Variola Minor in Braganca Paulista County, 1956: Overall description of the epidemic and of its study

Transcription

1 International Journal of Epidemiology Oxford University Press 1976 Vol.5 No. 4 Printed in Great Britain Variola Minor in Braganca Paulista County, 1956: Overall description of the epidemic and of its study JUAN J. AIMGULO 1 Angulo, J. J. (Institute Adolfo Lutz CP 7027, Sao Paulo, SP, Brazil). Variola minor in Braganpa Paulista County, 1956: overall description of the epidemic and of its study. International Journal of Epidemiology 1976, 5: An overall description of the epidemic of variola minor (alastrim) affecting Braganca Paulista County (Brazil) in 1956 is given. A total of 484 cases were recorded for 210 households, one hotel and one boarding school. At least 95 per cent of the households with cases of the disease in the capital city were surveyed and 90 per cent of the households in the rural districts. An orphanage, an old folk's home, the County jail and 10 schools operating in the capital city and some other social groups without cases were also surveyed, as well as 125 households without cases but with one or more contacts with the disease. An overall attack rate of 1924 cases, 267 cases and 781 cases per 100,000 inhabitants was obtained, respectively, for the capital city, the rural environment and the whole County. Clinical, epidemiological, serological and environmental surveys were conducted. Evidence on identification of the epidemic disease is presented. Study of the epidemic was made at these levels: the disease itself (frequency and severity); the disease in the individual (association of characteristics of persons with occurrence and clinical severity of variola minor); the disease in social units (occurrence and spread in households and school classes); the disease in small communities (occurrence and spread in housing projects, city blocks and farms); the disease in two large communities with contrasting socio-economic characteristics (the capital city and the rural environment); and the disease in the County as a whole. INTRODUCTION In spite of the abundant literature on variola, the epidemiology of this disease is little known because: (a) the bulk of knowledge on variola epidemiology was obtained when epidemiological methodology was poorly developed; (b) except for the aetiological agent, the clinical and pathologic pictures and the effects of vaccination, the knowledge on variola consists of inappropriately documented statements and personal impressions, which frequently contradict each other, transmitted from one textbook to another; (c) most reports on variola epidemics present highly biased pictures of the clinical characteristics of the epidemic, and of the disease itself, because they are based on highly selected samples (individuals decide, based on severity of illness, if admission to hospital or notification of health departments is necessary); (d) in studying the epidemiology of communicable diseases, particularly of variola, advantage has not been taken of the efficient analytical 1 Associate in the Emory University Computer Centre, Atlanta, Georgia, USA. Present address: Instituto Adolfo Lutz, CP 7027, S3o Paulo, SP, Brazil. tools provided by modern statistical methods. In 1956, the Brazilian county of Braganca Paulista was the site of a large epidemic of variola minor (alastrim). This epidemic enabled study of various topics of the occurrence and spread of the disease previously unknown or obscure, and confirmation and extension of known topics. A series of papers describes the occurrence and spread of variola minor during the epidemic and reports a search for associations which might be causal (1^1). This paper presents an overall description of the epidemic and the general plan of study. The Bragcmga Paulista County The County is in the north-eastern portion of the state of Sao Paulo, one of the south-eastern states of Brazil. The County has borders with the state of Minas Gerais, where variola was endemic until its eradication several years after the study epidemic took place. The area of the County is 1006 km 2. Interpolation between censuses in 1950 and 1960 gives an estimated population of 61,948 inhabitants for 1956, the year of the study epidemic. Of these people, 19,241 lived in the capital city (including its semi-urban periphery). The remaining 42,

2 360 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY inhabitants were sparsely distributed in rural settlements (mostly farms and some large estates). Some tiny villages are present in the seven rural districts of the County as well as large farms and estates but only one of these villages was affected, while the disease occurred on many large farms in six of the districts. People of Italian descent made up a considerable proportion of the population, mainly living in the capital city. People of mixed Portuguese, African and Indian ancestry, plus some Japanese and Portuguese made up the remainder. Agriculture, mostly coffee growing, was the main activity of the County population; there were a few small industries in the urban environment. The capital city of the County is connected by rail and road to the capitals of the states of Sao Paulo and Minas Gerais: dirt roads connected the city with the seven rural districts. Roads in rural areas were in poor condition with little motorized traffic. THE EPIDEMIC At the beginning of June 1956, an alarming number of cases of variola minor were occurring in the city of Braganca Paulista, administrative capital of the Braganca Paulista County. The public became aware that a serious situation could arise and health officials were compelled to conduct a mass vaccination campaign. Four patients were sent to an isolation hospital outside the County, while two patients living in a boarding school and another living alone were isolated in the local hospital, Santa Casa de Miseric6rdia. Tracing the chain of contagion revealed that the epidemic originated from a city dweller who visited relatives in a neighbouring county who had contracted the disease. The city dweller developed variola minor a few days after his return to the city on, approximately, 10 November, Transmission to a contact and to other persons occurred sporadically until the disease was introduced into the Jorge Tibirica (JT) School in March, In May, a number of cases occurred and the existence of an epidemic was recognized by both health officials and the public. Cases also occurred in farms in rural districts, particularly during June. In July, August and September there was a continuous decline of the epidemic in the County and the last case occurred on 1st October, Exhaustive search failed to reveal any cases occurring later than 10th October, Inquiries among public health officials and medical practitioners in the surrounding counties failed to disclose evidence of outbreaks, except for sporadic cases; the epidemic had been limited to the Braganca Paulista County. A total of 484 cases was recorded among the inhabitants of the County. An overall attack rate of 781 cases per 100,000 inhabitants was found for the whole County, while for the capital city a figure of 1924 cases per 100,000 inhabitants was obtained. The corresponding rate for the rural environment was 267 cases per 100,000 inhabitants. Clearly, the weight of the epidemic was heavily placed on the population of the capital city of the County. Two deaths were recorded in the County but they could not be directly attributed to variola but to concurrent disease in both cases. Several people suffering from the disease moved within the city or travelled from the city to the rural districts and gave rise to household or farm outbreaks. In some reports on the epidemic these patients may have been counted as members of more than one household but in this report moving or travelling patients were counted in only one household. This explains why slight differences in figures may be found in other reports. No restrictive measure, such as quarantine, was enforced by public health authorities. They advised limited contact with patients but the advice was not followed. This was partly because the eruption was not usually accompanied by any constitutional upset and partly because the authorities announced that the epidemic was not smallpox but alastrim. The meaning of the latter was unknown to lay people, many of whom considered the disease as a type of chickenpox. Consequently, most patients were nursed at home. Only a few patients were isolated from outsiders or relatives. Some young children were moved into their parents' bedrooms for better care, but even in these instances contact with other relatives was not restricted. Schools were not closed and many students attended while still suffering from the disease. Furniture and clothing in contact with patients was only disinfected in two cases. Previous variola minor epidemic in Braganga Paulista County In numerous households persons who had lived in the capital city for more than 20 years referred to 1936 when a sizeable epidemic had previously occurred. The severity of the previous epidemic was confirmed by private practitioners and health officials. Many people reported previous attacks of variola in other counties in years other than 1936 (particularly in the neighbouring state of Minas

3 Gerais). In the rural districts of the County, no one referred to the 1936 epidemic, perhaps because communications between the city and the rural districts were still very limited. On the other hand, small outbreaks in the rural districts were reported. Apparently these outbreaks resulted from importations from Minas Gerais, where variola minor was endemic until its eradication from Brazil. Variola minor in Brazil There is ample evidence that the disease prevalent in Brazil, in 1956, was variola minor (alastrim); the mortality rate reported in other studies (5-9) was typical of variola minor. This finding is supported by the data collected during the Campaign for Eradication of Variola in hundreds of epidemics occurring all over Brazil in the first half of the century. Striking evidence that, early in the 20th Century, the mortality rate from variola suddenly and markedly fell in Rio de Janeiro has been presented by Thibau Jnr (5). The mortality rate among patients admitted to the Emilio Ribas Isolation Hospital, of Sao Paulo City, was much higher early in the present century (11). Since 1930 case fatality rates have constantly been typical of variola minor in Brazil (5-10). These epidemiological findings have been supported by laboratory studies of strains of Poxvirus variolae isolated in S3o Paulo (12), in the northern jungle of Brazil (13) and in several states from northern, southern, eastern and central Brazil (14). In 1956, variola minor was endemic in the whole of Brazil, even in the five most southern states, Rio de Janeiro, Sao Paulo, Parana, Santa Catarina and Rio Grande do Sul, where vaccination had been common practice for years. This striking finding was apparently due to the steady influx of immigrants infected in central and north-eastern Brazil or in Paraguay, where vaccination was restricted to the few urban centres, and variola was endemic in these areas. The campaign aimed at eradication of variola from Brazil is now over; it has apparently been successful (7, 10). A recurrence of epidemicity in Brazil occurred in the study year, In northern and northeastern states, where variola was a common endemic disease, the high incidence caused public alarm. Health authorities from the southern state of Santa Catarina reported numerous epidemics flooding the valley of the Itajai River. In the central state of Minas Gerais, where variola was common, the number and extent of epidemics in 1956 compelled public officials to request help from VARIOLA MINOR IN BRAGANfA PAULISTA COUNTY, official agencies outside the state. In the state of Sao Paulo, bordering Minas Gerais, several outbreaks were reported by health officials from Santos County and outbreaks were reported in an unusually high number of other counties, particularly in the Paraiba River Valley. In the heavily populated city of Sao Paulo a large number of small, circumscribed outbreaks increased the incidence of disease over that of previous years. One of these outbreaks was studied by the author and associates (6, 15-17). The Biritiba Mirim County, close to the city of Sao Paulo, suffered an epidemic originating in a village and extending into the surrounding rural districts; it was also studied by the author and associates (18). Identification of the epidemic disease Characteristics of the epidemic provide convincing evidence that the disease was variola in its'minor'form: (a) the biphasic nature of the disease which included an initial pockless period of systemic manifestations distinct from the characteristic eruption of smallpox itself; (b) the constant occurrence of this pre-eruptive phase and its typical duration, severity and composition; (c) the clear occurrence of well-defined symptoms and the very close agreement of these symptoms with those defined by Dixon (19); (d) the distribution on the body, the individual characteristics of the pocks and the clinical course were typical of variola minor (alastrim); (e) the fatality rate was much less than one per cent, even if an unconfirmed case definitely had variola; (f) there were no other epidemics of exanthematous disease in Braganca Paulista County during the study period apart from small outbreaks of measles and chickenpox when the variola minor epidemic was in decline. This significantly increased the accuracy of identification of cases. The 'external' character of variola and two distinct main phases also made clinical diagnoses fairly reliable; physicians in charge of case identification also had previous experience in this task; (g) the similarity of observed clinical and epidemiological features with those of a simultaneous epidemic occurring in the closely located city of SSo Paulo, where one third of cases had supporting aetiological (laboratory) evidence (6). About 20 per cent of cases during the epidemic

4 362 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY of Braganca Paulista County had supporting laboratory data; (h) many other epidemics occurred in the same and other states of Brazil in 1956, a year of epidemic recurrence; (i) variola virus was the only disease agent isolated by the procedure employed (inoculation of the chick-embryo chorioallantois); (j) significant antibody titres for Poxvirus were found in the sera of most persons with variola, particularly of those without a previous history of the disease or vaccination against smallpox; (k) no pocks were found on any of the persons with a history of previous variola but who currently were exposed to one or more patients with variola; (1) frequent occurrence of variola in persons with a history of varicella but no history of variola; (m) the marked influence of a previous successful vaccination against smallpox on occurrence and clinical severity of the disease; (n) the pattern of transmission in the household, particularly the clear occurrence of waves of cases and the periodicity of these waves. STUDY OF THE EPIDEMIC Plan of study Study of the epidemic was designed to investigate the following factors; (a) the disease itself, description of the clinical manifestations of variola that occurred and analysis of their frequency and intensity; (b) the disease in the individual, association of personal characteristics with: (i) occurrence of variola minor (ii) clinical severity (iii) antibody response of variola minor (including subclinical variola), to vaccination against smallpox and to mixed infections by variola and vaccinia viruses; (c) the disease in social units, such as a household or a school class, where a group of people represent, in study of the disease, more than just a collection of individuals. In a group, study of disease spread would be added to the study of occurrence and clinical severity of the disease would be examined in individuals and for the group; (d) the disease in small communities, the community is taken as a group of social units in a distinct geographical area where relations between the members of the community were closer than the relations with members of other communities. Examples of such communities are a housing project, a farm, a city housing block or groups of blocks. Both individuals and social units in the small community comprised units for study for disease occurrence, spread and clinical severity; group antibody levels (social serology) were also studied. (e) the disease in two large communities with contrasting socio-economic characteristics, such as the communities occupying the capital city and the rural environment of the County. Both communities had similar collections of the smaller communities of (d). They were geographically, sociologically and hygienically distinct from each other; (f) the disease in the whole of the County, it could be regarded as a single community. It is a separate unit in a social and administrative sense, the disease affected the entire county and did not spread into surrounding counties. Time, space, person and social-unit factors possibly influencing the rise and decline of the epidemic were studied as a synthesis of the analyses of disease occurrence and spread in the lower (smaller) units. Among these factors were control measures (such as mass vaccination, school closure). The antibody response of large population groups of the County was another objective. This 'social serology' could be used to study the progress of the epidemic in time and space. The field study To achieve the aims of the plan of the study, collection of data consisted of: (a) a clinical survey identifying cases and grading the clinical severity of cases; (b) an epidemiological survey for characterization of members of every social unit with cases and the number of units without cases (Table I). The behaviour of every member of these social units was also investigated. Besides the persons clinically examined and/or interviewed which appear in Table I, over 50 contacts with cases were recorded but their households (without cases) were not surveyed. Two factories were kept under surveillance but no case occurred among staff who had been compulsorily vaccinated at close intervals; these people were not recorded. Over 11,500 individual records were made, of which more than 4,000 were the students of the 10 schools in the capital city (no rural school was surveyed). School children from households with

5 VARIOLA MINOR IN BRAGANCA PAULISTA COUNTY, TABLE I Overall composition of the study population sample Social groups Households with cases* Households without cases JG School JT School** AG School SCJ School** SL School RB School County jail Orphans' asylum Old folks' asylum Number of units Number of people *** Number of cases 484 Some records were incomplete, hence the actual number was probably higher ** Buildings used by more than one school on different shifts *** Average for the epidemic period cases were recorded twice, at home and at school; (c) a serological survey of various groups from the city and semi-urban and rural districts. Persons with and without variola minor had blood tests, the majority of them only once. A total of 1,392 blood specimens were collected; (d) collection of pathological material for isolation and identification of the infectious agent. This collection was done in 15 cases of variola minor in the city, semi-urban and rural districts; (e) an environmental survey of the County. This survey included mapping of 157 household dwellings, including sleeping arrangements and size of the rooms; maps of 21 classrooms from three elementary schools to show the relative positions of the desks occupied by infecting students and their contacts, including the immunity status of each student; maps of three schools with several cases of disease to determine the spatial relations of classes with and without cases; maps showing the land ownership of all city subdivisions, two housing projects and four farms where cases occurred, showing dwellings with cases and a map of the County showing the limits of rural districts, roads and dwellings with cases; investigation of the living habits of people in city subdivisions, housing projects, rural districts, farms and elementary schools to discover the influence of socio-economic factors on spread of the epidemic; investigation of the hygienic and other conditions in most accommodation with cases of disease and in number of houses without cases but located very close to others with cases. DISCUSSION The initial aim of the study was to discover the mechanism of spread of a large epidemic of variola minor. Epidemics in countries where variola is endemic can be expected to faithfully reflect the natural history of the disease because there is less interference from preventive measures such as vaccination, or isolation of patients and contacts. Up to 1956, modern studies on epidemics of variola were made in countries where variola had been eradicated and outbreaks studied were brought in from endemic countries. The resulting pattern clearly revealed an abnormal situation. It is worth noting that, in 1956, Brazil was the major endemic source of variola in the American continent as well as the major endemic source of variola minor in the World. Study of the epidemic was conceived as an intensive and extensive study. A previous exhaustive review of the literature on variola epidemiology led to the conclusion that knowledge of it was underdeveloped and often of unknown accuracy. Studies on the best known communicable diseases, such as measles, infectious hepatitis, some acute respiratory diseases and mumps, were also reviewed. A list of all the factors investigated revealed numerous gaps in the systematic treatment of occurrence and spread of the better studied diseases. The study of the 1956 epidemic of Braganca Paulista County was designed to provide data for examination and analysis of the known factors as well as several others not yet studied. Analysis of data was extended later according to findings made by the author and associates in

6 364 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY studies of other epidemics of variola minor (3, 6, 15-17, 21). Among the items so included are: personal characteristics which influenced the clinical severity of variola minor characteristics of persons which influenced the length of incubation period, infectious period, latency period and the serial interval between successive cases; seating pattern of infectious and susceptible students and its influence on spread in the school; space-time interactions among cases in classrooms and in household dwellings; time distribution of cases in each school class, each classroom population; each shift and each school population; determination of the population unit showing the pattern of occurrence or spread in schools and in household dwellings the influence of sleeping pattern of infected people and others on spread of disease within households; comparative occurrence and spread of variola minor in people living on both sides of one or more city blocks where all accommodation was surveyed; interaction of personal characteristics with the occurrence or spread of variola minor in schools and households; interactions of personal characteristics, space distribution of cases, time distribution of cases and spacetime interactions among cases in the occurrence and in the spread of variola minor in school or household. Some remarks on the sampling procedure are needed. 'Study of the disease itself will be made on all cases recorded, in households with or without complete surveillance, hence representing practically all cases which really occurred during the 1956 epidemic of Braganca Paulista County. "The disease in the individual' will be studied in cases from households where all members were characterized and surveillance was complete. All but one of the households where one or more cases occurred were carefully surveyed. Since only a small proportion was incompletely surveyed this is a representative sample. Random sampling of the general population of the city or County was not chosen since complete survey of all affected persons, households and school classes seemed feasible. The influence of a given characteristic can only be assessed accurately when the frequency of disease is measured in individuals with or without the given characteristic under presumably equal risk of disease. In the population of households with cases, study of individuals who, under presumably equal risk, did not develop variola minor, was the best way of discovering which characteristics were influencing occurrence of the disease. The field study left no doubt that this epidemic affected only two types of social unit, the household and the school class. Only one case occurred in a hotel and two cases in a boarding school. Other social units such as the County jail, hospitals, asylums, factories, shops and offices were not affected. In the two schools with a number of cases all classes were thoroughly surveyed the other schools had too few cases to obtain meaningful results. All the housing projects where cases occurred were plotted on maps showing land ownership and all persons living in these 'small communities' were characterized and their behaviour during the epidemic studied. This survey was also made in a population group from the most densely inhabited portion of the city capital of the County. Here blocks selected included more houses with cases than any other group of blocks. As a matter of fact, there were less households with cases of disease on the left side of the street. This contrast will presumably be rewarding in a study of personal and environmental factors responsible for marked differences in attack rates. Because of the difficulty of surveying every dwelling in the city and in the rural environment, the comparison between these 'large communities' was based only on the two groups of dwellings affected in these communities. That is, the comparison will be made on groups of dwellings naturally selected for attack by the epidemic. On the one hand, limitations of communications and transportation made surveying of the rural districts less complete than surveying of the capital city. On the other hand, poor communications and long distances between dwellings in the rural districts probably limited spread of disease. Hence, the rural households surveyed probably did not fall far short of the total number affected. The study consisted of examination of as many epidemic aspects as possible with the data available on personal characteristics and the environment. Special emphasis was put on the spatial factor most studies of epidemics describe the latter in terms of time but the influence of space is not usually examined while space-time interaction has occasionally been analysed in communicable-disease studies (1, 2). This interaction seems to be more important than time alone (22,23). Spatial factors included sleeping arrangements in household dwellings, seating arrangement in classrooms and localization of dwellings with cases marked on maps of the city and rural districts, crowding in household dwellings was also noted as it presumably increased person-toperson transmission of disease. Another spatial aspect of disease occurrence and spread was

7 progression of the disease in a person occupying one bed to someone in another either in the same or a different room. The progression of the disease from the occupant of one desk to that of another in the same or different classroom was another aspect of the influence of space and the same applies to progression from one house to another in a block or within the city at large, and from the capital city to the rural environment. The study also aimed to discover the orientation, in relation to the magnetic north, of movement of the epidemic through the city and the rural environment. Since no previous study of orientation of epidemic progression was known, a mathematical approach used in the study of locust movement (24) is being tried. The field study of the epidemic was evidently a follow-up study, that is, a prospective study. The data collected is thus free from the known biases inherent in retrospective studies. The control population is ideal as it consisted of all the 'homemates' who did not develop the disease, it was neither selected nor sampled but self- or 'naturally' selected. The control population came from households with cases, hence many sources of bias are eliminated and it is unusually representative of the study population. Even if the control population differed from the experimental population (all cases from all households with cases) in certain properties, matching is an additional resource which will be used in certain analyses. A further control population consisted of a sample of households without cases of disease. This was not done randomly but through selection based on criteria such as neighbourhood (both physical and social) or of repeated contacts between members of the households with and without cases. There was little voluntary selection of members of unaffected households since whenever frequent contacts with outside were reported in the household with cases, the corresponding household without cases was surveyed and entered in the records. In the study schools with a considerable number of cases, classes with or without cases were surveyed, without exception. Hence the control population of school-population units with cases were all the classes without cases from the corresponding school. ACKNOWLEDGEMENTS Professors W. Buell Evans and Charles Santos- Buch kindly provided facilities for preparing this report. Doctors Armado R. Taborda, Laura C. Taborda and Evandro Pimenta de Campos also VARIOLA MINOR IN BRAGANCA PAULISTA COUNTY, provided facilities. Invaluable help in the field study of the epidemic was given by Mr Januario delle Cave and Mr Luiz Pereira. The cooperation of the principals and teachers of Braganca Paulista schools and, particularly of the staff of the local health unit, is gratefully acknowledged. REFERENCES (1) Klauber, M. R. and Angulo, J. J.: Variola minor in Bragan9a Paulista County, 1956: Space-time interactions among variola-minor cases in two elementary schools. American Journal of Epidemiology 99: 65, (2) Klauber, M. R. and Angulo, J. J.: Variola minor in Braganga Paulista Country, 1956: Lack of evidence indicating the influence of contaminated classrooms on spread of the disease. Journal of Hygiene 11: 281, (3) Pederneiras, C. A. A., Angulo, J. J. and Megale, P.: Variola minor in Braganga Paulista County, 1956: flow of the epidemic through the schools of the city capital of the Country. Zentralblatt fur Bakteriologie erste Abteilung: Originale Reihe B 160: 180, (4) Klauber, M. R. and Angulo, J. J.: Variola minor in Braganca Paulista County, 1956: attack rates in various population units of two schools including most students with the disease. American Journal of Epidemiology 103: 112, (5) Thibau Jr.: Epidemiologia e profilaxia de variola, alastrim e varicela. Brasil-Medico. (Rio de Janeiro) 16: 503, (6) Rodrigues-da-Silva, G. Rabello, S. I. and Angulo, J. J.: Epidemic of variola minor in a suburb of Sao Paulo. Public Health Reports 78: 165, (7) Carvalho Fo, E. S., Morris, L., Lemos, A. L., Ponce de Leon, J., Escobar, A., and Silva, O. J.: Smallpox eradication in Brazil, Bulletin of the World Health Organization 43: 797, (8) Morris, L., da Silva, and Martinez, A. V.: Epidemiological investigation of a smallpox outbreak in a town reported 100 per cent vaccinated. American Journal of Epidemiology 92: 294, (9) Arnt, N. and Morris, L.: Smallpox outbreak in two Brazilian villages: epidemiological characteristics. American Journal of Epidemiology 95: 363,1972. (10) Campanha de Erradicacjio da Variola: Casos e 6bitos de variola notificados, taxa de incidencia e razao 6bitocaso de 1956 a Boletin Campanha de Erradicacao da Variola (Rio de Janeiro) 5 (no 2), (11) Pereira-Barreto, L, personal communication, (12) Downie, A. W., Dumbell, K. R., Ayrosa Galvao, A. P., and Zatz, I.: Alastrim in Brazil. Tropical Geographical Medicine 15: 25, (13) Briceflo Rossi, A. L.: Las diferencias del virus de alastrim. Boletin de la Oficina Sanitaria Panamericana 54: 419, (14) De Salles-Gomes, L. F., Conceicao, B. C, Galluzzi, Y. D., da Fonseca, Y. M., Weigl, D. R. and Figueiredo, M. E. F.: Variola; diagnostico etiol6gico de 1967 a 1970 no Instituto Adolfo Lutz. Revista Instituto Adolfo Lutz. (S5o Paulo) 31: 5, (15) Angulo, J. J., Rodrigues-da-Silva, G. and Rabello, S. I.: Variola minor in a primary school. Public Health Reports 79: 355, 1964.

8 366 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY (16) Angulo, J. J., Rodrigues-da-Silva, G. and Rabello, S. I.: Spread of variola minor in households. American Journal of Epidemiology 86:479,1967. (17) Angulo, J. J., Rodrigues-da-Silva, G. and Rabello, S. I.: Sociologic factors in the spread of variola minor in a semi-rural school district. Journal of Hygiene 66: 7, (18) Pederneiras, C. A. A. and Angulo, J. J.: Unpublished observations in (19)Dixon, C. W.: Smallpox in Tripolitania, 1946; an epidemiological and clinical study of 500 cases, including trials of penicillin treatment. Journal of Hygiene 46: 351, (20) Dixon, C. W.: Smallpox, London, Churchill Ltd (21) De Salles-Gomes, L. F., Angulo, J. J., Menezes, E. and Zamith, V. A.: Clinical and subclinical variola minor in a ward outbreak. Journal of Hygiene 63:49,1965. (22) Knox, G.: Detection of low epidemicity. British Journal of Preventive andsocial Medicine 17:121,1963. (23) Mantel, N.: The detection of disease clustering and a generalized regression approach. Cancer Research 27: 209, (24) Clark, D. P., Ashall, C, Waloff, Z. and Chimmick, L.: Field Studies on the Australian Plague Locust. Anti-Locust Bulletin 44: London, Anti-Locust Research Centre, {received 22 April 1976)