Preterm delivery is a public health priority in Martinique, a French oversees territory located in the Caribbean, with a rate of 10,5% substantially higher than in Continental France (7,5%). Bacterial vaginosis is an imbalance of vaginal flora known to be associated with preterm delivery risk. Studies driven in the United States have pointed out that 50% of the excess in preterm birth rate observed in Afro-American population compared to Caucasian population may be attributable to infection. Bacterial vaginosis appears much more prevalent in our population than in Continental France (approximatively 30% versus 7%) and may constitute an explanation to the discrepancy in preterm delivery rate.

Bacterial vaginosis during pregnancy is considered as a risk factor for preterm birth since the 1990's. A meta-analysis realized in 2003 found an Odds ratio of 2.4 for preterm delivery among women with bacterial vaginosis carriage, culminating at 7.5 when carriage before 17 weeks gestation was considered. Preterm delivery is a pregnancy complication known to be more prevalent among patients of African descent. Various hypothesis have been tested to account for this observed discrepancy and several authors consider that up to 50% of the difference in preterm delivery rate between Afro-american and Caucasians may have and infectious origin, with up to 30% attributable to a very different bacterial vaginosis carriage among patients of various ethnic origin. Several factors may interfere in the link observed between bacterial vaginosis and preterm delivery, which are also unevenly distributed among women of different ethnic origin. Exposure to stress, low socio-economic status, single marital status or low access to prenatal care are among those factors. Thus analyzing a relation between bacterial vaginosis and preterm delivery mandates to control for such factors. The present study will proceed to a vaginal sampling before 17 gestational age (GA) to diagnose a bacterial vaginosis through the Nugent score which is the standard diagnostic method for this pathology. At the same time, a written survey concerning stressful events in the past 12 months is submitted to volunteer patients in order to use the social readjustment rating scale by Holmes and Rahe and to account for the role of stress as a potential confounding factor. Socio-demographic data, obstetrical history and complications during pregnancy were also recorded and patients followed-up prospectively until delivery. Caregivers were blinded to the results of initial vaginal sampling. No alteration of the normal pregnancy follow-up was required by the study design.

Inclusion Criteria:

• All women followed-up since the beginning of pregnancy in the study center and supposed to deliver in this same center

Exclusion Criteria:

• Previous preterm delivery,
• Previous late abortion,
• Multiple pregnancy,
• Cervical cerclage before inclusion

Bacterial vaginosis appears much more prevalent in Martinique than in Continental France (approximatively 30% versus 7%) and may constitute an explanation to the discrepancy in preterm delivery rate.