Introduction Growth monitoring is a cardinal precept of paediatric practice. The UK 1990 reference curves for preterm infants, based on cross sectional birth data from live and stillborn infants at different gestational ages, have been incorporated into the UK-WHO growth charts. These and related charts cannot be considered a standard as the optimum pattern of growth after preterm birth is unknown.

Aims We aimed to construct birthweight centiles and illustrate the range of contemporaneous preterm postnatal growth to provide a population-based benchmark and a baseline for future examination of temporal trends.

Methods With National Research Ethics Service approval, we constructed descriptive growth curves for babies born <32 weeks gestation using electronic data on longitudinal weight measurements entered from 40 neonatal units in England. After extensive cleaning, data were analysed for each week of gestation separately, using the SITAR method. This provides a summary cubic spline growth curve per gestation group, and a set of three random effect parameters per infant to transform the summary curve to match their own data. The birth data were also analysed to construct birthweight centiles using the LMS method.

Results The birthweight centiles, based on 3172 infants, were similar to but slightly lower than UK-WHO. The gestation-specific growth curves were based on 57 746 weights from 2577 infants, with 96–514 infants and 2446–8107 weights per week of gestation (with 22/23 weeks combined). The figure 1 shows the fitted gestation-specific growth curves plotted against postmenstrual age. Weight consistently fell after birth with a greater fall for later gestations. Weight velocity stabilised after 3 weeks. Growth curves then tracked parallel to and between the second and ninth UK-WHO birthweight centiles, at a higher level in more mature infants.

• Download figure
• Open in new tab
• Download powerpoint

Abstract P7 Figure 1

Pattern of postnatal weight gain by week of gestation.

Conclusion The pattern of postnatal growth in very preterm infants is complex, dependent on gestation, and quite different from that represented by the birthweight chart. Longitudinal monitoring based on birthweight charts is likely to lead to inappropriate treatment decisions. A new form of chart is urgently needed to provide a basis for longitudinal monitoring of preterm growth.