Cohort Profile: The 1982 Pelotas (Brazil) Birth Cohort Study

Pelotas is a city in the extreme south of Brazil, near the border of Uruguay, with 214 000 urban inhabitants in 1982. At the time, we were assistant professors, each working in one of the two medical schools in the city, and both undergoing post-graduate training at the University of London. We were inspired by the findings of the British perinatal study, and one of us (FCB) decided to do a similar study for his doctoral thesis. The lack of reliable data on perinatal mortality in Brazil, due to poor registration of births and deaths—particularly stillbirths—justified the launch of the study. Funding from the International Development Research Center (Canada) was obtained for the perinatal survey, which led to FCB's PhD thesis.1,2 While perinatal data collection was underway, we obtained a grant to visit a sub-sample of these newborns at the age of 12 months. Later funds were obtained for two visits to the entire cohort, at the approximate ages of 2 and 4 years. Little did we imagine that our study would eventually become one of the largest and longest running birth cohorts in the developing world.3 Initially, the study focus was on perinatal, infant, and early childhood morbidity and mortality. We were particularly interested in breastfeeding patterns and nutritional status, as well as social and environmental factors. Deaths of cohort members were identified by regular visits to all hospitals, cemeteries, offices of civil registrations, and local health authorities, since 1982. By mid-childhood, the study shifted in emphasis to child care, utilization of health services, selected morbidity indicators, and child development. A random sub-sample of 360 four-year-olds was selected for an in-depth study of psychomotor development. In adolescence, issues related to sexual and reproductive behaviours (including teenage pregnancies), habits such as smoking and alcohol drinking, mental health, and education became the focus of the investigation. A sub-study investigated oral health in a random sample of 900 adolescents, and an ethnographic study of 96 cohort members, stratified by sex and socioeconomic status, has included repeated in-depth visits from the age of 15 to 23 years, aimed at understanding the role of adolescent development in influencing high-risk behaviours. In more recent phases, with cohort members being young adults, the main emphasis has shifted to risk factors for chronic disease (including smoking, diet, physical exercise, and overweight), reproductive history, and mental health. During the whole of 1982, the three maternity hospitals in the city were visited daily and 7392 births were recorded. Of these, 6011 infants were born to mothers living in the urban area of Pelotas. Using data from birth registration and from a city census, we identified another 46 children who were delivered at home in 1982, so that our hospital sample accounts for 99.2% of all births in the city. The 5914 live born infants constitute our original cohort. Brazil is a country with wide social disparities, and a population-based sample covering the entire social spectrum allows the detailed study of long-term consequences of poverty on health. Follow-up procedures have been somewhat haphazard, depending on the availability of funding. Table 1 summarizes the main visits to the cohort. Earlier publications provide further details about each visit.4,5 Main phases of the Pelotas birth cohort study (reproduced with minor changes from ref. 5, with permission) This includes those known to have died for the 1983, 1984, 1986, 2000, and 2005 visits, and 27% of those known to have died for the 1997 and 2001 visits. Main phases of the Pelotas birth cohort study (reproduced with minor changes from ref. 5, with permission) This includes those known to have died for the 1983, 1984, 1986, 2000, and 2005 visits, and 27% of those known to have died for the 1997 and 2001 visits. In early 1983, the available funding was barely sufficient for visiting one-third of the cohort children. We opted for examining those born from January to April 1982, who were ∼1 year old at the time. This visit was funded by the World Health Organization. Using the addresses available from hospital records, we examined 1457 children who, added to those known to have died, comprised ∼80% of the target group (Table 1). We became quite worried about the loss of one-fifth of the cohort in a single year, and decided to change the search strategy in the next visits. The 1984 and 1986 visits were planned well in advance, and received substantial funding from the United Kingdom's Department for International Development (then known as Overseas Development Administration). The objective was to examine every cohort child who was still living in the urban area. Rather than relying on addresses, in early 1984 we decided to visit every household in the city in search of children born in 1982. This led to 87% of the original cohort being traced, a substantial improvement over the follow-up rate obtained a year earlier. Of those located, 45% of families had already changed addresses since the cohort child had been born, indicating a very high rate of mobility. The same approach was used in early 1986, leading to a follow-up rate of 84%. At that stage, we had a massive amount of data to analyse as a small research unit. Because our primary interest was in child health, we assumed that the cohort study had been completed and set out to write up our results and become involved in other studies. There was no new data collection until 1995, when we were approached by the United Nations Children's Fund (UNICEF) and Development Fund for Women (UNIFEM) to collect information on issues related to adolescent sexuality. Limited funds were available, and the addresses obtained in 1986 were used to trace a random sample of 1100 cohort members. Only 70% of them could be located. This visit renewed our interest in the cohort, and in 1997 we decided to apply a similar approach to that used in the city censuses of 1984 and 1986. As funds—this time from the Brazilian government—were not available to cover the whole city, we systematically selected 70 census tracts (27% of the total) and visited every household in those tracts. This led to 72% of the cohort members expected to be living in those tracts to be traced. A special opportunity for follow-up was provided by the compulsory Army recruitment examination, held in 2000. All cohort males who were still living in the city were legally required to attend a local Army base to undergo a physical examination in August and September. Our research team was deployed to the base and was able to examine 79% of males from the cohort. Funds from the Brazilian government supported this examination. 2001 marked the end of adolescence for the cohort. We had special interest in investigating the high rates of teenage childbearing in the cohort, and obtained funding from the World Health Organization for this purpose. Using the national computerized birth registration system put in place in the late 1990s, we identified over 400 of cohort women who had delivered an infant up to March 2001, and visited them at home. To obtain a comparison group for a case–control analysis, we revisited all households in the 70 census tracts from the 1997 sample and examined all cohort subjects living in these tracts. An extensive interview was carried out with women and a shorter version with men, who had already been examined in the Army in 2000. The follow-up rate was 69%. At this time, it was evident that we had the largest running birth cohort outside high-income countries, and our interest was renewed. We were able to obtain substantial funding from the Wellcome Trust to visit the whole cohort once again. From October 2004 to August 2005, we visited all 98 000 households in the city and located 3924 cohort members. The system for monitoring mortality had identified 282 deaths. For those who had not been located and were not known to have died, we used the last known address and existing databases (including universities, secondary schools, telephone directories) for another attempt. This allowed us to interview 4297 subjects. Added to those known to have died, these represent a follow-up rate of ∼77% in relation to the original cohort. We have carried out several analyses of attrition rates according to baseline characteristics.5 The pattern of follow-up varies slightly according to the search strategy used, but in general subjects born to middle-class families are easier to trace than those born in either the upper or lower ends of the social distribution. There are no consistent differences in follow-up according to sex, birthweight, or skin colour. Subjects born to unmarried mothers are consistently harder to trace than those born in wedlock. The 1982 questionnaire was extremely short. Owing to our lack of experience, we worried about having several sheets of paper that might become separated and decided to used the longest sheet of paper commercially available (a bit longer than A4 size) and restrict the questionnaire to both sides of this sheet. All the information contained in this form took 80 columns in a punch card, which also made it rather convenient. Table 2 shows the main categories of variables collected in the early phases of the study. In the hospital interview, mothers answered questions on socioeconomic, demographic, and health-related variables. Their infants were weighed with regularly calibrated paediatric scales (Filizolla, Brazil) to the nearest 10 g. Birth length was not recorded. Mothers were weighed and measured. Main variables collected in the early phases of the cohort study (1982–86) (reproduced with minor changes from ref. 5, with permission) Main variables collected in the early phases of the cohort study (1982–86) (reproduced with minor changes from ref. 5, with permission) As we grew more confident, our questionnaires grew longer—possibly too much so—and the examinations more thorough. Table 3 shows the main categories of variables included in recent visits. The questionnaires now include two forms, one applied by an interviewer and another self-applied confidential form that is identified only by a questionnaire number. Main variables collected in the late phases of the cohort study (1995–2005) (reproduced with minor changes from ref. 5, with permission) Males only (Army examination in 2000). Main variables collected in the late phases of the cohort study (1995–2005) (reproduced with minor changes from ref. 5, with permission) Males only (Army examination in 2000). In 2000, biological materials were collected from males (blood samples from which sera were extracted and frozen at −70°C) and in 2005 for both sexes (extracted DNA samples as well as sera). In 2000, males were weighed with an electronic Tanita Body Fat Analyzer scale (model TBF-305; Tokyo, Japan), which also provided information on body composition through bio-impedance. These results were validated in a sub-sample of 48 subjects by comparison with total body water estimated through isotopic methods.6 Ethical requirements evolved considerably during the study period. In the early phases, verbal consent was obtained, and there were no local ethical review committees. Recent phases comply with current requirements of ethical review and include written informed consent. Special provisions are made for the ethical use of biological materials. A full list of all publications to date from the study is available as supplementary data at IJE Online. The first publications from the cohort addressed perinatal issues, highlighting the magnitude of the problems of perinatal mortality and low birthweight, and describing risk factors for these outcomes.1,7,8 The extremely high rate of caesarean sections (28%) was also highlighted in the publications,9 as were infant mortality levels, causes, and risk factors.10–12 Infant feeding patterns and their influence on health has been a major theme in our study.13–16 Given the prospective nature of our data, it has been possible to avoid some of the biases arising from using recalled breastfeeding duration in retrospective cohorts, and to investigate the long-term effects of breastfeeding on health and educational attainment, while adjusting for several early life factors.17–19 During the childhood phase, we investigated issues related to malnutrition and infection,20,21 as well as the associations between health and nutrition outcomes with maternal education22 and birth spacing.23 A cross-cutting theme in our cohort study has been the effect of social inequalities on health.24 In 1988, our book entitled ‘Epidemiology of Inequality’ came out in Portuguese, addressing social differentials in terms of perinatal, infant, and childhood outcomes.25 Five thousand copies were sold, and the Spanish version was published by the Pan-American Health Organization in 1992. More recent publications address issues of adolescent health, including the current epidemics of overweight,26 adolescent pregnancy,27 and asthma.28 Much of our current work is aimed at investigating the effects of low birthweight (which we can separate into intrauterine growth restriction and preterm delivery) and growth in childhood on several outcomes. These include blood pressure,29 overweight,26 and lung function.28 Analyses are forthcoming on blood lipids, glycaemia, body composition, oral health, and educational achievement. Cohorts from low-income and middle-income countries are needed because findings are likely to differ from those obtained in developed countries. Some exposures may have different characteristics—for example, most physical activity among males in our cohort is from occupation, rather than from leisure time activities. Another example is breastfeeding—whereas non-breastfed babies in high-income countries are likely to receive infant formula, in our setting cow's milk is the main breast milk substitute so the results of comparisons of breastfed and artificially fed infants may differ between populations. Also, some exposures—for example low birthweight or childhood malnutrition—are much more common than in developed countries, and their long-term effects can be studied with greater precision. Finally, confounding factors may act in different directions in rich and poor countries. For example, while breastfeeding tends to be associated with high socioeconomic status in wealthy populations, the reverse is often the case in low-income and middle-income countries. As a consequence, residual confounding—a critical issue in the study of the effects of breastfeeding on adult health—may operate in different directions. A review by Harpman et al.3 identified our study as the largest and longest running prospective birth cohort study in a developing country. About 4000 variables are available for subjects seen in all phases of the study, including anthropometric measurements not only at birth but also at different ages in childhood. Other strengths include the population base and relatively high follow-up rates. Although the latter are considerably lower than some of the studies from developed countries, we have had to face the challenge of tracing people actively, rather than passively through national databases. Our success is largely related to the characteristics of the city. Pelotas is a middle-size city with relatively low rates of in-migration and out-migration. The number of annual births provides sufficient statistical power for the study, while being still logistically manageable. Also, concerns with personal and home security are not as manifest as in larger Latin American cities and refusals are rare. On the negative side, there are many things that we—given the benefit of hindsight—would have done differently. At the perinatal interview, we should have measured birth length and assessed gestational age through physical examination (∼20% of the mothers were unable to recall the dates of their last menstrual period). We collected data on family income as a grouped variable rather than recording it as a continuous variable. In retrospect, obtaining information on the whole cohort at a smaller number of visits would probably have been better than using sub-samples. To date, we have data on the entire cohort for the original perinatal interview, the 1984, 1986, and 2004–05 follow-up visits, as well as for the 2000 Army examination for males. The 1983, 1995, 1997, and 2001 visits were based on sub-samples, with different sampling approaches. This means that only a few hundred subjects have complete data from all follow-up visits. Some lessons were learned from the Pelotas cohort that may be relevant to other studies. A critical issue in all cohorts is that of attrition. Two successful strategies were used in our study: household sampling and Army enlistment. The first entailed visiting all, or a sample of, the city's households to identify individuals born in 1982, and later tracing them to their cohort records. The second included taking advantage of the compulsory Army enlistment process. Attempts at locating cohort members using available addresses, both in 1983 and in 1995, led to high rates of attrition. Lessons can also be learned regarding administrative and financial aspects of the study. Funding agency fatigue means that few are prepared to support more than one to two rounds of the study. This precluded a more regular schedule of visits, and sampling fractions were sometimes determined by availability of funds rather than by scientific principles, as for example in the 1983 and 1995 follow-ups. Large birth cohort studies present specific funding issues that should deserve special treatment by grant-making agencies. As cohort members reach adult age, it is likely that they will increasingly move out of the city, where job opportunities are scarce. This may lead to higher attrition in the near future. Alternative approaches will have to be conceived, including passive follow-up through death certificates and nested case–control studies, among others. Nevertheless, the large amount of data already available from the study will certainly lead to many additional analyses on health, behaviour, and development in childhood and adolescence. Eleven years after the first cohort, we started another study—the 1993 Pelotas Birth Cohort—including all births in the city. The original plan was to launch it in 1992, a decade later than the first study, but of course funding was delayed. This study is underway and 87% of them were traced at the age of 12 years. In years the 2004 birth cohort was The of three birth cohorts is a comparison of in child health, and in the in adolescent and adult Research involved in cohort studies have that it is not to the studies while at some of the are still We are it not only to our cohorts but also to from a new one every years. this for the Pelotas Birth We analyses of the cohort We have with from the of and of and and of as well as several Brazilian Our most have been from having doctoral or to Pelotas for a few at a time, to analyse data and For interested young from Latin we a post-graduate in in 2005, which now includes and PhD from the who receive full to work on our are For further information our at or This was supported by the Wellcome entitled for Latin on Health of Earlier phases of the 1982 cohort study were funded by the International Development Research Center the World Health Organization of and Health and and the Overseas Development the United Nations Development Fund for the for of the Research and the of Health Special to who us launch the study in early We would also to our many who in the several phases of the study, particularly and

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.