The aim of this study was to determine the developmental profile of mannose-binding protein (MBP) in preterm infants. MBP was measured in 885 longitudinally collected serum samples from 168 preterm infants, and 63 were genotyped with respect to the codon 54 mutation in the MBP gene. MBP level/codon 54 genotyping were also determined on the cord blood of 146/123 term infants and 138/123 adults, respectively. The best cut-off values of MBP for dividing preterm, term infants and adults into 'low' and 'high' MBP groups were 400 ng/ml (55 low, 113 high), 700 ng/ml (35 low, 111 high) and 750 ng/ml (33 low, 105 high), respectively, by achieving the least number of misclassifications according to the codon 54 mutation. The relative risk of the 'low' groups for presence of the codon 54 mutation compared with 'high' groups were 42.4, 67.9 and 22.9 for preterm, term infants and adults, respectively (P < 0.00001). The gestational age and birth weight of the 'low' (n = 55) and 'high' (n = 113) MBP groups of the 168 preterm infants were 29.5 +/- 2.8 weeks, 30.5 +/- 2.8 weeks (P = 0.03) and 1230 +/- 317 g, 1277 +/- 289 g (P = 0.35), respectively. The mean MBP levels of these two groups of preterm infants were different (P < 0.001) at all ages measured. As a whole group, the MBP level rose from a mean of 500 ng/ml at 25 weeks gestation to 1700 ng/ml at 20 weeks post full-term. The mortality rates of 'low' and 'high' MBP groups of preterm infants were 22% and 12%, respectively (P = 0.113). This difference in mortality was due to gestational age and birth weight standard deviation score (SDS) after adjusting for length of gestation and gender (P = 0.0001) rather than to low MBP levels (P = 0.65). MBP levels were not related to birthweight SDS score (P = 0.26). The mean +/- s.d. MBP levels for preterm, term infants and adults without the codon 54 mutation were 1225 +/- 701 ng/ml (n = 45), 2064 +/- 829 ng/ml (n = 88) and 2473 +/- 1395 ng/ml (n = 95), respectively; the corresponding values for those with the codon 54 mutation were 130 +/- 275 ng/ml (n = 18), 533 +/- 665 ng/ml (n = 35) and 330 +/- 225 ng/ml (n = 28), respectively. Intra-uterine growth retardation in preterm infants does not influence MBP levels. For those without the codon 54 mutation, there is a significant difference in MBP level between the three age groups. For those with the codon 54 mutation, there is a significant difference between preterm and term infants, but not between term infants and adults. We conclude that there is a maturation in MBP levels for preterm infants, and that a moderately low MBP phenotype does not affect survival. We cannot exclude an effect of profoundly reduced MBP levels (characteristic of individuals homozygous for the codon 54 mutation), since no such preterm infant was identified in this study.