Predictive index for adverse perinatal outcome in pregnancies complicated by fetal growth restriction
J E Powel,E W Zantow,M F Bialko,L G Farley,M L Lawlor,S J Mullan,L K Vricella,T M Tomlinson
DOI: https://doi.org/10.1002/uog.26044
Abstract:Objectives: To develop and validate an index predictive of adverse perinatal outcome (APO) in pregnancies meeting the consensus-based criteria for fetal growth restriction (FGR) endorsed by the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG). Methods: This was a retrospective analysis of consecutive singleton non-anomalous gestations meeting the ISUOG-endorsed criteria for FGR at a single tertiary care center from November 2010 to August 2020. The dataset was divided randomly into a development set (two-thirds) and a validation set (one-third). The primary composite APO comprised one or more of: perinatal demise, Grade III-IV intraventricular hemorrhage (IVH), periventricular leukomalacia (PVL), seizures, hypoxic ischemic encephalopathy (HIE), necrotizing enterocolitis (NEC), sepsis, bronchopulmonary dysplasia (BPD) and length of stay in the neonatal intensive care unit (NICU) > 7 days. Regression analysis incorporated clinical factors readily available at the time of FGR diagnosis. The sum of β coefficient-based weights yielded an index score, the performance of which was assessed in the validation set. Score cut-offs were selected to identify 'high-risk' and 'low-risk' ranges for which positive (PPV) and negative (NPV) predictive values and positive (LR+) and negative (LR-) likelihood ratios were calculated. Results: Of the 875 consecutive pregnancies that met the criteria for FGR and were included in the study cohort, 405 (46%) were complicated by one or more components of the composite APO, including 54 (6%) perinatal deaths, 22 (3%) neonates with Grade III-IV IVH and/or PVL, nine (1%) with seizures and/or HIE, 91 (10%) with BPD, 57 (7%) with sepsis, 21 (2%) with NEC, and 361 (41%) who remained in the NICU > 7 days. In addition, 270 (31%) pregnancies were delivered by Cesarean section for non-reassuring fetal status, 43 (5%) were admitted to the NICU for < 7 days, 79 (9%) had 5-min Apgar score < 7, 125/631 (20%) had a cord gas pH ≤ 7.1 and 35/631 (6%) had a base excess ≥ 12 mmol/L. The predictive index we developed included seven factors available at the time of FGR diagnosis: hypertensive disorder of pregnancy (HDP) (+8 points), chronic hypertension without HDP (+4 points), gestational age ≤ 32 weeks (+5 points), absent or reversed end-diastolic flow in the umbilical artery (+8 points), prepregnancy body mass index ≥ 35 kg/m2 (+3 points), isolated abdominal circumference < 3rd percentile (-4 points) and non-Hispanic black race (-2 points). The bias-corrected bootstrapped (1000 replicates) area under the receiver-operating-characteristics curve (AUC) of the predictive index for composite APO in the validation group was 0.88 (95% CI, 0.84-0.92), which was similar to that in the development group (AUC, 0.86 (95% CI, 0.82-0.89); P = 0.34). In the total cohort, 40% of pregnancies had a low-risk index score (≤ 2), associated with a NPV of 85% (95% CI, 81-88%) and a LR- of 0.21 (95% CI, 0.16-0.27), and 23% had a high-risk index score (≥ 10), associated with a PPV of 96% (95% CI, 93-98%) and a LR+ of 27.36 (95% CI, 14.33-52.23). Of the remaining pregnancies that had an intermediate-risk score, 50% were complicated by composite APO. Conclusion: An easy-to-use index incorporating seven clinical factors readily available at the time of FGR diagnosis is predictive of APO and may prove useful in counseling and management of pregnancies meeting the ISUOG-endorsed criteria for FGR. © 2022 International Society of Ultrasound in Obstetrics and Gynecology.