRAL GASTROENTERITIS 59 tests and compare them with EM. The four tests were (i) the Abbott Rotazyme TT, a polyclonal antibody-based EIA; (ii) the Pathfinder EIA, a monoclonal antibody-based EIA; (iii) a polyclonal-based EIA, using reagents obtained from the National Institutes of Health; and (iv) Rotalex LA. The sensitivities were 73, 95, 57, and 61%, respectively, while the specificities were 88, 95, 96, and 98%, respectively. Gerna et al. (125) used 151 specimens and compared the Pathfinder monoclonal antibody EIA with SPIEM as a reference test. They found Pathfinder to have a sensitivity of 98.7% and a specificity of 98.5%. Cevenini et al. (43) compared the Rotazyme I EIA and the Rotalex LA test (Finland) with EM and found the sensitivity to be 96% for both and the specificity to be 89% for the Rotazyme EIA and 86% for the Rotalex LA test. Sambourg et al. (313) tested 204 samples for rotavirus by four techniques: two ETAs, Enzygnost and Rotazyme T; and two LA tests, the Slidex Rota-Kit (Biomerieux) and Rotalex (Medical Technology Corp.). The positive rates were 47, 38, 37, and 34%, respectively. However, 12 specimens positive by the Enzygnost test only and 3 specimens positive by the Rotalex test only could not be confirmed positive by EM. Brandt et al. (27) found the Slidex Rota-Kit to be 82% sensitive and 100% specific. Miotti et al. (256) compared 122 samples by three commercial tests and their own reference microplate EIA. The three tests were the Rotazyme I EIA, the Bio-EnzaBead EIA, and the Rotalex LA test. The sensitivity was determined to a great extent by the time after the onset of illness during which the specimens were collected. There was no significant difference in the three tests when they were run on specimens collected early in the patients' illnesses. However, lower degrees of sensitivity were seen with the Rotazyme and Rotalex on specimens obtained later in the patients' illnesses. The lower sensitivity of the Abbott Rotazyme was not statistically significant, although the lowered sensitivity of the Rotalex LA was statistically significant. In addition, these authors found that a % tissue culture infective dose per 0.1 ml of virus suspension was detected at a 1:300 dilution by Bio-EnzaBead, at a 1:30 dilution by the Abbott Rotazyme, and at a 1:10 dilution by the Rotalex LA. There were no false-positive results with any of the three commercial tests, and this was seen with newborn as well as other specimens. Pai et al. (280) compared Rotalex (Finland) with Rotazyme I and EM with 165 stools from children and neonates. Rotalex had a sensitivity of 82% and a specificity of 96% compared with EM and was slightly more sensitive and specific than Rotazyme T. These authors also found that the sensitivity with Rotalex was dependent on the time of collection of stool samples relative to onset of symptoms. Sensitivities of Rotalex were 100, 96, 60, and 33% during 1 to 4, 5 to 7, 8 to 10, and 11 to 18 days, respectively, after onset of symptoms, and similar results were observed with Rotazyme T. Of 214 EM-negative specimens from asymptomatic newborns, the false-positive rates were 3.3% (7 of 214) for Rotalex and 4.2% (9 of 214) for Rotazyme I. Several workers have primarily evaluated the Rotalex LA test. Hammond et al. (145) compared Rotalex (Finland) with EM and found a sensitivity of 92% and a specificity of 98%. However, 19 of 218 specimens could not be evaluated since 10 of them gave equivocal results and 9 -of them caused agglutination of the control latex. Julkunen et al. (178), using 570 specimens stored frozen at -20'C, compared Rotalex LA and a noncommercial EIA used in their laboratory with EM results obtained from these

10 60 CHRISTENSEN samples prior to freezing. They found that their EIA was more sensitive than EM (168 versus 145 positive). Of 570 specimens, 30 were EM negative and EIA positive, and these specimens were positive in a confirmatory EIA. Sixteen (2.8%) of the LA-positive specimens were negative by both EM and EIA, and 15 of 16 were only slightly positive. They concluded that the LA test was good for screening, and definitely positive results were reliable. Moosai et al. (258) compared the Rotalex LA test with EM, their own EIA, and PAGE of viral RNA on specimens frozen at -70TC. Although they found LA the least complex to perform, it lacked sensitivity and specificity. They suggested four modifications to improve the test, including diluting the specimens 1:100 rather than 1:10 and reading at 20 min, not at 2 min. Shinozaki et al. (330) compared Rotalex LA with four other methods: PAGE, EM, SPIEM, and a commercial reverse passive hemagglutination test. The positive rates for the five methods were 61% (LA), 63% (PAGE), 59% (EM), 59% (SPIEM), and 57% (reverse passive hemagglutination). DeSilva et al. (79) studied Rotalex (Finland) in 90 children with diarrhea and found 90% (80 of 89) agreement with the "established" method(s) of EM alone or in conjunction with the Enzygnost EIA. Ten percent (9 of 89) were considered false-negative by Rotalex since they were positive by EM alone or by EM with EIA that could be blocked. Various problems have been associated with ETAs. Yolken and Stopa (414) initially reported problems with nonspecific reactions in ETAs. The nonspecific activity was markedly reduced by pretreatment of the specimens with reducing agents, normal goat serum, and anti-human IgM. The authors concluded that it was likely that the specimens contained an IgM antibody capable of reacting nonspecifically with other components of the assay. Although pretreatment with the mild reducing agent N-acetylcysteine markedly reduced this nonspecific activity, such treatment did not reduce the specific EIA activity due to rotavirus. Studies by Hovi et al. (167) suggested that false-negative results might result from fecal protease activity. This problem could be alleviated by adding 1 to 5% bovine serum to dilution buffers or by using a synthetic broad-spectrum serine-type protease inhibitor. Hogg and Davidson (162) evaluated false-positive results. They found that, when they incorporated preimmunization serum-coated wells as control wells in their EIA, 9.7% of specimens giving positive results were eliminated as falsepositives. Several investigators have reported problems with falsepositive EIA results on specimens from neonates. These occurred especially in earlier studies with Rotazyme I, in which lower positive cutoff values were used. When the cutoff value was raised, some of the problems appeared to be eliminated. Krause et al. (208) reported that 22% (79 of 358) of neonatal stool specimens from both asymptomatic and diarrheal neonatal patients were positive by Rotazyme I, although of 61 of 79 positives that were analyzed by confirmatory tests, only 7% (4 of 61) were confirmed as positive. This was compared with 77% that were confirmed Rotazyme-positive specimens from children and adults. Suggested causes of these false-positive Rotazyme tests in neonates included nonspecific binding of rotavirus antibody to bacteria or staphylococcal protein A in neonatal stools. However, Pai and Mayock (279), using Rotazyme I to study specimens from infants under the age of 4 months, found that only 9.8% (21 of 214) that were negative by EM were positive by EIA and thus presumed to be falsepositives. However, only 4.5% (9 of 202) were positive when visual readings of >1+ were considered positive, as recommended by the manufacturer. Similar results were seen by Rand et al. (294), who studied stool specimens from diarrhea-free infants in a neonatal intensive care unit with Rotazyme T. None had rotavirus by EM. By EIA, only 6.8% (10 of 147) were considered either low-level positive or suspect positive. Excluding the suspect positives, which were negative on repeat testing, the false-positive rate was 4.1% (6 of 147). With five repeatedly positive specimens with sufficient quantity to retest, heating to 560C for 30 min eliminated binding to the Rotazyme bead; heating had no effect on the Rotazyme-positive control. One highly falsepositive result was not changed by heat or other treatment. Thus, the investigators concluded that heat treatment of positive samples from neonates could eliminate most of the false-positives, although false-positives may result from more than one cause. Chrystie et al. (58) found a 15% (8 of 53) false-positive rate on specimens from 5-day-old babies, using Rotazyme T. However, they used an initial cutoff value specified in early Rotazyme directions; had a revised, higher cutoff value been used, only 1.9% (1 of 53) of the neonatal samples would have been falsely positive (295, 365). In addition, weakly positive and borderline Rotazyme reactions correlated poorly with direct EM findings (12, 50, 295). Rotbart et al. (309) obtained rectal swabs from symptomatic and asymptomatic babies in a neonatal intensive care unit in which an outbreak of necrotizing enterocolitis and hemorrhagic gastroenteritis occurred. A total of 4.0% (19 of 475) of specimens were positive by Rotazyme T, 2.1% (10 of 475) from symptomatic babies and 1.9% (9 of 475) from asymptomatic babies. Confirmatory tests were positive in 80% (8 of 10) of the specimens from symptomatic babies, while confirmatory tests were positive in only 33% (3 of 9) of the specimens from asymptomatic babies. Differences in Rotbart's results and those of Krause et al. CLIN. MICROBIOL. REV. (208) may have been due to Rotbart's use of swabs, which may have contained less inhibitory substances, or inhibitory substances may have been less stable on swabs than in the stool samples used by Krause et al., or both. Since all of the studies were carried out at different locales and times, year-to-year and lot-to-lot variations in key reagents might also have accounted for some differences. Rotbart et al. recommended that the then current Rotazyme I test not be used for screening asymptomatic infants. They also suggested the inclusion of some type of confirmatory testing in the Rotazyme kit, e.g., reaction with nonimmune serum or use of monoclonal antibodies. Rudd and Carrington (311), in screening babies in a neonatal intensive care unit, found that 2.9% (5 of 170) of babies had specimens positive by the Rotazyme I test. Two of these babies had necrotizing enterocolitis, one had bloody diarrhea, and two were asymptomatic. Thus, a high falsepositive rate did not occur in this study. Giaquinto et al. (126) prospectively studied 500 fecal specimens from neonates in an obstetrical ward. Supernatant fluids after 3,000-rpm centrifugation of specimens were used in the Enzygnost EIA, and 5% (25 of 500) were positive. Of these, 52% (13 of 25) were confirmed positive by a blocking EIA. All (100%) positive specimens from babies with diarrhea were also positive in the confirmatory blocking test, whereas only 33% (6 of 18) from asymptomatic patients were positive in blocking tests. The authors concluded that the Enzygnost test is suitable only in neonates with symp-

11 VOL. 2, 1989 toms, and they did not recommend it for screening asymptomatic neonates. They suggested that the Enzygnost and Rotazyme test kits include a confirmatory reaction in their test kits. Herrmann et al. (151) compared both Rotazyme and their own EIA, which used a monoclonal detector antibody, with EM. They evaluated specimens from three types of patients: neonates, children, and adults. The sensitivity of their monoclonal EIA was 100% for all, while that of Rotazyme was 100, 86.2, and 50.0% in the three types of patients, respectively. The specificity of the monoclonal EIA was 100, 95.6, and 96.5%, while that of the Rotazyme was 37, 95.6, and 89.7% respectively. Thus, they concluded that, especially for specimens from children and adults, greater sensitivity and specificity could be achieved with monoclonal antibody as detector antibody as compared with Rotazyme. Also, use of the monoclonal antibodies eliminated the false-positive reactions seen in specimens from neonates when Rotazyme was used. Herrmann et al. thought that, by using monoclonal antibodies as detector antibodies, the test sensitivity may be decreased, since these antibodies react with only one epitope of a given antigen. This problem could be diminished by using monoclonal antibody as a capture antibody and a polyclonal serum as the detector antibody, as these investigators had reported for a rotavirus RIA. The advantage of using the monoclonal antibody for capture is that it may pick up a significant amount of an antigen with several different epitopes. These investigators considered that the polyclonal sera may contain a number of cross-reacting components and may be the cause of false-positives in neonates, since false-positives occur in neonatal specimens in other rotavirus ETAs besides Rotazyme. The disadvantage of using polyclonal antibodies as a detector is that they may react nonspecifically with antigen adsorbed to the solid phase. This may occur even if inhibitors such as serum, serum fractions, or gelatin are included in the diluents to block nonspecific reactions. For example, they found that in stools containing high titers of rotavirus there was sufficient reactivity to give a positive test with microtiter plates coated with preimmune sera, even though there was two to three times greater reactivity with plates coated with immune sera. None of the studies to date has found false-negative Rotazyme results in neonates. Other detection methods. Other testing procedures were also developed to detect rotaviruses or rotaviral antigens. Some of the earlier tests included counterimmunoelectrophoresis (237, 250, 251, 339, 366), CF (420, 422), various immunofluorescence tests (39, 358, 415), and RIAs (181, 249). Counterimmunoelectrophoresis and the CF test are rarely used now, since they are relatively insensitive. The RIA is sensitive (14, 319) and suitable for the study of large numbers of specimens, but the need for radioactive reagents and expensive radiation-counting equipment makes it impractical in many situations. Other techniques for detecting rotaviruses in stool samples have been developed. They are primarily research techniques and are not used routinely. Among the in vitro tests are an immune adherence hemagglutination test (262), a reverse passive hemagglutination test (314, 315), and a solid-phase aggregation of coupled erythrocytes technique (26, 388), and PAGE (258). Various methods to detect rotavirus or rotavii-al proteins in infected cell cultures have been developed and include a collodial gold-protein A-IgG technique (289), a radioimmunofocus assay (220), and an immunoperoxidase procedure (44). More recently, several sensitive and specific research techniques for the detection VIRAL GASTROENTERITIS 61 of rotavirus in clinical specimens have been developed, which include a dot-immunobinding assay (299), a dotavidin-biotin-amplified immunobinding assay (299), an avidin-biotin RIA (407), and several dot hybridization techniques (82, 100, 116, 217). Antibody detection. Various methods have been used to detect antibodies to rotaviruses. These have included 1EM (186), the indirect fluorescent-antibody test (23), CF (185), counterimmunoelectrophoresis (62), hemagglutination inhibition (HAI) (239), immune adherence hemagglutination (243), reverse passive hemagglutination inhibition (316), neutralization (359), and EIA (417). Neutralization tests have been used extensively for serotyping rotaviral isolates and are still the standard, although typing now can be done by EIA. Several types of tests have been developed to detect rotavirus IgG and IgM antibodies in human sera as well as to detect secretary IgA (siga) in sera and stools. These can be used for several purposes: (i) to evaluate the use of serologic tests for viral diagnosis, (ii) to study the immune response to rotavirus infection, (iii) for epidemiologic purposes, and (iv) to evaluate potential rotavirus vaccines. Yolken et al. developed an EIA test to detect rotavirus IgG and IgM levels in sera (417) and secretary antibody in human milk (410). They immobilized rotavirus antigens on microtiter wells by preadsorbing guinea pig antirotavirus serum onto the wells, since adsorption of rotavirus antigens directly onto wells did not give reliable results. McLean et al. (246) developed an EIA test to measure IgG, IgM, and IgA in serum and mucosal secretions (colostrum, milk, and fecal extracts). Grauballe et al. (129) also described an EIA to detect siga in patient sera. For this test, microtiter plates were precoated with rabbit anti-human rotavirus serum, followed by the addition of purified human rotavirus as the capture antigen. Rabbit anti-human secretory component, conjugated to horseradish peroxidase, was the detector antibody. Inouye et al. (173) described an EIA for detecting IgA in stool samples. Initially they found that direct adsorption of purified rotavirus antigen to a solid phase was unsatisfactory. However, they achieved good results by disrupting virions into small subunits, using the chaotropic agent NaSCN or guanidine HCl, prior to their direct adsorption of antigen onto microtiter plates. Zentner et al. (419) developed an indirect immunoperoxidase assay to detect rotavirus IgG antibody in sera. They used SA11-infected MA-104 cells on glass slides and antihuman IgG peroxidase conjugate. Heimer and Cubbit (147) described an indirect immunofluorescent-antibody test to detect IgG and IgM antibodies in sera. Bovine rotavirus-infected LLC-MK2 cells grown in microtiter plates were used as the substrate. Patient sera were then added, followed by anti-human IgG or IgM fluorescein conjugate. These tests are primarily used as research tools. They are not commercially available at this time, and there is probably no great demand for them since their primary use is not as diagnostic tests. Immunology A number of studies have been carried out in the field of rotaviral immunology. This area of study of the immune response to rotavirus infection has been of great interest, since rotavirus infections appear to be repetitive. What role various immunoglobulins and other factors play in the pro-

12 62 CHRISTENSEN CLIN. MICROBIOL. REV. TABLE 2. Serum rotavirus immunoglobulin response to infection Type of antibody Time antibody first detected Length of time antibody Reference(s) persisted Neutralizing antibody During convalescence Protection for <1 yr 52 Neutralizing antibody During convalescence Not given 59, 292 IgM In acute phase, elevated Decreased in convalescence 75, 301 IgG In convalescence, elevated Detectable at least 6 to 12 mo 155, 159 IgA Not given -12 mo 75 IgA Within 1st 2 wk.6 mo 155, 159 ScIg Within 1 to 2 wk <4 mo 155, 159 siga 4 to 10 days 4 to 10 days 129 tection against rotavirus infection is important to know, since oral vaccines are being developed that may confer immunity to rotavirus infection. Most of the antibody studies discussed below were carried out by EIAs, although RIA and indirect fluorescent-antibody procedures were used in a few studies. The temporal antibody response in the sera of patients with rotavirus infection has been studied by several groups of investigators (Table 2). Chiba et al. (52) reported that protection against rotavirus gastroenteritis seemed to be serotype specific and related to neutralizing antibody levels. The protective effect was of short duration (<1 year), which they concluded was a probable explanation for recurrent attacks of gastroenteritis caused by the same serotype. However, Clark et al. (59) studied the neutralizing antibody response to four human rotavirus types in eight infants after rotavirus infection. Five originally seronegative infants and three originally seropositive infants showed an increase in neutralizing antibody to two or more serotypes. No consistent pattern of response to different serotypes was detected. Similarly, Puerto et al. (292) studied neutralizing antibody responses in 36 convalescent children and found that 19 seroconverted to two or more serotypes. Conversion to two or more types occurred in both initially seronegative and seropositive infants. The investigators suggested that heterotypic responses may be due to the existence of epitopes that induce antibodies capable of neutralizing viruses belonging to different serotypes, since this has been seen with monoclonal antibodies. Riepenhoff-Talty et al. (301) and Davidson et al. (75) found that, in sera, rotavirus-specific IgM first appeared during the acute phase of infection. This was replaced in convalescence by rotavirus-specific IgG, while IgM decreased. Convalescent serum IgG titers were significantly lower in severely ill infants as compared with moderately ill babies (301). Most children still had detectable serum IgG titers after 12 months. Hjelt et al. (155, 159) also found that rotavirusspecific IgG was elevated as long as 6 months postinfection. They found that the severity of the illness correlated only with the increase of IgG in the serum, a result opposite what was reported above (301). Studies on the role of secretary antibody have been of special interest. Two studies by Hjelt et al. (155, 159) determined that IgA increased in serum within the first 2 weeks of illness and persisted in the serum for at least 6 months postinfection. Similarly, Davidson et al. (75) found that serum IgA dropped significantly or was not detectable by 12 months. Hjelt et al. (155, 159) found that the immunoglobulin secretory component (ScIg) increased in serum 1 to 2 weeks after disease onset; however, it had disappeared by 4 months. Grauballe et al. (129) detected siga in patient sera for only 4 to 10 days after rotavirus was detected in stools, and this correlated well with recent infection. These differences may have been due to differences in the sensitivities of the EIA tests used. Hjelt et al. (155, 159) found that the amount of ScIg in the serum after 1 week correlated with the amount found in the duodenal fluid. Rotavirus IgA in feces still persisted in most patients at 6 months postinfection, whereas ScIg had disappeared in the feces of all patients except one. The intestinal antibody response has been studied by other investigators as well (Table 3). In fecal specimens, Riepenhoff-Talty et al. (301) found rotavirus-specific siga in both acute and convalescent samples. Davidson et al. (75) reported that IgA in duodenal secretions was significantly higher in convalescent-phase secretions than in acute-phase specimens, although duodenal IgM levels were the reverse. In patients with severe or prolonged infection, convalescent duodenal siga levels were higher than in patients with mild or moderate disease (301). Sonza and Holmes (338) studied fecal antibodies in four small children after a family outbreak. Although low levels of fecal IgG, IgM, and IgA were observed in most cases at the time of onset, the levels started increasing by 1 to 2 weeks, peaked by 3 to 5 weeks, and quickly declined so that none were detectable after 2 months. Stals et al. (344) found that maximum excretion of fecal IgA occurred about day 7 of illness, when virus excretion, which peaked at 2 to 5 days, was subsiding. They concluded that IgA limits the duration of diarrhea and plays a major role in intestinal resistance to infection. Shinozaki et al. (331) detected fecal IgA at day 9 postinfection in their patients; excretion of rotavirus had subsided by day 8 in these patients. Fecal IgA reached a maximum titer by 2 to 6 weeks and then declined. Yamaguchi et al. (406) found that in primary infection fecal IgA appeared at 1 week after the onset of illness, and titers reached a plateau between 3 and 5 weeks and then declined gradually. In reinfection, IgA titers increased more rapidly and were maintained at higher levels for a longer duration, typical of an anamnestic response. To determine whether IgA appearing after primary infection limits the infection and plays a role in resistance to reinfection, challenge studies would probably have to be carried out in volunteers or prospective studies would have to be done. Thus, in a prospective study, Hjelt et al. (158) determined preexisting concentrations of serum rotavirus-specific IgA and IgG prior to the rotavirus "season." They found that, although preexisting IgA did not protect from infection, children with preexisting IgA developed milder cases of gastroenteritis. Concentrations of IgA measured by EIA were similar in symptomatic and asymptomatic groups. Rotavirus IgG did not have any protective effect. However, age had a protective effect in that older children had milder disease. There have been several studies on the levels of rotaviral immunoglobulins in the serum, milk, and colostrum of

13 VOL. 2, 1989 VIRAL GASTROENTERITIS 63 TABLE 3. Intestinal rotavirus immunoglobulin response to infection Location of Type of Time antibody first Length of time antibody Reference(s) antibody antibody detected persisted Duodenal fluid ScIg Within 1 wk; level similar Not given 155, 159 to that in serum Feces IgA Not given.6 mo 155, 159 Feces ScIg Not given <6 mo 155, 159 Feces IgA Acute phase Throughout convalescence 301 Duodenal secretions IgA Acute phase, low Throughout convalescence, higher 75 than in acute phase Duodenal secretions IgM Acute phase, high Throughout convalescence, lower 75 than in acute phase Feces IgA, IgM, IgG Low levels at onset; Peaked at 3 to 5 wk; lasted <2 mo 338 increased by 1 to 2 wk Feces IgA Maximum level by 7 days Not given 344 Feces IgA Detected by 9 days Peaked at 2 to 6 wk; then declined 331 Feces IgA Primary response: de- Peaked at 3 to 5 wk; then gradual 406 tected by 7 days decline Anamnestic response: Peak levels lasted longer detected sooner normal mothers and in cord blood and stools of their were low. In bottle-fed infants, no specific IgA was found in newborn infants (Table 4). These studies have been impor- fecal supernatant fluids. tant because of the possible role of immunoglobulins in the Hjelt et al. (157) found that the levels of IgA and ScIg in protection of infants from rotavirus infection. milk declined from 3 to 4 days postpartum to 2 weeks later McLean and Holmes (245) studied 92 mother-infant pairs. and then remained unchanged for 2 months. In duodenal Rotavirus-specific IgG, but not IgM, was detected in all fluid, rotavirus IgA and ScIg were seen only in a minority of maternal and cord sera tested, and this was the only immu- infants at 3 to 4 days and 2 weeks after birth. However, 80% noglobulin that crossed the placenta. Rotavirus-specific IgA of fecal samples contained rotavirus-specific IgA and ScIg, was found in 100% of maternal sera tested, while specific although they disappeared after lactation ceased. Interest- ScIg was found in 53% of the maternal sera. In another ingly, IgA and ScIg can survive proteolysis in the gut. Thus, study, Hjelt et al. (157) found that all of the mothers tested frequent breast meals have a possible protective effect. had IgG and IgA in their sera and 78% had low levels of ScIg Yolken et al. (412) studied secretary antibody (ScIg) to in their sera 3 to 4 days postpartum. Infant serum IgG levels rotavirus in human colostrum and milk. They found rotacorrelated well with those of their mothers, but they had no virus-specific ScIg in all colostral samples and most of the rotavirus IgA or ScIg in their sera. milk samples. Peak ScIg levels in colostrum fell to low but These same two groups of investigators also evaluated the detectable levels in milk 1 to 2 weeks postpartum and protective role of immunoglobulins in the colostrum and remained at these levels for 2 years after delivery, similar to milk in the same two groups of normal mothers. Rotavirus- results reported above. There were no significant differences specific IgG and IgM were found in the colostrum and milk in samples from different geographical areas of the world. of many of the mothers by McLean and Holmes (245), and Cukor et al. (71) studied rotavirus-specific siga in moththese levels dropped off significantly by 3 to 5 days postpar- ers' milk. They found that 80% of the mothers were positive tum. Rotavirus-specific IgA and ScIg were found in all for IgA <1 week after parturition, 50% were positive 7 to 14 colostrum and milk samples, except for two mothers who days postpartum, and 24% were positive at 1 month postwere IgA deficient. Specific IgA and ScIg levels in colos- partum. However, as late as 6 to 9 months postpartum, 56% trum-milk dropped off rapidly immediately postpartum to a of those tested were positive. This was due to mothers who steady low level at 3 to 4 days postpartum. In breast-fed were initially negative eventually becoming antibody posiinfants, specific IgA was found in fecal supernatant fluids tive, possibly because of a subclinical rotaviral infection. after 2 days, even when maternal milk IgA concentrations Thus, two routes of passively transferred rotavirus antibody TABLE 4. Rotavirus immunoglobulins in normal mothers and newborns Location of Type of Time antibody first Length of time antibody Reference antibody antibody detected persisted Colostrum/milk IgG, IgM Not given Dropped off by 3 to 5 days postpartum 245 Colostrum/milk IgA, ScIg Not given Dropped off by 3 to 4 days postpartum to 245 steady low level Feces of breast-fed infants IgA At 2 days Not given 245 Colostrum/milk IgA, ScIg Not given Declined from 3 to 4 days postpartum to wk later; then remained unchanged Infant duodenal fluid IgA, ScIg At 3 to 4 days At least 2 wk 157 Colostrum/milk ScIg Not given Fell to low but detectable levels 1 to 2 wk 412 postpartum and remained unchanged for 2 wk Colostrum/milk IgA Not given Continued dropping off for 7 to 30 days 71

14 64 CHRISTENSEN were identified by these studies, serum and colostrum-milk. However, serum antibodies have doubtful protective capacity (118). In addition to specific antibodies, nonspecific factors in milk and in the gut may play a role in the protection of infants against rotavirus infection. Studies by McLean and Holmes (247) suggested that both siga and trypsin inhibitors in human milk may protect neonates against rotavirus infection during the first 5 days of life. They found that breast-fed babies were less likely to become infected with rotavirus and showed lower stool tryptic activity or higher antirotaviral siga antibody, or both, than bottle-fed babies. While the above studies concentrated on the levels of various rotavirus-specific immunoglobulins in body fluids, these levels were not correlated with their possible role in the protection against, or modification of, rotavirus infection and disease. Thus, several studies have been done on the role of specific rotaviral antibody in rotavirus infection. Chrystie et al. (57) studied rotavirus infection in an endemic situation in a neonatal nursery. They found that breast-fed babies excreted rotavirus significantly less frequently than those who were formula fed. Also, if infection did occur among breastfed infants, they excreted less virus than infected formulafed infants. More specifically, in a second publication, this group (363) reported on the role of specific antibody in cord blood and breast milk in this same outbreak of rotavirus gastroenteritis. The ranges, distributions, and geometric mean titers of serotype 2-specific IgG in cord blood and serotype 2-specific IgG and IgA (97% positive) in the mothers' breast milk were similar among rotavirus-positive and rotavirus-negative neonates. Interestingly, there was no correlation between the amount of milk IgA antibody and the amount of neonatal viral excretion. These studies suggest that factors other than rotavirus antibody in breast milk are of importance in preventing rotavirus infection in neonates. To determine which nonspecific factors may play a role in resistance to rotavirus infection, these same workers (364) evaluated the role of neutralizing activity, nonimmunoglobulin antiviral activity, and a1-antitrypsin activity in the mothers' milk. They were unable to correlate any of these factors with protection from rotavirus infection. However, they found that rotavirus in infected stools from breast-fed infants was frequently clumped, although that in stools from formula-fed infants was rarely clumped (361). Thus, when they specifically studied human lacteal rotavirus antibodies by IEM (361), they found that the IgA- and IgG-containing fractions in human and bovine milk samples caused aggregation of rotavirus particles. Clumping did not occur with antibody-negative milk or with the nonimmunoglobulin fractions of antibody-positive milk. They suggested that adsorption to susceptible cells of viral aggregates formed in the presence of low levels of antibody and containing unneutralized virus was one mechanism by which infection in the presence of antibody might occur. Similar results were found by Weinberg et al. (392). In 50 infants who had rotavirus gastroenteritis within the first year of life, 64% had been breast-fed and 70% had not. There were no significant differences between the two groups in the average age of infection, the mean duration of diarrhea, the mean number of bowel movements in 24 h, or the frequency of fever or irritability. The only apparent difference between the groups was that the frequency of vomiting was significantly decreased in the breast-fed children. Their results suggested that breast-feeding offered little protection against rotavirus gastroenteritis. Similarly, Duffy et al. (90) found CLIN. MICROBIOL. REV. that, although rotavirus attack rates were similar between breast- and bottle-fed infants, rotavirus infection in breastfed infants was milder and of shorter duration than in bottle-fed babies. In addition to studies on the role of passively acquired maternal antibody in neonatal infection, studies on the role of actively acquired antibody in preventing or modifying recurrent infection have been carried out. Bishop et al. (17) studied the ability of neonatal rotavirus infection to confer immunity to postneonatal rotavirus infection. Fifty-five percent of infants with neonatal infection and 54% of those without infection developed rotavirus infection during the following 3 years. Babies with reinfection had symptoms that were significantly less severe than those who had no previous neonatal infection. They concluded that neonatal rotavirus infection does not confer immunity against reinfection, but does protect against the development of clinically severe disease during reinfection. In addition to the study of natural passive immunity, studies on the use of artificial passive immunity for prevention and treatment of rotavirus infection have been carried out. Barnes et al. (8) evaluated the use of human gamma globulin or placebo given in each feed during the first week of life to 75 low-birth-weight infants in a nursery where rotavirus was endemic. Twenty-five of the babies excreted rotavirus during the first 2 weeks of life and were considered the "challenge" group. Gamma globulin administration was associated with delayed excretion of rotavirus and with milder symptoms. The diarrhea necessitated low-lactose feeds in 55% (6 of 11) of placebo babies but in only 7% (1 of 14) gamma globulin-treated babies. The authors concluded that oral human gamma globulin seemed to protect lowbirth-weight infants from rotaviral diarrhea. In another study (223) two children and one adolescent with primary immunodeficiency syndromes and prolonged excretion of rotavirus were given a single oral dose of human serum immunoglobulins. The result was the generation of rotavirus-specific immune complexes in the stools with a subsequent decrease in the presence of uncomplexed rotavirus antigen. The clinical efficacy of this treatment was not determined. However, the authors concluded that the oral administration of immunoglobulins with specific reactivities had potential for prevention or treatment of gastrointestinal infections. Ebina et al. (96, 97) immunized pregnant cows with human rotavirus to prepare rotavirus-specific IgA-rich cow colostrum. In a therapeutic trial, the rotavirus-specific cow colostrum had no effect on the duration of diarrhea, bowel movements, or virus shedding. However, when used prophylactically, good results were obtained. Only 17% (one of six) of infants developed rotavirus diarrhea, whereas 86% (six of seven) of controls given market milk as a placebo developed diarrhea. Similarly, Brussow et al. (38) immunized pregnant cows with four human rotavirus serotypes. They then prepared freeze-dried lactation milk in which rotavirus-neutralizing activity was 10 times higher than that of pooled human immunoglobulins. When the milk concentrate was used to treat infants with rotavirus gastroenteritis (154), there was a reduction in the duration of virus excretion and of diarrhea. Yolken et al. (411) found detectable levels of rotavirus IgG in both raw and pasteurized milk, but little or no antibody in commercial infant formulas. In vitro, the milk samples inhibited the replication of human and bovine rotaviruses in tissue culture. In vivo, the milk samples decreased or

15 VOL. 2, 1989 prevented diarrhea in infant mice infected with virus-milk mixtures. Offit and Clark (274) found that, when pregnant mice were inoculated with rotavirus and their newborn mice were challenged, the maternal antibody protection was dependent on both antibody titers and the serotype, with protection being primarily to the homologous strain. Similarly, Losonsky et al. (225) found that incubating murine rotavirus with homologous mouse antisera prior to infection prevented symptomatic disease, whereas heterologous antisera did not neutralize the virus as effectively. Asymptomatic mice, however, developed an immune response. In contrast to the number of studies on the role of antibodies in rotavirus infection, there are few reports on the cellular immune response. Totterdell et al. (362) studied the lymphoproliferative responses to rotavirus antigen in whole blood taken from several types of individuals. These included healthy adults, elderly patients at geriatric institutions, adult renal transplant patients, one child with severe cellular immunodeficiency who had been excreting rotavirus, and cord blood from healthy babies. Using a lymphocyte transformation assay, they found that elderly and transplant patients had significantly lower lymphoproliferative responses to rotavirus antigen as compared with healthy adults. However, after acute rotavirus infection, the one infected transplant patient and the elderly patients had good, though transient, lymphoproliferative responses and good specific antibody response. The cord blood samples showed no lymphoproliferative responses to rotavirus, although they contained specific antibodies. The immunodeficient child had neither an immunoproliferative nor an antibody response to rotavirus antigen. Results similar to those seen in the healthy adults in the above study were seen in newborn mice infected with rotavirus in another study (303). A virus-specific cell-mediated response appeared in splenic lymphocytes 2 days after infection and peaked at 10 days postinfection. This peak of activity closely followed the cessation of diarrhea, suggesting that cellular immunity may function as a limiting factor in disease (303). However, Eiden et al. (99) found that neonatal T-cell-deficient (athymic) mice developed a self-limited rotavirus infection identical to that in immunocompetent mouse controls (99). Their athymic mice developed no serum or intestinal antibody, whereas the normal controls did. They suggested that recovery from rotavirus infection in mice is not dependent on functional T cells or on specific antibody, but rather is mediated by nonimmune mechanisms or macrophages. Thus, the mechanisms involved in recovery and protection from rotavirus infection are complex, involving an interplay between local and systemic and humoral and cell-mediated responses, and also nonspecific factors (362). Prevention Vaccines. Since rotavirus is a serious problem in developing nations and is the single most important etiologic agent of acute gastroenteritis requiring hospitalization in developed countries (188), there is a need to control this disease by use of a rotavirus vaccine. Vaccines would be administered orally, since this is the natural route of infection and this route would stimulate local intestinal IgA antibody (188). Several groups have been developing experimental oral vaccines for human use. In Europe, a collaborative group of Finnish and Belgian investigators ( , 382) have been evaluating an attenuated bovine rotavirus vaccine strain, RIT The rationale for a heterologous rotavirus vaccine VIRAL GASTROENTERITIS 65 has been (i) the ease with which animal rotaviruses can be propagated in vitro and (ii) their sharing of a common group antigen with most human rotavirus strains (group A) (378). This group antigen is the major inner capsid protein VP6, which is also the distinct subgroup antigen. RIT 4237 is a subgroup I rotavirus. The RIT 4237 vaccine strain was passed 147 times in primary bovine kidney cells, was then passed 7 times in primary Cercopithecus monkey kidney cells, and was used at pass 154 (362) with a mean titer of % tissue culture infective doses per oral dose. After evaluation of the safety of the vaccine in adults, 17 seronegative 2-year-old children were given a single oral dose. In this preliminary study, 70% of the children seroconverted unequivocally and 18% had a possible seroconversion, for a total seroconversion rate of 88%. None of the children had a major clinical reaction to the vaccine; i.e., no gastrointestinal or constitutional symptoms occurred. None of the children excreted rotavirus, although three individual stools were positive for antigen by EIA. In a larger group of 178 pediatric patients aged 8 to 11 months given the vaccine during the rotavirus season, the protective effect of the vaccine was evaluated (379). Of the seronegative infants, 47% of the vaccine recipients seroconverted, but there was an 88% protection rate from clinically significant rotavirus diarrhea. In this study, although it appeared that serum antibody could be used as a marker of successful vaccination, protection against clinical diarrhea could not be strongly associated with these serum antibodies. Vaccination and subsequent naturally acquired immunity protected against clinical diarrhea. However, vaccination did not protect against subclinical infection, as determined by a booster response or by antibody seroconversion. In a pre-rotavirus season vaccination study carried out in the autumn, infants aged 6 to 12 months were given either two doses of the RIT 4237 vaccine or a placebo (377). Results similar to those of the previous study were seen. Vaccinees who failed to seroconvert had less rotavirus diarrhea than placebo recipients, suggesting that immunity may be mediated by factors other than serum EIA antibody. The protection rates against clinically significant diarrhea caused by serotypes 1, 2, and 3 were 72, 100, and 100%, respectively. The RIT 4237 vaccine appeared to protect against several human rotavirus serotypes. In one study, the subgroup I RIT vaccine offered protection during an outbreak of subgroup II rotavirus (379), which is the most prevalent subgroup seen in the United States. In another study, limited evidence suggested that the RIT vaccine offered protection against serotypes 1, 2, and 3 (377). Preliminary data also indicated that this vaccine may offer protection for at least 2 years (376). Since the RIT 4237 strain is relatively acid labile and may lose infectivity if exposed to high gastric acidity occurring between feedings in infants, the investigators felt that an appropriate acid-neutralizing substance such as milk might increase the seroconversion rate. Thus, a RIT 4237 vaccine with % tissue culture infective doses was given to breast-fed, formula-fed, and fasting infants (382). A neutralizing antibody response occurred in 69% of the breast-fed, 100% of the formula-fed, and 63% of the fasting infants, with an overall response rate of 77%. The breast milk was somewhat inhibitory, since it may give partial protection against naturally occurring rotavirus diarrhea. Poorer responses occurred when the vaccine was used containing or % tissue culture infective doses per dose.

16 66 CHRISTENSEN Basically, the European bovine RIT 4237 vaccine was safe, but not as immunogenic as desired, lacking immunogenicity in some infants and children. This vaccine has not been marketed commercially. Another candidate rotavirus vaccine has been developed and evaluated at the National Institutes of Health (5, 55, 184, 224). The vaccine strain is also an animal virus, rhesus rotavirus-1 strain MMU The virus was derived from the stool of a 3-month-old rhesus monkey with diarrhea. It was passed nine times in primary or secondary monkey kidney cell cultures and seven times in a semicontinuous fetal rhesus monkey diploid lung cell strain, DBS-FRhL-2, developed by the Division of Biologic Standards (184). The major neutralizing protein is closely related or identical to VP7, the major outer capsid protein detected by crossneutralization (380) of human serotype 3 (184, 224). This vaccine is more immunogenic than the RIT 4237 vaccine, but it also produces more side reactions. In initial studies, the MMU strain was found to be attenuated in newborn and juvenile rhesus monkeys. It was first evaluated in human young adults and then children aged 2 to 12 years and was found to be immunogenic in >80%, but caused no side reactions (224). The vaccine, containing 105 PFU per dose, was then tested in 14 infants aged 5 to 20 months; 13 infants were given placebo. It was administered after the infants received formula buffered with sodium bicarbonate. There was a fourfold or greater rise in serum antibody titer in all of the vaccinees, but none in the placebo controls. The primary side reaction to the vaccine was the development of fever above F (37.80C) 3 to 4 days postvaccination. All vaccinated children excreted rotavirus in their stools within 1 week of vaccination. In a similar study of 13 vaccinated infants and 10 controls, significant rises in serum antibody titers occurred in all vaccine recipients, although they also exhibited side effects to the vaccine (5). These included rhinorrhea, passage of a larger than usual number of stools, and passage of more semiformed or unformed stools as compared with controls. All vaccinees shed rotavirus in their stools. The MMU vaccine appeared to be more immunogenic but caused more side reactions in American populations than the RIT 4237 vaccine did in Finnish recipients. One difference between the two populations was that, in the Finnish studies, most of the vaccine recipients were initially seronegative as determined by EIA, and many of the remaining vaccinees had only low levels of detectable antibody, whereas higher percentages of the American study group populations had detectable rotavirus antibody prior to vaccination. Since the vaccine populations differed, a collaborative study was carried out comparing the two vaccines in one study population in Finland (380). Results of both vaccines were similar, with the RIT vaccine being less immunogenic but showing fewer side reactions, while the MMU vaccine was more immunogenic with more adverse side effects. Adverse reaction rates to the MMU vaccine were lower, however, in a Venezuelan study. This was due perhaps to differences in preexisting immunity due to maternally transferred antibody in the vaccinees in that country. Thus, the MMU vaccine might be suitable for this type of population (380). The MMU vaccine might also be useful in developing countries, when the only chance to vaccinate might be simultaneously with the oral polio vaccine, which might interfere to some extent with the rotavirus vaccine. The MMU vaccine might also be effective in early infancy, when passively acquired antibodies are at a high CLIN. MICROBIOL. REV. level and when less side reactions may occur. In such a population, a more attenuated vaccine may not take, as was shown to be a problem with the RIT 4237 vaccine (380). However, the RIT 4237 vaccine did seem to offer protection against infection in infants in highly developed countries when the infants had little or no preexisting antibody. The high level of protection seen in spite of the lack of serum antibody response may be due to protection caused by factors such as local IgA antibody and cell-mediated immunity. Other rotavirus vaccines are also being considered. Clark et al. (60) reported on a vaccine developed from bovine rotavirus strain WC3, adapted to grow in cultures of CV-1 cells, a line of African green monkey kidney cells. The vaccine, used after the 12th cell culture passage, contained 3 x 107 PFU per dose and was initially tested in adults and older children. After it was tested in 52 infants and children aged 5 months to 6 years, no clinical sequelae occurred, and viral shedding was found in only 30% of the vaccinees. Serum neutralizing antibody was induced to the WC3 strain in 95% of the infants aged 5 to 11 months and to human serotype 3 virus in 50% of the vaccinees. Preexisting antibody to human serotype 1 or 3 frequently exhibited a booster response. The use of reassortant vaccines has also been proposed by the National Institutes of Health group (254, 255), since rotaviruses have a segmented genome which can undergo genetic reassortment in vitro (188). This group coinfected cell cultures with (i) the cultivatable bovine rotavirus strain UK and (ii) one of several "noncultivatable" human rotavirus strains representing serotypes 1, 2, and 3. Monoclonal antibodies to the major outer capsid neutralization glycoprotein VP7 were used to select reassortants with human rotavirus-neutralizing specificity. The technique yielded many reassortants which received only the gene segment coding for the major neutralization protein for the human parent strain and the remaining genes for the animal rotavirus parent strain. Reassortants of this type represent potential vaccine strains. Theoretically, additional reassortant rotavirus vaccines containing VP4 and VP7 from two antigenically distinct rotavirus parents might also protect against disease induced by two or more serotypes (276). Epitope-specific immune responses to rotavirus vaccine have been studied by Shaw et al. (326). They used serotypespecific monoclonal antibodies directed at VP7 in a competitive-binding EIA to measure epitope-specific immune responses to serotypes 1, 2, and 3 in sera of children who received a serotype 3 vaccine. Antibodies to serotype 3 were detected in 72% of the sera, and antibodies to serotypes 1 and 2 were detected in only 11% of the sera. Also, specific monoclonal antibody to VP4 (originally called VP3), which neutralizes three serotypically distinct strains of rotavirus, was used to detect the presence of similar antibodies in 56% of the test sera. This finding suggests a mechanism of heterotypic immunity to rotavirus vaccination. Chemical disinfection. The prevention and control of rotavirus infection by chemical disinfection have been studied by Springthorpe et al. (342). They studied the effects of 76 chemical disinfectants on suspensions of virus in the presence or absence of tryptose phosphate broth (peptides and inorganic salts) or fecal matter. Thirty-two percent of the disinfectants were considered highly or moderately effective and inactivated at least 106 PFU of virus in 1 min. However, the remaining 68% were effective only in the absence of organic matter or were completely ineffective. The same investigators (221) also evaluated 27 disinfectants for their

17 VOL. 2, 1989 ability to inactivate rotavirus on inanimate objects, using disks of stainless steel, glass, and plastic contaminated with 107 PFU in fecal matter which was allowed to dry. Only 9 of the 27 disinfectants reduced the rotavirus titer by 6 logo, while the others were ineffective. These nine disinfectants are used for specific purposes in a variety of products for home, hospital, and food service use. They include hydrochloric acid, peracetic acid, isopropyl alcohol, chlorhexidine gluconate, glutaraldehyde, chloramine-t, povidone-iodine complex, sodium o-benzyl-p-chlorophenate, and a quaternary ammonium compound. Since rotavirus was resistant to a wide range of commonly used chemical disinfectants, it is important to use those chemicals that are effective to prevent and control outbreaks of rotavirus disease. Treatment Although oral gamma globulin and bovine colostrum have been used experimentally in the treatment of rotavirus gastroenteritis as described above, these are not the mainstay of treatment of viral gastroenteritis. Morbidity and mortality in patients with rotaviral and other viral gastroenteritides are primarily due to dehydration and electrolyte imbalance (90, 206). Thus, the primary purpose of therapy is to correct these problems by providing adequate hydration to maintain blood volume, electrolyte homeostasis, and acid-base balance (90). Replacement of water and electrolytes in dehydrated patients can be done with either intravenous therapy or oral electrolyte solutions, depending on the clinical situation. A child with diarrhea must first be assessed for his/her degree of dehydration. This includes evidence of weight loss, changes in skin turgor, sunken eyes, dry mucous membranes, absence of tears, decreased urine output, and changes in mental status or vital signs or both (90). Intravenous fluid therapy for significant dehydration and fasting have been the standard therapy in developed nations for over 40 years (90, 218). Following this, a period of fluid maintenance and a gradual return to a normal diet is advocated. The rationale for this approach is to provide bowel rest and to prevent the possibility of food-induced malabsorption (360). Intravenous therapy is needed for severely sick children, who are (i) in shock and unable to drink fluids, (ii) persistently vomiting, and (iii) having stool losses of >100 ml/kg per h and (iv) who cannot tolerate oral fluids (360, 396). However, the routine use of intravenous hydration for less severely ill children has been questioned recently. Often oral sugar-electrolyte rehydration solutions are used in the United States for maintenance therapy of mild diarrhea with dehydration, as an interim feeding prior to grading to a normal diet, or are incorporated with early refeeding to restore nutritional losses (218, 360, 396). Oral glucose-electrolyte solutions have been used for 25 years in the treatment of dehydration due to acute infantile diarrhea including rotavirus diarrhea. Two basic types of oral electrolyte solutions are now used in the United States. The first type is "rehydration" fluids with a sodium content of 60 to 90 meq of sodium per liter, and the second is "maintenance" fluids of 30 to 50 meq of sodium per liter. The glucose content in both solutions is 2.0 to 2.5% (grams per deciliter); higher amounts are undesirable since they can produce osmotic diarrhea (396). Glucose-coupled sodium transport is the physiological basis for oral rehydration solutions, since glucose enhances sodium transport in the small intestine. VIRAL GASTROENTERITIS 67 The oral glucose-electrolyte rehydration fluids are effective, inexpensive, and widely available for hospital or home use. Other advantages are that they avoid the cost and risk of intravenous therapy and they stimulate recovery of small bowel mucosal absorptive function (396). In one study, Listernick et al. (218) studies 15 patients treated with oral solution as outpatients in an emergency room holding room and 14 matched controls treated as inpatients on intravenous therapy. Rotavirus was detected in 11 of 15 orally dehydrated patients and in 10 of 14 intravenously rehydrated patients. The group found that 13 of 15 of the orally treated patients were successfully rehydrated in the emergency room in 10.7 h at a cost of $275 versus h at $2,300 for the inpatients. In underdeveloped or developing countries where there are limited medical resources and where malnutrition is common, oral rehydration therapy and continued feeding have been advocated (360). For marginally nourished or malnourished children, diarrhea associated with starving can have deleterious effects. When a person has fasted for 3 to 5 days, depletion of intestinal digestive enzymes and gut mass occurs, and absorption and digestion of glucose, salt, water, amino acids, and disaccharide is substantially reduced (360). Besides the negative effects of fasting, there are some positive effects of feeding a child who has diarrhea in that intraluminal nutrients induce intestinal digestive enzymes and cell proliferation (360). Thus, in underdeveloped countries, continued feeding is an important adjunct to oral rehydration therapy. Older children with mild to moderate dehydration can be treated at home with oral fluids. Sugar-containing soft drinks, juices, and tea can be used. Since these drinks usually have a low sodium content (0.5 to 20 meq/liter), additional sodium supplement from salted crackers, bouillon, or soup is needed. However, in a small infant with severe diarrhea, more precise fluid and electrolyte replacement is necessary (396). Children with acute diarrhea are often fasted to reduce consequences of malabsorption, acidosis, fluid loss, and depleted bowel and mucosal injury. However, advocates of early feeding believe that they can prevent or minimize the deficit of calories and protein, maintain or stimulate the repair of the intestinal mucosa, and maintain the brush border by avoiding a prolonged fast (396). The American Academy of Pediatrics recommends that, if the child is stable, refeeding should not be delayed for more than 24 h unless there is significant dehydration, severe vomiting, or abdominal distention (396). Feedings should begin with breast milk or dilute formula and advance slowly to fullstrength formula over 2 or 3 days. Nonlactose carbohydrate bland solids such as rice, cereal, and potatoes should be reintroduced as soon as they are tolerated. Bland foods are thought to have a fast emptying time in the stomach and less likelihood of stimulating peristalsis. Also, frequent small feedings are less likely to distend the stomach and cause subsequent vomiting. Small feedings also present the small intestine with less lactose load at any one time (396). Infants require close observation for evidence of transient, mild, secondary lactose intolerance when starting feeding or formula, especially with rotaviral disease. Evidence of lactose intolerance includes a marked increase in stools, abdominal distension, frothy stools, or flatulence. An acidic stool ph and a positive Clinitest (Ames Div., Miles Laboratories, Inc., Elkhart, Ind.) for sugar further suggest lactose intolerance (396).

18 68 CHRISTENSEN ADENOVIRUSES Introduction Adenoviruses are 70- to 75-nm, nonenveloped, icosahedral, ds deoxyribonucleic acid (DNA) viruses classified in the family Adenoviridae. Adenoviruses infect most species of mammals, birds, and amphibians (386). Human and other mammalian adenoviruses are in the genus Mastadenovirus. There are 41 known distinct serotypes of human adenoviruses, now considered to be 41 species (386). The first adenovirus was originally isolated from adenoid tissue, and adenoviruses have been associated primarily with respiratory disease, as well as with ocular and genitourinary tract infections. Early reports also associated adenoviruses with diarrhea in infants and children, but actual causation was not determined (177, 257). These adenoviruses associated with diarrhea could be isolated in routine cell cultures and consisted of lower serotypes. During the last 13 years, adenoviruses have been detected in stools of patients with gastroenteritis (33, 108, 110, 119, 197, 232, 297, 300, 324), using the technique of EM (Fig. 1B), but often, they could not be isolated in routine cell cultures. These viruses that are not readily cultivable are called enteric or fastidious adenoviruses and represent new types, 40 and 41 (77). Some investigators have suggested using the term "fastidious" adenovirus for those adenoviruses that are detected by EM, in EIAs specific for types 40 and 41, or in special cell cultures such as 293 cells, but which cannot be propagated in commonly used routine cell culture (228). The general term "enteric" adenovirus is often used for any adenovirus observed in stool by EM, in which cell culture propagation is not attempted. This terminology will be followed in this review. CLIN. MICROBIOL. REV. Molecular Biology and Classification Most of the research on the molecular biology of adenoviruses has been carried out on adenovirus serotypes 1 to 39. However, much of the basic molecular biology of these serotypes applies to the fastidious types 40 and 41 as well. Adenoviruses, including the fastidious ones, are 70- to 75-nm icosahedral viruses made up of 252 capsomeres. The outer capsid is composed of two main capsomere types. The first type is the 240-nonvertex hexon capsomeres and the second type is the 12-penton base capsomeres which form the 12 vertices of the icosahedran. A fiber vertex projection protrudes from the base of each penton (387). The dsdna core contains the linear dsdna with a molecular weight of 20 x 106 to 25 x 106 (387). The capsid components have antigenic determinants with a wide range of specificities. The hexons are responsible for group, subgroup, and type specificities (387). Mammalian adenoviruses share a common group-specific determinant that can be measured by the CF test. The antigens that specify the adenovirus serotypes are located on the hexon and are measured in the neutralization test. The penton capsomeres carry group and subgroup specificities. This group specificity is also shared by mammalian (human, simian, and canine) adenoviruses. The fibers carry subgroup and type specificities. This type specificity is measured by the HAI test (387). There are 41 different adenovirus serotypes (77) considered to be species by Wadell (386). The 41 types share the group antigen, which can be detected by CF tests. The different serotypes were distinguished initially by neutralization tests, although RIAs and EIAs have been developed to distinguish types 40 and 41. The adenoviruses were first classified into groups or subgroups in several ways (168, 308, 387). Huebner (168) divided the human adenoviruses into subgroups based on their oncogenicity for newborn hamsters. Subgroups A and B contained "highly" and "weakly" oncogenic adenoviruses, respectively. Nononcogenic adenoviruses that transformed rodent cells in vitro were classified into two subgroups, C and D (244). Interestingly, this classification correlates with the subdivision of adenoviruses based on the G+C content of their genomes (130, 290). Since the above grouping of adenoviruses is based on properties representing <6% of the genome, Wadell et al. (386, 387), studied the adenovirus relationships based on their virion polypeptides. These polypeptides are encoded by a major portion of the adenovirus genome. Based on their polypeptide pattern, adenovirus types 1 to 39 fell into subgroups A to D, with the exception of type 4, whose pattern did not fit and thus was classified into subgroup E. A fastidious adenovirus isolated from diarrheal stools studied by Wadell also had its own distinctive characteristic polypeptide pattern and was therefore placed in subgroup F (type 40) Ġreen et al. (131) also identified five groups, A to E, based on the DNA relatedness among 31 adenovirus types. Wadell et al. (386, 387) compared the restriction endonuclease digest patterns of their subgroup F fastidious adenovirus with other subgroups. They found that their fastidious adenovirus displayed a unique restriction pattern with no resemblance to other adenovirus types and thus should be relegated to its own subgroup. De Jong et al. (77) and Uhnoo et al. (369), using restriction endonuclease analysis, determined that fastidious adenoviruses formed two different species (serotypes) and subgroups. Serotype 40 was placed in subgroup F and serotype 41 was placed in subgroup G. However, these two types have biological similarities and are indistinguishable by HAI tests (192). Thus, types 40 and 41 appear to have very similar DNA sequences coding for their antigenic determinants even though their genomes are apparently different as seen by restriction endonuclease digestion (194). Takiff et al. (351) carried out physical mapping of types 40 and 41 and found that, although their restriction profiles were different, they appeared to have several cleavage sites in common. Cross-hybridization studies showed considerable homology between types 40 and 41, but much less homology occurred between these two types and adenovirus type 2. Van Loon et al. (375) found that the DNA homology between the two types was 62 to 69% and recommended that adenovirus type 41, previously classified as subgroup G, be classified together with type 40 in subgroup F. These subgroups have also been referred to as subgenera (271, 375). Epidemiology Adenoviruses were among the first viral agents associated with acute nonbacterial infectious gastroenteritis (177, 257). In the 1960s, Moffet et al. (257) detected adenovirus in 17% of infants <2 years of age with diarrhea compared with 5% of control infants. However, these adenoviruses were types 1 to 31, with lower-numbered types predominating, and all were cultivated in routine cell cultures. Kapikian et al. (185) detected adenovirus by EM in 11% (16 of 143) of infants and young children with diarrhea, although in this early study culture was not carried out and there was no control group.

19 VOL. 2, 1989 Spratt et al. (341) found adenovirus by EM in 17% (7 of 41) of infants with diarrhea versus 5% (4 of 74) of asymptomatic controls. CPE was detected in six of eight inoculated HEp-2 cultures, and the viruses isolated were types 2 to 7. In the late 1970s, several groups reported that some adenoviruses in diarrheal stools of pediatric patients that were detected by EM could not be cultivated. Flewett et al. (110) first reported a nosocomial outbreak of gastroenteritis in 32% (6 of 19) of patients in a long-stay children's ward of an orthopedic hospital in England. Adenovirus was detected in four of six sick children, but in none of the asymptomatic controls. Attempts to culture the virus in a number of cell culture types failed except that subtle CPE was seen in some HEp-2 cultures 10 to 20 days after inoculation. These effects were not seen on passage. In 3 of 30 patients in Johannesburg, South Africa, Schoub et al. (324) detected, by EM, adenoviruses that were fastidious. Although they produced CPE with typical nuclear inclusion bodies in human embryonic fibroblast cultures, they could not be passaged or typed. By EM, Middleton et al. (253), in Toronto, Ontario, Canada, found adenoviruses in 12.7% (85 of 669) of viruspositive stools from pediatric patients with gastroenteritis. The authors did not claim that all of these adenoviruses were true causative agents of disease since no control population was studied. However, adenoviruses were seen more frequently by EM than could be isolated in cell culture. The authors also noted that adenoviruses were sometimes accompanied by what appeared to be adenovirus satellite viruses. Madeley et al. (232), in Glasgow, Scotland, detected adenovirus in 11.5% (21 of 183) of babies under 2 years of age with diarrhea; 9 adenoviruses were detected only by cell culture, 10 were detected only by EM, and 2 were detected by both methods. Those detected only by EM showed some evidence of growth in rhesus monkey kidney, human amnion, and/or human embryonic kidney (HEK) cells by CPE, but they could not be passed or typed; EM of the cell cultures did not suggest viral replication. It appeared to these investigators that the chance of growing adenovirus was inversely related to the amount of virus seen by EM. This observation regarding fecal adenoviruses has been made by others. Bryden et al. (39) found that many fecal samples containing adenoviruses that could not be isolated in cell culture contained large numbers of particles, up to 1011 virions per ml. In a four year study in Washington, D.C., Brandt et al. (33), detected significantly more uncultivatable adenoviruses in pediatric inpatients with gastroenteritis (5.1%, or 31 of 604, only 5 of which grew in cell culture) than from inpatient controls (1.9% or 10 of 522, 6 of which grew in cell culture). Similar results were seen in outpatients with gastroenteritis. This strongly suggested that these noncultivable viruses played a role in acute enteric disease. Similarly, Richmond et al. (300) reported a 2-month outbreak of gastroenteritis in 17 of 24 children under 2 years of age at a Royal Air Force station in the United Kingdom in which poorly cultivatable adenoviruses were detected by EM. None of the adenoviruses could be serially propagated in baboon kidney, HeLa, human embryonic lung, or HEK cells. However, a transient CPE typical of adenovirus developed in the HEK cultures inoculated with stool samples containing large numbers of adenovirus particles, and the affected cells were stained by indirect immunofluorescence with specific adenovirus antiserum. In a retrospective, 4-year study of 2,606 stool specimens from patients with acute gastroenteritis in Toronto, Retter et VIRAL GASTROENTERITIS 69 al. (297) detected 392 (15%) adenoviruses by EM. Of these 392, 216 (55%) grew in HAE-70 cells, a continuous line of human amnion cells, whereas 176 (45%) did not produce CPE. However, by immunofluorescence of the non-cpeproducing strains, only isolated cells fluoresced, with no spread to adjacent cells, indicating an abortive infection. Yolken et al. (409), in Baltimore, Md., and Washington, D.C., found fastidious adenoviruses in 14 of 27 (51.9%) cases of diarrhea during a 12-week period and one fastidious adenovirus in 1 of 72 (1.4%) children without diarrhea. Although nonfastidious adenoviruses were found in stools of 2 of 27 (7.4%) diarrheal patients, such virus also occurred in 5 of 72 (6.9%) controls. Fastidious adenoviruses were identified by their growth in 293 cells and by using rabbit antiserum specific for fastidious adenoviruses in a typespecific EIA. In a 2-year period in Glasgow, 159 stools from 71 children under 3 years of age were found to contain adenovirus by EM in a study by Kidd et al. (195). Established adenovirus types (1 to 39) were isolated from 81 of the stools from 40 of the children, 7 of whom shed two to three adenovirus types simultaneously. Thirty-six children shed fastidious adenoviruses in 64 specimens, and nine of these patients shed both fastidious adenoviruses and known serotypes at different times. The fastidious viruses produced little or no CPE in routine cultures or could not be typed or both; they were identified by CPE in Chang's conjunctival cells and by neutralization tests with antiserum produced against fastidious adenovirus. It was proposed by de Jong et al. (77) that two antigenically related, fastidious adenovirus variants that had no relationship to the 39 known human adenovirus species be called adenovirus types 40 and 41. De Jong's group studied 200 of these fastidious adenoviruses, which failed to replicate serially in conventional human embryonic fibroblast and HEK cells, but many of which could be established in Graham's 293 cells, Chang's conjunctival cells, or sublines of HeLa or cynomolgus monkey kidney cells that were treated in special ways. None of the 200 were related to types 1 to 39 by either neutralization or HAI tests. The variants were identical in HAI tests, but DNA restriction enzyme analyses showed that the two species, 40 and 41, had considerably different genomes. The reference strains are Tak and Dugan, respectively. Fastidious adenoviruses are ubiquitous with worldwide incidence. From the above studies, it can be seen that fastidious adenoviruses have been found in many parts of the world. They are probably the second most important cause of infantile gastroenteritis after rotavirus. Adenoviruses were found to be the second most common enteric virus detected by EM in a number of studies from the United States, United Kingdom, Scandinavia, and South Africa (30, 32, 33, 110, 185, 198, 324, 367). It was the third most common virus in other studies from the United States (284, 302), Canada (253), India (235), and South Africa (89) and the fourth most common virus in a Scottish study (232), although none were detected in one South African study (322). In many instances, fastidious adenovirus infections may occur throughout the year with no seasonal variation, according to reports from the United States, Canada, and Scotland (30, 31, 195, 253, 302), although there is a tendency toward more cases in the warmer months in reports from the United States (33, 185), United Kingdom (108), Scandinavia (381), Japan (51), and South Africa (89, 198, 324). Fastidious adenovirus infection tends to be endemic, rather than epidemic, although outbreaks in hospital nurseries have oc-

20 70 CHRISTENSEN curred. Fastidious adenoviruses usually infect and cause symptoms in children up to the age of 3 years, but primarily in those 2 years or younger (110, 195, 381). Clinical Features The symptoms associated with fastidious adenovirus infection have been studied by a number of investigators. Diarrhea appears to be the predominant symptom, with (195, 300, 370) or without (110) vomiting. The severity has ranged from mild, afebrile illness (110) to a fatal or "comatose state" (395) due primarily to dehydration. Reports of the duration of symptoms have been 2 to 4 days (110), a mean of 5 days (324), a mean of 7 days (300), and 4 to 8 days (195). In a Swedish study (370), diarrhea lasted 4 to 23 days with a mean of 9 days for those infected with type 40 and 2 to 47 days with a mean of 12 days for those with type 41 infection, with one-third of the type 41 patients having symptoms for 14 days or longer. Adenovirus-infected patients have milder disease with less pronounced vomiting and less fever and diarrhea than rotavirus-infected patients (367). Respiratory symptoms have also been associated with gastroenteritis in which adenoviruses have been detected in the patients' stools. In one study (381), clinical signs of respiratory tract infection was present in 31% of patients with diarrhea in which adenovirus was detected in their stools. In this study, adenovirus was detected by a RIA test in which the antisera used were prepared against a type 2 hexon antigen, so that the group antigen was detected. In still another study (409), the stools of 14 of 27 patients with diarrhea were found to contain the common hexon antigen of adenovirus, using an EIA test. Of these 14 patients, 13 also had respiratory symptoms such as cough, wheezing, or rhinorrhea; 6 of these had X-ray evidence of pneumonia; and 3 had bilateral conjunctivitis. Thus, it was concluded that the enteric types not only cause acute gastrointestinal disease, but also can be associated with a high rate of respiratory disease. The role of fastidious adenoviruses (types 40 and 41) in causing respiratory symptoms and that of respiratory adenoviruses (lower-numbered types) in causing diarrhea have been of great interest, but also a problem for the interpretation of data. In some of the earlier studies, adenoviruses were detected in stools by EM or by EIA or RIA tests with antisera to an adenovirus group antigen. Thus, one cannot determine in which outbreaks or with what incidence fastidious adenoviruses were found and what disease was attributable to lower-numbered, nonfastidious types. In addition, the term enteric refers only to adenoviruses detected in stools, without indicating their fastidiousness or type. Thus, it has been recommended that the term fastidious refer only to those adenoviruses that cannot be propagated in routine cell culture or are typed as 40 or 41 or both (228). With techniques becoming available to differentiate between the fastidious and nonfastidious adenoviruses in stools, the role of both groups in causing respiratory and gastrointestinal symptoms could be more thoroughly evaluated. In one study by Uhnoo (370), 21% of the patients with adenovirus types 40 and 41 also had respiratory symptoms, i.e., tonsillitis, pharyngitis, otitis, coryza, or cough. Patients who had diarrhea due to "established" adenovirus types, i.e., types 1 to 39, had diarrhea of shorter duration and higher fevers, and 79% had respiratory symptoms. Similarly, Leite et al. (215) detected adenovirus in stools of 39 of 746 children with gastroenteritis by IEM or EIA or both. Of these 39, 25 could be propagated in HEp-2 cells and CLIN. MICROBIOL. REV. neutralized by one of the antisera to types 1 to 18. The remaining 14 of the 39 could be propagated only in special 293 cells and were not neutralized by antisera to types 1 to 31. Thus, they concluded that "respiratory" adenoviruses could probably be responsible for gastroenteritis. Since fastidious adenoviruses have been associated with respiratory symptoms in some reports, the evaluation of the presence of adenoviruses in stools is a difficult diagnostic dilemma (215). However, in another study, Uhnoo et al. (367) found respiratory symptoms to occur only rarely in patients infected with types 40 and 41 and concluded that these adenoviruses were restricted to the intestinal tract. Other problems with interpretation of adenoviruses observed in diarrheal stools include (i) long-term shedding of adenoviruses in the stool and (ii) simultaneous infection with different enteric and respiratory tract adenoviruses. For example, Fox et al. (117) found that adenoviruses 1, 2, and 5 could be excreted from the intestinal tract without symptoms for at least 2 years after a primary infection. Other investigators have also found that types 1 and 2 especially may be shed in the stools for many months (195). Kidd et al. (195), in a 2-year study, obtained stool samples containing adenoviruses from 71 children, and serial samples were obtained from 35 of these children. Adenoviruses were detected by EM and by routine cell culture. However, there was no certainty that an adenovirus seen by EM was the one that grew in routine cell culture. The longest interval in which the same serotype (type 2) was isolated was 231 days. The investigators found that adenoviruses of different established serotypes may be shed by one child in succession or simultaneously over a period of days. Four children shed two adenovirus types and three children shed three adenovirus types over a period of 1 to 5 months. In two of these cases, two types were isolated from the same stool. Thus, dual and triple infections are probably common in children, with prolonged and overlapping infection by different serotypes. Of the 71 children, at least 36 shed fastidious adenovirus, and 9 of these children shed fastidious adenovirus and known lower serotypes at different times. Although common established serotypes were shed over several weeks or more, excretion of the fastidious adenoviruses was not seen for longer than 8 days, at least in the amount needed to be observed by EM. Brandt et al. (29) addressed the problem of simultaneous infections with different enteric and respiratory tract viruses. Infants and young children in whom adenovirus or rotavirus was visualized in their stools were tested for the simultaneous presence of respiratory viral pathogens in their respiratory tracts. Nearly 11% had dual respiratory and enteric viral infections. Non-adenovirus respiratory tract pathogens were seen in 8.5% (4 of 47) of gastroenteritis inpatients, 50% (8 of 16) of respiratory inpatients, and 19% (12 of 63) of total inpatients with visualized fecal adenovirus. Respiratory syncytial virus was the most common respiratory tract pathogen isolated from these dually infected patients, followed by parainfluenza virus type 3. To summarize, in cases of gastroenteritis, it is probably important to use a diagnostic test that specifically detects types 40 and 41 per se. Since a type 40- and 41-specific EIA is now commercially available (Adenoclone-Type 40/41; Cambridge BioScience, Worcester, Mass.), this test could be used in conjunction with one of the commercially available rotavirus tests on stool specimens from cases of gastroenteritis.

21 VOL. 2, 1989 Laboratory Diagnosis There are a number of methods for detecting adenoviruses and adenoviral antigen or DNA in stools. Some methods are less specific, e.g., EM and some RIAs and EIAs, in that differentiation between fastidious and nonfastidious adenoviruses cannot be done. Other tests, including type-specific ETAs and some hybridization techniques, are specific for types 40 and 41. In addition, special cell types can be utilized to propagate the fastidious adenoviruses. Initially adenoviruses were detected in stools by routine cell culture (nonfastidious types) and by direct EM (both fastidious and nonfastidious types). In earlier studies, adenoviruses in stool extracts could also be detected by counterimmunoelectrophoresis (288) and immunoelectroosmophoresis (174). RIAs and ETAs to detect adenovirus in stools have also been described. Halonen et al. (141) described a highly sensitive and specific four-layer RIA. In this test, adenovirus type 2 hexon antigen was used as the immunizing antigen for producing hyperimmune sera. The same group (319) also described a similar four-layer EIA. Johansson et al. (175) described two EIAs for adenovirus, one group specific and one type specific for subgroup F, type 40 adenovirus only. The group-specific test detected 35 established adenovirus types plus the subgroup F adenovirus. In both tests, the capture antibody was a groupspecific antibody, whereas the detector antibody was either group specific or type 40 specific, depending on the test. Thus by the use of both tests, all adenovirus types in the stool could be detected by the first test, and those of type 40 could be differentiated by the second test. In addition, Johansson et al. (176) described a similar type-specific EIA for adenovirus type 41. Singh-Naz and Naz (333) developed monoclonal antibodies to types 40 and 41 for use in an EIA. Two of these monoclonal antibodies recognized a single antigen of 17,000 molecular weight, which is probably polypeptide VII, the DNA core polypeptide. One of the two monoclonal antibodies reacted with both types 40 and 41, while the other reacted only with type 40, probably due to reaction with different antigenic determinants. These antibodies confirmed the close antigenic relationship as well as the distinctiveness of types 40 and 41. Type-specific ElAs for types 40 and 41 are also commercially available (Cambridge BioScience). Hierholzer et al. (153) described a monoclonal timeresolved fluoroimmunoassay, using a monoclonal antibody to adenovirus type 3 hexon antigen. This test was more sensitive than an all-monoclonal biotin-avidin EIA and a polyclonal capture biotin-avidin EIA, but it did not differentiate types 40 and 41 from other adenovirus types in specimens. However, Wood and Bailey (398) were able to detect types 40 and 41 specifically in stool by IEM, using hyperimmune rabbit sera to these two types. In addition to these serology-based tests, nucleic acid hybridization tests have been described recently. Stalhandske et al. (343) used nucleic acid extracted from 18 RIA-positive stool specimens which were spotted on nitrocellulose filters (spot hybridization). These were analyzed with 32P-labeled probes to type 2 DNA and to a cloned BamHI G fragment of adenovirus type 41. Fifteen of the 18 stools were positive to the type 2 probe. Five of the 18 were positive to the type 41 probe and were detected by the type 2 probe as well. Thus, by the use of both tests, fastidious adenoviruses could be distinguished from nonfastidious ones. Takiffet al. (352) described a similar dot blot hybridization VIRAL GASTROENTERITIS 71 test, using extracted stool specimens dotted onto nitrocellulose and hybridized with 32P-labeled probes to EcoRI and BglII fragments of types 40 and 41. This test was quite specific and sensitive, detecting as little as 20 pg of adenoviral 40 and 41 DNA. Kidd et al. (196) described a similar dot blot test, using only clarified or extracted stool suspensions and hybridization with 32P-labeled probes to cloned PstI fragments of types 40 and 41. The test was relatively fast, taking 48 h. Hammond et al. (143) simplified the hybridization detection test by spotting samples directly from stool specimens rather than extracting the stools. Niel et al. (271) described a type 40 and 41 hybridization test, using a nonradioactive peroxidase-labeled probe. Several of the detection tests described above were developed because fastidious adenovirus types 40 and 41 from stools could not be propagated in routine cell cultures in the laboratory. However, Takiff et al. (353) found that these fastidious adenoviruses could be propagated in Graham's 293 cells (128), HEK cells that were transformed by exposing the cells to sheared fragments of adenovirus type 5 DNA (128). In routine cell types, such as HeLa cells, the in vitro replication of fastidious adenoviruses is probably blocked at a very early stage in the viral growth cycle, whereas this block does not occur in the 293 cells (353). In 293 cells inoculated with stools containing fastidious adenovirus, small foci of progressive adenovirus-like CPE appear after several days (353). The specificity of the CPE can be determined by direct fluorescent-antibody test (353). However, the pattern of granular nuclear and cytoplasmic fluorescence differs from the fluorescent pattern seen with lower-numbered adenoviruses (353). Fastidious adenoviruses grow better when they are seeded into 293 cells that are subconfluent (36). Graham's 293 cells have also been used in suspension culture for isolating fastidious adenoviruses (329). One problem with the use of 293 cells is that lowernumbered non-fastidious adenoviruses also grow well in them (35, 36). In one study by Brown et al. (35), 4 of 15 stools positive for type 40 adenovirus also contained a lower-numbered adenovirus type, and these lower-numbered types overgrew the type 40 strain. In addition to the 293 cells, Kidd and Madeley (197) reported that 42% of fastidious adenoviruses from stools could be propagated in Chang's human conjunctival cell line, with development of typical adenoviral CPE. However, not all fastidious adenoviruses could be propagated in this cell type. Immunology Although extensive immunologic studies have been carried out on lower-numbered adenoviruses, there have been only a few reports on the immunity to adenoviruses 40 and 41 per se. Kidd et al. (193) obtained 377 serum samples from children under the age of 12 years living in the United Kingdom, New Zealand, Hong Kong, Guatemala, Gambia, and Kuwait. By a neutralization test, they found that at least 33% of the sera from the United Kingdom, Hong Kong, and the isolated country of Gambia had neutralizing antibody to 40 and 41. Sixty percent of the New Zealand sera was positive. The Kuwaiti sera had a lower percentage (15%) of positive sera, while all of the Guatemalan sera were negative. In a study of the age distribution of antibody in Japan, Shinozaki et al. (328) found that 20% of 1 to 6 month olds and

22 72 CHRISTENSEN 50% of 37 to 48 month olds were antibody positive to 40 and 41 by neutralization tests. Of adults, 48% of 18 to 20 year olds and 10% of people over 70 years of age were type 40 and 41 positive. All of the sera were positive to the adenovirus common antigen. Prevention and Treatment Experimental vaccines to types 40 and 41 have not yet been developed. Problems involve the relatively poor replication of fastidious adenoviruses in cell culture. The use of 293 cells as a substrate would be inappropriate, since these are transformed cells. In addition, both types 40 and 41 have been shown to have transforming ability in baby rat kidney cells (374). Fastidious adenovirus infection is not as common or as serious as rotavirus infection, and therefore the need for an adenovirus vaccine is not as pressing as that for a rotavirus vaccine. Treatment includes maintaining proper electrolyte balance and fluid levels, as in rotavirus infection. NORWALK AND NORWALK-LIKE VIRUSES Introduction Whereas rotavirus primarily causes disease in infants and young children, Norwalk and Norwalk-like viruses primarily infect and cause disease in older children and adults. The Norwalk virus is a small, round virus, 25 to 30 nm in diameter, which was first detected by Kapikian et al. (187) by EM examination of stools of patients from an outbreak of winter vomiting disease occurring in Norwalk, Ohio, in Several other viruses resemble Norwalk virus by their morphology, the type of clinical illness they cause, and their low concentration in infected stools. However, some are antigenically similar and others are antigenically distinct from the Norwalk agent. These other agents, discovered in the 1970s, include the Hawaii agent (401), the Montgomery County agent (356), discovered in Maryland, and the Snow Mountain (SM) agent, discovered in Colorado (87). These viruses, including the Norwalk agent, cannot be cultivated in vitro, so that little is known of their biophysical and biochemical characteristics. The viruses can be propagated only in human volunteers and in chimpanzees (133). Since very small amounts of these viruses are excreted in the stools of infected persons, the viruses can be detected by IEM, using convalescent sera from recovering patients. Frequently, they may not be detected at all. In such cases, a diagnosis may be made by detecting a significant serologic rise in antibody titer. Other small, round viruses which were originally thought to be Norwalk-like viruses are now considered to be candidate parvoviruses and include the Cockle agent and the Ditchling agent (6, 42). Physical Characteristics Norwalk and Norwalk-like agents observed by EM or IEM are 27 to 30 nm in diameter and have been described as picorna/parvovirus-like, as well as resembling caliciviruses (described below). However, their fine morphology is difficult to distinguish because of their small size and because, by IEM, the virions are coated with antibody. Since these agents cannot be cultivated in vitro, biophysical and biochemical studies have been carried out on virus purified from diarrheal stools of infected patients. The Norwalk agent is resistant to ether, since it lacks a lipid-containing outer envelope (22, 84). The Norwalk agent is relatively resistant to acid (ph 2.7 at room temperature), and this characteristic may protect the agent in its passage through the stomach (22, 84). The Norwalk agent is also relatively heat stable; after a viral suspension is heated to 60'C for 30 min, the suspension can still produce disease in some volunteers (22, 84). Norwalk virions, purified by Greenberg et al., appear to contain a single primary structural protein with a molecular weight of 59,000 (133). In addition, this group detected a soluble viral protein with a molecular weight of 30,000 in the stools of Norwalk-infected patients. The protein structure of the virion is similar to that of the Caliciviridae, which has one major structural protein of 60,000 to 71,000 molecular weight (133, 233). Parvoviruses differ in that they have three different structural proteins (133). The Norwalk agent has a buoyant density in CsCl of 1.38 to 1.40 g/cm3. The Caliciviridae size (35 to 40 nm) and buoyant density in CsCl (1.36 to 1.39 g/cm3) are similar, though not identical, to those of the Norwalk agent. The Hawaii agent has a density in the range of 1.38 to 1.40 g/cm3 in CsCl (356). The SM agent has been studied more extensively and compared with feline calicivirus (233). In CsCl, the SM agent peaked at 1.34 g/cm3 with a shoulder at 1.37 g/cm3 compared with the calicivirus, which peaked at 1.37 g/cm3 with a smaller peak at 1.33 g/cm3. On potassium tartrate-glycerol gradients, SM virions of two densities, 1.21 to 1.22 and 1.29 g/cm3, were observed, whereas the calicivirus peaked at 1.29 g/cm3 with a smaller peak at 1.21 g/cm3. In analyzing virion-associated proteins, a major SM agent structural polypeptide of 62,000 molecular weight was observed, similar in molecular weight to the major calicivirus structural protein of 65,000. These results suggested that SM, like Norwalk agent, has properties resembling that of the calicivirus group and may suggest a relationship between the two groups. The buoyant densities of Hawaii agent were found to be 1.37 to 1.39 g/cm3, with most virions recovered at 1.38 g/cm3 (356). The buoyant densities of the Montgomery County agent complete virions had a range of 1.37 to 1.41 g/ cm3, with the majority of virions found at 1.39 g/cm3 (356). Empty particles were seen at 1.30 g/cm3 (356). Epidemiology CLIN. MICROBIOL. REV. Norwalk virus has been responsible for a large number of outbreaks of acute infectious nonbacterial gastroenteritis. In one report, Greenberg et al. (136) obtained paired sera from patients in 25 outbreaks of gastroenteritis occurring over a 12-year period. Using the sera in a RIA test, they determined that 32% (8 of 25) of outbreaks were due to Norwalk or antigenically related viruses. One of the outbreaks was due to rotavirus, but the causes of the remaining outbreaks in this study could not be determined. The Norwalk outbreaks occurred in a grade school, colleges, a family, and cruise ships at sea. In a more extensive study by Greenberg et al. (137), 34% of 70 outbreaks of gastroenteritis were associated with Norwalk or antigenically related agents. In still another study, Kaplan et al. (189) reported that 38 of 81 outbreaks of gastroenteritis were caused by Norwalk or antigenically related viruses and gave detailed information on these occurrences. Ten outbreaks occurred in camps and recreational areas; seven, in elementary schools; four, in colleges; three, in small families; two, in larger communities; four, in restaurants; four, in nursing homes; and four, on ships at sea. Three occurred in other countries, while the remainder occurred in the United States. In this last study, the sources of infections included municipal water systems in two out-

23 VOL. 2, 1989 breaks, semipublic water supplies in seven, recreational swimming in two, and stored water on cruise ships in two; eating contaminated oysters in two and eating salad in two; and primary person-to-person transmission in seven. In addition, secondary person-to-person transmission occurred in a number of the outbreaks. The outbreaks lasted an average of 7 days, with a range of 1 day to 3 months. In seven outbreaks, successive weekly outbreaks occurred among newly introduced populations as on cruise ships or at camps. The largest outbreaks occurred in communities, schools, recreational areas, and on cruise ships, with an average of 348 people involved and with a range of 19 to 2,000 cases. Smallest outbreaks occurred in families and nursing homes. One nationwide outbreak was associated with eating raw oysters that were shipped around the country. The epidemiologic and clinical characteristics of the non-norwalk outbreaks suggest that they were caused by agents resembling, but antigenically distinct from, the Norwalk agent. They were similar in their incubation period, the prevalence of symptoms, and the duration of illness. In two of these outbreaks, 27-nm particles resembling Norwalk virus were seen in the stools of affected patients. Although the role of Norwalk virus and antigenically related viruses in large and small outbreaks of gastroenteritis is well documented, little is known about their endemic role in disease causation. This is due in part because patients usually are not hospitalized, and routine diagnostic laboratories do not have the reagents available for detecting these viruses or viral antibodies. Since Norwalk and similar viruses tend to infect older children and adults in developed nations, one would expect that these older populations would have a higher prevalence of antibodies than in small children. This is in contrast to rotavirus, in which antibody development occurs early in life, due to early infection with this virus. Several studies have addressed the issue of the development of antibody to Norwalk virus throughout different age groups which would reflect the acquisition of Norwalk infection. Blacklow et al. (21) found that acquisition of antibody to Norwalk virus in American subjects was minimal during childhood and was in the range of only 5% before the age of 12 years. Throughout adolescence and early adulthood, the prevalence of antibody rose rapidly, reaching 45% in 18 to 35 year olds and reaching to 55 to 60% in 45 to 65 year olds, before declining slowly in the elderly. In studying different populations, Greenberg et al. (134) found that antibody acquisition to Norwalk virus occurred much earlier in less developed nations. In underdeveloped countries such as Bangladesh and Ecuador, antibody levels ranged from 75 to 100% throughout childhood, while in Yugoslavia the rate of acquisition was between that of the less developed countries and that of the United States. Cukor et al. (72) found the rates of acquisition of antibody to Norwalk virus to be slow in the United States and in Taiwan, similar to Greenberg's rates, with rates in the Philippines higher, approaching that found in Yugoslavia by Greenberg et al. The report of Greenberg et al. (134) also noted the ubiquitousness of antibody to Norwalk virus from around the world. They looked at antibody levels in blood obtained from volunteer blood donors and found that 54 to 77% of blood from urban United States, Belgium, Switzerland, and Yugoslavia and from rural Bangladesh and Nepal had antibody to Norwalk virus. VIRAL GASTROENTERITIS 73 Clinical Features and Pathogenesis Clinical features. The clinical features of the Norwalk and similar agents have been reported from both the original outbreaks of these viruses and from studies in volunteers. The infections produced by Norwalk virus (85), the Hawaii agent (356), the Montgomery County agent (356), and the SM agent (87) are similar. The incubation periods were usually 24 to 48 h, with sudden onset. The most predominant features were nausea and vomiting, often severe. Low-grade fever and diarrhea usually occurred, but the latter tended to be relatively mild. Diarrheal stools did not contain blood, mucus, or leukocytes (86). Other symptoms included moderate to mild abdominal pain or cramps, headache, and malaise. Laboratory tests were within normal limits in volunteers inoculated with Norwalk agents, except that one had a transient elevation of leukocytes. The attack rate with these agents may be high, affecting 50% or more of the members of a school, camp, or other institution or group where an outbreak occurs. Respiratory symptoms do not appear to be a manifestation of Norwalk or Norwalk-like infections, which differs from that of rotavirus and possibly adenovirus infection. Children are more likely to have vomiting, whereas adults are more likely to experience diarrhea (189). The vomiting may be caused by a decrease in gastric mobility, which may be more pronounced in children than in adults (189). Although the disease seldom requires hospitalization, serious illness has occurred in rare instances. Hospitalization of a total of three middle-aged persons for severe dehydration was reported in two outbreaks (189). In one nursing home outbreak, three patients needed intravenous fluids, but they were not hospitalized. In another nursing home outbreak, two elderly debilitated patients died after the onset of gastroenteritis, but supposedly the deaths were due to diffuse atherosclerosis (189). Pathogenesis. The amount of virus needed to initiate an infection in human volunteers has been estimated. In one study, the Norwalk agent had a titer of at least 500 human disease-producing doses in the first human passage filtrate, whereas the disease was produced with only undiluted stool filtrate from a second human passage (22). In a study with the SM agent, Dolin et al. (87) reported that five of six volunteers given 1.0 to 0.5 ml of undiluted stool suspension became ill, two of two volunteers given 1.0 ml of a 1:10 dilution became ill, two of two given a 1:100 dilution became ill, and none given a 1:1,000 dilution became sick. Evaluating the respiratory route as a means of transmission of Norwalk infection, Blacklow et al. reported that nasopharyngeal washings from a volunteer acutely ill with Norwalk infection failed to induce disease in other volunteers (22). Thus, they concluded that the respiratory route is not a major route by which this virus could be transmitted. Jejunal biopsies from young adult volunteers obtained before, during, and after administration of suspensions of Norwalk virus have been described by Agus et al. (2) as follows: by light microscopy, abnormalities in the intestinal mucosa at the height of illness included partial villous flattening, broadening of the villi, and disorganization of the epithelial lining cells. Moderate infiltration of the lamina propria by mononuclear cells also occurred, and there were focal areas of epithelial vacuolization. EM of the jejunal epithelial cells showed dilation of the rough and smooth endoplasmic reticulum and an increase in multivesiculate bodies. The microvilli were shortened, and the intercellular spaces were widened and filled with an amorphous electron-

24 74 CHRISTENSEN dense material. No definite viral particles were seen. Convalescence biopsies carried out 2 weeks later showed that the mucosa had returned to normal in all volunteers. Agus et al. (2) also showed that, in addition, jejunal brush border enzyme activities, including alkaline phosphatase, sucrase, and trehalase levels, were decreased. Convalescence enzyme levels determined 2 weeks later returned to the base-line rates in all volunteers. These reduced enzyme levels may contribute to the pathogenesis of Norwalk disease. However, in contrast to the intestinal lesions seen with invasive bacterial agents, such as Shigella spp. and enteropathogenic Escherichia coli, in Norwalk-induced disease the mucosa remained intact. Rectal biopsies taken at the height of illness showed a normal histologic pattern. There was an absence of fecal leukocytes in Norwalk-induced illness. Thus, the colonic mucosa is relatively spared in this syndrome as compared with bacterial gastroenteritis. Inoculated volunteers who were asymptomatic had normal jejunal biopsies. Schreiber et al. (325) described the pathologic lesions seen in volunteers who ingested stool filtrates containing the Hawaii agent. The volunteers developed intestinal mucosal lesions in the proximal small intestine that were similar to those lesions seen in Norwalk infection. However, these two agents are immunologically distinct. Lesions from symptomatic volunteers showed shortening of the villi and an increased cellularity of the lamina propria. Many villous absorptive cells were decreased in height, and some had vacuolated cytoplasm. Increased numbers of mononuclear cells and some polymorphonuclear leukocytes infiltrated the intercellular spaces between epithelial cells. Many polymorphonuclear leukocytes and increased numbers of mononuclear cells were seen throughout the lamina propria. Two of four volunteers without symptoms developed mucosal lesions indistinguishable from those of sick volunteers 48 h after ingestion of virus. The remaining two of four asymptomatic volunteers had normal mucosal epithelia (325). Dolin et al. (86) also described jejunal biopsies from Hawaii agent-infected volunteers and found that they were similar, but not identical, to those seen in Norwalk virus infection. Light microscopy of biopsy specimens showed an intact mucosa, moderately blunted villi, and a moderate inflammatory cell infiltrate in the lamina propria, consisting of both polymorphonuclear and mononuclear cells. EM showed distorted and markedly shortened microvilli on intact epithelial cells, enlarged pale mitochondria containing indistinct membranes, and moderate widening of intercellular spaces filled with an amorphous electron-dense material. The biopsies of the asymptomatic volunteers remained normal. By 2 to 3 weeks postinfection, the lesions had disappeared. Laboratory Diagnosis The laboratory diagnosis of Norwalk and similar agents has been hindered by the inability to propagate these viruses in vitro. Reagents are available in research laboratories, but are not available for routine diagnostic laboratories. In addition, routine laboratory diagnosis of these agents is not carried out since patients do not usually require hospitalization, where testing is usually done. Cell and organ culture. Attempts at propagating the Norwalk agent in a number of types of cell cultures has failed. However, its replication has been attempted in organ cultures of human fetal intestine (22, 84). Full-thickness explants approximately 2 mm2 in size can be maintained in the laboratory for up to 3 weeks. They have differentiated cells CLIN. MICROBIOL. REV. and tissues, including villi, that maintain their morphologic integrity and their in vivo relationship to each other. Inoculation of Norwalk-containing stool suspensions into these organ cultures did not cause any morphologic alteration in the cultures that were seen by light or dissecting microscopy. However, fluids from the organ cultures ingested by volunteers did cause illness in some volunteers. Whether or not the virus actually replicated, or whether virus from the original inoculum was responsible for the illness, was not known. In vitro assays. The first assay used to detect Norwalk and Norwalk-like viruses, and to semiquantitate antibodies to them, was IEM. By this method, virus could be detected in stool suspensions by mixing a small amount of the suspension with sera from patients recently recovering from Norwalk infection. After placing the mixture on a grid and negatively staining with phosphotungstic acid, clumped Norwalk particles could be visualized. Similarly, known Norwalk-containing suspensions could be used to detect antibodies in sera of patients recovering from gastroenteritis. This technique was useful in initially detecting these viruses and studying the development of immunity. However, for extensive epidemiologic and other studies, in which large numbers of samples are examined, it was somewhat slow and cumbersome. Compared with tests developed later, IEM also lacked sensitivity. Since the Norwalk and Norwalk-like viruses cannot be propagated in vitro, investigators developing in vitro assays have had to rely on patient stool specimens for sources of virus and their convalescent sera for sources of antibody. Virus from stool samples cannot be purified in sufficient quantity to produce hyperimmune animal sera or monoclonal antibodies. Because of the lack of reagents, assays have been developed only in a few research laboratories, and no reagents are available commercially. The few assays that have been developed are the RIA, the EIA, and the biotinavidin-eia. Greenberg et al. (138) developed two similar solid-phase RIAs for detecting either Norwalk viral antigen or antibody. In both assays, patient convalescent serum was used as a capture antibody and acute-phase serum was used as a control antibody to coat microtiter plate wells. To use the RIA to detect viral antigen, suspect stool suspensions were added to both acute- and convalescent-phase serum-coated wells. This was followed by the addition of detector antibody consisting of IgG purified from convalescent-phase serum that was labeled with To detect the presence of antibody in patient sera, Greenberg et al. (138) used a blocking RIA. In this test a stool suspension known to contain Norwalk virus was added to the capture-antibody-coated wells. The patient sera to be tested for Norwalk antibody were then added. Lastly, a 125I-labeled detector antibody was added. A similar RIA for Norwalk virus has been described by Blacklow et al. (21). Dolin et al. (88) have described a similar RIA for the SM agent. These RIAs were shown to be highly specific and sensitive. However, other problems have been associated with them. The primary problem has been the rapid loss of reactivity of the 125I-labeled detector antibody, which is good for <1 week after labeling. In addition, there are the usual costs, hazards, and disposal problems associated with the use of radioactive substances. Because of the inherent problems with the RIA, an EIA for Norwalk antigen detection has been developed (149, 150). This test is similar to the RIAs. Again, the reagents were human clinical materials and differed only in that the

25 VOL. 2, 1989 IgG fraction from convalescent serum used as detector antibody was labeled with horseradish peroxidase rather than 125I. The major advantage of the EIA is the stability of the detector reagent. In addition, the EIA's ability to detect viral antigen appeared to be slightly more sensitive than that of the RIA. The test was quite sensitive in that there was no cross-reactivity between the two antigenically distinct Norwalk and Hawaii agents. Two biotin-avidin ElAs have been described that detect Norwalk or SM antigen or antibodies (120, 234). These two agents have been shown to be antigenically distinct. These EIAs are similar to the RIAs and EIA described except that the IgG detector antibody is labeled with biotin, rather than peroxidase or In the final steps of the tests, avidinperoxidase, an enzyme substrate, H202, and a chromogen are added as an indicator. These EIAs were as sensitive or more sensitive than comparable RIAs from these laboratories for the detection of antigen. All of these tests have been used in the epidemiologic and immunologic studies reported in this treatise. Immunology Since individuals have repeated episodes of gastroenteritis, it was of interest to determine the length of time one remains immune after infection with the Norwalk agent or similar viruses. It was also of interest to determine one's cross-immunity to Norwalk-like viruses and to determine the antigenic relationships among the various Norwalk-like viruses. First, to study homologous immunity to Norwalk virus, Blacklow et al. (22) reported infecting volunteers with Norwalk agent and rechallenging these volunteers 6 to 14 weeks later. None became ill, indicating that at least shortterm homologous immunity occurred after Norwalk virus infection. Similarly, Wyatt et al. (401) found immunity in volunteers when they were rechallenged 9 to 14 weeks after experimental infection with Norwalk agent. In addition, when volunteers were infected with the Hawaii agent and challenged 6 to 7 weeks later with the same agent, none became ill. When volunteers were infected with the Montgomery County agent and challenged 7 weeks later with the homologous virus, none became ill. This indicated at least short-term homologous immunity for these three viruses. Heterologous challenges by Wyatt et al. (401) were also carried out. Three of six volunteers given Norwalk agent and challenged 7 to 15 weeks later with Hawaii agent became ill, indicating that the Norwalk and Hawaii agents were antigenically dissimilar. When six volunteers given Norwalk virus were challenged 6 to 13 weeks later with the Montgomery County agent, none became ill, but the reverse was not true. One of three volunteers given Montgomery County agent and challenged 11 to 12 weeks later with Norwalk agent became ill. Thus, Norwalk infection appeared to protect against Montgomery County agent, although Montgomery County infection may not protect against Norwalk infection. However, when four volunteers given Hawaii agent were challenged 8 weeks later with Montgomery County agent, none became ill, indicating an antigenic relationship between these two viruses. Parrino et al. (281) orally inoculated 12 volunteers with Norwalk virus and found that six became ill with vomiting or diarrhea or both. Sera were collected from five of the volunteers. By IEM, four of five had initial low-level antibody which increased after challenge and then declined in the interim time between the first inoculation and a second, reinoculation 27 to 42 months later. When reinoculated with VIRAL GASTROENTERITIS 75 Norwalk virus 27 to 42 months later, the same six people became sick, indicating that immunity to Norwalk agent was relatively short-lived. Again, antibody levels increased after challenged in the same four individuals. A fifth individual who became sick each time he was challenged had persistently high antibody levels before and after each challenge, indicating that serum antibody alone does not protect against infection with Norwalk virus, at least in some individuals. When the volunteers were challenged a third time 4 to 8 weeks after the second challenge, all except one remained well, indicating that at least some immunity occurs for a short period of time. The six remaining volunteers did not become sick the first time they were inoculated with Norwalk virus. When rechallenged 31 to 34 months later, they again resisted infection. Serum samples obtained from three of these volunteers had low levels of antibodies detected by IEM which remained low before and after both challenges. From these studies of Parrino et al. (281), there appeared to be two forms of Norwalk immunity. The first group was susceptible to virus infection and developed only short-term immunity. The second group maintained long-term immunity, although they showed only low-level antibody to Norwalk or possible cross-reactivity to Norwalk due to antibodies to a related virus. However, protective mechanisms other than serum antibody appeared to protect them against disease. It is possible that individuals in the second group lacked a receptor for Norwalk virus on their intestinal mucosal cells that are usually infected by this virus. In a similar study, Blacklow et al. (21) found that the majority of volunteers given Norwalk virus who became sick had preexisting Norwalk antibody, whereas the majority of volunteers who did not become ill did not have preexisting Norwalk antibody or had very low levels of antibody. In addition, all of those volunteers who had preexisting antibody in their duodenal fluids developed illness, whereas only one-half of those with no duodenal antibody became sick. Thus, clinical immunity to Norwalk virus appears to be complex. Some individuals appear to be susceptible to repeated infections with Norwalk virus, whereas other individuals appeared to be resistant to Norwalk infection (21). From immunologic studies, the possible relationship of Norwalk virus to calicivirus is unclear. Madore et al. (234), using a biotin-avidin-eia, did not detect any cross-reaction between feline caliciviruses and sera from Norwalk or SM virus-infected individuals. However, Nowak et al. (N. A. Nowak, W. D. Cubitt, J. E. Herrmann, and N. E. Blacklow, Abstr. Annu. Meet. Am. Soc. Microbiol. 1986, C73, p. 340) studied 18 paired serum samples from four human calicivirus outbreaks, using a Norwalk RIA. From two calicivirus type UK4 outbreaks, 7 of 10 paired sera showed a significant antibody rise to Norwalk virus. The Norwalk antibody titers were similar to those seen in sera from outbreaks of Norwalk disease. Thus, the latter data provide evidence for at least a one-way serologic cross-relatedness between Norwalk virus and calicivirus (88). Prevention and Treatment Since Norwalk and Norwalk-like viruses cannot be grown in vitro, no vaccines are available. In addition, vaccines may not be necessary since infections usually occur sporadically or in isolated outbreaks and clinical illness is short-lived and relatively mild in older children and adults who become infected. There are a variety of measures to prevent outbreaks and sporadic infection due to these agents. These include good personal hygiene to prevent primary and sec-

26 76 CHRISTENSEN ondary person-to-person contact, good personal hygiene among food handlers, not allowing food handlers with gastroenteritis to work, and not ingesting raw or undercooked shellfish. In addition, outbreaks can be prevented by proper sewage and water treatment, maintenance of proper water supplies, and avoidance of contamination of water supplies with sewage systems, particularly in rural or underdeveloped areas. Treatment includes maintaining electrolyte and fluid levels in patients. CALICIVIRUSES Caliciviruses were known to infect mammals before they were discovered in humans. They were first described in humans by Madeley and Cosgrove (231) and Flewett and Davies (112), who found them by EM in the stools from babies with diarrhea. Caliciviruses were originally considered to be in the family Picornaviridae but have now been placed in the new family Caliciviridae. They are RNA viruses about 30 nm in diameter. Caliciviruses derive their name from the 32 cup-shaped depressions on the surface of their virions. This gives their periphery a spiky appearance and causes the formation of a Star of David configuration at certain rotations of the virion (Fig. 1C). The Star of David configuration is formed by six peripheral hollows surrounding a seventh central hollow (227, 331). The buoyant density of human calicivirus in CsCl ranges from 1.38 to 1.40 glcm3, with a peak at 1.39 g/cm3 (346). Caliciviruses primarily cause gastroenteritis in infants and young children. They can cause gastroenteritis in the general pediatric population (67, 267, 346) as well as outbreaks in institutions including schools (248), orphanages (53, 54), a mother-baby psychiatric unit (69), and nosocomially in pediatric wards in hospitals (67, 341). Sporadic cases have occurred in the United Kingdom, Norway, Australia, Canada, United States, India, and Bulgaria (68). Most individuals are infected with caliciviruses at a very young age. In Madeley and Cosgrove's first report, all but two children were 2 months old or younger. In other reports, patients were 2 to 20 and 1 to 27 months old in orphanage outbreaks (53, 54); 1 to 18 months old in a mother-baby psychiatric unit (69); 6 weeks to 13 years, with a peak at 1 to 6 months old in a general hospitalized population (67); and 4 to 6 years old in a school outbreak (248). However, in one outbreak occurring in a residential home for the elderly, 18 affected residents were 77 to 102 years old, and the 32 affected staff members were 16 to 68 years old (70). Calicivirus infections were found to occur throughout the year, with peaks in the winter months, according to one 30-month study (67). Sporadic as well as more epidemic infection can occur (67). The symptoms of calicivirus infection are about as severe as those seen with rotavirus infection and may be indistinguishable from them (67). In two studies, diarrhea was reported as the predominant feature (341, 346). In one study, infants had either diarrhea only or diarrhea plus vomiting (53). In one 30-month study of hospitalized pediatric patients (67), diarrhea was the most common symptom in 89%, 50% had vomiting in addition to diarrhea (usually in the 4 to 6 year olds), one-third had upper respiratory tract symptoms, and about 20% had fever. One patient had projectile vomiting (67). In 4 to 6 year olds in a school, most patients had nausea and vomiting, although some also had diarrhea (248). In initially well babies in the mother-baby psychiatric ward, CLIN. MICROBIOL. REV. some had mild vomiting and diarrhea lasting only 1 day, some had severe vomiting and diarrhea lasting 7 to 9 days, and one had very severe projectile vomiting (69). Two orphanage outbreaks lasted 4 to 6 days, and viral shedding was correlated with the days of illness, which rarely occurred for more than 10 days (54). In the 30-month study (67), the illness lasted an average of 4 days, with a range of 2 to 11 days. In the initial report of Flewett and Davies (112), calicivirus was found in the small bowel obtained at the necropsy of a 22-month-old child who died of gastroenteritis, but the investigators were not sure of its significance. Caliciviruses have also been found in the stools of well individuals (53, 231), and in the outbreak in babies in the mother-baby psychiatric unit, none of the mothers or staff developed symptoms (69). However, in the elderly in the residential home (70), 100% of the sick residents and staff had diarrhea and 50% of both groups had vomiting. A few of the staff members also reported headaches, general aches, and malaise, although the elderly residents did not. In the 30-month study of patients hospitalized with gastroenteritis in the United Kingdom, 6.6% was due to calicivirus (67), while in another 3-year study in Japan, 1.2% was due to calicivirus (346). Thus, caliciviruses are responsible for only a minority of viral gastroenteritis among hospitalized patients, although they have worldwide prevalence (see below). In several studies there was good correlation between enteric disease and the presence of caliciviruses in stool seen by direct EM (53, 54, 69, 70). In addition, paired acute and convalescent serum samples were obtained from the patients, and rises in antibody titers to caliciviruses were seen by 1EM (53, 69, 70, 265). Caliciviruses in patient stools were used as the source of virus for IEM. No one has been able to isolate the human caliciviruses in routine cell cultures (53, 70, 346), although other mammalian caliciviruses such as feline calicivirus can readily be propogated in cells of their species (231). In addition to EM for the detection of caliciviruses in stools, a solid-phase RIA has been developed by Nakata et al. (264). They prepared guinea pig hyperimmune serum for the test by using as an antigen calicivirus particles purified from patient stools. They found the RIA to be much more sensitive than direct EM for detecting caliciviruses in stools. With a blocking RIA, their test was shown to be specific for calicivirus. This RIA is also advantageous in that caliciviruses are smaller and present in lower numbers in stools than rotaviruses and thus are harder to detect by EM. In another report, Nakata et al. (266) described a similar EIA that was as sensitive as their RIA. To determine the prevalence of antibody to calicivirus in the general population in Japan, Sakuma et al. (312) used IEM to detect antibody in 83 serum specimens. The agerelated prevalence of antibody was 23% in 0 to 5 month olds, 30% in 6 to 23 months olds, 65% in 2 to 5 year olds, and 90% in school children and adults. Similar results were found by Nakata et al. (266) in sera from healthy children in the United States. All (100%) infants 0 to 3 months old had calicivirus antibody, whereas the percentage dropped to 25% for 4 to 11 month olds; by the ages of 4 to 6 years, however, 100% of children had developed antibody. Thus, calicivirus appears to infect the population early in life. Nakata et al. (263) used their blocking RIA to determine antibody prevalence in 390 sera from adults in Japan, Indonesia, Singapore, and Papua New Guinea. The positivity rates were 87, 88, 70, and 94%, respectively. The results indicate that the virus is a common infectious agent in Japan and Southeast Asia.

27 VOL. 2, 1989 A study of the worldwide prevalence of antibody to calicivirus strain UK1 was carried out by Cubitt and Mc- Swiggen (68). High antibody titers were seen in all batches of pooled immune globulin from the United Kingdom, Belgium, Switzerland, Canada, United States, and Japan. The percentage of antibodies in this calicivirus study increased with age in patients from the United Kingdom and Japan and in Saudi Arabian Bedouins. All older children and adolescents in Tanzania had antibodies; sera from younger children were not available. The only major difference in antibody prevalence was seen in Australian aborigines. Here there was little evidence of infection before the age of 13 years, although most adults had antibody to calicivirus. Thus, the virus causes infection worldwide, occurring first in young children and infants. At least two strains of calicivirus have been recovered from children. However, the agent responsible for the outbreak in the home for the elderly was antigenically distinct from the two strains isolated from children (70). To date, vaccines cannot be developed to control calicivirus infection, since the virus cannot be propagated in vitro. Treatment consists primarily of maintaining the proper electrolyte balance and fluid balance, as in rotavirus gastroenteritis. This is important since calicivirus infection is probably as serious as rotavirus gastroenteritis. ASTROVIRUSES Human astroviruses were first discovered by EM in the stools of infants with diarrhea by Madeley and Cosgrove (229, 230). Shortly thereafter, astroviruses were found to infect various species of mammals, including cattle, sheep, and pigs (209, 400), causing diarrhea, especially in the young of these species. Astroviruses have been isolated in the United Kingdom (7, 210, 229, 230) Norway (199), Federal Republic of Germany (226), Japan (204), and North America. They are responsible for about 5% of infantile gastroenteritis, including nosocomial infections. The Marin County agent, discovered in elderly patients with diarrhea in a convalescent hospital in California, is also an astrovirus (148, 278). Their appearance has been described in detail by Madeley (227). They are 28- to 30-nm, round virus particles with a smooth edge, as opposed to caliciviruses which have a rough edge. Most of the virus particles have a star-shaped configuration on their surface (Fig. 1D). This star configuration can have either five or six points, whereas caliciviruses have only six-pointed stars. The point of the astrovirus star radiates our from a central point or knob and has no hollow center as does the calicivirus. The astrovirus buoyant density in cesium chloride has been reported to be 1.39 to 1.40 (204) and 1.33 to 1.34 (226) g/cm3. There appear to be at least five distinct serotypes of human astroviruses (199, 209, 214). These have been distinguished by an immunofluorescence test and by immunosorbent EM. In the former test, acetone-fixed astrovirus-infected LLC-MK2 or HEK cells were used. In the latter test, astrovirus antisera against the various strains were raised in rabbits and used to coat Formvar-coated EM grids. The astrovirus suspensions were added to these grids. The grids were examined and the number of virus particles was counted. The only cross-reaction seen was with serotypes 1 and 3. There were no cross-reactions of any of the five human strains with bovine or ovine astroviruses. From strains collected over a 10-year period in the United Kingdom, serotype 1 appears to be the most prevalent, account- VIRAL GASTROENTERITIS 77 ing for 77% of the astroviruses. The remaining types 2 to 5 had a similar incidence, ranging from 5 to 7%. In determining the incidence of astrovirus infection in Japan, Konno et al. (204) found that 50% of young adults tested by IEM had antibodies to astroviruses. They also found astrovirus antibodies in the commercial gamma globulin preparation they tested. By IEM, they found no crossreaction with caliciviruses, Norwalk virus, polioviruses, or a coxsackievirus B. Children, from infancy to 5 to 7 years of age, are most likely to develop symptomatic disease with astroviruses (7, 204, 210), although exposed adults can also develop mild disease, but with less frequency (204, 210, 278). In one study (210), older children were also found to be less likely to develop disease than younger children. Asymptomatic infection can also occur among infants and young children (7, 229). The incubation period appears to be 24 to 36 hours, similar to those for other gastroenteritis viruses (204). Symptoms can last from 12 h to 3 to 4 days (7, 204, 210). In one study, all symptomatic children had watery diarrhea, while only one-third had vomiting (210). However, in another study, some children had vomiting only, while the remainder suffered from both vomiting and diarrhea (7). In a third study, the rates of symptoms were as follows: vomiting, 74%; diarrhea, 30%; abdominal pain, 49%; and fever, 30% (204). Clinical illness appeared to be less severe than with rotavirus (210). Astroviruses can be detected in stool suspensions by direct EM (7, 210, 229, 230). Konno et al. (204) found that the star configuration was much more distinct if the particles were stained with phosphotungstic acid rather than with uranyl acetate. Astroviruses, including the Marin County agent (278), can be isolated and propagated in HEK cells to which trypsincontaining medium is added (148, 213, 214). Astrovirusinfected HEK cells can be used for immunofluorescence tests and EIAs for serologic studies (148, 209, 214). Antibodies can be detected by IEM. Antibody rises can be observed in paired acute and convalescent sera obtained from patients (204, 210). In addition, astrovirus-specific IgM can be detected by IEM after fractionation of patient sera by sucrose density gradient centrifugation (7). There is no specific treatment for astrovirus infection, similar to other viral gastroenteritides. Nonspecific treatment consists of maintaining fluids and electrolyte balance, although astrovirus infection appears to be less severe than rotavirus infection (210). CORONAVIRUSES Coronaviruses are enveloped, medium-sized, positivestranded RNA viruses that were first discovered in patients with upper respiratory tract illness. These viruses have also been associated with diarrheal diseases in mammals, especially newborns. They may be involved in human gastrointestinal disease (296), since by direct EM coronavirus-like particles (CVLP) have been seen in stools of patients with acute nonbacterial gastroenteritis and necrotizing enterocolitis. Coronaviruses are approximately 80 to 150 nm in diameter. They are seen as rounded or pleomorphic particles that have club-shaped projections on their surface, called peplomers (296). These projections on the rounded particles give the appearance of the corona of the sun, hence, the virions' name.

28 78 CHRISTENSEN Mortensen et al. (260) found that CVLP in stools in infants with diarrhea were morphologically distinct from respiratory coronaviruses in that the fecal CVLP had a flexible-appearing fringe whereas the respiratory coronaviruses had a more rigid-appearing fringe. The CVLP in stools also had more closely spaced peplomers. Two respiratory coronaviruses, OC43 and 229E, are responsible for common colds and have been well characterized. CVLP from stools have not been well characterized, since most investigators have not been able to cultivate them. However, Gerna et al. (11, 122, 123) purified CVLP directly from the stools of two patients with gastroenteritis, using sucrose density gradient ultracentrifugation. This group also used the purified particles to study biophysical characteristics. The two strains of purified human enteric coronavirus (HEC) of Gerna et al., HECV-24 and HECV-35, had a buoyant density of 1.20 g/cm3, a value in agreement with those reported for human and animal coronaviruses. In several epidemiologic studies, CVLP have been observed in the stools of normal individuals as well as in diarrheal or necrotizing enterocolitis stools (334), so that the role of CVLP in gastrointestinal disease has been questioned. However, in some studies, a positive correlation has been made between the incidence of fecal CVLP and gastrointestinal illness. In a prospective study, Gerna et al. (122, 123) found that, by EM, 16% of 208 gastroenteritis patients had CVLP in their stools compared with only 2% of 182 controls. Similarly, Mortensen et al. (248) found that 39% (49 of 126) of diarrheal stools were positive for CVLP by EM. In addition to the correlation or lack of correlation of CVLP with disease, some controversy or questions have arisen over the nature of the CVLPs. This is due in part to the inability of many investigators to propagate CVLP in vitro. In addition, in some instances the CVLP seen may not have been true virus particles, but rather, artifactual (283, 298, 322). In some instances, the CVLP may be normal intestinal epithelial cell components, consisting of R-bodies (rodcontaining bodies) and C-bodies (coccoid or glycocalyceal bodies), which resemble CVLP (283). The possibility has also been considered that these CVLP may be mycoplasmas. However, efforts to propagate them on mycoplasma media have failed (260). Viruses from human diarrheal stools have also been observed (by EM) that resemble coronaviruses, but are morphologically more similar to the newly described toroviruses (122). The human toro-like viruses are antigenically similar to the Breda virus. The toroviruses include Breda virus isolated from calves with diarrhea and Berne virus, which was isolated from a horse. The toroviruses are morphologically similar to coronaviruses in that they are medium-sized viruses with a peplomer-bearing envelope. However, their polypeptides are of different molecular weights than those of coronaviruses, and they appear to be negative-stranded RNA viruses, whereas the coronaviruses are positive stranded. Thus, in humans, some CVLP seen by EM may actually be members of the torovirus group (122). In summary, some investigators believe that the CVLP are a heterogeneous group of particles, some of which are true coronaviruses. Enteric coronavirus infections, or infections by CVLP, tend to occur in the cooler, drier months of the years. In southern Arizona, where a large percentage of gastroenteritis appears to be caused by coronavirus, most of the cases occur in the fall and early winter (260). Most cases of gastroenteritis, as well as necrotizing enterocolitis, occur in infants. Mortensen et al. (260) found that 49 patients with CLIN. MICROBIOL. REV. CVLP-associated gastroenteritis ranged in age from 1 month to 12 years. Of these 49 patients, 88%o were <2 years old and 71% were <1 year old. The median duration of the illness was 7 days. Diarrhea occurred in 94%; vomiting, in 51%; fever, in 63%; and occult blood, in the stools of 18%. Eighteen percent had at least one other identifiable enteric pathogen, including Salmonella, Shigella, and Campylobacter spp. and one enterovirus. Mortensen et al. (260) suggested that nursing might have a protective effect on preventing illness due to coronavirus-like agents, since weaned infants seemed to be at risk for severe CVLP-associated disease. Serologic studies have been carried out for various purposes: (i) for epidemiologic purposes, (ii) to study the immune response to enteric coronavirus infection, and (iii) to study possible cross-reactivity of fecal coronaviruses with known respiratory coronaviruses. In a retrospective study, Gerna et al. (122, 123) found that 27% of sera from 62 patients with gastroenteritis reacted with coronavirus OC43 by neutralization and HAI tests, whereas only 2% of control sera reacted. They also immunized guinea pigs and mice with the CVLP purified from the stools of two gastroenteritis patients by sucrose density gradient ultracentrifugation. In an IEM test, they used the animal sera to the two patients' strains, animal sera to OC43, and one patient's convalescent serum. In this test, they found a two-way cross between OC43, and their two human enteric coronavirus strains. Thus, their fecal isolates appeared to share a common antigen(s) with respiratory coronaviruses. However, Mortensen et al. (260) found that their CVLP in stools did not cross-react with respiratory coronavirus OC43 or 229E by immunodiffusion or by IEM. Thus, there may be different antigenic forms of CVLP. The laboratory diagnosis of enteric coronavirus infection or infection with CVLP is carried out primarily by direct EM, since other laboratory tests are generally not available. Most investigators have not been able to propagate enteric coronaviruses, and no one has been able to propagate them in routine cell types. Mortensen et al. (260) could not propagate CVLP from stools in (i) a number of routine cell types, (ii) cell lines of human embryonic intestine or human rectal tumor, or (iii) four strains of human embyronic tonsil fibroblasts. In some of the special cell types, five blind passes were made, with negative results: Coronavirus vaccines have not been developed for several reasons. One, the role of enteric coronavirus in disease causation has not been firmly established in different geographic areas of the world. Two, where coronavirus infection may play a role, it is usually of minor importance, except in the southwest United States. Three, enteric coronaviruses and CVLP are difficult to propagate in vitro. Four, long-term immunity to coronavirus infection has not been studied. Only short-term studies on convalescent sera have been carried out. Treatment consists of maintaining fluid and electrolyte balance in infected infants, similar to that used for rotavirus and other serious viral gastroenteritis, when dehydration and electrolyte imbalance can occur. ACKNOWLEDGMENTS I thank Cynthia Howard and Joseph Zientarski for providing the excellent electron photomicrographs and Donna Tieken for excellent editorial assistance. LITERATURE CITED 1. Agliano, A. M., A. Rossi, and A. Sanna Isolation from faecal specimens of new strains of human rotavirus primarily

29 VOL. 2, 1989 cytopathic for stationary cell culture without trypsin. Arch. Virol. 84: Agus, S. G., R. Dolin, R. G. Wyatt, A. J. Tousimis, and R. S. Northrup Acute infectious nonbacterial gastroenteritis: intestinal histopathology. Histologic and enzymatic alterations during illness produced by the Norwalk agent in man. Ann. Intern. Med. 79: Albert, M. J., Y. Soenarto, and R. F. Bishop Epidemiology of rotavirus diarrhea in Yogyakarta, Indonesia, as revealed by electrophoresis of genome RNA. J. Clin. Microbiol. 16: Albert, M. J., L. E. Unicomb, and R. F. Bishop Cultivation and characterization of human rotaviruses with "super short" RNA patterns. J. Clin. Microbiol. 25: Anderson, E. L., R. B. Belshe, J. Bartram, F. Crookshanks- Newman, R. M. Chanock, and A. Z. Kapikian Evaluation of rhesus rotavirus vaccine (MMU 18006) in infants and young children. J. Infect. Dis. 153: Appleton, H., M. Buckley, B. T. Thom, J. L. Cotton, and S. Henderson Virus-like particles in winter vomiting disease. Lancet i: Ashley, C. R., E. 0. Caul, and W. K. Paver Astrovirusassociated gastroenteritis in children. J. Clin. Pathol. 31: Barnes, G. L., L. W. Doyle, P. H. Hewson, A. M. L. Knoches, J. A. McLellan, W. H. Kitchen, and R. F. Bishop A randomised trial of oral gammaglobulin in low-birth-weight infants infected with rotavirus. Lancet i: Barron-Romero, B. L., J. Barreda-Gonzalez, R. Doval-Ugalde, J. Zermeno-Equia Liz, and M. Huerta-Pena Asymptomatic rotavirus infection in day care centers. J. Clin. Microbiol. 22: Bastardo, J. W., J. L. McKimm-Breschkin, S. Sonza, L. D. Mercer, and I. H. Holmes Preparation and characterization of antisera to electrophoretically purified SA11 virus polypeptides. Infect. Immun. 34: Battaglia, M., N. Passarani, A. DiMatteo, and G. Gerna Human enteric coronaviruses: further characterization and immunoblotting of viral proteins. J. Infect. Dis. 155: Beards, G. M., A. D. Campbell, N. R. Cottrell, J. S. M. Peiris, N. Rees, R. C. Sanders, J. A. Shirley, H. C. Wood, and T. H. Flewett Enzyme-linked immunosorbent assays based on polyclonal and monoclonal antibodies for rotavirus detection. J. Clin. Microbiol. 19: Bellamy, K., P. S. Gardner, M. H. Hambling, S. Rice, and A. F. Bradburne Enzyme-linked immunosorbent assay for the detection of human rotavirus in stools. J. Virol. Methods 7: Birch, C. J., N. I. Lehmann, A. J. Hawker, J. A. Marshall, and I. D. Gust Comparison of electron microscopy, enzymelinked immunosorbent assay, solid-phase radioimmunoassay, and indirect immunofluorescence for detection of human rotavirus antigen in faeces. J. Clin. Pathol. 32: Birch, C. J., F. A. Lewis, M. L. Kennett, M. Homola, H. Pritchard; and I. D. Gust A study of the prevalence of rotavirus infection in children with gastroenteritis admitted to an infectious disease hospital. J. Med. Virol. 1: Birch, C. J., S. M. Rodger, J. A. Marshall, and 1. D. Gust Replication of human rotavirus in cell culture. J. Med. Virol. 11: Bishop, R. F., G. L. Barnes, E. Cipriani, and J. S. Lund Clinical immunity after neonatal rotavirus infection. A prospective longitudinal study in young children. N. Engl. J. Med. 309: Bishop, R. F., G. P. Davidson, I. H. Holmes, and B. J. Ruck Virus particles in epithelial cells of duodenal mucosa from children with acute non-bacterial gastroenteritis. Lancet ii: Bishop, R. F., G. P. Davidson, I. H. Holmes, and B. J. Ruck Detection of a new virus by electron microscopy of faecal extracts from children with acute gastroenteritis. Lancet i: Black, R. E., M. H. Merson, A. S. M. M. Rahman, M. Yunis, VIRAL GASTROENTERITIS 79 A. R. M. A. Alim, I. Huq, R. H. Yolken, and G. T. Curlin A two-year study of bacterial, viral and parasitic agents associated with diarrhea in rural Bangladesh. J. Infect. Dis. 142: Blacklow, N. R., G. Cukor, M. K. Bedigian, P. Echeverria, H. B. Greenberg, D. S. Schreiber, and J. S. Trier Immune response and prevalence of antibody to Norwalk enteritis virus as determined by radioimmunoassay. J. Clin. Microbiol. 10: Blacklow, N. R., R. Dolin, D. S. Fedson, H. DuPont, R. S. Northrup, R. B. Hornick, and R. M. Chanock Acute infectious nonbacterial gastroenteritis: etiology and pathogenesis. Ann. Intern. Med. 76: Blacklow, N. R., P. Echeverria, and D. H. Smith Serologic studies with reovirus-like enteritis agent. Infect. Immun. 13: Bohl, E. H., L. J. Saif, K. W. Theil, A. G. Agnes, and R. F. Cross Porcine pararotaviruses: detection, differentiation from rotavirus, and pathogenesis in gnotobiotic pigs. J. Clin. Microbiol. 51: Both, G. W., L. J. Seigman, A. R. Bellamy, N. Ikegami, A. J. Shatkin, and Y. Furuichi Comparative sequence analysis of rotavirus genomic segment 6, the gene specifying viral subgroups 1 and 2. J. Virol. 51: Bradburne, A. F., J. D. Almeida, P. S. Gardner, R. B. Moosai, A. A. Nash, and R. R. A. Coombs A solid-phase system (SPACE) for the detection and quantification of rotavirus in faeces. J. Gen. Virol. 44: Brandt, C. D., C. W. Arndt, G. L. Evans, H. W. Kim, E. P. Staalings, W. J. Rodriguez, and R. H. Parrott Evaluation of a latex test for rotavirus detection. J. Clin. Microbiol. 25: Brandt, C. D., H. W. Kim, W. J. Rodriguez, J. 0. Arrobio, B. C. Jeffries, and R. H. Parrott Rotavirus gastroenteritis and weather. J. Clin. Microbiol. 16: Brandt, C. D., H. W. Kim, W. J. Rodriguez, J. 0. Arrobio, B. C. Jeffries, and R. H. Parrott Simultaneous infections with different enteric and respiratory tract viruses. J. Clin. Microbiol. 23: Brandt, C. D., H. W. Kim, W. J. Rodriguez, J. 0. Arrobio, B. C. Jeffries, E. P. Stallings, C. Lewis, A. J. Miles, R. M. Chanock, A. Z. Kapikian, and R. H. Parrott Pediatric viral gastroenteritis during eight years of study. J. Clin. Microbiol. 18: Brandt, C. D., H. W. Kim, W. J. Rodriguez, J. 0. Arrobio, B. C. Jeffries, E. P. Stallings, C. Lewis, A. J. Miles, M. K. Gardner, and R. H. Parrott Adenovirus and pediatric gastroenteritis. J. Infect. Dis. 151: Brandt, C. D., H. W. Kim, W. J. Rodriguez, L. Thomas, R. H. Yolken, J. 0. Arrobio, A. Z. Kapikian, R. H. Parrott, and R. M. Chanock Comparison of direct electron microscopy, immune electron microscopy, and rotavirus enzymelinked immunosorbent assay for detection of gastroenteritis virus in children. J. Clin. Microbiol. 13: Brandt, C. D., H. W. Kim, R. H. Yolken, A. Z. Kapikian, J. 0. Arrobio, W. J. Rodriguez, R. G. Wyatt, R. M. Chanock, and R. H. Parrott Comparative epidemiology of two rotavirus serotypes and other viral agents associated with pediatric gastroenteritis. Am. J. Epidemiol. 110: Bridger, J. C., S. Pedley, and M. A. McCrae Group C rotaviruses in humans. J. Clin. Microbiol. 23: Brown, M Selection of nonfastidious adenovirus species in 293 cells inoculated with stool specimens containing adenovirus 40. J. Clin. Microbiol. 22: Brown, M., M. Petric, and P. J. Middleton Diagnosis of fastidious enteric adenoviruses 40 and 41 in stool specimens. J. Clin. Microbiol. 20: Brown, S. E., III, K. T. Sauer, M. Nations-Shields, D. S. Shields, J. G. Araujo, and R. L. Guerrant Comparison of paired whole milk and dried filter paper samples for antienterotoxin and anti-rotavirus activities. J. Clin. Microbiol. 16: Brussow, H., H. Hilpert, I. Walther, J. Sidoti, C. Mietens, and

30 80 CHRISTENSEN P. Bachman Bovine milk immunoglobulins for passive immunity to infantile rotavirus gastroenteritis. J. Clin. Microbiol. 25: Bryden, A. S., H. A. Davies, R. E. Hadley, T. H. Flewett, C. A. Morris, and P. Oliver Rotavirus enteritis in the West Midlands during Lancet ii: Cameron, D. J. S., R. F. Bishop, A. A. Veenstra, and G. L. Barnes Noncultivable viruses and neonatal diarrhea: fifteen-month survey in a newborn special care nursery. J. Clin. Microbiol. 8: Carlson, J. A. K., P. J. Middleton, M. T. Szymanski, J. Huber, and M. Petric Fatal rotavirus gastroenteritis. An analysis of 21 cases. Am J. Dis. Child. 132: Caul, E. O., and H. Appleton The electron microscopal and physical characteristics of small round human fecal viruses: an interim scheme for classification. J. Med. Virol. 9: Cevenini, R., F. Rumpianesi, R. Mazzaracchio, M. Donati, E. Falcieri, and R. Lazzari Evaluation of a new latex agglutination test for detecting human rotavirus in faeces. J. Infect. 7: Cevenini, R., F. Rumpianesi, R. Mazzaracchio, M. Donati, E. Falcieri, and I. Sarov A simple immunoperoxidase method for detecting enteric adenovirus and rotavirus in cell culture. J. Infect. 8: Champsaur, H., M. Henry-Amar, D. Goldszmidt, J. Prevot, M. Bourjouane, E. Questiaux, and C. Bach Rotavirus carriage, asymptomatic infection, and disease in the first two years of life. II. Serological response. J. Infect. Dis. 149: Champsaur, H., E. Questiaux, J. Prevot, M. Henry-Amar, D. Goldszmidt, M. Bourjouane, and C. Bach Rotavirus carriage, asymptomatic infection, and disease in the first two years of life. I. Virus shedding. J. Infect. Dis. 149: Chanock, S. J., E. A. Wenske, and B. N. Fields Human rotaviruses and genome RNA. J. Infect. Dis. 148: Chen, G.-M., T. Hung, J. C. Bridger, and M. A. McCrae Chinese adult rotavirus is a Group B rotavirus. Lancet ii: Chernesky, M., S. Castriciano, J. Mahony, and D. DeLong Examination of the Rotazyme II enzyme immunoassay for the diagnosis of rotavirus gastroenteritis. J. Clin. Microbiol. 22: Cheung, E. Y., S. I. Hnatko, H. Gunning, and J. Wilson Comparison of Rotazyme and direct electron microscopy for detection of rotavirus in human stools. J. Clin. Microbiol. 16: Chiba, S., I. Nakamura, S. Urasawa, S. Nakata, K. Taniguchi, K. Fujinaga, and T. Nakao Outbreak of infantile gastroenteritis due to type 40 adenovirus. Lancet ii: Chiba, S., S. Nakata, T. Urasawa, S. Urasawa, T. Yokoyama, Y. Morita, K. Taniguchi, and T. Nakao Protective effect of naturally acquired homotypic and heterotypic rotavirus antibodies. Lancet ii: Chiba, S., Y. Sakuma, R. Kogasaka, M. Akihara, K. Horino, T. Nakao, and S. Fukui An outbreak of gastroenteritis associated with calicivirus in an infant home. J. Med. Virol. 4: Chiba, S., Y. Sakuma, R. Kogasaka, M. Akihara, H. Terashima, K. Horino, and T. Nakao Fecal shedding of virus in relation to the days of illness in infantile gastroenteritis due to calicivirus. J. Infect. Dis. 142: Christy, C., H. P. Madore, J. J. Treanor, K. Pray, A. Z. Kapikian, R. M. Chanock, and R. Dolin Safety and immunogenicity of live attenuated rhesus monkey rotavirus vaccine. J. Infect. Dis. 154: Chrystie, I. L., B. Totterdell, M. J. Baker, J. W. Scopes, and J. E. Banatvala Rotavirus infections in a maternity unit. Lancet ii: Chrystie, I. L., B. M. Totterdell, and J. E. Banatvala Asymptomatic endemic rotavirus infections in the newborn. Lancet i: Chrystie, I. L., B. M. Totterdell, and J. E. Banatvala CLIN. MICROBIOL. REV. False positive Rotazyme tests on faecal samples from babies. Lancet ii: Clark, H. F., K. T. Dolan, P. Horton-Slight, J. Palmer, and S. A. Plotkin Diverse serologic response to rotavirus infection of infants in a single epidemic. Pediatr. Infect. Dis. 4: Clark, H. F., T. Furukawa, L. M. Bell, P. A. Offit, P. A. Perrella, and S. A. Plotkin Immune response of infants and children to low-passage bovine rotavirus (strain WC3). Am. J. Dis. Child. 140: Clark, H. F., Y. Hoshino, L. M. Bell, J. Groff, G. Hess, P. Bachman, and P. A. Offit Rotavirus isolate W161 representing a presumptive new human serotype. J. Clin. Microbiol. 25: Cook, D. A., A. Zbitnew, G. Dempster, and J. W. Gerrard Detection of antibody to rotavirus by counterimmunoelectrophoresis in human serum, colostrum, and milk. J. Pediatr. 93: Coulson, B. S., and I. H. Holmes An improved enzymelinked immunosorbent assay for the detection of rotavirus in faeces of neonates. J. Virol. Methods 8: Coulson, B. S., L. E. Unicomb, G. A. Pitson, and R. F. Bishop Simple and specific enzyme immunoassay using monoclonal antibodies for serotyping human rotaviruses. J. Clin. Microbiol. 25: Cruickshank, J. G., J. H. M. Axton, and 0. F. Webster Viruses in gastroenteritis. Lancet i: Cubitt, W. D Rotavirus infection: an unexpected hazard in units caring for the elderly. Geriat. Med. Today 1: Cubitt, W. D., and D. A. McSwiggan Calicivirus gastroenteritis in North West London. Lancet ii: Cubitt, W. D., and D. A. McSwiggan Seroepidemiological survey of the prevalence of antibodies to a strain of human calicivirus. J. Med. Virol. 21: Cubitt, W. D., D. A. McSwiggan, and S. Arstall An outbreak of calicivirus infection in a mother and baby unit. J. Clin. Pathol. 33: Cubitt, W. D., P. J. Pead, and A. A. Saeed A new serotype of calicivirus associated with an outbreak of gastroenteritis in a residential home for the elderly. J. Clin. Pathol. 34: Cukor, G., N. R. Blacklow, F. E. Capozza, Z. F. K. Panjvani, and F. Bednarek Persistence of antibodies to rotavirus in human milk. J. Clin. Microbiol. 9: Cukor, G., N. R. Blacklow, P. Echeverria, M. K. Bedigian, H. Puruggan, and V. Basaca-Sevilla Comparative study of the acquisition of antibody to Norwalk virus in pediatric populations. Infect. Immun. 29: Dai, G.-Z., M.-S. Sun, S.-Q. Liu, X.-F. Ding, T.-D. Chen, L.-C. Wang, D.-P. Du, G. Zhao, Y. Su, J. Li, W.-M. Xu, T.-H. Li, and X.-X Chen First report of an epidemic of diarrhoea in human neonates involving the new rotavirus and biological characteristics of the epidemic virus strain (KMB/R85). J. Med. Virol. 22: Davidson, G. P., and G. L. Barnes Structural and functional abnormalities of the small intestine in infants and young children with rotavirus enteritis. Acta Paediatr. Scand. 68: Davidson, G. P., R. J. Hogg, and C. P. Kirubakaran Serum and intestinal immune response to rotavirus enteritis in children. Infect. Immun. 40: Davidson, G. P., R. R. W. Townley, R. F. Bishop, I. H. Holmes, and B. J. Ruck Importance of a new virus in acute sporadic enteritis in children. Lancet i: de Jong, J. C., R. Wigand, A. H. Kidd, G. Wadell, J. G. Kapsenberg, C. J. Muzerie, A. G. Wermenbol, and R.-G. Firtzlaff Candidate adenoviruses 40 and 41: fastidious adenoviruses from human infant stool. J. Med. Virol. 11: Dennehy, P. H., and G. Peter Risk factors associated with nosocomial rotavirus infection. Am. J. Dis. Child. 139: DeSilva, L. M., M. Hanlon, and J. Stanton Rotavirus

31 VOL. 2, 1989 infection. Med. J. Aust. 2: de Torres, B. V., R. M. de lua, and J. Esparza Epidemiological aspects of rotavirus infection in hospitalized Venezuelan children with gastroenteritis. Am. J. Trop. Med. Hyg. 27: Dimitrov, D. H., M. K. Estes, S. M. Rangelova, L. M. Shindarov, J. L. Melnick, and D. Y. Graham Detection of antigenically distinct rotaviruses from infants. Infect. Immun. 41: Dimitrov, D. H., D. Y. Graham, and M. K. Estes Detection of rotaviruses by nucleic acid hybridization with cloned DNA of simian rotavirus SA11 genes. J. Infect. Dis. 152: Doern, G. V., J. E. Herrmann, P. Henderson, D. Stobbs-Walro, D. M. Perron, and N. R. Blacklow Detection of rotavirus with a new polyclonal antibody enzyme immunoassay (Rotazyme II) and a commercial latex agglutination tests (Rotalex): comparison with a monoclonal antibody enzyme immunoassay. J. Clin. Microbiol. 23: Dolin, R., N. R. Blacklow, H. DuPont, R. F. Buscho, R. G. Wyatt, J. A. Kasel, R. Hornick, and R. M. Chanock Biological properties of Norwalk agent of acute infectious nonbacterial gastroenteritis. Proc. Soc. Exp. Biol. Med. 140: Dolin, R., N. R. Blacklow, H. DuPont, S. Formal, R. F. Buscho, J. A. Kasel, R. P. Chames, R. Hornick, and R. M. Chanock Transmission of acute infectious nonbacterial gastroenteritis to volunteers by oral administration of stool filtrates. J. Infect. Dis. 123: Dolin, R., A. G. Levy, R. G. Wyatt, T. S. Thornhill, and J. D. Gardner Viral gastroenteritis induced by the Hawaii agent. Jejunal histopathology and serologic response. Am. J. Med. 59: Dolin, R., R. C. Reichman, K. D. Roessner, T. S. Tralka, R. T. Schooley, W. Gary, and D. Morens Detection by immune electron microscopy of the Snow Mountain agent of acute viral gastroenteritis. J. Infect. Dis. 146: Dolin, R., K. D. Roessner, J. J. Treanor, R. C. Reichman, M. Phillips, and H. P. Madore Radioimmunoassay for detection of the Snow Mountain agent of viral gastroenteritis. J. Med. Virol. 19: Dowling, J. M., and H. Wynne Role of enteric adenoviruses and rotaviruses in infantile gastroenteritis. Lancet ii: Duffy, L. C., T. E. Byers, M. Riepenhoff-Talty, L. J. La Scolea, M. Zielezny, and P. L. Ogra The effects of infant feeding on rotavirus-induced gastroenteritis: a prospective study. Am. J. Public Health 76: DuPont, H. L Rotaviral gastroenteritis-some recent developments. J. Infect. Dis. 149: Dyall-Smith, M. L., A. A. Azad, and I. H. Holmes Gene mapping of rotavirus double-stranded RNA segments by Northern blot hybridization: application to segments 7, 8, and 9. J. Virol. 46: Dyall-Smith, M. L., and I. H. Holmes Comparisons of rotavirus polypeptides by limited proteolysis: close similarity of certain polypeptides of different strains. J. Virol. 40: Dyall-Smith, M. L., and I. H. Holmes Gene-coding assignments of rotavirus double-stranded RNA segments 10 and 11. J. Virol. 38: Dyall-Smith, M. L., and I. H. Holmes Sequence homology between human and animal rotavirus serotype-specific glycoproteins. Nucleic Acids Res. 12: Ebina, T., A. Sato, K. Umezu, N. Ishida, S. Ohyama, A. Oizumi, K. Aikawa, S. Katagiri, N. Katsushima, A. Imai, S. Kitaoka, H. Suzuki, and T. Konno Prevention of rotavirus infection by oral administration of cow colostrum containing anti-human rotavirus antibody. Med. Microbiol. Immunol. 174: Ebina, T. A., K. Umezu, S. Ohyama, K. Aikawa, N. Katsushima, S. Kitaoka, A. Sato, N. Ishida, A. Ohizumi, S. Katagiri, A. Imai, H. Suzuki, and T. Konno Prevention VIRAL GASTROENTERITIS 81 of rotavirus infection by cow colostrum containing antibody against human rotavirus. Lancet ii: Echeverria, P., N. R. Blacklow, G. G. Cukor, S. Vibulbandhitkit, S. Changchawalit, and P. Boonthai Rotavirus as a cause of severe gastroenteritis in adults. J. Clin. Microbiol. 18: Eiden, J., H. M. Lederman, S. Vonderfecht, and R. Yolken T-cell-deficient mice display normal recovery from experimental rotavirus infection. J. Virol. 57: Eiden, J., S. Sato, and R. Yolken Specificity of dot hybridization assay in the presence of rrna for detection of rotaviruses in clinical specimens. J. Clin. Microbiol. 25: Elden, J., S. Vonderfecht, K. Theil, A. Torres-Medina, and R. H. Yolken Genetic and antigenic relatedness of human and animal strains of antigenically distinct rotaviruses. J. Infect. Dis. 154: Engvall, E., and P. Perlmann Enzyme-linked immunosorbent assay, ELISA. III. Quantitation of specific antibodies by enzyme-labeled anti-immunoglobulin in antigen-coated tubes. J. Immunol. 109: Espejo, R. T., E. Calderon, N. Gonzalez, A. Salomon, A. Martuscelli, and P. Romero Presence of two distinct types of rotavirus in infants and young children hospitalized with acute gastroenteritis in Mexico City, J. Infect. Dis. 139: Espejo, R. T., 0. Munoz, F. Serafin, and P. Romero Shift in the prevalent human rotavirus detected by ribonucleic acid segment differences. Infect. Immun. 27: Espejo, R. T., F. Puerto, C. Soler, and N. Gonzalez Characterization of human pararotavirus. Infect. Immun. 44: Estes, M. K., E. L. Palmer, and J. Obijeski Rotaviruses: a review. Curr. Top. Microbiol. Immunol. 105: Fernbach, S. K., and J. D. Lloyd-Still The radiographic findings in severe rotavirus-induced colitis. J. Can. Assoc. Radiol. 35: Flewett, T. H Diagnosis of enteritis virus. Proc. R. Soc. Med. 69: Flewett, T. H., A. S. Bryden, and H. Davies Virus particles in gastroenteritis. Lancet ii: Flewett, T. H., A. S. Bryden, H. Davies, and C. A. Morris Epidemic viral enteritis in a long-stay children's ward. Lancet i: Flewett, T. H., A. S. Bryden, H. Davies, G. N. Woode, J. C. Bridger, and J. M. Derrick Relation between viruses from acute gastroenteritis of children and newborn calves. Lancet ii: Flewett, T. H., and H. Davies Calicivirus in man. Lancet i: Flewett, T. H., and G. N. Woode The rotaviruses. Arch. Virol. 57: Flores, J., H. B. Greenberg, H. Myslinski, A. R. Kalica, R. G. Wyatt, A. Z. Kapikian, and R. M. Chanock Use of transcription probes for genotyping rotavirus reassortants. Virology 121: Flores, J., I. Perez, L. White, M. Perez, A. R. Kalica, R. Marquina, R. G. Wyatt, A. Z. Kapikian and R. M. Chanock Genetic relatedness among human rotaviruses as determined by RNA hybridization. Infect. Immun. 37: Flores, J., R. H. Purcell, I. Perez, R. G. Wyatt, E. Boeggeman, M. Sereno, L. White, R. M. Chanock, and A. Z. Kapikian A dot hybridization assay for detection of rotavirus. Lancet i: Fox, J. P., C. D. Hall, and M. C. Cooney The Seattle virus watch. VII. Observations on adenovirus infections. Am. J. Epidemiol. 105: Fragoso, M., A. Kumar, and D. L. Murray Rotavirus in nasopharyngeal secretions of children with upper respiratory tract infections. Diagn. Microbiol. Infect. Dis. 4: Gary, G. W., Jr., J. C. Hierholzer, and R. E. Black Characteristics of noncultivable adenoviruses associated with diarrhea in infants: a new subgroup of human adenoviruses. J.

32 82 CHRISTENSEN Clin. Microbiol. 10: Gary, G. W., Jr., J. E. Kaplan, S. E. Stine, and L. J. Anderson Detection of Norwalk virus antibodies and antigen with a biotin-avidin immunoassay. J. Clin. Microbiol. 22: Gerna, G., N. Passarani, M. Battaglia, and E. Percivalle Rapid serotyping of human rotavirus strains by solid-phase immune electron microscopy. J. Clin. Microbiol. 19: Gerna, G., N. Passarani, M. Battaglia, and E. G. Rondanelli Human enteric coronaviruses: antigenic relatedness to human coronavirus OC43 and possible etiologic role in viral gastroenteritis. J. Infect. Dis. 151: Gerna, G., N. Passarani, P. M. Cereda, and M. Battaglia Antigenic relatedness of human enteric coronavirus strains to human coronavirus OC43: a preliminary report. J. Infect. Dis. 150: Gerna, G., N. Passarani, A. Sarasini, and M. Battaglia Characterization of serotypes of human rotavirus strains by solid-phase immune electron microscopy. J. Infect. Dis. 152: Gerna, G., A. Sarasini, N. Passarani, M. Torsellini, M. Parea, and M. Battaglia Comparative evaluation of a commercial enzyme-linked immunoassay and solid-phase immune electron microscopy for rotavirus detection in stool specimens. J. Clin. Microbiol. 25: Giaquinto, C., G. Errico, E. Ruga, I. Naso, and R. D'Elia Evaluation of ELISA test for rotavirus diagnosis in neonates. J. Pediatr. 109: Goldwater, P. N., I. L. Chrystie, and J. E. Banatvala Rotaviruses and the respiratory tract. Br. Med. J. 2: Graham, F. L., J. Smiley, W. C. Russell, and R. Nairn Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J. Gen. Virol. 36: Grauballe, P. C., K. Hjelt, P. A. Krasilnikoff, and P. O. Schiotz Elisa for rotavirus-specific secretary IgA in human sera. Lancet ii: Green, M Oncogenic viruses. Annu. Rev. Biochem. 39: Green, M., J. K. Mackey, W. S. M. Wold, and P. Rigden Thirty-one human adenovirus serotypes (Adl-Ad3l) form five groups (A-E) based upon DNA genome homologies. Virology 93: Greenberg, H., V. McAuliffe, J. Valdesuso, R. Wyatt, J. Flores, A. Kalica, Y. Hoshino, and N. Singh Serological analysis of the subgroup protein of rotavirus, using monoclonal antibodies. Infect. Immun. 39: Greenberg, H. B., J. R. Valdesuso, A. R. Kalica, R. G. Wyatt, V. J. McAuliffe, A. Z. Kapikian, and R. M. Chanock Proteins of Norwalk virus. J. Virol. 37: Greenberg, H. B., J. R. Valdesuso, A. Z. Kapikian, R. M. Chanock, R. G. Wyatt, W. Szmaness, J. Larrick, J. Kaplan, R. H. Gilman, and D. A. Sack Prevalence of antibody to the Norwalk virus in various countries. Infect. Immun. 26: Greenberg, H. B., J. Valdesuso, K. Van Wyke, K. Midthun, M. Walsh, V. McAuliffe, R. G. Wyatt, A. R. Kalica, J. Flores, and Y. Hoshino Production and preliminary characterization of monoclonal antibodies directed at two surface proteins of rhesus rotavirus. J. Virol. 47: Greenberg, H. B., J. Valdesuso, R. H. Yolken, E. Gangarosa, W. Gary, R. G. Wyatt, T. Konno, H. Suzuki, R. M. Chanock, and A. Z. Kapikian Role of Norwalk virus in outbreaks of nonbacterial gastroenteritis. J. Infect. Dis. 139: Greenberg, H. B., R. G. Wyatt, A. R. Kalica, R. H. Yolken, R. Black, A. Z. Kapikian, and R. M. Chanock New insights in viral gastroenteritis. Perspect. Virol. 11: Greenberg, H. B., R. G. Wyatt, J. Valdesuso, A. R. Kalica, W. T. London, R. M. Chanock, and A. Z. Kapikian Solid-phase microtiter radioimmunoassay for detection of the Norwalk strain of acute nonbacterial, epidemic gastroenteritis virus and its antibodies. J. Med. Virol. 2: Grunow, J. E., S. F. Dunton, and J. L. Waner Human rotavirus-like particles in a hepatic abscess. J. Pediatr. 106: CLIN. MICROBIOL. REV Haikala, 0. J., J. 0. Kokkonen, M. K. Leinonen, T. Nurmi, R. Mantyjarvi, and H. K. Sarkkinen Rapid detection of rotavirus in stool by latex agglutination: comparison with radioimmunoassay and electron microscopy and clinical evaluation of the test. J. Med. Virol. 11: Halonen, P., H. Sarkkinen, P. Arstila, E. Hjertsson, and E. Torfason Four-layer radioimmunoassay for detection of adenovirus in stool. J. Clin. Microbiol. 11: Halvorsrud, J., and I. Orstavik An epidemic of rotavirus-associated gastroenteritis in nursing home for the elderly. Scand. J. Infect. Dis. 12: Hammond, G., C. Hannan, T. Yeh, K. Fischer, G. Mauthe, and S. E. Straus DNA hybridization for diagnosis of enteric adenovirus infection from directly spotted fecal specimens. J. Clin. Microbiol. 25: Hammond, G. W., G. S. Ahluwalia, F. G. Barker, G. Horsman, and P. R. Hazelton Comparison of direct and indirect enzyme immunoassays with direct ultracentrifugation before electron microscopy for detection of rotaviruses. J. Clin. Microbiol. 16: Hammond, G. W., G. S. Ahluwalia, and P. R. Hazelton Detection of human rotavirus in fecal specimens by a commercial latex-agglutination test. J. Infect. Dis. 149: Hasegawa, A., S. Matsuno, S. Inouye, R. Kono, Y. Tsurukubo, A. Mukoyama, and Y. Saito Isolation of human rotaviruses in primary cultures of monkey kidney cells. J. Clin. Microbiol. 16: Heimer, G. V., and W. D. Cubitt Improved immunofluorescence techniques with microplates for the detection of M and G immunoglobulins against rotavirus. J. Virol. Methods 6: Herrmann, J. E., R. W. Hudson, N. R. Blacklow, and W. D. Cubitt Marin County agent, an astrovirus. Lancet ii: Herrmann, J. E., G. P. Kent, N. A. Nowak, J. Brondum, and N. R. Blacklow Antigen detection in the diagnosis of Norwalk virus gastroenteritis. J. Infect. Dis. 154: Herrmann, J. E., N. A. Nowak, and N. R. Blacklow Detection of Norwalk virus in stools by enzyme immunoassay. J. Med. Virol. 17: Herrmann, J. E., D. M. Perron, N. R. Blacklow, G. Cukor, P. J. Krause, J. S. Hyams, H. J. Barrett, and P. L. Ogra Enzyme immunoassay with monoclonal antibodies for the detection of rotavirus in stool specimens. J. Infect. Dis. 152: Hieber, J. P., S. Shelton, J. D. Nelson, J. Leon, and E. Mohs Comparison of human rotavirus disease in tropical and temperate settings. Am. J. Dis. Child. 132: Hierholzer, J. C., K. H. Johansson, L. J. Anderson, C. J. Tsou, and P. E. Halonen Comparison of monoclonal timeresolved fluoroimmunoassay with monoclonal capture-biotinylated detector enzyme immunoassay for adenovirus antigen detection. J. Clin. Microbiol. 25: Hilpert, H., H. Brussow, C. Mietens, J. Sidoti, L. Lerner, and H. Werchau Use of bovine milk concentrate containing antibody to rotavirus to treat rotavirus gastroenteritis in infants. J. Infect. Dis. 156: Hjelt, K., P. C. Grauballe, L. Andersen, P. 0. Schiotz, P. Howitz, and P. A. Krasilnikoff Antibody response in serum and intestine in children up to six months after a naturally acquired rotavirus gastroenteritis. J. Pediatr. Gastroenterol. Nutr. 5: Hjelt, K., P. C. Grauballe, L. Henriksen, and P. A. Krasilnikoff Rotavirus infections among the staff of a general paediatric department. Acta Paediatr. Scand. 74: Hjelt, K., P. C. Grauballe, 0. H. Nielsen, P. 0. Schiotz, and P. A. Krasilnikoff Rotavirus antibodies in the mother and her breast-fed infant. J. Pediatr. Gastroenterol. Nutr. 4: Hjelt, K., P. C. Grauballe, A. Paerregaard, 0. H. Nielsen, and P. A. Krasilnikoff Protective effect of preexisting rotavirus-specific immunoglobulin A against naturally acquired rotavirus infection in children. J. Med. Virol. 21: Hjelt, K., P. C. Grauballe, P. 0. Schiotz, L. Andersen, and

33 VOL. 2, 1989 P. A. Krasilnikoff Intestinal and serum immune response to a naturally acquired rotavirus gastroenteritis in children. J. Pediatr. Gastroenterol. Nutr. 4: Hjelt, K., P. A. Krasilnikoff, P. C. Grauballe, and S. W. Rasmussen Nosocomial acute gastroenteritis in a paediatric department, with special reference to rotavirus infections. Acta Paediatr. Scand. 74: Hjelt, K., P. A. Krasilnikoff, P. C. Grauballe, and S. W. Rasmussen Clinical features in hospitalized children with acute gastroenteritis. Acta Paediatr. Scand. 74: Hogg, R. J., and G. P. Davidson Improved specificity of ELISA for rotavirus. Aust. Paediatr. J. 18: Holmes, I. H Rotaviruses, p In W. K. Joklik (ed.), The Reoviridae. Plenum Publishing Corp., New York Holmes, I. H., M. Mathan, P. Bhat, M. J. Albert, S. P. Swaminathan, P. P. Maiya, S. M. Pereira, and S. J. Baker Orbiviruses and gastroenteritis. Lancet ii: Hoshino, Y., M. M. Sereno, K. Midthun, J. Flores, A. Z. Kapikian, and R. M. Chanock Independent segregation of two antigenic specificities (VP3 and VP7) involved in neutralization of rotavirus infectivity. Proc. Natl. Acad. Sci. USA 82: Hoshino, Y., R. G. Wyatt, H. B. Greenberg, J. Flores, and A. Z. Kapikian Serotypic similarity and diversity of rotaviruses of mammalian and avian origin as studied by plaquereduction neutralization. J. Infect. Dis. 149: Hovi, T., V. Vaisanen, P. Ukkonen, and C.-H. von Bonsdorff Solid-phase enzyme immunoassay for rotavirus antigen: faecal protease activity as a reason for false-negative results. J. Virol. Methods 5: Huebner, R. J Adenovirus-directed tumor and T antigens. Perspect. Virol. 5: Hughes, J. H., A. V. Tuomari, D. R. Mann, and V. V. Hamparian Latex immunoassay for rapid detection of rotavirus. J. Clin. Microbiol. 20: Hung, T., G. Chen, C. Wang, Z. Chou, T. Chao, W. Ye, H. Yao, and K. Meng Rotavirus-like agent in adult nonbacterial diarrhoea in China. Lancet ii: Hung, T., G. Chen, C. Wang, H. Yao, Z. Fang, T. Chao, Z. Chou, W. Ye, X. Chang, S. Den, X. Liong, and W. Chang Waterborne outbreak of rotavirus diarrhoea in adults in China caused by a novel rotavirus. Lancet i: Hyams, J. S., P. J. Krause, and P. A. Gleason Lactose malabsorption following rotavirus infection in young children. J. Pediatr. 99: Inouye, S., S. Matsuno, and H. Yamaguchi Efficient coating of the solid phase with rotavirus antigens for enzymelinked immunosorbent assay of immunoglobulin A antibody in feces. J. Clin. Microbiol. 19: Jacobsson, P. A., M. E. Johansson, and G. Wadell Identification of an enteric adenovirus by immunoelectroosmophoresis (EIOP) technique. J. Med. Virol. 3: Johansson, M. E., I. Uhnoo, A. H. Kidd, C. R. Madeley, and G. Wadell Direct identification of enteric adenovirus, a candidate new serotype, associated with infantile gastroenteritis. J. Clin. Microbiol. 12: Johansson, M. E., I. Uhnoo, L. Svensson, C.-A. Pettersson, and G. Wadell Enzyme-linked immunosorbent assay for detection of enteric adenovirus 41. J. Med. Virol. 17: Joncas, J., A. Moisan, and V. Pavilanis Incidence of adenovirus infection: a family study. Can. Med. Assoc. J. 87: Julkunen, I., J. Savolainen, A. Hautanen, and T. Hovi Detection of rotavirus in faecal specimens by enzyme immunoassay, latex agglutination and electron microscopy. Scand. J. Infect. Dis. 17: Kalica, A. R., J. Flores, and H. B. Greenberg Identification of the rotaviral gene that codes for hemagglutination and protease-enhanced plaque formation. Virology 125: Kalica, A. R., H. B. Greenberg, R. G. Wyatt, J. Flores, M. M. Sereno, A. Z. Kapikian, and R. M. Chanock Genes of human (strain Wa) and bovine (strain UK) rotaviruses that code for neutralization and subgroup antigens. Virology 112: VIRAL GASTROENTERITIS Kalica, A. R., R. H. Purcell, M. M. Sereno, R. G. Wyatt, H. W. Kim, R. M. Chanock, and A. Z. Kapikian A microtiter solid-phase radioimmunoassay for detection of human reovirus-like agent in stools. J. Immunol. 118: Kalica, A. R., M. M. Sereno, R. G. Wyatt, C. A. Mebus, R. M. Chanock, and A. Z. Kapikian Comparison of human and animal rotavirus strains by gel electrophoresis of viral RNA. Virology 87: Kantharidis, P., M. L. Dyall-Smith, and I. H. Holmes Completion of the gene-coding assignments of SA11 rotavirus: gene products of segments 7, 8, and 9. J. Virol. 48: Kapikian, A. Z., J. Flores, Y. Hoshino, R. I. Glass, K. Midthun, M. Gorziglia, and R. M. Chanock Rotavirus: the major etiologic agent of severe infantile diarrhea may be controllable by a "Jennerian" approach to vaccination. J. Infect. Dis. 153: Kapikian, A. Z., H. W. Kim, R. G. Wyatt, W. L. Cline, J. 0. Arrobio, C. D. Brandt, W. J. Rodriguez, D. A. Sack, R. M. Chanock, and R. H. Parrott Human reovirus-like agent as the major pathogen associated with "winter" gastroenteritis in hospitalized infants and young children. N. Engl. J. Med. 294: Kapikian, A. Z., H. W. Kim, R. G. Wyatt, W. J. Rodriguez, S. Ross, W. L. Cline, R. H. Parrott, and R. M. Chanock Reovirus-like agent in stools: association with infantile diarrhea and development of serologic tests. Science 185: Kapikian, A. Z., R. G. Wyatt, R. Dolin, T. S. Thornhill, A. R. Kalica, and R. M. Chanock Visualization by immune electron microscopy of a 27-nm particle associated with acute infectious nonbacterial gastroenteritis. J. Virol. 10: Kapikian, A. Z., R. G. Wyatt, H. B. Greenberg, A. R. Kalica, H. W. Kim, C. D. Brandt, W. J. Rodriguez, R. H. Parrott, and R. M. Chanock Approaches to immunization of infants and young children against gastroenteritis due to rotaviruses. Rev. Infect. Dis. 2: Kaplan, J. E., G. W. Gary, R. C. Baron, N. Singh, L. B. Schonberger, R. Feldman, and H. B. Greenberg Epidemiology of Norwalk gastroenteritis and the role of Norwalk virus in outbreaks of acute nonbacterial gastroenteritis. Ann. Intern. Med. 96: Keswick, B. H., T. W. Hejkal, H. L. DuPont, and L. K. Pickering Evaluation of a commercial enzyme immunoassay kit for rotavirus detection. Diagn. Microbiol. Infect. Dis. 1: Keswick, B. H., L. K. Pickering, H. L. DuPont, and W. E. Woodward Prevalence of rotavirus in children in day care centers. J. Pediatr. 103: Kidd, A. H Genome variants of adenovirus 41 (subgroup G) from children with diarrhoea in South Africa. J. Med. Virol. 14: Kidd, A. H., J. E. Banatvala, and J. C. de Jong Antibodies to fastidious faecal adenoviruses (species 40 and 41) in sera from chidren. J. Med. Virol. 11: Kidd, A. H., F. E. Berkowtiz, P. J. Blaskovic, and B. D. Schoub Genome variants of human adenovirus 40 (subgroup F). J. Med. Virol. 14: Kidd, A. H., B. D. Cosgrove, R. A. Brown, and C. R. Madeley Faecal adenoviruses from Glasgow babies. Studies on culture and identity. J. Hyg. 88: Kidd, A. H., E. H. Harley, and M. J. Erasmus Specific detection and typing of adenovirus types 40 and 41 in stool specimens by dot-blot hybridization. J. Clin. Microbiol. 22: Kidd, A. H., and C. R. Madeley In vitro growth of some fastidious adenoviruses from stool specimens. J. Clin. Pathol. 34: Kidd, A. H., A. Rosenblatt, T. G. Besselaar, M. J. Erasmus, C. T. Tiemessen, F. E. Berkowtiz, and B. D. Schoub Characterization of rotaviruses and subgroup F adenoviruses from acute summer gastroenteritis in South Africa. J. Med. Virol. 18:

34 84 CHRISTENSEN 199. Kjeldsberg, E Small spherical viruses in faeces from gastroenteritis patients. Acta Pathol. Microbiol. Scand. Sec. B 85: Kjeldsberg, E Specific labelling of human rotaviruses and adenoviruses with gold-igg complexes. J. Virol. Methods 12: Kjeldsberg, E., and J. C. Siebke Use of immunosorbent electron microscopy for detection of rota- and hepatitis A virus in sucrose solutions. J. Virol. Methods 12: Knisley, C. V., A. J. Bednarz-Prashad, and L. K. Pickering Detection of rotavirus in stool specimens with monoclonal and polyclonal antibody-based assay systems. J. Clin. Microbiol. 23: Konno, T., H. Suzuki, A. Imai, T. Kutsuzawa, N. Ishida, N. Katsushima, M. Sakamoto, S. Kitaoka, R. Tsuboi, and M. Adachi A long-term survey of rotavirus infection in Japanese children with acute gastroenteritis. J. Infect. Dis. 158: Konno, T., H. Suzuki, M. Ishida, R. Chiba, K. Mochizuki, and A. Tsunoda Astrovirus-associated epidemic gastroenteritis in Japan. J. Med. Virol. 9: Konno, T., H. Suzuki, N. Katsushima, A. Imai, F. Tazawa, T. Kutsuzawa, S. Kitaoka, M. Sakamoto, N. Yazaki, and N. Ishida Influence of temperature and relative humidity on human rotavirus infection in Japan. J. Infect. Dis. 147: Koopman, J. S., V. F. Turkisk, A. S. Monto, F. E. Thompson, and R. E. Isaacson Milk fat and gastrointestinal illness. Am. J. Public Health 74: Kouvelos, K., M. Petric, and P. J. Middleton Comparison of bovine, simian and human rotavirus structural glycoproteins. J. Gen. Virol. 65: Krause, P. J., J. S. Hyams, P. J. Middleton, V. C. Herson, and J. Flores Unreliability of Rotazyme ELISA test in neonates. J. Pediatr. 103: Kurtz, J. B., and T. W. Lee Human astrovirus serotypes. Lancet ii: Kurtz, J. B., T. W. Lee, and D. Pickering Astrovirus associated gastroenteritis in a children's ward. J. Clin. Pathol. 30: Kutsuzawa, T., T. Konno, H. Suzuki, A. Z. Kapikian, T. Ebina, and N. Ishida Isolation of human rotavirus subgroups 1 and 2 in cell culture. J. Clin. Microbiol. 16: Laemmli, U. K Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature (London) 227: Lee, T. W. and J. B. Kurtz Serial propagation of astrovirus in tissue culture with the aid of trypsin. J. Gen. Virol. 57: Lee, T. W. and J. B. Kurtz Human astrovirus serotypes. J. Hyg. 89: Leite, J. P. G., H. G. Pereira, R. S. Azeredo, and H. G. Schatzmayr Adenoviruses in faeces of children with acute gastroenteritis in Rio De Janeiro, Brazil. J. Med. Virol. 15: Lewis, H. M., J. V. Parry, H. A. Davies, R. P. Parry, A. Mott. R. R. Dourmashkin, P. J. Sanderson, D. A. J. Tyrrell, and H. B. Valman A year's experience of the rotavirus syndrome and its association with respiratory illness. Arch. Dis. Child. 54: Lin, M., M. Imai, A. R. Bellamy, N. Ikegami, Y. Furuichi, D. Summers, D. L. Nuss, and R. Deibel Diagnosis of rotavirus infection with cloned cdna copies of viral genome segments. J. Virol. 55: Listernick, R., E. Zieserl, and A. T. Davis Outpatient oral rehydration in the United States. Am. J. Dis. Child. 140: Liu, M., P. A. Offit, and M. K. Estes Identification of the simian rotavirus SA11 genome segment 3 product. Virology 163: Liu, S., C. Birch, A. Coulepis, and I. Gust Radioimmunofocus assay for detection and quantitation of human rotavirus. J. Clin. Microbiol. 20: Lloyd-Evans, N., V. S. Springthorpe, and S. A. Sattar CLIN. MICROBIOL. REV. Chemical disinfection of human rotavirus-contaminated inanimate surfaces. J. Hyg. 97: Lopez, S., C. F. Arias, J. R. Bell, J. H. Strauss, and R. T. Espejo Primary structure of the cleavage site associated with trypsin enhancement of rotavirus SA11 infectivity. Virology 144: Losonsky, G. A., J. P. Johnson, J. A. Winkelstein, and R. H. Yolken Oral administration of human serum immunoglobulin in immunodeficient patients with viral gastroenteritis. J. Clin. Invest. 76: Losonsky, G. A., M. B. Rennels, A. Z. Kapikian, K. Midthun, P. J. Ferra, D. N. Fortier, K. M. Hoffman, A. Baig, and M. M. Levine Safety, infectivity, transmissibility and immunogenicity of rhesus rotavirus vaccine (MMU 18006) in infants. Pediatr. Infect. Dis. 5: Losonsky, G. A., S. L. Vonderfecht, J. Eiden, S.-B. Wee, and R. H. Yolken Homotypic and heterotypic antibodies for prevention of experimental rotavirus gastroenteritis. J. Clin. Microbiol. 24: Maass, G., H. G. Bauneister, R. Heigemoller, P. Jansen, and H. Schumacher Visualization by electron microscopy of 23-nm virus particles (astrovirus) in faeces in newborns with acute nonbacterial gastroenteritis. Zentralbl. Bakteriol. Parasitenkd. Infektionskr. Hyg. Abt. 1 Orig. Reihe A 242: Madeley, C. R Comparison of the features of astroviruses and caliciviruses seen in samples of feces by electron microscopy. J. Infect. Dis. 139: Madeley, C. R The emerging role of adenoviruses as inducers of gastroenteritis. Pediatr. Infect. Dis. 5:S63-S Madeley, C. R., and B. P. Cosgrove Viruses in infantile gastroenteritis. Lancet ii: Madeley, C. R., and B. P. Cosgrove nm particles in faeces in infantile gastroenteritis. Lancet ii: Madeley, C. R., and B. P. Cosgrove Caliciviruses in man. Lancet i: Madeley, C. R., B. P. Cosgrove, E. J. Bell, and R. J. Fallon Stool viruses in babies in Glasgow. 1. Hospital admissions with diarrhea. J. Hyg. 78: Madore, H. P., J. J. Treanor, and R. Dolin Characterization of the Snow Mountain agent of viral gastroenteritis. J. Virol. 58: Madore, H. P., J. J. Treanor, K. A. Pray, and R. Dolin Enzyme-linked immunosorbent assays for Snow Mountain and Norwalk agents of viral gastroenteritis. J. Clin. Microbiol. 24: Maiya, P. P., S. M. Pereira, M. Mathan, P. Bhat, M. J. Albert, and S. J. Baker Aetiology of gastroenteritis in infancy and early childhood in southern India. Arch. Dis. Child. 52: Maki, M A prospective clinical study of rotavirus diarrhoea in young children. Acta Paediatr. Scand. 70: Mankikar, S. D., P. J. Middleton, and M. Petric Counterimmunoelectrophoresis assay for detection of adenovirus antigen. J. Clin. Microbiol. 10: Marrie, T. J., S. H. S. Lee, R. S. Faulkner, J. Ethier, and C. H. Young Rotavirus infection in a geriatric population. Arch. Intern. Med. 142: Martin, M. L., G. W. Gary, Jr., and E. L. Palmer Comparison of hemagglutination-inhibition, complement-fixation, and enzyme-linked immunosorbent assay for quantitation of human rotavirus antibodies. Arch. Virol. 62: Mason, B. B., D. Y. Graham, and M. K. Estes In vitro transcription and translation of simian rotavirus SA11 gene products. J. Virol. 3: Mason, B. B., D. Y. Graham, and M. K. Estes Biochemical mapping of the simian rotavirus SA11 genome. J. Virol. 46: Matsumo, S., A. Hasegawa, A. Mukoyama, and S. Inouye A candidate for a new serotype of human rotavirus. J. Virol. 54: Matsumo, S., S. Inouye, A. Hasegawa, and R. Kono Assay of human rotavirus antibody by immune adherence hemagglutination with a cultivable human rotavirus as antigen.

35 VOL. 2, 1989 J. Clin. Microbiol. 15: McAllister, R. M., M. 0. Nicolson, G. Reed, J. Kern, R. V. Gilden, and R. J. Huebner Transformation of rodent cells by adenovirus 19 and other group D adenovirus. J. Nati. Cancer Inst. 43: McLean, B., and I. H. Holmes Transfer of antirotaviral antibodies from mothers to their infants. J. Clin. Microbiol. 12: McLean, B., S. Sonza, and I. H. Holmes Measurement of immunoglobulin A, G, and M class rotavirus antibodies in serum and mucosal secretions. J. Clin. Microbiol. 12: McLean, B. S., and I. H. Holmes Effects of antibodies, trypsin, and trypsin inhibitors on susceptibility of neonates to rotavirus infection. J. Clin. Microbiol. 13: McSwiggan, D. A., D. Cubitt, and W. Moore Calicivirus associated with winter vomiting disease. Lancet i: Middleton, P. J., M. D. Holdaway, M. Petric, M. T. Szymanski, and J. S. Tam Solid-phase radioimmunoassay for the detection of rotavirus. Infect. Immun. 14: Middleton, P. J., M. Petric, C. M. Hewitt, M. T. Szymanski, and J. S. Tam Counter-immunoelectro-osmophoresis for the detection of infantile gastroenteritis virus in conventional and germ-free piglets. Infect. Immun. 12: Middleton, P. J., M. Petric, C. M. Hewitt, M. T. Szymanski, and J. S. Tam Counter-immunoelectro-osmophoresis for the detection of infantile gastroenteritis virus (Orbi-group) antigen and antibody. J. Clin. Pathol. 29: Middleton, P. J., M. T. Szymanski, G. D. Abbott, R. Bortolussi, and J. R. Hamilton Orbivirus acute gastroenteritis of infancy. Lancet i: Middleton, P. J., M. T. Szymanski, and M. Petric Viruses associated with acute gastroenteritis in young children. Am. J. Dis. Child. 131: Midthun, K., H. B. Greenberg, T. Hoshino, A. Z. Kapikian, R. G. Wyatt, and R. M. Chanock Reassortant rotaviruses as potential live rotavirus vaccine candidates. J. Virol. 53: Midthun, K., Y. Hoshino, A. Z. Kapikian, and R. M. Chanock Single gene substitution rotavirus reassortants containing the major neutralization protein (VP7) of human rotavirus serotype 4. J. Clin. Microbiol. 24: Miotti, P. G., J. Eiden, and R. H. Yolken Comparative efficiency of commercial immunoassays for the diagnosis of rotavirus gastroenteritis during the course of infection. J. Clin. Microbiol. 22: Moffet, H. L., H. K. Shulenberger, and E. R. Burkholder Epidemiology and etiology of severe infantile diarrhea. J. Pediatr. 72: Moosai, R. B., R. Alcock, T. M. Bell, F. R. Laidler, J. S. M. Peiris, A. P. Wyn-Jones, and C. R. Madeley Detection of rotavirus by a latex agglutination test Rotalex: comparison with electron microscopy, immunofluorescence, polyacrylamide gel electrophoresis, and enzyme linked immunosorbent assay. J. Clin. Pathol. 38: Moosai, R. B., M. J. Carter, and C. R. Madeley Rapid detection of enteric adenovirus and rotavirus: a simple method using polyacrylamide gel electrophoresis. J. Clin. Pathol. 37: Mortensen, M. L., C. G. Ray, C. M. Payne, A. D. Friedman, L. L. Minnich, and C. Rousseau Coronaviruslike particles in human gastrointestinal disease. Am. J. Dis. Child. 139: Murphy, A. M., M. B. Albrey, and E. B. Crewe Rotavirus infections of neonates. Lancet ii: Nagayoski, S., H. Yamaguchi, T. Ichikawa, M. Miyazu, T. Morishima, T. Ozaki, S. Isomura, S. Suzuki, and M. Hoshino Changes of the rotavirus concentration in faeces during the course of acute gastroenteritis as determined by the immune adherence hemagglutination test. Eur. J. Pediatr. 134: Nakata, S., S. Chiba, H. Terashima, and T. Nakao Prevalence of antibody to human calicivirus in Japan and Southeast Asia determined by radioimmunoassay. J. Clin. VIRAL GASTROENTERITIS 85 Microbiol. 22: Nakata, S., S. Chiba, H. Terashima, Y. Sakuma, R. Kogasaka, and T. Nakao Microtiter solid-phase radioimmunoassay for detection of human calicivirus in stools. J. Clin. Microbiol. 17: Nakata, S., S. Chiba, H. Terashima, T. Yokoyama, and T. Nakao Humoral immunity in infants with gastroenteritis caused by human calicivirus. J. Infect. Dis. 152: Nakata, S., M. K. Estes, and S. Chiba Detection of human calicivirus antigen and antibody by enzyme-linked immunosorbent assays. J. Clin. Microbiol. 26: Nakata, S., M. K. Estes, D. Y. Graham, S. Wang, G. W. Gary, and J. L. Melnick Detection of antibody to group B adult diarrhea rotaviruses in humans. J. Clin. Microbiol. 25: Nakata, S., B. L. Petrie, E. P. Calomeni, and M. K. Estes Electron microscopy procedure influences detection of rotaviruses. J. Clin. Microbiol. 25: Nicolaieff, A., G. Obert, and M. H. V. van Regenmortel Detection of rotavirus by serological trapping on antibodycoated electron microscope grids. J. Clin. Microbiol. 12: Nicolas, J. C., J. Cohen, B. Fortier, M. H. Lourenco, and F. Bricout Isolation of an human pararotavirus. Virology 124: Niel, C., S. A. Gomes, J. P. G. Leite, and H. G. Pereira Direct detection and differentiation of fastidious and nonfastidious adenovirus in stools by using a specific nonradioactive probe. J. Clin. Microbiol. 24: Offit, P. A., and G. Blavat Identification of the two rotavirus genes determining neutralization specificities. J. Virol. 57: Offit, P. A., G. Blavat, H. B. Greenberg, and H. F. Clark Molecular basis of rotavirus virulence: role of gene segment 4. J. Virol. 57: Offit, P. A., and H. F. Clark Maternal antibody-mediated protection against gastroenteritis due to rotavirus in newborn mice is dependent on both serotype and titer of antibody. J. Infect. Dis. 152: Offit, P. A., H. F. Clark, W. G. Stroop, E. M. Twist, and S. A. Plotkin The cultivation of human rotavirus, strain 'Wa,' to high titer in cell culture and characterization of the viral structural polypeptides. J. Virol. Methods 7: Offit, P. A., R. D. Shaw, and H. B. Greenberg Passive protection against rotavirus-induced diarrhea by monoclonal antibodies to surface proteins vp3 and vp7. J. Virol. 58: Orstavik, I., K. J. Figenschau, and J. C. Ulstrup "Rotavirus" in stored specimens of faecal extracts. Lancet ii: Oshiro, L. S., C. E. Haley, R. R. Roberto, J. L. Riggs, M. Croughlan, H. Greenberg, and A. Kapikian A 27-nm virus isolated during an outbreak of acute infectious nonbacterial gastroenteritis in a convalescent hospital: a possible new serotype. J. Infect. Dis. 143: Pai, C. H., and D. Mayock Rotazyme test in neonates. J. Pediatr. 106: Pai, C. H., M. S. Shahrabadi, and B. Ince Rapid diagnosis of rotavirus gastroenteritis by a commercial latex agglutination test. J. Clin. Microbiol. 22: Parrino, T. A., D. S. Schreiber, J. S. Trier, A. Z. Kapikian, and N. R. Blacklow Clinical immunity in acute gastroenteritis caused by Norwalk agent. N. Engl. J. Med. 297: Paul, M. O., and E. A. Erinle Influence of humidity on rotavirus prevalence among Nigerian infants and young children with gastroenteritis. J. Clin. Microbiol. 15: Payne, C. M., and C. G. Ray Coronavirus? Pediatrics 74: Payne, C. M., C. G. Ray, V. Borduin, L. L. Minnich, and M. D. Lebowitz An eight-year study of the viral agents of acute gastroenteritis in humans: ultrastructural observations and seasonal distribution with a major emphasis on coronaviruslike particles. Diagn. Microbiol. Infect. Dis. 5: Pedley, S., J. C. Bridger, J. F. Brown, and M. A. McCrae Molecular characterization of rotaviruses with distinct

36 86 CHRISTENSEN group antigens. J. Gen. Virol. 64: Pedley, S., J. C. Bridger, D. Chasey, and M. A. McCrae Definition of two new groups of atypical rotaviruses. J. Gen. Virol. 67: Pereira, H. G., R. S. Azeredo, J. P. G. Leite, Z. P. Andrade, and L. De Castro A combined enzyme immunoassay for rotavirus and adenovirus (EIARA). J. Virol. Methods 10: Petric, M., P. J. Middleton, and S. D. Mankikar Detection of adenovirus antibody by counterimmunoelectrophoresis. J. Clin. Microbiol. 10: Petrie, B. L., H. B. Greenberg, B. Y. Graham, and M. K. Estes Ultrastructural localization of rotavirus antigens using colloidal gold. Virus Res. 1: Pina, M., and M. Green Biochemical studies on adenovirus multiplication. IX. Chemical and base composition analysis of 28 human adenoviruses. Proc. Nati. Acad. Sci. USA 54: Prince, S. D., C. Astry, S. Vonderfecht, G. Jakab, F.-M. Shen, and R. H. Yolken Aerosol transmission of experimental rotavirus infection. Pediatr. Infect. Dis. 5: Puerto, F. I., L. Padilla-Noriega, A. Zamora-Chavez, A. Briceno, M. Puerto, and C. F. Arias Prevalent patterns of serotype-specific seroconversion in Mexican children infected with rotavirus. J. Clin. Microbiol. 25: Radetsky, M Laboratory evaluation of acute diarrhea. Pediatr. Infect. Dis. 5: Rand, K. H., H. J. Houck, and H. M. Swingle Rotazyme assay in neonates without diarrhea. Am. J. Clin. Pathol. 84: Ratnam, S., A. M. Tobin, J. B. Flemming, and P. J. Blaskovic False positive Rotazyme results. Lancet i: Resta, S., J. P. Luby, C. R. Rosenfeld, and J. D. Siegel Isolation and propagation of a human enteric coronavirus. Science 229: Retter, M., P. J. Middleton, J. S. Tam, and M. Petric Enteric adenoviruses: detection, replication, and significance. J. Clin. Microbiol. 10: Rettig, P. J., and G. P. Altshuler Fatal gastroenteritis associated with coronaviruslike particles. Am. J. Dis. Child. 139: Reynolds, D., and J. H. Hughes Comparison of the Rotizyme assay with an avidin-biotin-amplified dot-immunobinding assay for detecting rotaviruses. J. Infect. Dis. 152: Richmond, S. J., S. M. Dunn, E. O. Caul, C. R. Ashley, and S. K. R. Clarke An outbreak of gastroenteritis in young children caused by adenoviruses. Lancet i: Riepenhoff-Talty, M., S. Bogger-Goren, P. Li, P. J. Carmody, H. J. Barrett, and P. L. Ogra Development of serum and intestinal antibody response to rotavirus after naturally acquired rotavirus infection in man. J. Med. Virol. 8: Riepenhoff-Talty, M., L. J. Saif, H. J. Barrett, H. Suzuki, and P. L. Ogra Potential spectrum of etiological agents of viral enteritis in hospitalized infants. J. Clin. Microbiol. 17: Riepenhoff-Talty, M., H. Suzuki, and P. L. Ogra Characteristics of the cell-mediated immune response to rotavirus in suckling mice. Dev. Biol. Stand. 53: Rodger, S. M., R. F. Bishop, C. Birch, B. McLean, and I. H. Holmes Molecular epidemiology of human rotaviruses in Melbourne, Australia, from 1973 to 1979, as determined by electrophoresis of genome ribonucleic acid. J. Clin. Microbiol. 13: Rodger, S. M., R. F. Bishop, and I. H. Holmes Detection of a rotavirus-like agent associated with diarrhea in infants. J. Clin. Microbiol. 16: Rodriguez, W. J., H. W. Kim, C. D. Brandt, M. K. Gardner, and R. H. Parrott Use of electrophoresis of RNA from human rotavirus to establish the identity of strains involved in outbreaks in a tertiary care nursery. J. Infect. Dis. 148: Rodriguez, W. J., H. W. Kim, C. D. Brandt, R. H. Yolken, M. Richard, J. 0. Arrobio, R. H. Schwartz, A. Z. Kapikian, R. M. CLIN. MICROBIOL. REV. Chanock, and R. H. Parrott Common exposure outbreak of gastroenteritis due to type 2 rotavirus with high secondary attack rate within families. J. Infect. Dis. 140: Rosen, L Hemagglutination inhibition technique for typing adenovirus. Am. J. Hyg. 71: Rotbart, H. A., R. H. Yolken, W. L. Nelson, D. Davis, M. H. Roe, and M. J. Levin Confirmatory testing of Rotazyme results in neonates. J. Pediatr. 107: Rubenstein, A. S., and M. F. Miller Comparison of an enzyme immunoassay with electron microscopic procedures for detecting rotavirus. J. Clin. Microbiol. 15: Rudd, R. T., and D. Carrington A prospective study of chlamydial, mycoplasmal, and viral infections in a neonatal intensive care unit. Arch. Dis. Child. 59: Sakuma, Y., S. Chiba, R. Kogasaka, H. Terashima, S. Nakamura, K. Horino, and T. Nakao Prevalence of antibody to human calicivirus in general population of Northern Japan. J. Med. Virol. 7: Sambourg, M., A. Goudeau, C. Courant, G. Pinon, and F. Denis Direct appraisal of latex agglutination testing, a convenient alternative to enzyme immunoassay for the detection of rotavirus in childhood gastroenteritis, by comparison of two enzyme immunoassays and two latex tests. J. Clin. Microbiol. 21: Sanekata, T., and H. Okada Human rotavirus detection by agglutination of antibody-coated erythrocytes. J. Clin. Microbiol. 17: Sanekata, T., Y. Toshida, and K. Oda Detection of rotavirus from faeces by reverse passive haemagglutination method. J. Clin. Pathol. 32: Sanekata, T., Y. Yoshida, K. Oda, and H. Okada Detection of rotavirus antibody by inhibition of reverse passive hemagglutination. J. Clin. Microbiol. 15: Santosham, M., R. H. Yolken, E. Quiroz, L. Dillman, G. Oro, W. C. Reeves, and R. B. Sack Detection of rotavrius in respiratory secretions of children with pneumonai. J. Pediatr. 103: Sarkkinen, H. K Human rotavirus antigen detection by enzyme-immunoassay with antisera against Nebraska calf diarrhoea virus. J. Clin. Pathol. 34: Sarkkinen, H. K., H. Tuokko, and P. E. Halonen Comparison of enzyme-immunoassay and radioimmunoassay for detection of human rotaviruses and adenoviruses from stool specimens. J. Virol. Methods 1: Sato, K., Y. Inaba, T. Shinozaki, R. Fujii, and M. Matsumoto Isolation of human rotavirus cell cultures. Arch. Virol. 69: Schnagl, R. D., F. Morey, and I. H. Holmes Rotavirus and coronavirus-like particles in aboriginal and non-aboriginal neonates in Kalgoorlie and Alice Springs. Med. J. Aust. 2: Schoub, B. D Enteric adenoviruses and rotaviruses in infantile gastroenteritis in developing countries. Lancet ii: Schoub, B. D., F. Cohen, D. Thompson, H. J. Koornhof, M. D. Miliotis, C. S. Still, F. E. Berkowitz, S. Miller, and E. Kushlick Variance in rotavirus infection rates in different urban population groups in South Africa. J. Med. Virol. 10: Schoub, B. D., H. J. Koornhof, G. Lecatsas, 0. W. Prozesky, I. Freiman, E. Hartman, and H. Kassel Viruses in acute summer gastroenteritis in black infants. Lancet i: Schreiber, D. S., N. R. Blacklow, and J. S. Trier The small intestinal lesion induced by Hawaii agent acute infectious nonbacterial gastroenteritis. J. Infect. Dis. 129: Shaw, R. D., K. J. Fong, G. A. Losonsky, M. M. Levine, Y. Maldonado, R. Yolken, J. Flores, A. Z. Kapikian, P. T. Vo, and H. B. Greenberg Epitope-specific immune responses to rotavirus vaccination. Gastroenterology 93: Shaw, R. D., D. L. Stoner-Ma, M. K. Estes, and H. B. Greenberg Specific enzyme-linked immunoassay for rotavirus serotypes 1 and 3. J. Clin. Microbiol. 22: Shinozaki, T., K. Araki, H. Ushijima, and R. Fujii Antibody response to enteric adenovirus types 40 and 41 in sera from people in various age groups. J. Clin. Microbiol. 25:

37 VOL. 2, Shinozaki, T., K. Araki, H. Ushijima, R. Fujii, and Y. Eshita Use of Graham 293 cells in suspension for isolating enteric adenoviruses from the stools of patients with acute gastroenteritis. J. Infect. Dis. 156: Shinozaki, T., K. Araki, H. Ushijima, B. Kim, T. Tajima, and R. Fujii Comparison of five methods for detecting human rotavirus in stool specimens. Eur. J. Pediatr. 144: Shinozaki, T., K. Araki, H. Ushijima, B. Kim, T. Tajima, R. Fujii, and M. Minamitani Coproantibody response to rotavirus in an outbreak in a day-care nursery. Eur. J. Pediatr. 144: Singh, N., M. M. Sereno, J. Flores, and A. Z. Kapikian Monoclonal antibodies to subgroup 1 rotavirus. Infect. Immun. 42: Singh-Naz, N., and R. K. Naz Development and application of monoclonal antibodies for specific detection of human enteric adenoviruses. J. Clin. Microbiol. 23: Sitbon, M Human-enteric-coronaviruslike particles (CVLP) with different epidemiological characteristics. J. Med. Virol. 16: Smith, M. L., I. Lazdins, and I. H. Holmes Coding assignments of double-stranded RNA segments of SA11 rotavirus established by in vitro translation. J. Virol. 33: Soenarto, Y., T. Sebodo, R. Ridho, H. Alrasjid, J. E. Rohde, H. C. Bugg, G. L. Barnes, and R. F. Bishop Acute diarrhea and rotavirus infection in newborn babies and children in Yogyakarta, Indonesia, from June 1978 to June J. Clin. Microbiol. 14: Sonza, S., A. M. Breschkin, and I. H. Holmes The major surface glycoproteins of simian rotavirus (SA11) contains distinct epitopes. Virology 134: Sonza, S., and I. H. Holmes Coproantibody response to rotavirus infection. Med. J. Aust. 2: Spence, L., M. Fauvel, R. Petro, and S. Bloch Comparison of counterimmunoelectrophoresis and electron microscopy for laboratory diagnosis of human reovirus-like agent associated with infantile gastroenteritis. J. Clin. Microbiol. 5: Spencer, E., F. Avendano, and M. Araya Characteristics and analysis of electropherotypes of human rotavirus isolated in Chile. J. Infect. Dis. 148: Spratt, H. C., M. I. Marks, M. Gomersal, P. Gill, and C. H. Pai Nosocomial infantile gastroenteritis associated with minirotavirus and calicivirus. J. Pediatr. 93: Springthorpe, V. S., J. L. Grenier, N. Lloyd-Evans, and S. A. Sattar Chemical disinfection of human rotaviruses: efficacy of commercially-available products in suspension tests. J. Hyg. 97: Stalhandske, P., R. Hyypia, A. Allard, P. Halonen, and U. Pettersson Detection of adenoviruses in stool specimens by nucleic acid spot hybridization. J. Med. Virol. 16: Stals, F., F. J. Walther, and C. A. Bruggeman Faecal and pharyngeal shedding of rotavirus and rotavirus IgA in children with diarrhoea. J. Med. Virol. 14: Suzuki, H., Y. Amano, H. Kinebuchi, E. G. Vera, A. Davila, J. Lopez, R. Gustabo, T. Konno, and N. Ishida Rotavirus infection in children with acute gastroenteritis in Ecuador. Am. J. Trop. Med. Hyg. 30: Suzuki, H., T. Konno, T. Kutsuzawa, A. Imai, F. Tazawa, N. Ishida, N. Katsushima, and M. Sakamoto The occurrence of calicivirus in infants with acute gastroenteritis. J. Med. Virol. 4: Suzuki, H., T. Konno, Y. Numazaki, S. Kitaoka, T. Sato, A. Imai, F. Tazawa, T. Nakagomi, 0. Nakagomi, and N. Ishida Three different serotypes of human rotavirus determined using an interference test with coxsackie virus B1. J. Med. Virol. 13: Suzuki, H., T. Sato, S. Kitaoka, F. Tazawa, T. Konno, Y. Amano, A. Alava-Alprecht, E. Gutierrez-Vera, J. Lopez-Villalta, Y. Numazaki, and N. Ishida Epidemiology of rotavirus in Guayaquil, Ecuador. Am. J. Trop. Med. Hyg. 35: VIRAL GASTROENTERITIS Svensson, L., M. Grandien, and C.-A. Pettersson Comparison of solid-phase immune electron microscopy by use of protein A with direct electron microscopy and enzyme-linked immunosorbent assay for detection of rotavirus in stool. J. Clin. Microbiol. 18: Svensson, L Identification of an outer capsid glycoprotein of human rotavirus by concanavalin A. J. Gen. Virol. 65: Takiff, H. E., W. Reinhold, C. F. Garon, and S. E. Straus Cloning and physical mapping of enteric adenovirus (candidate type 40 and 41). J. Virol. 51: Takiff, H. E., M. Seidlin, P. Krause, J. Rooney, C. Brandt, W. Rodriguez, R. Yolken, and S. E. Straus Detection of enteric adenoviruses by dot-blot hybridization using a molecularly cloned viral DNA probe. J. Med. Virol. 16: Takiff, H. E., S. E. Straus, and C. F. Garon Propagation and in vitro studies of previously non-cultivable enteral adenoviruses in 293 cells. Lancet ii: Tan, G. S., R. R. W. Townley, G. P. Davidson, R. F. Bishop, I. H. Holmes, and B. J. Ruck Virus in faecal extracts from children with gastroenteritis. Lancet i: Taniguchi, K., S. Urasawa, and T. Urasawa Preparation and characterization of neutralizing monoclonal antibodies with different reactivity patterns to human rotaviruses. J. Gen. Virol. 66: Thornhill, T. S., R. G. Wyatt, A. R. Kalica, R. Dolin, R. M. Chanock, and A. Z. Kapikian Detection by immune electron microscopy of 26- to 27-nm viruslike particles associated with two family outbreaks of gastroenteritis. J. Infect. Dis. 135: Thouless, M. E., G. M. Beards, and T. H. Flewett Serotyping and subgrouping of rotavirus strains by the ELISA test. Arch. Virol. 73: Thouless, M. E., A. S. Bryden, and T. H. Flewett Serotypes of human rotavirus. Lancet i: Thouless, M. E., A. S. Bryden, T. H. Flewett, G. N. Woode, J. C. Bridger, D. R. Snodgrass, and J. A. Herring Serologic relationships between rotaviruses from different species as studied by complement-fixation and neutralization. Arch. Virol. 53: Tolia, V. K., and R. S. Dubois Update of oral rehydration: its place in treatment of acute gastroenteritis. Pediatr. Ann. 14: Totterdell, B. M., J. E. Banatvala, and I. L. Chrystie Studies on human lacteal rotavirus by immune electron microscopy. J. Med. Virol. 11: Totterdell, B. M., J. E. Banatvala, I. L. Chrystie, G. Ball, and W. D. Cubitt Systemic lymphoproliferative responses to rotavirus. J. Med. Virol. 25: Totterdell, B. M., I. L. Chrystie, and J. E. Banatvala Cord blood and breast-milk antibodies in neonatal rotavirus infection. Br. Med. J. 280: Totterdell, B. M., K. G. Nicholson, J. MacLeod, I. L. Chrystie, and J. E. Banatvala Neonatal rotavirus infection: role of lacteal neutralising activity, alpha1 anti-trypsin and non-immunoglobulin antiviral activity in protection. J. Med. Virol. 10: Tronen, H False positive Rotazyme results. Lancet i: Tufvesson, B., and T. Johnsson Immunoelectroosmophoresis for detection of reo-like virus: methodology and comparison with electron microscopy. Acta Pathol. Microbiol. Scand. Sect. B 84: Uhnoo, I., E. Olding-Stenkvist, and A. Kreuger Clinical features of acute gastroenteritis associated with rotavirus, enteric adenoviruses, and bacteria. Arch. Dis. Child. 61: Uhnoo, I., and L. Svensson Clinical and epidemiological features of acute infantile gastroenteritis associated with human rotavirus subgroups 1 and 2. J. Clin. Microbiol. 23: Uhnoo, I., G. Wadell, L. Svensson, and M. Johansson Two new serotypes of enteric adenovirus causing infantile

38 88 CHRISTENSEN diarrhoea. Dev. Biol. Stand. 53: Uhnoo, I., G. Wadell, L. Svensson, and M. Johansson Importance of enteric adenoviruses 40 and 41 in acute gastroenteritis in infants and young children. J. Clin. Microbiol. 20: Urasawa, S., T. Urasawa, and K. Taniguchi Three human rotavirus serotypes demonstrated by plaque neutralization of isolated strains. Infect. Immun. 38: Urasawa, T., S. Urasawa, and K. Taniguchi Sequential passage of human rotavirus in MA-104 cells. Microbiol. Immunol. 25: Ushijima, H., K. Bosu, T. Abe, and T. Shinozaki Suspected rotavirus encephalitis. Arch. Dis. Child. 61: van Loon, A. E., R. Maas, R. T. M. J. Vaessen, A. M. C. B. Reemst, J. S. Sussenbach, and T. H. Rozijn Cell transformation by the left terminal regions of the adenovirus 40 and 41 genomes. Virology 147: van Loon, A. E., T. H. RoziJn, J. C. de Jong, and J. S. Sussenbach Physicochemical properties of the DNAs of the fastidious adenovirus species 40 and 41. Virology 140: Vesikari, T Progress in rotavirus vaccination. Pediatr. Infect. Dis. 4: Vesikari, T., E. Isolauri, A. Delem, E. D'Hondt, F. E. Andre, G. M. Beards, and T. H. Flewett Clinical efficacy of the RIT 4237 live attenuated bovine rotavirus vaccine in infants vaccinated before a rotavirus epidemic. J. Pediatr. 107: Vesikari, T., E. Isolauri, A. Delem, E. D'Hondt, F. E. Andre, and G. Zissis Immunogenicity and safety of live oral attenuated bovine rotavirus vaccine strain RIT 4237 in adults and young children. Lancet ii: Vesikari, T., E. Isolauri, E. D'Hondt, A. Delem, F. E. Andre, and G. Zissis Protection of infants against rotavirus diarrhoea by RIT 4237 attenuated bovine rotavirus strain vaccine. Lancet i: Vesikari, T., A. Z. Kapikian, A. Delem, and G. Zissis A comparative trial of rhesus monkey (RRV-1) and bovine (RIT 4237) oral rotavirus vaccines in young children. J. Infect. Dis. 153: Vesikari, T., M. Maki, H. J. Sarkkinen, P. P. Arstila, and P. E. Halonen Rotavirus, adenovirus, and non-viral enteropathogens in diarrhea. Arch. Dis. Child. 56: Vesikari, T., T. Ruuska, H. Bogaerts, A. Delem, and F. Andre Dose-response study of RIT 4237 oral rotavirus vaccine in breast-fed and formula-fed infants. Pediatr. Infect. Dis. 4: Voller, A., A. Bartlett, D. E. Bidwell, M. F. Clark, and A. N. Adams The detection of viruses by enzyme-linked immunosorbent assay (ELISA). J. Gen. Virol. 33: Vollett, J. J., III, H. L. DuPont, and L. K. Pickering Nonenteric sources of rotavirus in acute diarrhea. J. Infect. Dis. 144: von Bonsdorff, C.-H., T. Hovi, P. Makela, and A. Morttinen Rotavirus infections in adults in association with acute gastroenteritis. J. Med. Virol. 2: Wadell, G Molecular epidemiology of human adenoviruses. Curr. Top. Microbiol. Immunol. 110: Wadell, G., M.-L. Hammarskjold, G. Winberg, T. M. Varsanyi, and G. Sundell Genetic variability of adenoviruses. Ann. N.Y. Acad. Sci. 354: Wall, R. A., B. J. Mellars, P. Luton, and S. Boulding Comparison of ELISA, SPACE, and electron microscopy for the routine diagnosis of rotavirus infection. J. Clin. Pathol. 35: Walther, F. J., C. Bruggeman, M. S. M. Daniels-Bosman, S. Pourier, G. Grauls, F. Stals, and A. V. D. Boggard Symptomatic and asymptomatic rotavirus infections in hospitalized children. Acta Paediatr. Scand. 72: Ward, R. L., D. I. Bernstein, E. C. Young, J. R. Sherwood, D. R. Knowlton, and G. M. Schiff Human rotavirus studies in volunteers: determination of infectious dose and CLIN. MICROBIOL. REV. serological response to infection. J. Infect. Dis. 154: Ward, R. L., D. R. Knowlton, and M. J. Pierce Efficiency of human rotavirus propagation in cell culture. J. Clin. Microbiol. 19: Weinberg, R. J., G. Tipton, W. J. Klish, and M. R. Brown Effect of breast-feeding on morbidity in rotavirus gastroenteritis. Pediatrics 74: Weiss, C., and H. F. Clark Rapid inactivation of rotaviruses by exposure to acid buffer or acidic gastric juice. J. Gen. Virol. 66: White, L., I. Perez, M. Perez, G. Urbina, H. Greenberg, A. Kapikian, and J. Flores Relative frequency of rotavirus subgroups 1 and 2 in Venezuelan children with gastroenteritis as assayed with monoclonal antibodies. J. Clin. Microbiol. 19: Whitelaw, A., H. Davies, and J. Parry Electron microscopy of fatal adenovirus gastroenteritis. Lancet i: Williams, E. K., J. A. Lohr, and R. L. Guerrant Acute infectious diarrhea. II. Diagnosis, treatment, and prevention. Pediatr. Infect. Dis. 5: Wong, C. J., Z. Price, and D. A. Bruckner Aseptic meningitis in an infant with rotavirus gastroenteritis. Pediatr. Infect. Dis. 3: Wood, D. J., and A. S. Bailey Detection of adenovirus types 40 and 41 in stool specimens by immune electron microscopy. J. Med. Virol. 21: Woode, G. N., J. C. Bridger, J. M. Jones, T. H. Flewett, A. S. Bryden, H. A. Davies, and G. B. B. White Morphologic and antigenic relationships between viruses (rotaviruses) from acute gastroenteritis of children, calves, piglets, mice, and foals. Infect. Immun. 14: Woode, G. N., N. E. K. Gourley, J. F. Pohlenz, E. M. Leibler, S. L. Marhews, and M. P. Hutchinson Serotypes of bovine astrovirus. J. Clin. Microbiol. 22: Wyatt, R. G., R. Dolin, N. R. Blacklow, H. L. DuPont, R. F. Buscho, T. S. Thornhill, A. Z. Kapikian, and R. M. Chanock Comparison of three agents of acute infectious nonbacterial gastroenteritis by cross-challenge in volunteers. J. Infect. Dis. 129: Wyatt, R. G., H. B. Greenberg, W. D. James, A. L. Pittman, A. R. Kalica, J. Flores, R. M. Chanock, and A. Z. Kapikian Definition of human rotavirus serotypes by plaque reduction assay. Infect. Immun. 37: Wyatt, R. G., W. D. James, E. H. Bohl, K. W. Theil, L. J. Saif, A. R. Kalica, H. B. Greenberg, A. Z. Kapikian, and R. M. Chanock Human rotavirus type 2: cultivation in vitro. Science 207: Wyatt, R. G., H. D. James, Jr., A. L. Pittman, Y. Hoshino, H. B. Greenberg, A. R. Kalica, J. Flores, and A. Z. Kapikian Direct isolation in cell culture of human rotaviruses and their characterization into four serotypes. J. Clin. Microbiol. 18: Wyatt, R. G., A. R. Kalica, C. A. Mebus, H. W. Kim, W. T. London, R. M. Chanock, and A. Z. Kapikian Reoviruslike agents (rotaviruses) associated with diarrheal illness in animals and man. Perspect. Virol. 10: Yamaguchi, H., S. Inouye, M. Yamauchi, T. Morishima, S. Matsuno, S. Isomura, and S. Suzuki Anamnestic response in fecal IgA antibody production after rotaviral infection of infants. J. Infect. Dis. 152: Yolken, R. H Avidin-biotin radioimmunoassay for human rotavirus. J. Infect. Dis. 148: Yolken, R. H., H. W. Kim, T. Clem, R. G. Wyatt, A. R. Kalica, R. M. Chanock, and A. Z. Kapikian Enzyme-linked immunosorbent assay (ELISA) for detection of human reovirus-like agent of infantile gastroenteritis. Lancet ii: Yolken, R. H., F. Lawrence, F. Leister, H. E. Takiff, and S. E. Strauss Gastroenteritis associated with enteric type adenovirus in hospitalized infants. J. Pediatr. 101: Yolken, R. H., and F. J. Leister Evaluation of enzyme immunoassays for the detection of human rotavirus. J. Infect. Dis. 144: Yolken, R. H., G. A. Losonsky, S. Vonderfecht, F. Leister, and

39 VOL. 2, 1989 VIRAL GASTROENTERITIS 89 S.-B. Wee Antibody to human rotavirus in cow's milk. N. Engl. J. Med. 312: Yolken, R. H., L. Mata, B. Garcia, J. J. Urrutia, R. G. Wyatt, R. M. Chanock, and A. Z. Kapikian Secretory antibody directed against rotavirus in human milk-measurement by means of enzyme-linked immunosorbent assay (ELISA). J. Pediatr. 93: Yolken, R., and M. Murphy Sudden infant death syndrome associated with rotavirus infection. J. Med. Virol. 10: Yolken, R. H., and P. J. Stopa Analysis of nonspecific reactions in enzyme-linked immunosorbent assay testing for human rotavirus. J. Clin. Microbiol. 10: Yolken, R. H., R. G. Wyatt, A. R. Kalica, H. W. Kim, C. D. Brandt, R. H. Parrott, A. Z. Kapikian, and R. M. Chanock Use of a free viral immunofluorescence assay to detect human reovirus-like agent in human stools. Infect. Immun. 16: Yolken, R. H., R. G. Wyatt, and A. Z. Kapikian ELISA for rotavirus. Lancet ii: Yolken, R. H., R. G. Wyatt, H. W. Kim, A. Z. Kapikian, and R. M. Chanock Immunological response to infection with human reovirus-like agent: measurement of anti-human reovirus-like agent immunoglobulin G and M levels by the method of enzyme-linked immunosorbent assay. Infect. Immun. 19: Yolken, R. H., R. G. Wyatt, G. Zissis, C. D. Brandt, W. J. Rodriquez, H. W. Kim, R. H. Parrott, J. J. Urrutia, L. Mata, H. B. Greenberg, A. Z. Kapikian, and R. M. Chanock Epidemiology of human rotavirus types 1 and 2 as studied by enzyme-linked immunosorbent assay. N. Engl. J. Med. 299: Zentner, B.-S., M. Margalith, A. Galil, B. Halevy, and I. Sarov Detection of rotavirus-specific IgG antibodies by immunoperoxidase assay and enzyme-linked immunosorbent assay. J. Virol. Methods 11: Zissis, G., and J. P. Lambert Different serotypes of human rotavirus. Lancet i: Zissis, G., and J. P. Lambert Enzyme-linked immunosorbent assays adapted for serotyping of human rotavirus strains. J. Clin. Microbiol. 11: Zissis, G., J. P. Lambert, and D. Deicegel Routine diagnosis of human rotavirus in stools. J. Clin. Pathol. 31: Downloaded from on September 25, 2018 by guest