Immunoprophylaxis of Viral Hepatitis

Transcription

1 GASTROENTEROLOGY 77: , 1979 PROGRESS ARTICLE Immunoprophylaxis of Viral Hepatitis LEONARD B. SEEFF and JAY H. HOOFNAGLE The Medical Services, Veterans Administration Medical Center and Georgetown University School of Medicine; and the Liver Section Digestive Diseases Branch, National Institutes of Arthritis, Metabolic and Digestive Diseases, Bethesda, Maryland Historical Background The concept of immunotherapy and immunoprophylaxis for infectious diseases was introduced to clinical medicine in 1890, when von Behring and Kitasato reported that they had developed an antitoxin against diphtheria." The work they performed was so highly regarded that von Behring was selected to receive the Nobel Prize for Physiology and Medicine in 1901, the first year of the award. The immunologic basis of these studies remained valid, but the antitoxin itself proved to be of only limited use. Because it was a rabbit antiserum, 'it produced frequent and often life threatening hypersensitivity reactions in humans. Immunoprophylaxis with material derived from human sources was first used in the early 1930s. Among the studies performed was one by McKhann and Chu who demonstrated that a globulin fraction, prepared by ammonium sulfate precipitation of placental extract, could attenuate and even prevent the development of measles in susceptible persons! A relevant advance in methodology came in 1936 when Tiselius reported that serum proteins could be separated and purified by electrophoresis. 3 Shortly thereafter, he and Kabat showed that the serum antibodies were concentrated in the gamma globulin fraction of the proteins.- Human gamma globulin concentrated by this method was also found to be effective for the prophylaxis of measles. 5,6 Furthermore, because the globulin was derived from human sources, the problem of hypersensitivity was, found to be far less prevalent than had been noted with the heterologous antitoxins. Before the wide-scale use of gamma globulin Received October 24, Accepted March 20, Address requests for reprints to: Leonard B. Seeff, M.D., Veterans Administration Medical Center, Hepatitis Research (151W), 50 Irving Street, N.W., Washington, D.C We wish to thank Ms. Erma Price for her unfailing help and dedication in the preparation and typing of this manuscript. could become practical, however, it was necessary to develop more efficient and safe methods of plasma fractionation. Although several were devised,7 it was the procedure reported by Cohn et al. in 1944 that proved to be the most efficient, practical, and durable." Indeed, despite the many advances in protein chemistry and technology in the p ~ s30 t years, this method continues to be the one ~ o frequently employed in the United. States. In t h I ~ r o c e s t dure, the various plasma fractions are precipitated by adjusting the ionic strength, the temperature, and the ethanol concentration. The result is the production of gamma globulin (Cohn Fraction II) with antibody concentrated approximately 25 times that of the originap The antibody spectrum and concentration are enhanced also by the fact that large volumes of plasma can be handled by this process. An important attribute of the method is that the globulin produced almost never transmits hepatitis, even when derived from icterogenic plasma.1o Description of Gamma-Globulin Preparations Immune Serum Globulin (Human), ISG The product in use in the United States is licensed by the Bureau of Biologics, Food and Drug Administl'ation. The manufacturers of ISG derive their source material from out-dated donor blood, from commercial donors who undergo plasmapheresis specificahy for this purpose, or from placental material. The final product is a sterile solution, slightly opalescent, colorless, or light amber in color, obtained from a minimum of 1,000 donors. It contains approximately 165 mg/ml of gamma globulin, is stabilized with glycine, and has thiomerosal (a mercury derivative) added as a preservative. Federal regulations require that ISG be standardized byhaving defined titers of antibody against diphtheria, measles, and poliovirus, type I. Standards for titers of antibody to the hepatitis A virus (anti-hav) and

2 162 SEEFF AND HOOFNAGLE GASTROENTEROLOGY Vol. 77, No.1 the hepatitis B virus (anti-hbs) have not yet been established. Recent studies indicate that most ISG preparations contain anti-ha V, but that the titer varies considerably between lots and manufacturers, both in this country" and abroad. 12 The titers, measured by immune adherence hemagglutination (IAHA), have ranged from 1: 220 to 1: 16,000. Surveys in the United States have demonstrated that the prevalence of anti-hav is highest among older persons from low socioeconomic backgrounds It is evident that this fact will need to be borne in mind in the future when selecting donors for ISG preparation. The minimum effective anti-ha V titer has yet to be established. In three studies, permitting retrospective analysis, ISG with IAHA titers of 1: 10,000,'5 1: 3200,'6 and 1: 2000,17 respectively, were found to be effective in preventing type A hepatitis. Other studies are in progress to better define the necessary anti-ha V titer. Since 1972, Federal regulations have required that all blood and plasma donors be screened for the presence of the hepatitis B surface antigen (HBsAg). However, even before this time, ISG prepared by Cohn fractionation rarely contained detectable HBsAg. Screening of over 1000 samples submitted during a 15-yr period to the Bureau of Biologics 'revealed that only 0.8% of ISG lots were HBsAg-positive, all of them in low titer.'b Furthermore, early experiments involving fractionation of strongly HBsAg-positive plasma demonstrated that <0,1% of the original HBsAg is precipitated in the cold ethanol fraction II with most being distributed to later fractions. '9. 2o Although the HBsAg-reactivity may be specific (Holland PV, personal communication), its presence in ISG should not necessarily be equated with infectivity, since HBsAg itself is not the hepatitis B virus and exists predominantly in a noninfectious form. Relevant to this are the data from a recent study in. which four plasma units with high HBsAg and hepatitis B e antigen (HBeAg) titers, and containing specific DNA polymerase activity and visible Dane particles (the presumptive hepatitis B virus), were subjected to cold ethanol fractionation. 21 The resultant fraction II was found to be free of detectable HBsAg, HBeAg, DNA polymerase, and Dane particles; these HBV reactivities were found primarily in fraction III. Nevertheless, there have been rare instances of viral hepatitis attributed to the use of ISG. The best characterized was an outbreak of type B hepatitis in Italy thought to have been caused by an anti allergy drug which contained ISG!2 The ISG, produced many years ago in this country, has been found to contain HBsAg. Most current lots of ISG have detectable anti HBs.'B.23 Studies show that the prevalence and tit.er of this antibody have increased since 1972 from a mean of 1 : 4 to a mean of 1 : 64, as measured by passive hemagglutination (PHA) (Figure 1). This increase coincided with the introduction of routine screening of donor blood for HBsAg. It has been speculated that the previous low prevalence and titer of anti-hbs might have resulted from its having been complexed with HBsAg.'B Hepatitis B Immune Globulin. HBIG The theoretical possibility that high titers of anti-hbs might prove more effective in preventing type B hepatitis has led to the production of HBIG, using donors with known high anti-hbs titers. Titers of anti-hbs in these preparations have ranged from 1: 25,000 to 1: 1,000,000 by PHA. The current United States standard has a titer of approximately 1: 100,000 (Gerety RJ, personal communication). Half-Life of ISG and HBIG The rate of ISG degradation is important as a guide to planning its frequency of administration. Dixon et ai., using I'31-labeled ISG, determined the half-life after intramuscular injection to be 13.1 days in adults and 20,3 days in children!4 Havens et ai., using a similar technique, reported a half-life of 13.7 days in normal adults and 9.1 days in adults with cirrhosis. 25 A turnover time of days was found by Armstrong et ai. with intravenous I'31-labeled material, but it was days with ingestion of S35_la_ beled yeast!6 Although similar results were obtained by Vol wiler et al. with oral and intravenous infusions of S35-labeled cystine, they believed that the available techniques did not permit accurate determinations. 27 Indeed, Waldman and Strober have indicated that in early studies, proteins were damaged by the isolation and radiolabeling procedures, and consequently, survival was found to be shorter than that observed with biosynthetic labels!8 In their summary of several later studies in which the IgG was isolated and labeled by a more sophisticated technique!9 they note that the mean half-life of IgG was found to be 21 days. Biologic assay of anti-hbs aft,er injection of HBIG suggests that its half-life is approximately 27 days.30 Adverse Reactions Adverse reactions to ISG are uncommon when it is administered by the recommended intramuscular or subcutaneous routes. Janeway reported the development of adverse reactions in 1.2% of persons receiving intramuscular injections for measles prevention. 31 The frequency of adverse reactions

3 July 1979 IMMUNOPROPHYLAXIS OF VIRAL HEPATITIS 163 <Il CO I, i= ~ 100 a:: a LL ~ 80 E (J) a ~ 60 I- a...j CD (J)4Q LL a I- z a:: w a.. ~ 20 $: m ~ Z ::0 648 ij ::0 a n 32 ~ r -i -i m ::0 16 a "'T1 (f) C) 8 5 -i (f) ij I 4 ~ OL-L- L- ~...JL -L ~ ~ L ~ YEAR OF SUBMISSION Figure 1. Prevalence and mean titer of anti-hbs in commercial ISG lots by year of submission to the Bureau of Biologics, Food and Drug Administra tion. 18 I among 2551 ISG recipients in a nation-wide transfusion hepatitis study was 1%.32 These included pain at the injection site, hematomas, arthralgias, urticaria, rash, and unexpl.ained fever. Similar mild reactions were noted in 0.8% of 1094 ISG recipients in a Veterans Administration cooperative study of posttransfusion hepatitis. 33 However, the same reactions were reported by 0.7% of 1110 patients in that study who had received an albumin placebo. In children with IgA deficiency, allergic reactions are more common, particularly when ISG is administered frequently.34. In contrast, the toxicity rate is very high when ISG is administered intravenously. Barandun and coworkers, in studies of intravenous administration of ISG, reported the occurrence of toxic reactions in 30% of normal people and in 92% of those with antibody deficiency.35 The sequence of events noted in, their studies was an initial feeling of uneasiness, followed by tachycardia, tachypnea, flushing of the face, a feeling of oppression in the chest, and lumbar pain with progression in 1-2 hr to chills, fever, pallor, and fatigue. Occasionally, anaphylactic shock supervened. It is on the basis of such findings that the intravenous administration of ISG is proscribed. Although ISG can be modified for safe intravenous use,36 these preparations are not available in the United States. Rarely, anaphylactic shock and angioneurotic edema have been reported even after intramuscular or subcutaneous injections of ISG It is possible that some instances of development of this reaction might be accounted for by inadvertent intravenous injection. It has also been suggested that the preservative in commercial ISG preparations-thiomerosal (Merthiolate)-might cause an allergic response.'9 The use of placentally derived globulin may bring its own side effects. Because this form of ISG contains blood group substances, there is the danger of isoimmunization, with its attendant risks. There are no definitive contraindications to the use of ISG or HBIG. Of importance is that they both appear to be s a for f ~ use in pregnant women. There has been theoretical concern that the administration of HBIG to persons who are already HBsAg-positive might induce i m m u n ~ - diseases. c o m p Instances l e x of fever which have followed the intravenous administration of HBIG to fibv carriers might indeed represent an immunologic reaction However, as pointed out recently by Grady, there have been a number of fortuitous circumstances in which HBIG has been given inadvertently to HBV carriers without obvious ill effects.43 Nevertheless, because of the threat of this potentially severe adverse reaction, HBIG should not be given to a known HBV carrier. Another concern that had been expressed was that

4 164 SEEFF AND HOOFNAGLE GASTROENTEROLOGY Vol. 77, No.1 HBIG might favor..the progression of acute to chronic hepatitis, and perhaps the development of the HBV -carrier state. 44 However, this too has proved to be unfounded. Clinical Uses of Gamma Globulin During the past 30 years, gamma globulin has been evaluated both for the treatment and the prevention of several viral and bacterial diseases. However, it has proved to be less successful as a therapeutic than as a prophylactic agent. Examples of this are its demonstrated lack of benefit for the treatment of acute self-limited hepatitis,45 acute fulminant hepatitis and chronic hepatitis At present, the accepted indications for use of ISG are for the prophylaxis of viral hepatitis and measles, and as replacement therapy in persons with hypo- and agammaglobulinemia. Hyperimmune globulins are also available for the prevention and/or treatment of certain specific disorders, e.g., tetanus, rabies, pertussis, Rh incompatibilities, etc. In its role as a prophylactic agent, ISG can be given after exposure has already occurred, or it may be administered in anticipation of exposure. The former situation is referred to as postexposure prophylaxis, and the latter as preexposure prophylaxis. Gamma Globulin for Prophylaxis of Viral Hepatitis The present discussion is confined to a review of the use of ISG and HBIG for the prophylaxis of viral hepatitis. It has been organized by assembling the data separately for each type of hepatitis (types A, B, and non-a, non-b hepatitis). However, it must be remembered that the early studies predated the availability of specific serologic tests. Accordingly, investigators had to rely on the imprecise method of epidemiologic.analysis to define the hepatitis types. It is now recognized that a distinction on this basis was frequently in error. This short-coming must be taken into account when interpreting reported data. Prophylaxis of Type A Hepatitis Stokes and Neefe were the first investigators to evaluate ISG for the prophylaxis of viral hepatitis.47 The setting they chose was an extensive hepatitis outbreak which occurred in 1944 in a summer camp in the Pocono Mountains near Philadelphia. The source of the infection was later traced to contaminated well water.4b The investigators had sufficient ISG available to permit administration to 53 young campers and their adult counselors. During a follow-up period, hepatitis developed significantly less frequently in the inoculated individuals than in a group of 278 boys, girls, and adults not given ISG (Table 1). This positive result was supported by the data of two other studies undertaken in the same year. In one of them, Gellis et al. reported ISG to be effective in preventing hepatitis in adults in the United States Armed Forces,49 and, in the other, conducted by Havens and Paul, it was shown to prevent hepatitis among institutionalized children 50 (Table 1). Thus, the effectiveness of ISG in preventing epidemic (type A) viral hepatitis seemed established within a short period of time. Subsequent studies have concentrated on defining the circumstances of spread of type A hepatitis and hence the indications for use of ISG, the optimum dose, the degree of effectiveness, and the mechanism of action of ISG. Circumstances of Transmission of Type A Hepatitis It has been demonstrated repeatedly that type A hepatitis is spread by the fecal-oral route Accordingly, for transmission of the disease to occur, it is necessary that there be contact with the infected stool. This may result from an exposure which is direct (person-to-person exposure) or indirect, through ingestion of contaminated food or water (common vehicle exposure). It is therefore apparent that the disease is most likely to affect persons who have close and prolonged contact with a patient who is incubating or who has type A hepatitis. This includes all household members and large groups of institutionalized persons living together in enforced close proximity, e.g., in institutions for the mentally retarded, army camps, prisons, etc. Furthermore, in some of these situations, the chances of spread are greatly enhanced by poor personal hygiene and by substandard sanitation. The period of greatest danger in person-to-person exposure is just before the onset of overt illness, since fecal excretion of virus occurs from approximately 2 wk before to about 1 wk after the onset of jaundice. 55 Official r e c o m ~ e nfor d ause t iof oisg n s in prevention of type A hepatitis. POSTEXPOSURE PRO PHYLAXIS. At present, p o s t ~ ~ prophylaxis p o s u r e for type A hepatitis is recommended by the United States Public Health Service (PHS, US DHEW) only for contacts in private households-both permanent and temporary household members having been shown to be at risk-and for contacts in large institutions, such as prisons or institutions for the mentally retarded (Table 2).56 Indeed, numerous studies have demonstrated the efficacy of ISG administered for these forms of exposure!5, By contrast, transmission of the disease is uncom-

5 July 1979 IMMUNOPROPHYLAXIS OF VIRAL HEPATITIS 165 Table 1. Efficacy of Ise in the Prevention of Type A Hepatitis Initial Trials Incidence of icteric hepatitis ('Yo) Investigators Group studied ISG dose Stokes and Neefe 47 Summer campers 0.3 ml/kg Gellis et al,, 9 U.S. soldiers 10.0ml Havens and PauP o Institutionalized 0.16 ml/kg children No. patients NoISG ISG , mon to casual contacts of a single case in schools, offices, factories, or hospitals. Accordingly, ISG is not recommended for routine use in this setting, unless an obvious epidemic develops (Table 2). Although the use of ISG routinely had been recommended in the past 64 if a hepatitis outbreak could be traced to apparently contaminated water65 or foodstuffs,66.67 this is no longer the case. 56 This altered decision resulted from the evidence that organized inoculation programs have often been unproductive because of the prolonged lag time in identifying the infectious source, and because of the complex logistics required to mount a large-scale inoculation program. 67 PREEXPOSURE PROPHYLAXIS. In 1961, Hillis reported an outbreak of hepatitis at an Air Force Base among human handlers of chimpanzees. 66 By 1973, 198 cases had been identified as havi.ng been spread in this manner, and it was estimated that hepatitis developed in approximately 5% of direct handlers of these animals o The species involved has been the higher apes, and the illness identified as type A hepatitis. It has therefore been recommended that all persons working directly with newly imported primates receive ISG at 4-6 mo intervals (Table 2).56 Another defined high risk group is travelers to developing countries in which viral hepatitis is endemic because of substandard sanitation. The likelihood of transmission is greater if the stay is prolonged, as has been demonstrated in studies involving United States Armed Forces personnel, State Department employees, Peace Corps workers, and missionaries Of note is that the attack rate appears to remain constant for the first 10 yr of residence abroad. 76 The incidence is low in travelers to Europe, but is higher in persons residing for prolonged periods in North Africa, the Middle East, India, the Philippines, or Central and South America. Immunoprophylaxis has been documented to afford protection in these circumstances,71-76 and it is therefore recommended that travelers to tropical areas and developing countries who bypass the usual tourist routes receive ISG at 4-6 mo intervals 56 (Table 2). Optimum Dose In the first study by Stokes and Neefe of a hepatitis outbreak in a summer camp, an arbitrarily chosen dose of 0.15 mljlb (equivalent to 0.3 mljkg*) was found to be 90% effective in reducing the hepatitis incidence. 47 Because of the high cost and limited supply of ISG in the 1950s and 1960s, other investigations were undertaken to define a lower effective dose. A later study by Stokes and co-workers of an epidemic in a training school for children,77 and by Hsia et al. of family contacts of children with acute hepatitis,58 showed that a dose of 0.02 mljkg was equally effective. However, Krugman et al. found that this dose was only partially effective in protecting inmates of the Willowbrook State School against the development of endemic hepatitis and not at all effective for the adult employees. 78 The in- * All future doses will be reported as milliliters per kilograms. Table 2. Ise for the Prevention of Type A Hepatitis Recommendation of the U.S. Public Health Service Advisory Committee on Immunization Practices 56 Recommended for use Type of exposure Circumstance of exposure Dose (ml/kg) Frequency of administration Yes No Postexposure Preexposure Postexposure Contact with persons with acute type A hepatitis (households, institutions for mentally retarded, prisons) Workers with nonhuman primates Travelers to tropical and developing countries Visit for <3 mo Visit for >4 mo Casual school, hospital, office and factory contacts (unless overt epidemic develops) Common vehicle exposure Once 4-6 monthly Once 4-6 monthly

6 166 SEEFF AND HOOFNAGLE GASTROENTEROLOGY Vol. 77, No.1 mates were protected only when a higher dose of 0.12 ml/kg was used. The initial poor response was ascribed to the low dose used and to the differential degrees of severity of hepatitis between adults and children. However, it has since been recognized that the endemic hepatic disease at the Willowbrook School is not only type A hepatitis, but is also type B, and probably non-a, non-b hepatitis. 79 More recent trials suggest that larger doses are necessary if there is to be any chance of success in preventing the non-a forms of viral hepatitis Three studies involving clear-cut outbreaks of type A hepatitis, one an epidemic in an institution for the mentally retarded,81 another, an epidemic in the general population,82 and the third, a study of family contacts of young children with hepatitis,83 have all shown protection with an even lower dose of 0.01 ml/kg body weight. However, the evidence in one of them was that greater protection was afforded by a dose of 0.02 ml/kg.81 Subsequent field experience suggests that effective postexposure prophylaxis is achieved with a dose of 0.02 ml/kg body weight (Table 2). A widely accepted approach is to give 0.5 ml to children weighing <50 pounds, 1.0 ml to those between 50 and 100 pounds, and 2.0 ml to persons >100 pounds in weight. 54 The dose for pre exposure prophylaxis varies according to body weight and length of exposure, as shown in Table 2. It is apparent that the aims of treatment are to administer antibody in an amount sufficient to protect against the disease. Now that serologic tools are available to measure hepatitis antibody levels in gamma globulin and in serum, further studies will permit a more rational approach to the selection of dose, taking into account the antibody titer of the ISG. Ultimately, it is likely that the recommended dose will be expressed in terms of units or micrograms of specific antibody. Degree of Efficacy of ISG Efficacy is defined here as the ability to protect against development of overt clinical viral hepatitis. Not taken into account by this formulation, however, is the prevention of subclinical viral hepatitis and the induction of long-lasting immunity. Although all of these factors are of great importance, it is the prevention of overt disease that should be the prime measure of efficacy. Indeed, until recently, this had been the only measure available. Even though it has not been possible until recently to test for anti-ha V titers in ISG, field trials have established ISG to be highly effective in the prophylaxis of type A hepatitis. Using the formula: Efficacy (%) = Attack rate of untreated - Attack rate of treated Attack rate of untreated X 100 to evaluate the results of trials conducted in the United States, it has been shown that ISG is capable of reducing the incidence of type A hepatitis by 80-90%.84 Similar results have come from studies conducted abroad. In one of these, reported from New Zealand, ISG was estimated to be 80% effective. 85 In another, a British Research group observed that ISG preparations used in 87 controlled trials undertaken in that country had an efficacy ranging between 84 and 87%.86 Factors that might influence efficacy include the dose of ISG and its anti-ha V titer, the timing of administration, the method of preparation of ISG, and its stability. The role of dose and of anti-hay titer has already been discussed. With regard to timing, it is generally accepted that the efficacy of passively transferred antibody is highest the closer in time it is given to the moment of exposure. It must be apparent that mass inoculation is unlikely to produce total protection. Invariably, there will be exposed individuals who are already too far along in the incubation period for virus neutralization to occur or to be effective. Cases of hepatitis that develop within 7-10 days after inoculation probably fall into this category.a7 It is presently the view of the Advisory Committee on Immunization Practices (PHS, US DHEW), that the use of ISG more than 2 wk after exposure is not indicated. 56 However, Mosley and Galambos suggest that it is worth considering the use of ISG up to 6 wk after known exposure. 88 That the method of preparation might influence the efficacy is suggested by the results of a study reported by Mosley et al. 89 They evaluated two lots of ISG and found that one of them, prepared by the Cohn fractionation method, had high levels of several measurable viral antibodies and an 87% protective efficacy agajnst type A hepatitis, whereas the other lot prepared by ammonium sulfate precipitation had low levels of the same antibodies and only a 47% protective e.fficacy against type A hepatitis. The same group of investigators demonstrated that ISG fragmentation did not reduce the efficacy and also that ISG manufactured in one part of the world protected against hepatitis that develops in another part. 90 Mechanism of Action In 1951, Stokes et al. reported that ISG given during a hepatitis epidemic that occurred in three in-

7 July 1979 IMMUNOPROPHYLAXIS OF VIRAL HEPATITIS 167 stitutiohs for the mentally retarded provided protection for 5-9 mo, despite the fact that the inoculated inmates continued to be exposed to cases of hepatitis that developed among the inmates not given ISG. 77 The investigators later used known infectious material to challenge 50 of the inmates who had received the ISG, as well as 4 previously inoculated prison volunteers. Clinical hepatitis developed in 6 (12%) of the former and in all (100%) of the latter group. Using these data and those derived from two other hepatitis outbreaks that they had studied, they attempted to define the mechanism by which the ISG might have afforded protection. The three mechanisms that they considered were passive immunity, active immunity, and a combination of the two that they referred to as passive-active immunity. In passive immunization, protection is conferred by the transmission of neutralizing antibody which blocks both clinical disease and subclinical infection. Because both disease and infection are prevented, long-lasting immunity does not develop. Stokes et al. argued that this mechanism was unlikely because of the prolonged protection of 5-9 mo, far exceeding the half-life of ISG, and because only a small number of previously inoculated children developed hepatitis after challenge with infectious material. In active immunization, protection is induced by the active stimulation of antibody as the result of the presence in the ISG of inactivated, noninfectious virus. This too seemed an unlikely mechanism, in the view of these investigators, since hepatitis could be induced in the previously inoculated persons. Also, it was thought doubtful that the ISG contained sufficient hepatitis A antigen to stimulate an antibody response. Accordingly, they concluded that ISG had probably conferred protection by the mechanism of passive-active immunization. According to this concept, passively transferred antibody from the ISG attenuates the illness, so that infection occurs, but clinical disease is prevented. As the result of the subclinical infection, endogenous long-lasting antibody develops. It must be stressed that in order for ISG to induce passive-active immunization, an infecting exposure must also have occurred. Further compelling evidence supporting t,he view that ISG does not induce active immunization came from the observations of Wehrle and Hammon. 91 They noted that persons who had been given either ISG or gelatin several years earlier during a field trial of prophylaxis for poliomyelitis, subsequently developed hepatitis with equal frequency. Also, support for the concept of induction of passive-active immunization by ISG came from two other studies that were conducted in institutions for the mentally retarded. In one of them, Drake and. Ming found that icteric hepatitis developed less frequently in inmates who had been given ISG than in those who were untreated, but that anicteric hepatitis occurred with equal frequency.81 Similar results were observed by Krugman and co-workers.78 ISG given to patients newly admitted to the Willowbrook State School reduced the incidence of icteric hepatitis for a period of 5 mo. The incidence of anicteric hepatitis (as judged by routine serum aminotrasferase testing) was not altered. However, after 5 mo, icteric hepatitis developed with equal frequency in the inoculated and in the uninoculated children who were continuously exposed to hepatitis in this high risk setting. The investigators reasoned that the incidence of late hepatitis should have been higher in the inoculated than in un inoculated children if the ISG had conferred protection by passive immunization alone. It is apparent that passive-active immunization will result only if the anti-ha V titer in the ISG is just sufficient to suppress the clinical disease, but is not so great as to entirely neutralize the virus and inhibit the infection. Since anti-ha V titers in ISG lots vary, and since ISG is administered at different time intervals in the incubation period, it can be predicted that the mechanism of ISG protection will not be the same in all instances. 92 This seems to be confirmed by the results of recent studies using specific assays for anti-ha V. An example is the study of Krugman, in which he reanalyzed stored sera from his earlier experimental trial of ISG prophylaxis of type A hepatitis at the Willowbrook State School for the mentally retarded '6 (Table 3). He found that all 14 susceptible inmates given 0.1 ml of a 1: 2 dilution of acute-phase serum from a child with type A (MS- 1) hepatitis developed clinical hepatitis and anti HA V. In contrast, none of 8 susceptible children given a mixture of acute-phase serum and ISG (0.2 ml/kg) developed clinical hepatitis, but 2 developed anti-hay. Thus, only 2 of 8 (25%) developed subclinical infection without disease (passive-active immunity), whereas 6 of 8 (75%) were protected from developi1}g both clinical disease and infection (pas- Table 3. Efficacy of ISG in Prevention of Experimentally Ind4ced Type A Hepatitis o No. No. who developed No. susceptible Group subjects subjects b Hepatitis Anti-HAV Controls (6)" 8 (6)" ISG a Krugman, Reference 16. b No pre-existing anti-hav. " Eight developed hepatitis within 45 days, suggesting primary infection from the inoculation, and the remaining 6 developed hepatitis days after inoculation, suggesting secondary infection from exposure to the 8 children with primary infection.

8 168 SEEFF AND HOOFNAGLE GASTROENTEROLOGY Vol. 77, No.1 sive immunization). Remarkably similar results have since been reported by Rakela et ai. 17 Prophylaxis of Type B Hepatitis The earliest studies of prophylaxis of type B hepatitis were conducted in transfused individuals in the mistaken belief that posttransfusion hepatitis and type B hepatitis were synonymous. The first trial of ISG for the prevention of presumed type B hepatitis unrelated to transfusions was performed in 1948 by Stokes et ai. 93 They showed that both ISG and convalescent plasma were ineffective in protecting against hepatitis in volunteers challenged with "icterogenic" plasma. Similarly, in a study reported 5 yr later by Drake et ai., ISG prepared from volunteers convalescing from experimentally induced type B hepatitis failed to protect others challenged with the same virus strain. 90 The anti-hbs titer of the globulin preparations used in these two studies is, of course, not known. In this regard, Kuhns et al. have shown that ISG prepared from donors with a history of hepatitis does not necessarily contain high titers of anti-hbs.g5 A study involving United States soldiers assigned to military duty in Korea was the first to produce evidence that ISG with measurable levels of a'nti HBs might be effective in preventing type B hepatitis.flo In this double-blind trial, 107, 803 soldiers received two injections, 5-7 mo apart, of either 2, 5, or 10 ml of ISG, or of a placebo. The recipients of the 5 ml and 10 ml dose of ISG developed HBsAg-positive and negative icteric hepatitis significantly less frequently than did those who received the 2 ml dose or the placebo. The ISG was later recognized to contain anti-hbs in a titer similar to that found in current ISG preparations. oo In 1971, HBIG was tested for the first time by Krugman et ap7 It had been prepared at the New York Blood Center from a single donor with hemophilia, and contained anti-hbs in a titer by PHA of 1: 256, The HBIG was given to 10 children in a dose of 0.04 ml/kg 4 hr after the experimental injection of MS-2 (HBsAg-positive) serum. Six of the 10 children were completely protected, and 1 of the remaining 4 developed an attenuated illness (60-70% protection). Three of 5 children given MS-2 and conventional ISG (titer of 1: 16) developed hepatitis (40% protection). All 11 children given MS-2 serum without HBIG developed type B hepatitis. Other effects noted in HBIG recipents included a prolongation of the incubation period of hepatitis, and a lower frequency of development of the HBV carrier state (HBIG, 20% vs. controls, 45%). The HBIG recipients all developed early anti-hbs (presumably passively transmitted), which disappeared days (mean of 42 days) later. Anti-HBs reappeared subsequently in 5 of the 6 children who did not develop type B hepatitis, suggesting that passive-active immunization had resulted after HBlG administration. The same HBIG preparation was later evaluated in a trial conducted in three institutions for the mentally retarded. 9 9 In this study, 44 patients newly admitted to these institutions received the HBIG in a dose of 0.02 ml/kg, and 37 received conventional ISG (anti-hbs titer by PHA of 1 : 16). The injections were given at 4 mo intervals, starting 1-2 wk after admission, and the patients received between three and six of them. Serologic and biochemical tests were performed at monthly intervals. Fifty-two patients admitted to one of the institutions and not accepted into the trial were similarly followed as a "control" group, but blood specimens were obtained from them at 3-5 mo intervals. During the follow-up period of mo, type B hepatitis (all cases anicteric) and/or HBsAg-seropositivity only, developed in approximately 14% of the HBIG recipients, in 8% of the ISG recipients, and in 25% of the untreated group. None of the globulin recipients and 14% of the placebo recipients became HBsAg carriers. Antibody seroconversion occurred significantly more frequently in the ISG than in the HBIG group. Thus, these data suggested that HBIG and ISG were equally effective in preventing the development of overt type B hepatitis and of the HBV carrier state. However, the degree of benefit of the globulin preparations could not be determined from this study because a comparably selected and evaluated control group was not examined. The investigators inferred from the high seroconversion rate in the ISG recipients that it had permitted development of passiveactive immunity, whereas HBIG appeared to have blocked infection and hence subsequent passive-active immunization. In light of these provocative results, it became apparent that specific controlled trials would be needed to define more clearly the efficacy of HBIG for prophylaxis against type B hepatitis. Accordingly, a series of stushes was initiated in various parts of the world. Among the eight conducted to date, four have examined HBIG for pre exposure prophylaxis,,00-,03 and four for post exposure prophylaxis The salie'nt features of these studies are outlined in Tables 4 and 5. With the exception of the study of Kleinknecht et al.,102 the follow-up evaluation in all was conducted in a double-blind fashion. It was hoped that the results of these carefully designed trials would be complementary and would provide definitive and consistent information regarding prophylaxis of type B hepatitis, but instead they have conflicted and, if anything, have created greater confusion and uncertainty. The different re-

9 July 1979 IMMUNOPROPHYLAXIS O F VIRAL HEPATITIS 169 Table 4. EffiC(1cy of ISG and HBIG in Prevention of Type B Hepatitis Preexposure Prophylaxis No. Months Investigators Group studied patients followed Desmyter et al 100 Dialysis Patients Iwarson et ai. 'OI High Risk Kleinknecht et al 102 Medical Staff Dialysis Patients Prince et al,03 Dialysis Patients Dialysis Staff Incidence of type B hepatitis Amount and (Incidence of HBsAg) (%) No. of injections HBIG ISG NIL 5ml x2 (13.3) (71.4) 3ml x2 (6.9) (7.8) 3-5ml X9-17 (0.0) (77) 3ml x2 (18.7) (21.7) 3ml X2 (5.6) (11.3) a ISG with no detectable anti-hbs. suits have come, in part, from differences in study design, and the populations studied, in the methods, intensity and duration of follow-up evaluation, and in the assays and criteria chosen to diagnose hepatitis. More important, however, is that the globulin preparations evaluated varied considerably. In the first place, comparison of anti-hbs titers of the HBIG used in the several studies has uncertain meaning, since test methods in different laboratories might not be comparable. Secondly, most of the studies were launched before it had become apparent that conventional ISG manufactured in the United States was undergoing a change, from having no detectable anti-hbs to acquiring slowly increasing titers. Some of the studies chose, quite appropriately, to use the the.n-available anti-hbs-negative ISG as a control globulin. Since this type of ISG was to become obsolete in the United States, the applicability of the results would ultimately be called into question. Results of Preexposure Prophylaxis Studies of HBlG (Table 4) The first randomized, double-blind trial involving dialysis patients was reported in 1975 from Belgium by Desmyter et al. 'oo In this study, HBIG with an anti-hbs titer by PHA of 1: 25,000 was compared with ISG, devoid of detectable anti-hbs, both A second randomized, double-blind trial involving hemodialysis and laboratory staff members was reported by Iwarson et al. from Sweden.,ol Participants received two injections at 3 rna intervals either of HBIG (anti-hbs titer by PHA of 1: 355,000) or of ISG (anti-hbs titer of 1: 100). A group of persons working in the same location who were unwilling to participate in the trial constituted a "control" group. Clinical hepatitis occurred in 1 of the HBIG group (1.7%), in 2 of the ISG group (3.8%), and in 13 of the untreated group (10.4%). The difference between the HBIG and untreated groups was statistically significant, but not between the ISG and the untreated group. There were no cases of persistent antigenemia. As the authors pointed out, the decision to interpret the results of this study as showing benefit of both globulin preparations required comparison with a control group not randomly selected and not followed in the same manner as the inoculated groups. Also, the numbers of persons entered into each inoculation group and the number of cases of hepatitis that developed were too small to have definitive meaning. A third trial in hemodialysis units, conducted in France by Kleinknecht et al., examined the effect of multiple injections of HBIG.102 Fifteen randomly selected patients.received HBIG (anti-hbs titer by PHA of 1: 64,000) in a dose of 0.08 ml/kg every 5 wk for a 6-mo period and then every 2 mo thereafter un the conclusion of the study. Thirteen randomly given at 6 rna intervals. HBsAg and hepatitis d e v~ e ltil oped significantly less frequently in the HBIG than selected patients not given immunoglobulin served in the ISG group. None of the HBIG recipients but as a "control" group. Type B hepatitis developed in 36% of the ISG recipients were still HBsAg-positive 77% of the uninoculated group, but in none of the at the conclusion of the study. The authors suggested that HBIG had both modified and prevented group who developed type B hepatitis became HBIG recipients. Half of the patients in the control hepatitis B infection, with a protection rate of 80%. HBsAg carriers. The conclusion reached was that The shortcomings of this trial were that only a small multiple injections of HBIG appeared to be highly number of persons were studied and that HBIG was protective for chronically exposed hemodialysis patients. The major drawback of this study, in addition not compared with ISG containing current levels of anti-hbs. to its small size, was the absence of a treated control

10 170 SEEFF AND HOOFNAGLE GASTROENTEROLOGY Vol. 77, No.1 group, thus precluding double-blind evaluation and comparison with current ISG. Although the study suggests that multiple injections were highly effective, it should be noted that none of the HBIG recipients developed long-lasting anti-hbs, and thus they remained susceptible unless the injections were continued. The fourth randomized, double-blind trial in hemodialysis units was conducted in the United States by Prince et al. 103 Patients and medical staff received two injections, at an interval of 4 mo, either of high titer (anti-hbs by PHA of 1: 500,000), intermediate titer (anti-hbs of 1: 5,000), or low titer (anti-hbs of 1: 50) globulin. At 8 months after injection, the incidence of type B hepatitis among the patients was 3.3% (high titer), 11.9% (intermediate titer), and 12.0% (low titer). HBsAg alone developed in 2.2%, 5.0%, and 6.5% of them. However, this highly significant effect of the high titer globulin disappeared by the 12th mo, when the combined incidence of type B hepatitis and HBsAg-seropositivity alone was found to be about equal in all three groups (18.7%, 20.8%, and 21.7%, respectively). Similar results were noted for staff members. At 8 months after the injection, high titer globulin was statistically significantly superior to both intermediate and low titer globulin. However, at 12 mo, the difference became statistically' nonsignificant. Nevertheless, the incidence of type B hepatitis in staff members was considerably lower in recipients of high as compared with low titer globulin (4.5% vs. 10.3%). The investigators concluded that HBIG was no better than current ISG in preventing hepatitis among dialysis patients. They could not determine whether the prolonged incubation of type B hepatitis in the HBIG recipients was a result of the HBIG or of reexposure of type B hepatitis. If the latter, they surmised that multiple injections given to these constantly exposed individuals would have been preferable. Although there was a trend to \Yard benefit from HBIG among dialysis staff, the investigators concluded that its routine use was not warranted in view of the high cost that this would entail and of the available evidence that careful preventive measures and the use of home dialysis without immunoprophylaxis are highly effective in curtailing hepatitis in dialysis units. lob While this was the largest and most carefully designed of the four preexposure prophylaxis trials, the study had an important omission, stated to be for ethical reasons. This was the lack of a placebotreated group. Accordingly, since it was reported that there was no significant difference in the degree of protection afforded by the different titered globulin preparations, it is not possible to determine whether they were equally effective or equally ineffective. Results of Postexposure Prophylaxis Studies of HBIG (Table 5) Redeker et al. studied spouses who had had sexual contact with persons with acute hepatitis within 2-4 wk of onset. 104 Persons who entered the study were randomly selected to receive a single injection either of HBIG (anti-hbs titer by PHA of 1: 200,000) or of ISG (prepared from anti-hbs negative plasma units) within 7-30 days of the recognition of type B hepatitis in their partners. Acute type B hepatitis developed in 4.0% of susceptible HBIG recipients and in 27% of susceptible ISG recipients, a significant difference. Thus, this study produced highly compelling data showing benefit of HBIG. However, there have been misgivings expressed about the study. These include the methods used for follow-up, considered by some to be too superficial, as well as the duration of follow-up evaluation. Because it was only a 5-mo period, it could not permit delineation between delayed onset of disease and its actual prevention. However, Mosley has since reported that continued evaluation for 7-12 mo failed to reveal additional cases." l9 There have been two trials designed to evaluate HBIG for prophylaxis in persons or medical person- Table 5. Efficacy of ISG and HBIG in Prevention of Type B Hepatitis Postexposure Pwphylaxis No. Months Investigators Group studied patients followed Redeker et al.' 4 Sexual contacts 58 >5 Grady et al. 105 Accidental 435 >9 needles tick Seeff et al. 106 Accidental 419 >9 needles tick Beasley et al.' 7 Neonates o ISG with no detectable anti-hbs. Amount and No. of injections 5ml Xl 3ml X2 5ml X2 0.2 ml Xl HBIG Incidence of type B hepatitis (Incidence of HBsAg) (%) ISG NIL (3.2) (5.9) (55.9) (75.0)0 (64.1)

11 July 1979 IMMUNOPROPHYLAXIS OF VIRAL HEPATITIS 171 nel expmied accidentally by needle stick to HBsAgpositive blood. In a study conducted by Grady et al.,30,los one of three globulin preparations was given within 7 days of exposure and again days later. The globulin preparations were the same as those used in the hemodialysis study of Prince et al. '03 (high titer-1 : 500,000; intermediate titer- 1: 5,000; low titer-1: 50). At the 6-mo period after study initiation, the incidence of type B hepatitis in the high, intermediate, and low titer globulin groups was 4.5%, 10.2%, and 12.3%, respectively. Additional cases of hepatitis after 6 mo developed only in the HBIG recipents, and hence the significant difference between the high and low titer groups that had been noted at 6 mo changed to a nonsignificant difference at 9 mo (an incidence of 9.8%, 10.2%, and 12.3%, respectively). The late cases in the high titer group were not considered to be caused by a second exposure, but rather to be the result of an extended incubation period induced by the HBIG. A second "needle-stick" trial, a Veterans Administration cooperative study, involved 419 susceptible persons who received either HBIG (anti-hbs titer by PHA of 1: 100,000) or ISG (no detectable anti-hbs) within 7 days of exposure and again approximately 28 days later."l6 At the end of 6 mo, type B hepatitis had developed in 0.9% of HBIG and in 5.9% of ISG recipients. One additional case occurred in the HBIG group after 6 mo, so that at 9 mo the respective incidences were 1.4% and 5.9%, a statistically significant difference. HBsAg without clinical disease developed late in ~ HBIG recipients, but in none of the ISG recipients. All persons recovered, and none became chronic carriers. The data were interpreted as showing HBIG to be clearly superior to the ISG used in the study. Thus, the two needle stick studies, although similar in many regards (design, populations studied, size, execution, etc.), yielded apparently conflicting r ~ s u However, 1 t s. it must be stressed that there were important differences between them that prevents direct comparison. The study of Grady et al. compared a very high titered HBIG with current low titer ISG (no difference between the two being found), whereas the VA study examined a moderate titered HBIG and a placebo ISG, with a finding of superior benefit for HBIG. It may be tempting to combine the information from the two studies and conclude that HBIG is superior to placebo (as shown in the VA study), but that it is no better than currently available ISG (as shown by the study of Grady et al.). However, the two studies were sufficiently dissimilar so as to make extrapolation of data in this manner inappropriate. The major event responsible for the differing outcome of the two studies was that late cases of type B hepatitis developed in HBIG recipients in the study of Grady et al., but not in the Veterans Administration study. The reason for this is unknown. It has been suggested by Grady that the fragmentation of the HBIG that occurred during the course of that study might have interfered with its neutralizing ability.30 Although the two studies differed in regard to HBIG efficacy, there were other important results that were in agreement. These included: (a) the lack of development of type B hepatitis in exposed persons who had preexisting anti-hbs at the time of exposure; (b) a striking correlation between the development of hepatitis and exposure to HBsAg-positive blood that was also positive for HBeAg; and (c) the lack of development of hepatitis in persons exposed to blood from patients with fulminant hepatitis. Both needlestick studies employed a two dose schedule, the injections being given 1 mo apart. To determine whether this was really necessary, Klein et al., using the HBIG of the Veterans Administration trial, conducted another study to compare a single dose (10 ml once only) with a double dose (two 5 ml doses separated by 1 mo) schedule. 110 Type B hepatitis developed in 17% of persons given one injection and in 7% of those given two (P = 0.059). Although the results of this study were comprised by the high dropout rate, the trend was such that a two dose schedule is currently recommended. The fourth postexposure prophylaxis study, reported by Beasley and Stevens, was a double-blind, controlled trial involving newborn infants of chronic HBsAg carrier mothers.,o7 In this study, conducted in Taiwan, the infants were given a single injection within 7 days of birth of one of three preparations- 0.2 ml of HBIG (anti-hbs titer by PHA of 1: 200,(00), 1.0 ml of ISG (with no detectable anti-hbs by PHA), or 1.0 ml of heat sterilized albumin. Hepatitis B virus infection (the development of HBsAg or anti-hbs) occurred with similar frequency in all three injection groups, although its onset was significantly later in the HBIG recipients than in the other two groups. Persistence of antigenemia was also similar in the three gro ups, but in the HBIG recipients, it was influenced by the time of administration of HBIG. Forty-seven percent of those who received HBIG before 48 hr and who developed HBsAg became chronic carriers, whereas the figure was 100% for those who received the HBIG at 48 hr or more after birth. Of note was that transmission was most frequent when the mother was HBeAg-positive, less frequent when she was negative for both HBeAg and anti-hbe, and least frequent when she was only anti-hbe positive. The authors concluded that the single injection of HBIG simply prolonged the incubation period of development of HBsAg in infants