Objective To identify the incidence of in-hospital mortality among preterm infants by infant and health service characteristics, diagnoses and procedures, and congenital anomalies.

Setting Republic of Ireland

Population Preterm infants between 24 and 36 weeks gestation born in hospital and admitted to intensive care unit (ICU), 2005-2008

Methods Diagnostic and procedural codes from hospital discharge records were used to identify outcomes of interest. Nationally representative incidence rates and corresponding exact 95% confidence intervals were calculated for in-hospital mortality, co-morbidities, procedures, congenital anomalies and indications for delivery. Unadjusted risk ratios were derived to identify factors associated with in-hospital mortality.

Main outcomes measures in-hospital mortality; morbidity status; congenital anomalies

Results 6,599 preterm infants were admitted to ICU of whom 3.9% died prior to hospital discharge. Mortality rates were high among infants diagnosed with necrotising enterocolitis (29.8%), patent ductus arteriosus (12.4%), intra-ventricular haemorrhage (19.6%), or other brain haemorrhage (56.4%). The total major congenital anomaly in-hospital mortality rate was 14.9%. Markedly high in-hospital mortality rates (>25%) were identified for several congenital anomaly classifications, which subsequently translated into a five-fold risk (RR 5.1; 95% CI 4.0-6.6) of in-hospital mortality.

Conclusions Our population-based study ensures provision of unbiased and reliable estimates of in-hospital mortality among preterm infants admitted to Irish ICUs. A population focus using large-scale registers may be important in monitoring preterm infant mortality rates among infants admitted to ICU.