Short-term weight variability in infants and toddlers: an observational study.

Aims: To explore short-term weight variability in young children; (1) how it relates to expected weight gain and (2) how it is affected by age, time of day and dietary intakes and outputs.

Methods: Twenty healthy infants aged 2-10 months and 21 healthy toddlers aged 12-35 months were weighed at home by their parents six times over 3 days. The toddlers' parents also recorded whether they had eaten, drunk, urinated or passed stool in the previous 2 hours. The primary outcome was 'noise': the within-subject weight SD pooled separately for infants and toddlers, compared with their expected weight gain over 4 or 8 weeks. Analysis by successive pairs of weights was used to assess the extent of short-term weight gain and loss associated with time of day and eating, drinking and excretion.

Results: In infants, noise (117 g) was much less than the expected weight gain over 4 weeks (280-1040 g) but in toddlers, noise (313 g) was higher than the expected gain over 4 weeks (180-230 g) and around three-quarters the expected gain over 8 weeks (359-476 g). In toddlers, weight tended to fall overnight and rise by day, and recent eating and passage of stool were associated with increased weight gain, even after adjustment for time of day.

Conclusions: In toddlers, the recorded weight may be 300 g higher or lower than the underlying weight trajectory, so that their weight gain based on measurements collected fewer than 8 weeks apart will often be misleading.