Effectiveness and Costs of Molecular Screening and Treatment for Bacterial Vaginosis to Prevent Preterm Birth

The AuTop Randomized Clinical Trial

Florence Bretelle, MD, PhD^1,2; Sandrine Loubière, PhD³; Raoul Desbriere, MD⁴; et al

Anderson Loundou, PhD³; Julie Blanc, MD, PhD^3,5; Hélène Heckenroth, MD¹; Thomas Schmitz, MD, PhD⁶; Alexandra Benachi, MD, PhD^7,8; Bassam Haddad, MD^9,10; Franck Mauviel, MD¹¹; Xavier Danoy, MD¹²; Pierre Mares, MD¹³; Nawal Chenni, MD¹⁴; Jean-Pierre Ménard, MD, PhD¹⁵; Jean-François Cocallemen, MSc¹⁶; Nadia Slim, MD; Marie Victoire Sénat, MD, PhD^17,18; Céline Chauleur, MD, PhD¹⁹; Caroline Bohec, MD²⁰; Gilles Kayem, MD, PhD²¹; Cynthia Trastour, MD²²; André Bongain, MD, PhD²²; Patrick Rozenberg, MD, PhD^18,23; Valerie Serazin, MD, PhD^24,25; Florence Fenollar, MD, PhD^26,27; for the Groupe de Recherche en Obstetrique et Gynécologie (GROG) Investigators

Author Affiliations Article Information

JAMA Pediatr. Published online July 17, 2023. doi:10.1001/jamapediatrics.2023.2250

visual abstract icon

Visual

Abstract

Key Points

Question  What are the medical and economic effects of screen and treat for bacterial vaginosis using point-of-care quantitative real-time polymerase chain reaction during pregnancy?

Findings  In this randomized clinical trial including 6671 pregnant women enrolled before 20 weeks’ gestation assigned to screen and treat or usual care, the preterm birth rate was 3.8% and 4.6%, respectively, which was not significantly different. Total costs were also not significantly different.

Meaning  Compared with usual care, screen and treat for bacterial vaginosis did not reduce the risk of preterm birth; however, this strategy should be further evaluated in nulliparous and high-risk multiparous women.

Abstract

Importance  Bacterial vaginosis (BV) is a well-known risk factor for preterm birth. Molecular diagnosis of BV is now available. Its impact in the screening and treatment of BV during pregnancy on preterm births has not been evaluated to date.

Objective  To evaluate the clinical and economic effects of point-of-care quantitative real-time polymerase chain reaction screen and treat for BV in low-risk pregnant women on preterm birth.

Design, Setting, and Participants  The AuTop trial was a prospective, multicenter, parallel, individually randomized, open-label, superiority trial conducted in 19 French perinatal centers between March 9, 2015, and December 18, 2017. Low-risk pregnant women before 20 weeks’ gestation without previous preterm births or late miscarriages were enrolled. Data were analyzed from October 2021 to November 2022.

Interventions  Participants were randomized 1:1 to BV screen and treat using self-collected vaginal swabs (n = 3333) or usual care (n = 3338). BV was defined as Atopobium vaginae (Fannyhessea vaginae) load of 10⁸ copies/mL or greater and/or Gardnerella vaginalis load of 10⁹ copies/mL or greater, using point-of-care quantitative real-time polymerase chain reaction assays. The control group received usual care with no screening of BV.

Main Outcomes and Measures  Overall rate of preterm birth before 37 weeks’ gestation and total costs were calculated in both groups. Secondary outcomes were related to treatment success as well as maternal and neonate health. Post hoc subgroup analyses were conducted.

Results  Among 6671 randomized women (mean [SD] age, 30.6 [5.0] years; mean [SD] gestational age, 15.5 [2.8] weeks), the intention-to-treat analysis of the primary clinical and economic outcomes showed no evidence of a reduction in the rate of preterm birth and total costs with the screen and treat strategy compared with usual care. The rate of preterm birth was 3.8% (127 of 3333) in the screen and treat group and 4.6% (153 of 3338) in the control group (risk ratio [RR], 0.83; 95% CI, 0.66-1.05; P = .12). On average, the cost of the intervention was €203.6 (US $218.0) per participant, and the total average cost was €3344.3 (US $3580.5) in the screen and treat group vs €3272.9 (US $3504.1) in the control group, with no significant differences being observed. In the subgroup of nulliparous women (n = 3438), screen and treat was significantly more effective than usual care (RR, 0.62; 95% CI, 0.45-0.84; P for interaction = .003), whereas no statistical difference was found in multiparous (RR, 1.30; 95% CI, 0.90-1.87).

Conclusion and Relevance  In this clinical trial of pregnant women at low risk of preterm birth, molecular screening and treatment for BV based on A vaginae (F vaginae) and/or G vaginalis quantification did not significantly reduce preterm birth rates. Post hoc analysis suggests a benefit of screen and treat in low-risk nulliparous women, warranting further evaluation in this group.