Fentanyl's fivefold BPD signal still carries major confounding

In 142 matched ventilated preterm infants, BPD occurred in 13/50 fentanyl-exposed versus 5/92 unexposed infants; adjusted OR 5.49 (95% CI 1.72 to 17.54). But surfactant-treated RDS was 88.0% versus 43.5%, so confounding and reverse causation make the proposed 3-day threshold a hypothesis, not a stop rule.

THE HOT TAKE

The association is large. So is the imbalance in baseline lung disease.

A 2026 single-center cohort studied infants born before 34 weeks who received respiratory support and had documented moderate-to-severe pain. Of 315 eligible infants, 67 received continuous fentanyl and 248 did not. Matching on gestational age and birth-weight z-score produced an analytic cohort of 50 exposed and 92 unexposed infants. This was not a randomized comparison.

WHAT SURVIVED ADJUSTMENT

After adjustment for gestational age and fetal distress, fentanyl exposure remained associated with BPD: adjusted OR 5.49 (95% CI 1.72 to 17.54). Among 77 infants with at least one invasive-ventilation day, the duration model returned an adjusted exp(beta) of 2.25 (95% CI 1.54 to 3.29). Association is the correct word; neither model establishes causation.