Poor validity of the routine diagnosis of necrotising enterocolitis in preterm infants at discharge

Aim Necrotising enterocolitis contributes considerably to the mortality of preterm infants, but most questions remain unsolved after decades of extensive research. This Danish study investigated the validity of necrotising enterocolitis diagnoses at discharge according to Bell's staging system. Methods We conducted a retrospective single-centre cohort study of 714 preterm infants with a gestational age of less than 30 weeks born in 2006–2013. The infants were diagnosed with necrotising enterocolitis according to Bell's stages 2–3 at discharge and in retrospect by an expert panel, which served as our gold standard. Results The sensitivity of necrotising enterocolitis diagnosed at discharge was 0.72–0.75 depending on whether spontaneous intestinal perforation was included as necrotising enterocolitis or not. The positive predictive value of the diagnosis was 0.49–0.61. The incidence was significantly higher when diagnosed at discharge than when diagnosed by the expert panel (11.1 versus 9.0%, p = 0.03). The mortality rate for infants who were underdiagnosed at discharge was 50.0%, and it was 25.8% for infants who were overdiagnosed (p = 0.10). Conclusion We found poor validity for the discharge diagnosis of necrotising enterocolitis. In future, a better way of defining the disease is needed for large-scale epidemiologic research.