Maternal and biochemical predictors of spontaneous preterm birth among nulliparous women: a systematic analysis in relation to the degree of prematurity.

BACKGROUND: Nulliparous women are at increased risk of spontaneous preterm birth. Other maternal and biochemical risk factors have also been described. However, it is unclear whether these associations are strong enough to offer clinically useful prediction. It is also unclear whether the predictive power of these factors varies in relation to the degree of prematurity.
METHODS: The risk of spontaneous preterm birth associated with maternal characteristics and second trimester serum screening data was analysed in a dataset of 84 391 first births in Scotland between 1992 and 2001 using Cox and logistic regression. Variation in the relative risk of preterm birth over the period 24-36 weeks was assessed using a test of the proportional hazards assumption.
RESULTS: The risk of spontaneous preterm birth was positively associated with maternal serum levels of alpha-fetoprotein, socioeconomic deprivation, number of previous therapeutic abortions, smoking, and being unmarried and was negatively associated with height and body mass index. The risk of preterm birth at 24-28 weeks, but not later gestations, was increased in association with maternal levels of human chorionic gonadotrophin >95th percentile, maternal age <20, and two or more previous miscarriages. The area under the receiver operating characterise curve (95% CI) for models based on these factors was 0.67 (0.63-0.71) for 24-28 weeks, 0.65 (0.62-0.68) for 29-32 weeks, and 0.62 (0.61-0.63) for 33-36 weeks.
CONCLUSIONS: Time to event analytic methods can identify factors that are differentially associated with spontaneous preterm birth according to the degree of prematurity. However, models based on maternal and biochemical data perform poorly as a screening test for any degree of spontaneous preterm birth.