Caffeine and NEC: a 45,624-infant signal is not a timing rule

In a 45,624-infant Chinese cohort, caffeine was not associated with overall Bell stage IIA-or-higher NEC after adjustment, while first doses within 72 hours tracked with less surgical NEC. Baseline imbalance, missing dose data, and a reporting mismatch make that timing signal observational, not a protocol.

THE HOT TAKE

The overall NEC result is the anchor; the surgical timing signal is the hypothesis.

A five-year Chinese Neonatal Network cohort linked caffeine records with in-hospital NEC outcomes in 45,624 infants born before 32 weeks. After adjustment, any caffeine exposure was not associated with Bell stage IIA-or-higher NEC. First doses within 48 hours and from 48 to 72 hours were each associated with lower odds of surgical NEC than no caffeine, but this was not a randomized timing comparison.

The chart shows crude patient-level proportions, not adjusted effects or repeated event counts. NEC occurred in 2,315/36,514 caffeine-exposed infants and 544/9,110 unexposed infants; surgical NEC occurred in 849/36,514 and 263/9,110, respectively. The exposed group started out clinically different, so the small raw gaps cannot be read causally.