American Journal Of Obstetrics And Gynecology 2022

Non-invasive prediction models of intraamniotic infection in women with preterm labor.

Cobo T, Burgos-Artizzu XP, Collado MC, Andreu V, Sánchez-García AB, Filella X, Marín S, Cascante M, Bosch J, Ferrero S, Boada D, Murillo C, Rueda C, Ponce J, Palacio M, Gratacós E
Among women with preterm labor, those with intra-amniotic infection present the highest risk of early delivery and the most adverse outcomes. Identification of intra-amniotic infection requires amniocentesis, perceived as too invasive by women and physicians. Non-invasive methods for identifying intra-amniotic infection and/or early delivery are critical to focus early efforts on high-risk while avoiding unnecessary interventions in low-risk preterm labor women.We modeled the best performing models integrating biochemical data with clinical and ultrasound information to predict a composite outcome of intra-amniotic infection and/or spontaneous delivery within 7 days.From 2015-2020, we used data from a cohort of women admitted with diagnosis of preterm labor below 34 weeks at Hospital Clinic and Hospital Sant Joan de Déu, Barcelona, who had undergone amniocentesis to rule in/out intra-amniotic infection or inflammation. Transvaginal ultrasound, maternal blood and vaginal samples were prospectively performed at admission. Using high-dimensional biology, we explored vaginal proteins (by multiplex immunoassay), amino acids (by high-performance liquid chromatography) and bacteria (by 16S rRNA gene amplicon sequencing) to predict the composite outcome. We selected ultrasound, maternal blood and vaginal predictors that could be tested with rapid diagnostic techniques and developed prediction models employing Machine Learning that were applied in a validation cohort.We studied a cohort of 288 women with PTL below 34 weeks, of which 103 (35%) had a composite outcome of IAI and/or spontaneous delivery within 7 days. The sample was divided into derivation (n=116) and validation cohorts (n=172). Four prediction models were proposed, including ultrasound transvaginal cervical length, maternal C-reactive protein, vaginal IL-6 (using automated immunoanalyzer), vaginal pH (using pH meter), vaginal lactic acid (using reflectometer) and vaginal Lactobacillus genus (using quantitative-PCR), with areas under the curve ranging from 82.2% (+-3.1% 95% confidence interval) to 85.2%(+-3.1% 95% confidence interval), sensitivities ranging from 76.1 to 85.9% and specificities of 75.2 to 85.1%.These results provide proof-of-principle of how non-invasive methods suitable for point-of-care systems can select high-risk cases among women with preterm labor and might substantially aid in clinical management and outcomes, while improving use of resources and patient experience.Copyright © 2022. Published by Elsevier Inc.