点击显示 收起
1 From the Medical Research Council International Nutrition Group, London School of Hygiene and Tropical Medicine (SEM and AMP); the Department of Social and Preventive Paediatrics, King Edward Medical College and Mayo Hospital, Lahore, Pakistan (FJ and RA); the National Institutes of Health, Bethesda, MD (SCS); and the Department of Clinical Immunology, University of Göteborg, Göteborg, Sweden (LAH)
2 Supported by the Nestlé Foundation, with additional support from Aventis Pasteur, Lyon, France. 3 Address reprint requests to SE Moore, MRC International Nutrition Group, Nutrition and Public Health Intervention Research Unit, Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, United Kingdom. E-mail: sophie.moore{at}lshtm.ac.uk.
ABSTRACT
Background: Substantial evidence exists linking small size at birth to later-life susceptibility to chronic disease. Evidence is also emerging that some components of immune function may be programmed in early life. However, this evidence is limited and requires confirmation.
Objective: We investigated the association between size at birth and response to vaccination in a cohort of 257 adults (mean age: 29.4 y; 146 men) born in an urban slum in Lahore, Pakistan, during 1964-1978.
Design: A single dose of Vi polysaccharide vaccine for Salmonella typhi and 2 doses of rabies vaccine were given to each subject. Antibody titers were measured in prevaccination serum samples (Vi) and in postvaccination samples (Vi and rabies).
Results: The mean birth weight of the subjects was 3.24 kg; 14% of the subjects had low birth weights (<2.5 kg). Vaccine responses were not consistently associated with contemporary variables (month of study, sex, current age, or indicators of wealth). Response to typhoid vaccination was positively related to birth weight (anti-Vi immunoglobulin G: r = 0.138, P = 0.031; anti-Vi immunoglobulin M: r = 0.197, P = 0.034). Response to the rabies vaccine was not significantly associated with birth weight.
Conclusions: These findings add to a growing body of evidence suggesting that components of the immune system may be permanently programmed by events in early life. The contrasting effects on typhoid and rabies responses suggest that antibody generation to polysaccharide antigens, which have greater B cell involvement, is compromised by fetal growth retardation.
Key Words: Birth weight • fetal origins • vaccine response • rabies • typhoid • Pakistan
INTRODUCTION
Events in early life are now known to be important in the susceptibility to certain chronic diseases in later life. The "fetal origins of adult disease" hypothesis states that cardiovascular disease and type 2 diabetes originate through adaptations made by the fetus when it is undernourished. These adaptations, which include the slowing of growth, permanently change the structure and function of the body (1, 2). Evidence is now emerging to suggest that certain components of immune function may also be programmed in early life.
Recent work from The Gambia, West Africa, led to a heightened interest in the early-life programming of human immune function. It was found that young adults born during the hungry (wet) season are up to 10 times as likely to die from infectious diseases as are their counterparts born during the harvest (dry) season (3, 4). Subsequent work in this same rural population showed evidence of seasonal effects on infant thymic size (5) and on T cell subset profiles (6); these findings suggest that the defect may be the consequence of an early insult to thymic development. A study of adolescents participating in an ongoing longitudinal study in the Philippines has added to this evidence by showing that prenatal undernutrition is significantly associated with reduced thymopoietin production, and that growth in length during the first year of life is positively associated with adolescent thymopoietin production (7). In the same cohort, the predicted probability of mounting an adequate antibody response to a Salmonella typhi Vi vaccine was found to be lowest in adolescents who were born small for gestational age (SGA) and had the lowest body mass index (BMI) at age 14-15 y (8).
The hypothesis that immune function can be programmed during early life is most relevant to countries where infectious diseases still contribute to the majority of mortality, as in most developing countries. Unfortunately, however, opportunities for the long-term follow-up of individuals in such countries are limited by the rarity of early-life data. The current study utilized birth records collected during 1964-1978 from a cohort of individuals living in an urban slum in Lahore, Pakistan. The aim was to assess vaccine responses in relation to size at birth in the subset of these adults for whom birth weight was recorded.
SUBJECTS AND METHODS
Study population
The study population was drawn from a cohort of 2468 subjects born during 14 consecutive years, from 1964 to 1978, in an urban slum (Gowalmandi) in Lahore, Pakistan. Early published studies reported detailed data on the growth and health status of this population during infancy (9, 10). In brief, these infants and their mothers were followed from birth, partly longitudinally, by personal surveillance by a pediatrician (FJ). Weekly visits were made during the first month of life, then every month until 6 mo, and then every 3 mo. The infant mortality rate was 57/1000 live births, and neonatal mortality was 22/1000 live births: 40% of deaths were due to acute gastroenteritis and 21% to acute bronchopneumonia. Poverty was severe: 75% of the families had no tap water and no toilet of their own; in 42% of the families, 5 persons lived in one room, and in a further 12%, >8 persons lived in one room.
Most infants were born in the home, and after delivery, health workers notified the investigator so that a preliminary visit could be made. A cash payment system to the health workers was established to encourage the prompt notification of each birth. As soon as permission was granted, the family was visited and the infant measured. However, despite this system of notification, it was not possible to see all of the infants immediately after birth. At this first home visit, the investigator made detailed anthropometric measures by using regularly validated standard equipment. Weight was measured to the nearest 100 g with a portable infant weighing scale, crown-heel length was measured to the nearest mm with the infant in the supine position and by using a portable metal length board, and occipital-frontal head circumference was measured to the nearest mm with a narrow metal tape measure. Gestational age was assessed on the basis of the reported date of the mother’s last menstrual period.
At each follow-up visit after birth, growth and morbidity were measured and assessed for each infant. The key findings in relation to the data obtained are detailed elsewhere (9, 10). Feeding status was also estimated at these time points. In brief, these early data recorded a high frequency of early weaning, with almost 50% of mothers introducing a breast milk substitute (normally buffalo milk) from 1 mo of age (10). Even among the "completely" breastfed infants, a bottle was used for the administration of extra water during the hot season. The early data on growth, morbidity, and feeding status will not be shown or included in the present analysis.
Study protocol
The current study was part of a larger study looking at the effects of poverty, early-life malnutrition, and infections on adult health and mortality. Descriptions of the findings from the parallel and ongoing studies in this cohort will be detailed elsewhere. From the 2468 infants initially registered, 76% (1885) were retraced during 2001. Of these, 733 had infant data available and were available for follow-up within the framework of the main study. From these subjects, a smaller subcohort was invited to participate in the current investigation of immune function. These subjects were selected on the basis of their living not too far from Lahore and their willingness to participate in the more in-depth study. In addition, only subjects with measurements made within 28 d of birth were recruited.
Fieldwork was conducted during the months of April to September 2002. A total of 257 adults (146 men, 111 women) were successfully recruited. Each adult was seen on 3 separate occasions (days 0, 7, and 14). On the first visit (day 0), the subjects were brought to the hospital and their height was measured to the nearest 0.1 cm by using a wall-mounted stadiometer and their weight was measured to the nearest 100 g on Tanita electronic scales (Tanita UK Ltd, West Drayton, United Kingdom). In addition, each subject completed a short questionnaire on his or her household income, housing, and personal belongings. From the data collected, a scoring system was devised to give an indicator of socioeconomic status. For details of housing standard and family education level at birth, the same scoring system was used as reported in earlier publications (10). A sample of venous blood was then collected for the analysis of prevaccination serum antibody titers. A single dose of Vi polysaccharide vaccine for S. typhi (Aventis Pasteur, Lyon, France) and a single dose of rabies vaccine (Rabies vaccine BP; Verorab, Aventis Pasteur) were administered to each subject in opposing arms. Seven days later, a blood sample was collected to assess the antibody response to the rabies vaccine. After sampling, a second dose of the rabies vaccine was given to each subject. On the final day of the study (day 14), a blood sample was collected for the analysis of antibody response to the typhoid vaccine and to the second dose of the rabies vaccine.
Approval for the study was granted by the Medical Ethics Committee for Research, King Edward Medical College, Lahore, Pakistan, and by the Ethics Committee of Göteborg University, Sweden. The study was conducted with informed consent from all subjects.
Laboratory analyses
Rabies-neutralizing antibody titers were assayed at the Central Veterinary Laboratories, Surrey, United Kingdom, by using the fluorescent antibody virus neutralization method, as described by Cliquet et al (11). In brief, a constant dose of previously titrated challenge virus was incubated with serial dilutions of the sera to be titrated. Reference sera included on the test were Office International des Epizooties (OIE)-positive dog sera at 0.5 IU/mL, World Health Organization (WHO)-positive human sera at 0.5 IU/mL, and negative dog sera. After incubation of the serum-virus mix, a suspension of susceptible cells was added. After 2 d of incubation, the cell monolayer was fixed in acetone and stained with a fluorescent antibody to the nucleocapsid to detect the presence of nonneutralized virus (fluorescent foci). The titer of the test serum was calculated by the Spearman Karber method and was recorded in IU/mL by comparison of the value obtained for the test serum with that obtained for the standard reference serum. The lowest detectable assay result was 0.02 IU/mL.
Owing to a problem with the original protocol, blood was not drawn from 51 of the subjects after the first dose of the rabies vaccine. Only 3 subjects, however, did not have a blood sample after the second dose of the vaccine.
Anti-Vi immunoglobulin G (IgG) and IgM analyses was conducted at the Laboratory of Developmental and Molecular Immunity, National Institutes of Child Health and Human Development, Bethesda, MD. Briefly, microtiter plates were coated with Vi (0.2 µg/well) purified from Citrobactor freundii, and goat anti-human IgG (Jackson Immuno Research Laboratories Inc, West Grove, PA) or IgM (Sigma, St Louis) conjugated to alkaline phosphatase was used for enzyme-linked immunosorbent assay (ELISA) (12). The anti-Vi IgG standard was a plasma sample from an adult vaccinated with Vi polysaccharide typhoid vaccine (provided by Wendy Keitel, Baylor University, Houston). The Vi antibody content of this serum was also assayed by a radioimmunoassay by Pasteur Merieux Connaught. Serum from a typhoid carrier with high titer of anti-Vi IgM was used as the IgM reference. The antibody titers were expressed in ELISA units (EU), and the reference sera were assigned a value of 75 EU for IgG and 616 EU for IgM. All samples were run in duplicates. Antibody titers were calculated by using ELISA, version 12 (Centers for Disease Control and Prevention, Atlanta). The lowest detectable level of the assay for anti-Vi IgG was 0.1 EU and that for IgM was 1 EU.
Prevaccination anti-IgG samples were not available for 11 of the subjects. Only one subject for whom a prevaccination sample was available did not have a postvaccination sample. The mean (±SD) interval between vaccination and collection of samples for the postvaccination antibody titer measurement was 15.9 ± 5.6 d. Samples were analyzed in 2 batches: those arriving in September 2002 (batch 1) and those arriving November 2002 (batch 2). No significant differences in mean antibody titer were observed between the batches, for either the prevaccination (P = 0.4249) or the postvaccination (P = 0.2238) samples. The batches were therefore combined for statistical analysis. Anti-Vi IgM titers were measured for 117 pairs of sera from batch 1 only.
Before analysis, all data were log transformed, and the results are presented as geometric means. All antibody titers are expressed as EU. Anti-IgG and anti-IgM data were analyzed according to absolute values and the increase in value between the pre- and post-vaccination samples. In addition, positive seroconversion to the vaccine was considered as a 4-fold increase in serum antibody titers. This criterion is commonly used to assess the rate of seroconversion to the Vi capsular polysaccharide vaccine (for example, see reference 13).
Statistical analyses
Weight-for-age, height-for-age, and BMI-for-age SD scores were calculated by using stature, weight, and BMI reference curves from the WHO (14). Vaccine data are presented as geometric means and 95% CIs. Comparisons between group means were made by using two-sample t tests. Associations between early-life variables and vaccine responses were tested by analysis of variance. Probability values <0.05 were considered to be statistically significant. The data are presented graphically by using the point means of the exposure variables (birth weight) divided into groups of 4 (lowest to highest; ie, quartiles) plotted against the corresponding outcome variable (vaccine response) for each group. All statistical analyses were performed by using DATADESK, version 6 for WINDOWS (Data Description Inc, Ithaca, NY).
RESULTS
Subject characteristics
Descriptive characteristics of the 257 subjects recruited into the study at the time of the first measurement made after birth are detailed in Table 1. All of the subjects had been visited as infants within 28 d of delivery. The age distribution of the infants when this first postpartum measurement was made is shown in Table 2. The mean weight of the subjects was 3.24 kg, their mean length was 51.7 cm, and their mean head circumference was 39.1 cm. A sex difference was observed in weight, length, and head circumference at birth, with male infants being consistently larger than female infants. Forty-seven (19.3%) of the infants were born before 37 wk of gestation and can be considered to be premature. A total of 36 of the infants (14%; 16 male, 20 female) weighed <2.5 kg at their first measurement and can be defined as having low birth weight. Only 5 of these infants were premature. The true prevalence of low birth weight might have been slightly higher if all of the infants had been weighed at birth.
View this table:
TABLE 1. Subject characteristics at the first measurement after birth1
View this table:
TABLE 2. Age at the first visit after birth1
Most of the infants were born at home by vaginal delivery (225, or 87.5%), with only 25 (9.7%) requiring specific medical intervention. Delivery details are not known for only 7 (2.7%) of the subjects.
The subjects’ characteristics at follow-up are shown in Table 3. The subjects’ mean age at follow-up was 29.4 y. The men were significantly heavier and taller than the women, but the difference in BMI between the sexes was not significant. With the use of the proposed classification of weight by BMI (in kg/m2) in adult Asians (15), 12.5% of the subjects were classified as underweight (BMI < 18.5), 32.8% had an average weight (BMI between 18.5 and 22.9), 16.0% were overweight and at risk (BMI between 23 and 24.9), 26.2% were obese stage I (BMI between 25 and 29.8), and the remaining 12.5% were obese stage II (BMI 30).
View this table:
TABLE 3. Subject characteristics at follow-up1
Vaccine responses in relation to adult variables at the time of vaccination
Rabies
The WHO recommends that satisfactory seroconversion to rabies vaccination has been achieved when antibody titers are >0.5 IU/mL (16). The absolute mean (±SD) antibody titer after the first dose of the vaccine was 4.94 ± 58.1 IU/mL, and after the second dose, it had risen to 117.7 ± 191.1 IU/mL. Many of the study subjects (206; 80.2%) had not achieved satisfactory seroconversion after the first dose of the vaccine, whereas only 18 subjects (7.0%) had not achieved this after the second dose of the vaccine.
Response to rabies vaccination was not related to the subjects’ sex, after either the first dose (P = 0.6444) or the second dose (P = 0.1806). A slight negative relation was observed between the age of the subjects and their response to the second dose of the rabies vaccine (P = 0.0235). However, no associations were observed with age and the response to either the first dose of the vaccine or the increase in response between the 2 doses. The response to the first dose of the rabies vaccine was not related to month of vaccination (P = 0.8038). However, both the response to the second dose of the vaccine and the increase in response between the 2 doses were associated with the month of vaccination [response to dose 2, P = 0.0011 (Figure 1); increase in response, P = 0.0665]. The highest titers were observed during July and the lowest titers in April and September. Response to the rabies vaccine was not significantly associated with current anthropometric measurements.
FIGURE 1.. Geometric mean (95% CI) antibody responses to the second dose of the rabies vaccine, by month of vaccination. n = 254: April, 47; May, 40; June, 47; July, 74; August, 39; and September, 7. P for trend = 0.0011 (ANOVA).
Anti-Vi IgG analysis
Geometric mean prevaccination antibody titers were 0.53 EU (95% CI: 0.45, 0.61), and this increased to 5.94 EU (95% CI: 4.88, 7.23) after vaccination. The mean increase between doses was 11.3 EU (95% CI: 9.49, 13.4). A total of 172 (66.9%) of the subjects responded to the vaccine with a >4-fold increase in anti-Vi antibody titer at 2 wk. Anti-Vi antibody titers were not significantly related to sex, age, adult anthropometry, or month of vaccination.
Anti-Vi IgM analysis
The geometric mean prevaccination titer was 7.68 EU (95% CI: 6.82, 8.65). After vaccination, this increased to 30.7 EU (95% CI: 26.6, 35.4). The mean increase in titer between samples was 4.00 EU (95% CI: 3.48, 4.59), and 42 (16.3%) subjects showed a >4-fold increase in antibody titers from preimmune levels.
Prevaccination IgM anti-Vi antibody titers were significantly higher in the women than in the men (2.19 compared with 1.92 EU; P = 0.0277), but there was no significant difference in postvaccination titers between the sexes. The increase in titer from baseline was significantly greater in the men (P = 0.0531), but there was no significant sex difference in the number of subjects who responded with a >4-fold increase (P = 0.1212). Antibody titers in both the prevaccination and postvaccination samples decreased with increasing age of the subjects (prevaccination, P = 0.0068; postvaccination, P = 0.0124). However, the increase in antibody titers between the prevaccination and postvaccination samples and the number of subjects with a >4-fold increase in response showed no significant association with the subjects’ age (P = 0.7918 and P = 0.9396, respectively). There was no significant association between IgM anti-Vi antibody titers and the month of vaccination.
Vaccine responses in relation to early-life variables
Rabies
No significant associations were observed between birth weight and the antibody response to rabies vaccination (response to dose 1, P = 0.4166; response to dose 2, P = 0.5416; increase in titer between doses, P = 0.4419). This remained the case when the analysis was adjusted for the number of days after birth at which the weights were measured (response to dose 1, P = 0.3152; response to dose 2, P = 0.7039; increase in titer between doses, P = 0.5883). Similarly, no significant associations were found with birth length or head circumference. No significant associations were observed between month of birth and response to rabies vaccination.
Anti-Vi IgG
The unadjusted postvaccination IgG antibody response was not significantly related to size at birth (P = 0.1771), and this remained the case when adjusted for the number of days after birth when the infants were measured (P = 0.2354). The relation between birth weight and the increase in antibody response between baseline and after vaccination approached significance (ß = 0.259, P = 0.0659), but became nonsignificant when adjusted for the number of days after delivery when the measurement was made (P = 0.1026). The number of subjects showing a >4-fold increase in antibody titers was significantly related to birth weight (ß = 0.1006, P = 0.031), and this remained significant when adjusted for the number of days after delivery when the measurement was made (ß = 0.0949, P = 0.0449).
Birth weight according to responder group is shown in Figure 2; subjects showing a >4-fold increase in anti-Vi IgG titers after vaccination were significantly heavier at birth than those with a <4-fold increase in response (3.30 kg versus 3.11 kg; P = 0.031). No consistent associations were observed between birth length or head circumference and the anti-Vi IgG response, and no significant associations were observed between month of birth and the anti-Vi IgG response.
FIGURE 2.. Mean (±SEM) birth weight by antibody response to typhoid vaccine in subjects showing a <4-fold increase in anti-Vi antibody titers () and in those showing a 4-fold increase (). The significance of the birth weight difference (tested by ANOVA) was as follows: for immunoglobulin (Ig) G (n = 73 and 172), P = 0.031; for IgM (n = 75 and 42), P = 0.034.
Anti-Vi IgM
There were no significant associations between birth weight and the IgM antibody titer after vaccination (P = 0.1750). However, a significant positive association was seen with the increase in antibody titers after vaccination (ß = 0.2452, P = 0.0279), and this remained significant when adjusted for the number of days after delivery when birth weight was measured (ß = 0.2318, P = 0.034). The number of subjects showing a >4-fold increase in antibody titers was significantly related to birth weight (ß = 0.1495, P = 0.034), and this remained significant when adjusted for the number of days after delivery when the first weight measurements were made (ß = 0.1484, P = 0.034).
Birth weight according to responder group is shown in Figure 2; subjects showing a >4-fold increase in antibody titer after vaccination were significantly heavier at birth than were those with a <4-fold increase in response (3.38 compared with 3.12 kg; P = 0.034). No consistent associations existed between anti-Vi IgM titers and birth length or head circumference, and no significant associations were observed between month of birth and the anti-Vi IgM response.
DISCUSSION
Small size at birth was significantly negatively associated with response to typhoid vaccination in this cohort of adults from Lahore, Pakistan. This relation existed independently of current factors that may influence response to vaccination (eg, month of vaccination, current body size, sex, and current socioeconomic status score) and independently of other factors from around the time of birth (eg, gestational age, father’s socioeconomic status score, and month of birth). No associations, however, were found between size at birth and the antibody response to rabies vaccine. This finding adds to the evidence from existing studies that the antibody response to certain vaccines may be permanently programmed early in life.
The current study supports the findings of McDade et al (8), who looked at the antibody response to vaccination in a cohort of adolescents from the Cebu Longitudinal Health and Nutrition Survey (CLHNS) in the Philippines. Using the same typhoid vaccine, McDade et al measured the predicted probabilities of mounting a positive antibody response for adolescents born small for gestational age or appropriate for gestational age in combination with size (BMI) at 14-15 y of age. They found that the predicted probability of mounting a positive antibody response for adolescents who were prenatally and currently undernourished was 0.32, compared with probabilities of 0.49-0.70 for adequately nourished adolescents (P = 0.023) (8). The Lahore study, however, has the additional strengths that the sample size was larger (257 compared with 96) and the association between size at birth and response to typhoid vaccination was apparent without the adjustment for postnatal factors.
The main limitation of the current protocol is the timing of the first measures after birth. Owing to the community-based nature of the initial study and the difficulty of working in such an environment at that time (10), it was difficult to obtain measures in the period immediately after delivery. For this reason, the mean age of the infants when first measured is 11.5 d; hence, the usefulness of these data as a measure of birth weight must be regarded with some caution. However, the fact that a statistically significant relation existed between these early postnatal weights and response to vaccination suggests that the true relation against birth weight might be stronger if fetal growth retardation is indeed the critical causal factor.
The 257 subjects recruited into the current study were drawn from a total of 2468 infants originally studied. Although this represents a minority of the original group, it is unlikely that selection bias influenced our results. The subjects recruited into the current study were chosen on the basis of their availability for follow-up, having infant data available, living in close proximity to Lahore, and having had a first weight measured within 28 d of delivery.
The 2 vaccines used were chosen specifically because of their differential ability to drive a primarily T cell-dependent (rabies virus protein) or a potentially T cell-independent (typhoid capsular polysaccharide) immune response. Of interest, a previous study conducted in The Gambia found no detectable differences in the antibody response to either the same rabies vaccine or a 23-valent pneumococcal capsular polysaccharide vaccine in 6-10-y-old children, according to their weight at birth (17). The current study suggests that early-life defects in the immune response appear to be mediated through the T cell-independent arm of antibody generation. The lack of an association in the comparative study from The Gambia also suggests that such defects may only be detected when vaccine responses are tested beyond childhood. Understanding these observations could help pinpoint the precise nature of the immunologic defect and will be pursued in future studies.
The Vi polysaccharide, like other polysaccharide vaccines, is not as effective in infants and young children as in older individuals. Generally, children respond to bacterial polysaccharides with more IgG1 antibodies and adults with more IgG2 antibodies (18, 19). In children <18 mo of age, there is a high proportion of nonresponders with IgG2 antibodies (20, 21). The detailed background of these age differences is not really understood, but further work in this area may help to explain the lower response to the typhoid vaccine in subjects born small for gestational age.
It is possible that the present observation of an impaired vaccine response in adult age relating to birth weight may be the long-term consequence of abnormalities in the placentas of these infants. Immunologic abnormalities have been noted in the placentas of intrauterine-growth-retarded Swedish newborns with significantly low levels of messenger RNA of the immunosuppressive cytokine IL-10 (22), a finding which suggests that additional immunologic abnormalities could exist in this condition. If such abnormalities then lead to immune defects beyond infancy, this could explain the link between size at birth and lower response to the typhoid vaccine. Indeed, ongoing work has confirmed this observation in placentas from deliveries affected by intrauterine growth retardation collected in the same area of Pakistan as the present study (Hahn-Zoric et al, unpublished observations, 2004). Further work is required to confirm the long-term effects of these observed defects.
At birth, the immune system is small and expands primarily on exposure to external antigenic material, especially the microbes colonizing the gut from delivery and later foods other than breast milk (23). Protective immune responses develop, as does immunologic tolerance to harmless material, such as foods. Postnatal antigen exposure is required for the normal development and expansion of the immune system. Because exposure to both benign and pathogenic microbes has been a long-standing feature of human neonatal experience, such exposures are required for instructing the immune system to ignore or tolerate benign organisms (such as those inhabiting the intestinal tract), while priming the neonate to be able to initiate a functionally robust immune response to potentially dangerous pathogens (24). Against this background, it is therefore possible that a defect in immune function could occur during early development of immune system cells, such as the antigen-presenting cells, mainly the dendritic cells, and the T and B lymphocytes, or at the stage of antigen exposure and clonal proliferation, which takes place mainly after delivery. However, the observation in the current study that the relation with size at birth was strongest in the infants measured before 28 d postdelivery, and that this relation did not strengthen after adjustment for postnatal factors, suggests that prenatal events are responsible for the observed defect in antibody response. The lack of any association between month of birth in Lahore and response to vaccination adds strength to the suggestion that the observed association is a consequence of poor maternal nutrition leading to small size at birth, rather than the consequence of any seasonal insults to the developing immune system (such as seasonal infections in either the mother or the infant).
No associations were observed between month of birth and response to vaccination. However, an interesting observation was the strong association between vaccine response and month of vaccination. A similar association was observed in previous studies with a defined seasonal pattern (17), although the precise etiology has yet to be defined. There are 2 distinct seasons in Lahore. The first is hot (initially dry and then rainy, April to September), and the second is cool (October to March). Highest titers were observed during the peak summer month of July, and lowest titers at the start and end of the hot season (April and September). It is possible that these monthly differences are the consequence of methodologic problems directly related to the month, such as differences in ambient temperature affecting the cold chain of the vaccines. However, this is considered unlikely because all of the vaccines were stored in refrigerators at the study site (Department of Social and Preventive Pediatrics), the subjects were brought to the department on the day of vaccinations, and the vaccines were removed immediately before vaccination. The 2 seasons in Lahore are known to have significant effects on the health and growth of young children and may determine the timing of exposure to potentially critical challenges, such as enteric or respiratory viral infections (25, 26). It is therefore possible that this seasonal increase in antigenic exposure acts as an adjuvant to the vaccine and primes the immune system, thus increasing the antibody response during the peak months of exposure.
In summary, the findings of the current study support the hypothesis that immune function may be programmed by events early in life that lead to small size at birth. Maternal undernutrition, or other factors leading to fetal growth restraint, seems a likely contributory factor. If true, this has implications for the risk of infectious diseases in areas of the world where large numbers of infants are born with a low birth weight and adds further importance to public health interventions focused on maternal and neonatal health. In addition, the efficacy of certain vaccines in vulnerable individuals needs to be considered. Follow-up studies are planned in this cohort of adult Pakistanis. These studies together with parallel studies in other populations with more detailed data from around the time of birth may help to elucidate the mechanisms involved.
ACKNOWLEDGMENTS
We thank the study subjects for their willing participation in this ongoing research program, and we are grateful to the field staff in Lahore for their help with this study. We also thank Tony Fooks and Trudie Goddard from the Central Veterinary Laboratories for their help and advice with the analysis of the rabies antibody titers.
FJ initiated and collected all the data for the 1964-1978 infant study. SEM, FJ, AMP, and LÅH designed the follow-up study of vaccine response. FJ and RA coordinated the study. SCS coordinated the Vi antibody laboratory analysis and provided intellectual input to the study. SEM conducted the statistical analyses and wrote the article. All authors contributed to the final version of the article. None of the authors had a conflict of interest to report.
REFERENCES