A study investigating the relation between fetal growth and subsequent mortality of Swedish residents born in Uppsala during 1915-29 identified slow fetal growth rate rather than small size at birth as the factor associated with increased risk of ischaemic heart disease.1 The study also showed an expected association between birthweight and infant mortality. Surprisingly, birthweight was also associated with mortality in children aged 1-14 years. This association has not been previously observed, so we sought confirmation of the relation in the 1958 British birth cohort study. Because childhood death is now rare, we also examined whether birthweight had long term effects on disability.
Subjects, methods, and results
The 1958 cohort includes about 17 000 people born on 3-9 March 1958 in England, Scotland, and Wales followed up to 1991.2 Birthweight was recorded in ounces and classified into five categories (see table). Birthweight for gestational age was defined as a sex specific Z score for each week of gestation. A total of 423 deaths occurring between the ages of 1 month and 38 years were notified to the NHS central register, with 19% (82) of these occurring in the first year. Disability or chronic illness and social class were ascertained at ages 7, 16, 23, and 33 years. We applied logistic regression to calculate the risk of death and disability associated with birthweight and birthweight for gestational age.
For males, infant mortality was inversely associated with birthweight (odds ratio 2.13 (95% confidence interval 1.48 to 3.07) per 1000 g reduction in birthweight). For females, infant mortality was higher for those with birthweights under 3250 g, but not significantly, possibly because of the small numbers of deaths (table).
We found no clear associations with mortality after age 1 year. Significant inverse associations were found between birthweight and newly identified cases of disability at ages 7 (both sexes), 16 (males), and 23 (females). Risk of disability at ages 23 and 33 was raised for those with birthweights under 2500 g, although not always significantly. Birthweight for gestational age showed similar associations with mortality and disability to those described for birthweight. Confounding by social class did not account for the relations.
Comment
Although we found the well established relation between birthweight and infant mortality, there was no evidence of a relation with childhood mortality, as seen in the Swedish sample born in 1915-29.1 Fewer deaths occurred in our 1958 cohort, and this might contribute to the discrepant findings. More plausibly, the inconsistency is due to differences in the main causes of death. Causes related to poor intrauterine growth may have been common earlier this century but have become less so as overall death rates have fallen. Unfortunately, we did not have sufficient numbers of deaths to examine separate causes. Nevertheless, our failure to detect a relation suggests that the effect of birthweight on childhood mortality has weakened.
The risk of disability in low birthweight infants has been demonstrated previously, notably for growth, neuromotor, and sensory impairments.3–5 Our study suggests that the association is graded, with risk of disability at age 7 years increasing as birthweight falls, mirroring the trend for infant mortality. Moreover, effects of birthweight seem to persist, as suggested by associations with incident disability at later ages (16 for men and 23 for women). Risk of later disability among those weighing <2500 g at birth is raised at most ages, highlighting the vulnerability of this group. Improvements in neonatal care since the 1960s4 may have reduced the consequences of low birthweight in recent cohorts, but our study suggests that contemporary adults manifest long term sequelae related to their intrauterine growth.
Table.
Odds ratios (95% confidence interval) for all cause mortality and disability by birthweight and sex
| Age (years) | No of deaths/ with disability* | Birthweight (g)
|
P value for trend† | ||||
|---|---|---|---|---|---|---|---|
| <2500 (n=743) | 2500-3249 (n=3503) | 3250-3749 (n=5657) | 3750-4249 (n=4668) | >4249 (n=2111) | |||
| Mortality | |||||||
| Males: | |||||||
| <1 | 46 | 8.41 (2.03 to 34.7) | 4.51 (1.27 to 34.8) | 2.27 (0.63 to 8.18) | 1.58 (0.42 to 6.01) | 1.00 | <0.001 |
| 1-15 | 68 | 1.15 (0.31 to 4.25) | 1.07 (0.47 to 2.43) | 0.95 (0.45 to 2.02) | 1.01 (0.48 to 2.12) | 1.00 | 0.71 |
| 16-38 | 143 | 0.45 (0.13 to 1.51) | 1.23 (0.74 to 2.04) | 0.61 (0.36 to 1.03) | 0.77 (0.47 to 1.26) | 1.00 | 0.94 |
| Females: | |||||||
| <1 | 36 | 2.01 (0.27 to 14.9) | 2.94 (0.65 to 13.4) | 1.37 (0.29 to 6.44) | 1.62 (0.33 to 7.89) | 1.00 | 0.12 |
| 1-15 | 51 | 2.02 (0.39 to 10.4) | 2.11 (0.59 to 7.50) | 1.18 (0.33 to 4.27) | 2.31 (0.66 to 8.09) | 1.00 | 0.52 |
| 16-38 | 79 | 1.45 (0.45 to 4.70) | 1.09 (0.44 to 2.66) | 1.09 (0.47 to 2.56) | 1.22 (0.51 to 2.94) | 1.00 | 0.91 |
| Disability* | |||||||
| Males: | |||||||
| 7 | 403 (7005) | 1.78 (0.94 to 3.37) | 1.42 (1.00 to 2.01) | 1.24 (0.90 to 1.70) | 0.96 (0.69 to 1.34) | 1.00 | 0.02 |
| 16 | 541 (5746) | 1.51 (0.81 to 2.80) | 1.79 (1.32 to 2.42) | 1.18 (0.88 to 1.57) | 1.03 (0.76 to 1.38) | 1.00 | <0.001 |
| 23 | 197 (5017) | 3.65 (1.77 to 7.49) | 1.33 (0.80 to 2.22) | 1.03 (0.65 to 1.64) | 1.20 (0.76 to 1.89) | 1.00 | 0.09 |
| 33 | 212 (4387) | 2.32 (0.97 to 5.55) | 1.19 (0.72 to 1.98) | 1.30 (0.84 to 2.01) | 1.14 (0.73 to 1.78) | 1.00 | 0.15 |
| Females: | |||||||
| 7 | 274 (6672) | 2.35 (1.19 to 4.62) | 1.44 (0.88 to 2.37) | 1.32 (0.82 to 2.13) | 1.07 (0.64 to 1.77) | 1.00 | 0.001 |
| 16 | 392 (5545) | 0.71 (0.35 to 1.42) | 0.82 (0.56 to 1.19) | 0.90 (0.63 to 1.27) | 0.74 (0.51 to 1.08) | 1.00 | 0.41 |
| 23 | 158 (5267) | 2.24 (0.93 to 5.37) | 1.49 (0.80 to 2.75) | 1.01 (0.55 to 1.87) | 1.00 (0.52 to 1.89) | 1.00 | 0.02 |
| 33 | 242 (4741) | 2.06 (0.94 to 4.51) | 1.37 (0.80 to 2.36) | 1.28 (0.76 to 2.14) | 1.55 (0.91 to 2.63) | 1.00 | 0.17 |
Newly identified cases from school doctor and health visitor assessments of disabling “abnormal conditions” ages 7-16, and from individual reports of a limiting longstanding illness ages 23 and 33; the denominator is all those with data minus subjects with disability at any previous age.
Trend tests were performed with continuous measures of birthweight. The P values show the significance of linear trend except for disability at 7 for males, which has a quadratic trend.
Footnotes
Funding: Department of Health for England and Wales.
Competing interests: None declared.
References
- 1.Leon DA, Lithell HO, Vågerö D, Koupilová L, Mohsen R, Berglund L. Reduced fetal growth rate and increased risk of death from ischaemic heart disease: cohort study of 15 000 Swedish men and women born 1915-29. BMJ. 1998;317:241–245. doi: 10.1136/bmj.317.7153.241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ferri E. Life at 33: the fifth follow-up of the national child development study. London: National Children's Bureau; 1993. [Google Scholar]
- 3.Scottish Low Birthweight Study Group. The Scottish low birthweight study. I. Survival, growth, and neuromotor and sensory impairment. Arch Dis Child. 1992;67:675–681. doi: 10.1136/adc.67.6.675. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Stewart AL, Reynolds EOR, Lipscomb AP. Outcome for infants of very low birthweight: survey of world literature. Lancet. 1981;i:1038–1040. doi: 10.1016/s0140-6736(81)92198-x. [DOI] [PubMed] [Google Scholar]
- 5.Pharoah POD, Stevenson CJ, Cooke RWI, Stevenson RC. Clinical and subclinical deficits at 8 years in a geographically defined cohort of low birthweight infants. Arch Dis Child. 1994;70:264–270. doi: 10.1136/adc.70.4.264. [DOI] [PMC free article] [PubMed] [Google Scholar]