(469) Identifying Culture-Positive Intrapartum Fever: Risk Factors from a Retrospective Cohort and CHAID Analysis

A retrospective cohort study included all deliveries (2011-2024) at a single, university affiliated, tertiary medical center (~13,000 annual deliveries). Women with intrapartum fever (maternal temperature of ≥38.0°C (100.4°F) during labor) were categorized based on bacterial culture results. Cultures were obtained from blood, urine, and placenta. Univariate analysis was followed by multivariable logistic regression. Moreover,  a CHAID (Chi-squared Automatic Interaction Detector) decision tree was constructed using only variables that remained significant in the multivariable model. Women with intrapartum fever who had missing data regarding bacterial culture sample were excluded from study.

Results:
1. During the study period, 2,845/146,999 (3.9%) women had intrapartum fever, with 352 (12.4%) yielding at least one positive culture: blood (39%), placenta (37.5%), and urine (23.4%).
2. In the univariate analysis, positive cultures were associated with older maternal age, earlier gestational age, lower gestational weight gain, nulliparity, IVF, multiple gestation, cesarean delivery and maternal fever >39°C (p< 0.05).
3. In the multivariable logistic regression model, preterm delivery (OR=2.47, 95% CI 1.51–4.05, p< 0.001), IVF (OR=1.71, 95% CI 1.16–2.5, p=0.007), multiple gestation (OR=2.68, 95% CI 1.4–5.1, p=0.003), and maternal fever >39°C (OR=1.67, 95% CI 1.11–2.51, p=0.013), were found to be independent risk factors for the study outcome.
4. In the CHAID model, preterm birth was the strongest predictor of positive cultures (p< 0.001). Among preterm births, age >35 and nulliparity further increased risk. In term births, fever >39°C and weight gain >12 kg were linked to higher positivity rate.

Conclusion:
Positive cultures in intrapartum fever are associated with identifiable clinical risk factors. The CHAID decision tree offers a practical tool for infection risk stratification and targeted management.