Diagnosis-Day Prothrombin Time and Transaminase as Exploratory Risk-Stratification Indicators for Severe Neonatal Outcomes in Acute Fatty Liver of Pregnancy.

OBJECTIVES

Acute fatty liver of pregnancy (AFLP) is a life-threatening condition associated with high neonatal morbidity. This study aimed to develop a minimal, clinically feasible prediction model using diagnosis-day laboratory indicators to stratify the risk of severe adverse neonatal outcomes (SANO).

METHODS

In this single-center retrospective cohort (2004-2021), 18 pregnancies with AFLP confirmed by Swansea criteria were analyzed. Firth penalized logistic regression assessed associations between the first laboratory values on the diagnosis day and SANO (intrauterine death/severe asphyxia). A two-variable model combining prothrombin time (PT) and peak transaminase was developed. Model performance was evaluated via area under the receiver operating characteristic curve (AUC) and leave-one-out cross-validation (LOOCV).

RESULTS

The incidence of AFLP was 0.71 per 10,000 pregnancies. SANO occurred in 4 cases (17.39%). Diagnosis-day coagulopathy, especially prolonged PT, was the strongest predictor. The PT+transaminase model showed excellent discriminatory performance (AUC=0.982) and robustness upon internal validation (LOOCV AUC=0.964), with a negative predictive value of 1.000 for ruling out SANO.

CONCLUSIONS

Coagulation dysfunction at diagnosis is a key indicator for neonatal risk in AFLP. The proposed PT-based model offers a practical bedside tool for early obstetric risk stratification, aiding in urgent clinical decision-making for delivery planning and neonatal resuscitation preparedness.