THE PRETERM PREDICTION STUDY - RISK-FACTORS IN TWIN GESTATIONS

OBJECTIVE: Our purpose was to determine the association between the pr esence of bacterial vaginosis, fetal fibronectin, and a short cervix a nd the risk of spontaneous preterm birth of twins. STUDY DESIGN: We pr ospectively screened 147 women with twins at 24 and 28 weeks' gestatio n for more than 50 potential risk factors for spontaneous preterm birt h. We also measured cervical length with ultrasound scans and tested f or the presence of bacterial vaginosis. Fetal fibronectin level was ev aluated every 2 weeks from 24 to 30 weeks' gestation. Outcomes include d spontaneous preterm birth at <32 weeks, <35 weeks, and <37 weeks. RE SULTS: Among twin as compared with singleton pregnancies, a cervical l ength less than or equal to 25 mm was more common at both 24 and 28 we eks, a statistically significant difference. There were no significant differences in most other risk factors. Of the factors evaluated by m eans of univariate analysis at 24 weeks, only a short cervix (less tha n or equal to 25 mm) was consistently associated with spontaneous pret erm birth. The odds ratios and 95% confidence interval for spontaneous preterm birth at <32 weeks, <35 weeks, and <37 weeks were 6.9 (2.0 to 24.2), 3.2 (1.3 to 7.9), and 2.8 (1.1 to 7.7). At 28 weeks, a cervica l length less than or equal to 25 mm was not a strong predictor of spo ntaneous preterm birth. At both 28 weeks (odds ratio, 9.4; confidence interval, 1.0 to 67.7) and 30 weeks (odds ratio, 46.1; confidence inte rval, 4.2 to 1381), a positive fetal fibronectin result was significan tly associated with spontaneous preterm birth at <32 weeks. Bacterial vaginosis at 24 or 28 weeks was not associated with spontaneous preter m birth of twins. Multivariate analysis confirmed the association betw een cervical length less than or equal to 25 mm at the 24-week visit a nd spontaneous preterm birth and also confirmed that at 24 weeks the o ther risk factors were less consistently and often not statistically s ignificantly associated with spontaneous preterm birth. Of the risk fa ctors evaluated at 28 weeks, only a positive fetal fibronectin was ass ociated with a significantly increased risk for spontaneous preterm bi rth. CONCLUSIONS: Most known risk factors for spontaneous preterm birt h were not significantly associated with spontaneous preterm birth of twins. At 24 weeks, cervical length less than or equal to 25 mm was th e best predictor of spontaneous preterm birth at <32 weeks, <35 weeks, and <37 weeks. Of the risk factors evaluated at 28 weeks, fetal fibro nectin was the only statistically significant predictor of spontaneous preterm birth at <32 weeks.