other reasons (n=70), mostly due to technical difficulties in the suction system that were recognised before randomisation. 2514 infants were randomly allocated to receive suction (n=1263) and no suction (n=1251). Shortly after randomisation, 33 patients (18 suction, 15 no suction) did not meet inclusion criteria (because, for instance, of the presence
of congenital malformations). Nevertheless, we included these individuals in the intention-to-treat analysis.