RESUMO
OBJECTIVE: To assess the diagnostic accuracy and cost-effectiveness of fetal fibronectin (fFN) and cervical phosphorylated insulin-like growth factor binding protein-1 (phIGFBP-1) tests, individually and in combination, to predict preterm delivery within 48 hours, 7 days and 14 days in symptomatic women. METHOD: We selected women in Victoria, British Columbia, who presented between January 2008 and December 2017 at <34 weeks gestation at intermediate risk for labour (intact membrane, cervical dilatation <3 cm, and >6 contractions per hour). We calculated sensitivity, specificity, and positive and negative predictive values (PPV, NPV) for independent and concurrent testing and conducted a cost-effectiveness analysis to ensure appropriate test utilization. RESULTS: We identified 2911 cases. Both fFN and phIGFBP-1 tests showed high and comparable NPV in predicting risk of delivery within 48 hours, 7 days and 14 days (fFN: 99.3%, 98.5% and 97.3%; phIGFBP-1: 98.8%, 97.9% and 96.1%). In 1976 cases, samples for fFN and phIGFBP-1 tests were collected and analyzed concurrently. Concurrent analysis increased specificity (90.8%, 91.4%, and 91.8%) and PPV (11.8%, 19.8% and 24.2%). Independently, both tests had comparable sensitivity, while the fFN test had higher specificity. Concurrent testing offered the highest PPV. The net gain in PPV comes with a clinically insignificant net loss (<1%) in NPV when compared with either of the tests individually. CONCLUSION: Clinical usefulness of PPV for either test is limited. Routine concurrent testing comes with additional costs, and fFN has additional collection requirements. Point-of-care phIGFBP-1 testing has proven to be cheaper, simpler, and equally effective. Ordering physicians should be provided with education on how to interpret test results and should have protocols to guide clinical decision making.
