Prediction of preterm delivery in symptomatic women using PAMG‐1, fetal fibronectin and phIGFBP‐1 tests: systematic review and meta‐analysis

Abstract

To assess the accuracy of placental alpha microglobulin-1 (PAMG-1), fetal fibronectin (fFN) and phosphorylated insulin-like growth factor-binding protein-1 (phIGFBP-1) tests in predicting spontaneous preterm birth (sPTB) within 7 days of testing in women with symptoms of preterm labor, through a systematic review and meta-analysis of the literature. The test performance of each biomarker was also assessed according to pretest probability of sPTB ≤ 7 days. The Cochrane, MEDLINE, PubMed and ResearchGate bibliographic databases were searched from inception until October 2017. Cohort studies that reported on the predictive accuracy of PAMG-1, fFN and phIGFBP-1 for the prediction of sPTB within 7 days of testing in women with symptoms of preterm labor were included. Summary receiver-operating characteristics (ROC) curves and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and positive (LR+) and negative (LR-) likelihood ratios were generated using indirect methods for the calculation of pooled effect sizes with a bivariate linear mixed model for the logit of sensitivity and specificity, with each diagnostic test as a covariate, as described by the Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy. Bivariate mixed model pooled sensitivity of PAMG-1, fFN and phIGFBP-1 for the prediction of sPTB ≤ 7 days was 76% (95% CI, 57-89%), 58% (95% CI, 47-68%) and 93% (95% CI, 88-96%), respectively; pooled specificity was 97% (95% CI, 95-98%), 84% (95% CI, 81-87%) and 76% (95% CI, 70-80%) respectively; pooled PPV was 76.3% (95% CI, 69-84%) (P < 0.05), 34.1% (95% CI, 29-39%) and 35.2% (95% CI, 31-40%), respectively; pooled NPV was 96.6% (95% CI, 94-99%), 93.3% (95% CI, 92-95%) and 98.7% (95% CI, 98-99%), respectively; pooled LR+ was 22.51 (95% CI, 15.09-33.60) (P < 0.05), 3.63 (95% CI, 2.93-4.50) and 3.80 (95% CI, 3.11-4.66), respectively; and pooled LR- was 0.24 (95% CI, 0.12-0.48) (P < 0.05), 0.50 (95% CI, 0.39-0.64) and 0.09 (95% CI, 0.05-0.16), respectively. The areas under the ROC curves for PAMG-1, fFN and phIGFBP-1 for sPTB ≤ 7 days were 0.961, 0.874 and 0.801, respectively. In the prediction of sPTB within 7 days of testing in women with signs and symptoms of preterm labor, the PPV of PAMG-1 was significantly higher than that of phIGFBP-1 or fFN. Other diagnostic accuracy measures did not differ between the three biomarker tests. As prevalence affects the predictive performance of a diagnostic test, use of a highly specific assay for a lower-prevalence syndrome such as sPTB may optimize management. Copyright © 2018 ISUOG. Published by John Wiley & Sons Ltd.