Eight studies of asymptomatic women (n=124,647) were included. Four studies used a prospective cohort design (n=54,281), three used a retrospective cohort design (n=69,676), and one was a case-control study (n=690). No relevant studies of symptomatic women were identified.
None of the studies fulfilled all four quality criteria of an ideal study. Blinding of maternal anthropometric results and consecutive enrolment into the study were lacking in all but one of the included studies.
Accuracy of maternal pre-pregnancy BMI in predicting birth at <37 weeks' gestation (n=5).
Pre-pregnancy BMI was a poor predictor of pre-term birth. The positive LRs ranged from 0.96 to 1.75 and the negative LRs from 0.84 to 1.01.
Accuracy of maternal BMI in predicting birth at <32 weeks' gestation (n=2).
Pre-pregnancy BMI was also a poor predictor of pre-term birth at <32 weeks' gestation. The positive LRs ranged from 1.01 to 2.00 and the negative LRs from 0.83 to 1.00.
Accuracy of adequacy of pregnancy weight gain in predicting birth at <37 weeks' gestation (n=4).
Weight gain throughout pregnancy was a poor predictor of pre-term birth. The summary positive LR was 1.69 (95% CI: 1.48, 1.92) and the summary negative LR was 0.81 (95% CI: 0.77, 0.86).
Accuracy of maternal height in predicting pre-term birth (n=2).
These studies used thresholds of below the 25% quartile of the population or less than 152 cm in height, but neither was an accurate predictor of spontaneous pre-term birth. The positive LRs were 1.79 and 1.26, while the negative LRs were 0.75 and 0.96.