Relation between infants' birth weight and mothers' mortality: prospective observational study

BMJ 2000; 320 doi: (Published 25 March 2000) Cite this as: BMJ 2000;320:839
  1. George Davey Smith, professor of clinical epidemiology (zetkin{at},
  2. Seeromanie Harding, senior research officerb,
  3. Michael Rosato, senior research officerb
  1. a Department of Social Medicine, University of Bristol, Canynge Hall, Bristol BS8 2PR
  2. b Office for National Statistics, London SW1V 2QQ
  1. Correspondence to: G Davey Smith
  • Accepted 7 December 1999

Several studies have shown inverse associations between birth weight and incidence of cardiovascular disease in adulthood, 1 2 suggesting that development in early life may influence the risk of disease many years later. The existence of intergenerational influences on birth weight, illustrated by correlations between the birth weight of parents and the birth weight of their offspring, suggests that birth weight of offspring should be associated with the risk of mortality from cardiovascular disease among parents. So far only one small study has investigated this possibility.3 In that study each kilogram decrease in birth weight of the infant was associated with a doubling in the risk of parental mortality from cardiovascular disease. We have replicated this potentially important finding in the Office for National Statistics' longitudinal study, a follow up of 1% of the population of England and Wales from the 1971 census.4

Methods and results

In the longitudinal study, information from birth registrations of infants during 1976–97 is linked to data from the census and death registration (1976–97) for study members. Birth weight was collected from 1976 onwards, but during the 1970s the data were not complete. The first recorded birth weight of infants to 44 813 women aged 15–45 years at birth registration was used as the exposure measure. Mean birth weight was 3210 g (SD 566 g). The average birth weight changed little over the study period (4.6 g increase a year, 95% confidence intervals 3.6 g to 5.6 g). The difference between average male birth weight and average female birth weight (99 g) was subtracted from each male birth weight. Socioeconomic position (housing tenure and car access) of mothers was taken from the 1971 census.

Mortality was analysed by using Cox's proportional hazards model in relation to birth weight, with adjustment for socioeconomic position and whether the birth occurred inside or outside marriage (with additional adjustment for sole registration or joint registration if outside marriage).

The table shows a substantial association between infants' birth weight and mothers' mortality from all causes and from cardiovascular disease. The relative risk of maternal cardiovascular mortality associated with a 1 kg lower birth weight in offspring was closely similar to that seen in the previous study of this issue (2.00 (1.18 to 3.33)3); the corresponding associated risks for the other two major causes of death (153 deaths from cancer (1.33; 1.03 to 1.72) and 38 accidental or violent deaths (1.06; 0.61 to 1.85)) were considerably smaller than those for cardiovascular mortality.

Infants' birth weight and mothers' mortality, 1976-97, according to data from Office for National Statistics' longitudinal study (England and Wales). Values are hazard ratios (95% confidence interval) unless stated otherwise

View this table:


We found a strong inverse relation between infants' birth weights and mothers' mortality from cardiovascular disease. Three basic processes could generate this association. Firstly, poor social circumstances could lead to both lower birth weight and higher mortality risk. Secondly, maternal health, nutritional, and behavioural profiles could influence both birth weight and cardiovascular mortality. Thirdly, intergenerational factors—such as genomic and epigenetic processes that lead to a positive correlation between the birth weights of mothers and their offspring—could influence cardiovascular risk.

Adjustment for socioeconomic position and marital status had little influence on the findings in either the current study or the previous investigation of this issue, 3 rendering a simple explanation in terms of socioeconomic confounding unlikely. In the current study we had no data on health status, but in the earlier study adjustment for a wide range of measures of health and health related behaviours reduced only slightly the association between infants' birth weight and mothers' cardiovascular mortality.3 The magnitude of the association in the current study is too great to be generated plausibly by the known associations between birth weight and maternal smoking, alcohol drinking, or anthropometry.5

The marked similarity between the current findings and those from the previous study—which related to an earlier generation of women living in widely different circumstances—suggests that an important influence is being uncovered by our analyses. Possible intergenerational influences on birth weight and cardiovascular risk therefore merit further investigation.


Contributors: GDS formulated the hypothesis regarding infants' birth weight and mothers' mortality. All authors discussed and specified the analyses, which were carried out by SH and MR. The first draft of the paper was written by GDS, with SH and MR contributing to the final draft. GDS is the guarantor of the paper.


  • Funding None.

  • Competing interests None declared.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract