Birth Weight and Adult Hypertension, Diabetes Mellitus, and Obesity in US Men
Background Low birth weight has been associated with several chronic diseases in adults, including hypertension, diabetes mellitus, and obesity. Further study of these diseases in a large cohort with information on a wide variety of risk factors is essential to determine more precisely the risks associated with birth weight.
Methods and Results We examined the relation between birth weight and cumulative incidence of adult hypertension, incidence of non–insulin-dependent diabetes mellitus, and prevalence of obesity in a cohort of 22 846 US men (Health Professionals Follow-up Study). Birth weights, medical histories, family histories, and other factors were collected by biennial mailed questionnaires. Logistic regression was used to examine the association between birth weight and these chronic adult diseases. Low birth weight was associated with an increased risk of hypertension and diabetes; high birth weight was associated with an increased risk of obesity. Compared with men in the referent birth weight category (7.0 to 8.4 lb), men who weighed <5.5 lb had an age-adjusted odds ratio for hypertension of 1.26 (95% confidence interval [CI], 1.11 to 1.44) and for diabetes mellitus of 1.75 (95% CI, 1.21 to 2.54). There was no material change after controlling for adult body mass index and parental histories of hypertension and diabetes mellitus. Compared with men in the referent group, the age-adjusted odds ratio of being in the highest versus the lowest quintile of adult body mass index for men with birth weight ≥10.0 lb was 2.08 (95% CI, 1.73 to 2.50).
Conclusions These findings support the hypothesis that early life exposures, for which birth weight is a marker, are associated with several chronic diseases in adulthood.
Low birth weight has been associated with several chronic diseases in adults, including hypertension,1 diabetes mellitus,2 3 and obesity.4 When these important cardiovascular risk factors are present in childhood, they may persist into adulthood, suggesting origins in early life.5 The associations between birth weight and adult hypertension, diabetes mellitus, and obesity have been described in studies predominantly from Britain.1 2 6 These reports contained relatively small numbers of subjects with birth weights in the extreme categories and did not consistently adjust for other important factors such as BMI and family history of these diseases. To determine more precisely the risks associated with birth weight, we examined the relation between birth weight and adult hypertension, NIDDM, and obesity in a cohort of >22 000 US men.
The HPFS is a longitudinal study of diet and disease among 51 529 male dentists, optometrists, osteopaths, pharmacists, podiatrists, and veterinarians who were 40 to 75 years of age in 1986. The participants returned a mailed questionnaire in 1986 concerning diet, medical history, and medications. Follow-up questionnaires were mailed every 2 years to update information on exposures and newly diagnosed diseases.
Assessment of Birth Weight
The participants were asked on the 1994 questionnaire to provide their birth weights with the following categories of responses (in lb): <5.5, 5.5 to 6.9, 7.0 to 8.4, 8.5 to 9.9, and ≥10.0 and unknown. Of the 38 668 men responding to the 1994 long-form questionnaire, 22 846 (59%) provided information on birth weight. To assess the validity of the self-reported birth weights, a questionnaire was mailed to 6065 mothers of cohort members to inquire about the HPFS subject's birth weight. Birth weight information was provided by 86% of the mothers, and a weight was available from both the man and his mother for 3803 participants. The mean birth weight reported by the subjects was 7.65±1.25 lb (mean±SD); that reported by the mothers was 7.63±1.17 lb. The Spearman correlation coefficient for the subject's self-reported birth weight and the birth weight reported by the mother was .71 (P<.001). The frequency of the mother and subject reporting the exact same birth weight category was 68.6%. In 97.9%, the reported birth weight categories were within one category of each other.
Assessment of Hypertension and Blood Pressure
The baseline and biennial follow-up questionnaires inquired about physician-diagnosed hypertension. A study participant was considered to have hypertension if it was reported on any questionnaire. The 1990 questionnaire inquired about the subjects' blood pressure during the preceding 2 years. For men taking medication that might lower blood pressure, the reported values were for while they were on medication and not the prediagnosis untreated values. The categories of systolic blood pressure responses (in mm Hg) were <105, 105 to 114, 115 to 124, 125 to 134, 135 to 144, 145 to 154, 155 to 164, 165 to 174, and 175+ and unknown or not checked within the past 2 years. The categories of diastolic pressure responses (in mm Hg) were <65, 65 to 74, 75 to 84, 85 to 89, 90 to 94, 95 to 104, and 105+ and unknown or not checked within the past 2 years. In this cohort, the diagnosis of hypertension is reported reliably, and self-reported blood pressure is a strong predictor of stroke.7 Men who were taking medication that might lower blood pressure (such as β-blockers) but who did not report hypertension were not considered to have hypertension.
Assessment of Diabetes
The 1986 baseline questionnaire inquired about a history of physician-diagnosed diabetes mellitus, and biennial questionnaires mailed between 1988 and 1994 were used to identify newly diagnosed cases of diabetes. To confirm self-reports and to restrict the disease group to those with newly diagnosed NIDDM, a supplementary questionnaire was mailed to men reporting a diagnosis of diabetes between 1986 and 1994. Subjects were considered to have confirmed NIDDM if any of the following criteria were met: (1) one or more classic symptoms (thirst, polyuria, weight loss, hunger, or pruritus) and a raised fasting (≥7.8 mmol/L) or random (≥11.1 mmol/L) plasma glucose concentration, (2) elevated plasma glucose concentrations on at least two separate occasions (fasting, ≥7.8 mmol/L; random, ≥11.1 mmol/L; or after ≥2 hours on glucose tolerance testing, ≥11.1 mmol/L) in the absence of symptoms, or (3) treatment with a hypoglycemic drug. The validity of self-reported NIDDM by medical professionals using the same supplementary questionnaire has been documented in the Nurses' Health Study in which a substudy revealed that 98% of the self-reported cases were confirmed by medical record review.8
Assessment of BMI
BMI was calculated as weight in kilograms divided by the square of height in meters. Height was reported in 1986, and current weight was asked on each biennial questionnaire. In this cohort, self-reported weight was highly correlated with actual measurement (Pearson's r=.97).9
Assessment of Family History
Information was obtained in 1990 regarding family history of disease, including histories of hypertension and diabetes mellitus in the subjects' mothers, fathers, and siblings.
The analyses were adjusted for age according to 5-year age groups by direct standardization to the age distribution of the overall cohort. Multiple logistic regression analysis was used to compute ORs in the individual birth weight groups, with the middle birth weight group (7.0 to 8.4 lb) considered the referent category and simultaneous adjustment for potentially confounding variables.10 For the hypertension analyses, a subject was considered to have hypertension if it was reported on any questionnaire. For the diabetes analyses, only confirmed incident cases occurring after 1986 were included because the type of diabetes present before 1986 was not documented. A secondary analysis was performed that also included self-reported prevalent cases of diabetes on the 1986 questionnaire. The variables considered in the hypertension and NIDDM models were birth weight categories (including a category for missing), age (in 5-year categories), adult BMI in 1994 (10 categories), maternal and parental histories of hypertension (yes or no), maternal and parental histories of diabetes mellitus (yes or no), smoking status in 1986 (never, past, or current), and quintiles of physical activity. Effect modification by maternal or paternal history of diabetes and maternal or paternal history of hypertension was investigated by examination of separate logistic regression models within each stratum. For all ORs, we calculated 95% CIs.
Multiple linear regression analysis was performed to estimate the contribution of individual variables to systolic and diastolic blood pressures. The medians of the categories of reported systolic and diastolic blood pressures were used for those with and without hypertension. For the models including men with hypertension, this probably included blood pressure readings in men taking blood pressure–lowering medication at the time of the self-report. Systolic and diastolic pressures, in separate models, were the dependent variables; the independent variables included birth weight (continuous, with 5.0 for the lowest category, 10.5 for the highest, and the midpoints of the other categories), age (continuous), BMI (10 categories), and parental history of hypertension. For all linear regression estimates, we calculated 95% CIs.
Caucasians were the most common racial group, making up 97% of the cohort. The number of nonwhite respondents with reported birth weight data was too small to perform a subgroup analysis by race.
Table 1⇓ gives the age-adjusted characteristics of the cohort. Men with birth weights <5.5 lb made up 4.9% of the cohort with reported birth weights, and 7.6% had birth weights ≥10 lb. Compared with the whole cohort, men with birth weights <5.5 lb had a higher frequency of maternal and paternal hypertension and maternal diabetes. Men with birth weights ≥10.0 lb were on average 3.8 years older, had a slightly lower frequency of paternal history of hypertension and diabetes, and had a higher frequency of maternal diabetes. The percentage of subjects having physical examinations between 1988 and 1990 was 70.6% and was similar across all birth weight categories, suggesting that the screenings for hypertension and diabetes were similar. Men with missing birth weight data were on average 4 years older but otherwise were very similar to men in the middle birth weight category with respect to systolic and diastolic blood pressures, family history, and prevalence of hypertension and diabetes, both at baseline and in 1994.
The age-adjusted cumulative incidence of hypertension was higher in men with birth weights ≤6.9 lb than in men in the referent birth weight category (7.0 to 8.4 lb; Table 1⇑). Compared with men in the referent birth weight category, the age-adjusted ORs of hypertension were 1.26 (95% CI, 1.11 to 1.44) for men in the lowest birth weight category (<5.5 lb) and 1.12 (95% CI, 1.04 to 1.21) for men with birth weights of 5.5 to 6.9 lb (Table 2⇓). After adjustment for BMI (1994) and parental history of hypertension, there was no material change in the results for birth weights <10 lb. However, in men with birth weight ≥10 lb, the OR was 0.89 and was statistically significant (95% CI, 0.80 to 1.00).
Among men who reported their birth weights, the mean systolic and diastolic blood pressures reported on the 1990 questionnaire were 127.7 and 80.2 mm Hg, respectively. The mean age-adjusted systolic blood pressure decreased slightly across increasing birth weight categories (Table 1⇑).
The relation between birth weight and age-adjusted blood pressure was not linear (Table 1⇑). However, to compare our results with previous reports that proposed a linear relation,1 birth weight as a continuous variable was used in linear regression models to estimate the magnitude of the effect of birth weight on systolic and diastolic blood pressures. After controlling for age, BMI, and parental history of hypertension, systolic blood pressure decreased by 0.31±0.07 mm Hg (mean±SE) for each 1-lb increase in birth weight, and diastolic blood pressure decreased by 0.13±0.04 mm Hg. Because the reported blood pressure values included those for subjects with treated hypertension, we repeated this analysis after assigning a blood pressure of 150/95 mm Hg to all individuals with histories of hypertension. For each 1-lb increase in birth weight, systolic and diastolic blood pressures decreased by 0.48±0.08 and 0.28±0.05 mm Hg, respectively. After men with diagnosed hypertension were excluded from the analysis, the systolic and diastolic blood pressures decreased by 0.21±0.07 and 0.08±0.05 mm Hg, respectively, for each 1-lb increase in birth weight.
The age-adjusted cumulative incidence of diabetes in 1994 was higher in men with birth weights ≤6.9 lb than in men in the referent birth weight category (7.0 to 8.4 lb; Table 1⇑). Compared with men in the referent birth weight category, the age-adjusted ORs of incident diabetes (from 1986 to 1994) were 1.75 (95% CI, 1.21 to 2.54) for men in the lowest birth weight category (<5.5 lb) and 1.17 (95% CI, 0.91 to 1.49) for men with birth weights of 5.5 to 6.9 lb (Table 3⇓). There was no material change after adjustment for BMI, parental history of diabetes, smoking, and physical activity. The magnitude of this association was slightly greater when men who reported diabetes on the 1986 baseline questionnaire were included in the analysis (OR for men with birth weight < 5.5 lb, 1.88; 95% CI, 1.46 to 2.43). Because these prevalent self-reported cases were not validated by the supplementary questionnaire, they may have included some cases of insulin-dependent (type I) diabetes mellitus.
Among men who reported their birth weights, the mean BMI was 26.1 kg/m2 in 1994, and the age-adjusted BMI increased with increasing birth weight (Table 1⇑, the Figure⇓). To examine the association between birth weight and adult obesity, we calculated the OR of being in the highest quintile of adult BMI compared with the lowest quintile, with the middle birth weight category as the referent group (Table 4⇓). Compared with men in the referent group, men with birth weights of 8.5 to 9.9 lb had an age-adjusted OR for being in the highest (>28.2 kg/m2) versus the lowest (<23.2 kg/m2) quintile of adult BMI of 1.50 (95% CI, 1.31 to 1.70); for men with birth weights ≥10 lb, the OR was 2.08 (95% CI, 1.73 to 2.50). In contrast, men with birth weights ≤6.9 lb were significantly less likely to be in the highest quintile of adult BMI (Table 4⇓).
These findings support the hypothesis that prenatal exposures, for which birth weight is a marker, are associated with several important chronic diseases in adulthood. Low birth weight was associated with an increased risk of adult hypertension and diabetes; high birth weight was associated with an increased risk of adult obesity.
These results are consistent with previous reports,1 2 6 11 but our study differed in several respects. First, more stable estimates than in other reports were provided because the cohort contained an ample number of participants, particularly in extreme birth weight categories. Second, a wide geographic area was represented; cohort members come from all 50 states in the United States. Third, the mean systolic and diastolic blood pressures were substantially lower than those reported from Britain.1 Finally, for the diabetes analyses, we included as cases only participants with physician-diagnosed diabetes. In a previous report describing an association between low birth weight and glucose intolerance, subjects with diagnosed diabetes were excluded.2 When we also included prevalent cases of diabetes, the increased risk associated with low birth weight remained.
Although birth weight, adult height and weight, blood pressure, and the diagnoses of hypertension and diabetes were self-reported, validation studies of these factors demonstrated that they are reported reliably. Biased recall is unlikely because the association between birth weight and hypertension, diabetes, and obesity has only recently been reported. Detection bias also is unlikely because the proportion of participants having had routine physical examinations between 1988 and 1990 was similar across the birth weight categories.
In field studies by Barker and colleagues1 6 in Britain, blood pressure was measured by the investigators and was found to be inversely related to birth weight. In a recent study of men and women 64 to 71 years of age, systolic blood pressure was 5.2 mm Hg lower for each 1-kg (2.4 mm Hg/lb) increase in birth weight after adjustment for current BMI.1 Although the relation in our cohort was not linear, we performed a similar analysis for comparison (including hypertensive subjects) and found that the decrease was much smaller, only 0.31 mm Hg/lb. Interestingly, the mean systolic and diastolic blood pressures reported in the British series were substantially higher than those in our cohort. For example, in a recent study of British men with a mean age of 64 years, the mean systolic and diastolic blood pressures were 166 and 89.8 mm Hg, respectively,2 compared with 128 and 80.2 mm Hg in our cohort. The mean blood pressure values and the decrease per 1-lb increase in birth weight in our cohort were similar to those in a study of 50-year-old men in Sweden.12 The reason for the discrepancies in blood pressure values is uncertain but perhaps may be partially attributable to differences in the therapeutic approaches to hypertension between the United States and Britain.
The magnitude of the association between birth weight and hypertension may appear to be out of proportion to the small decrease in the mean systolic blood pressure per pound of birth weight. The blood pressure values reported may be lower because of treatment of the men with hypertension. In addition, the absolute difference in the age-adjusted cumulative incidence of hypertension between the lowest (38.0%) and middle (32.8%) birth weight categories in our cohort was 5.2%. This additional fraction of hypertensive men in the low birth weight category may have had only a slight impact on the mean systolic blood pressure, particularly if the hypertension had been treated.
The physiological basis for the apparent elevation in blood pressure associated with low birth weight remains unknown. Conceivably, one or more events in utero such as changes in fetal blood flow or hormonal variations may lead to abnormal development of a variety of organ systems involved in blood pressure control, including the kidneys,13 autonomic nervous system, endocrine glands, or the vasculature itself.6 In contrast to a recent report,12 we found no modification by adult height or BMI of the association between birth weight and adult hypertension.
We did not have information on duration of gestation and thus cannot determine whether the relation between low birth weight and hypertension was due to intrauterine growth retardation or premature birth with weight appropriate for gestational age. However, in two large female cohorts, we found that the risk of hypertension was similar in women who were born at term compared with those born 2 weeks or more prematurely.14
The mechanism for the association with diabetes also is unknown. Prenatal exposures may affect adult β-cell function and/or insulin resistance.2 A decrease in β-cell mass in low-birth-weight individuals15 similar to that seen in rats who were protein restricted in utero16 may also occur. Potentially, there may be a selective survival advantage for low-birth-weight infants with a predisposition to insulin resistance.17 Interestingly, adult BMI had no apparent effect on the association between low birth weight and risk of hypertension and diabetes, suggesting that in low-birth-weight infants the association between hypertension and NIDDM was independent of obesity.
Our results confirm that BMI increases with increasing birth weight,2 18 but this increase was not linear. Both genetic predisposition and factors associated with birth weight itself are likely to be determinants of adult obesity. It is well known that genetic influences on the development of obesity are important, but exactly how important is not understood. The mechanism for the potential independent association between birth weight and obesity also remains unclear.
Our findings are most directly generalizable to white men 40 years and older, but similar results have been reported for white women14 and Mexican American men.18 Low birth weight is not likely to be a major cause of hypertension in the United States because low-birth-weight infants make up <10% of the general population. It may be an important cause in subgroups with a higher frequency of low-birth-weight infants.19 Similarly, high birth weight is not likely to be a major cause of obesity in the United States. Nevertheless, our findings demonstrate independent associations between birth weight and hypertension, diabetes mellitus, and obesity, strongly suggesting that early life exposures play a role in the subsequent development of certain chronic diseases in adulthood.
Selected Abbreviations and Acronyms
|BMI||=||body mass index|
|HPFS||=||Health Professionals Follow-up Study|
|NIDDM||=||non–insulin-dependent diabetes mellitus|
This work was supported by research grants DK-45362, HL-35464, and CA-55075 from the NIH. We are indebted to the participants of the HPFS for their continuing cooperation and to Dr Sharon Curhan, Elaine Coughlan-Havas, Al Wing, Betsy Frost-Hawes, Mitzi Wolff, Kerry Pillsworth, Jan Vomacka, Jill Arnold, and Mira Koyfman for their expert help.
- Received April 23, 1996.
- Revision received July 15, 1996.
- Accepted July 27, 1996.
- Copyright © 1996 by American Heart Association
Law CM, de Swiet M, Osmond C, Fayers PM, Barker DJP, Cruddas AM, Fall CHD. Initiation of hypertension in utero and its amplification throughout life. BMJ. 1993;306:24-27.
Hales CN, Barker DJP, Clark PMS, Cox LJ, Fall C, Osmond C, Winter PD. Fetal and infant growth and impaired glucose tolerance at age 64. BMJ. 1991;303:1019-1022.
Barker DJP, Bull AR, Osmond C, Simmonds SJ. Fetal and placental size and risk of hypertension in adult life. BMJ. 1990;301:259-262.
Ascherio A, Rimm EB, Giovannucci EL, Colditz GA, Rosner B, Willett WC, Sacks F, Stampfer MJ. A prospective study of nutritional factors and hypertension among US men. Circulation. 1992;86:1475-1484.
Colditz GA, Willett WC, Stampfer MJ, Manson JE, Hennekens CH, Arky RA, Speizer FE. Weight as a risk factor for clinical diabetes in women. Am J Epidemiol. 1990;132:501-513.
Kleinbaum DG, Kupper LL, Muller KE. Applied Regression Analysis and Other Multivariable Methods. Boston, Mass: PWS-KENT Publishing; 1988.
Lithell HO, McKeigue PM, Berglund L, Mohsen R, Lithell U-B, Leon DA. Relation of size at birth to non-insulin dependent diabetes and insulin concentrations in men aged 50-60 years. BMJ. 1996;312:406-410.
Leon DA, Koupilova I, Lithell HO, Berglund L, Mohsen R, Vagero D, Lithell U-B, McKeigue PM. Failure to realise growth potential in utero and adult obesity in relation to blood pressure in 50 year old Swedish men. BMJ. 1996;312:401-406.
Curhan GC, Chertow GM, Willett WC, Spiegelman D, Colditz GA, Manson JE, Speizer FE, Stampfer MJ. Birth weight and adult hypertension and obesity in women. Circulation. 1996;94:1310-1315.
Cook JTE, Levy JC, Page RCL, Shaw JAG, Hattersley AT, Turner RC. Association of low birth weight with beta cell function in the adult first degree relatives of non-insulin dependent diabetic subjects. BMJ. 1993;306:302-306.
Swenne I, Crace CJ, Milner RDG. Persistent impairment of insulin secretory response to glucose in adult rats after limited period of protein-calorie malnutrition early in life. Diabetes. 1987;36:454-458.
McCance DR, Pettitt DJ, Hanson RL, Jacobsson THL, Knowler WC, Bennett PH. Birth weight and non-insulin dependent diabetes: thrifty genotype, thrifty phenotype, or surviving small baby genotype? BMJ. 1994;308:942-945.