Background Transfusion practices are highly variable between hospitals and previous studies have suggested that blood transfusions may increase the risk of necrotizing enterocolitis (NEC).

Aim To explore the association between blood transfusions and incidence of NEC in extremely preterm infants.

Methods We used data from a Swedish population-based study including extremely preterm infants (<27 weeks) born between 2004–2007, (n = 602). All data on blood transfusions and haemoglobin (Hb) concentrations up to 28 days of age was collected for survivors. We performed a nested case-control study where two controls were chosen for each case of NEC (n = 21).

Results During the first 28 days of life, infants received a median (25^th-75^th percentile) of 6 (3–9) blood transfusions resulting in 75 (44–120) ml/kg of blood. Predictors for receiving a higher volume of blood transfusions were days on respiratory support (R = 0.345, p < 0.001), hospital (R = 0.339, p < 0.001), low birth weight (R = -0.236, p < 0.001) and total steroid dose (R = 0.209, p < 0.001). Hb was not a significant predictor.

Overall NEC incidence was 5.8%. There was no significant difference between NEC cases and controls in number of blood transfusions (p = 0.420), volume of blood transfused from birth to NEC diagnosis (p = 0.274), or during the 48 h preceding NEC diagnosis (p = 0.459).

Conclusions Blood transfusions were given liberally in Sweden compared to other studied populations. Morbidity related variables, especially those related to respiratory illness, were significant predictors of blood transfusion. NEC incidence was comparable with other populations but no significant association was found between blood transfusions and NEC among these extremely preterm infants.