The influence of secular trend for height on ascertainment and eligibility for growth hormone treatment.

OBJECTIVE Assessment of short stature in many instances is based on a comparison with the Centers for Disease Control's (CDC) growth curves. The secular trend for height may limit the utility of CDC data for contemporary populations. We investigate the effect of the secular trend on Australian and US populations. DESIGN Describe CDC-defined height SDS distributions of contemporary populations for different ages and genders. Compare observed means and standard deviations (SDs) to expected values of 0 and 1. Compare frequency of individuals shorter than the CDC-1st centile to those shorter than 1st centile defined empirically from the contemporary population. SUBJECTS Healthy Kids Queensland Survey 2006: 1686 boys, 1822 girls. Australian National Children's Nutrition and Physical Activity Survey 2007: 2415 boys, 2379 girls. US National Health and Nutrition Examination Survey 2005-2006: 2160 boys, 2118 girls. MEASUREMENTS Means, SDs and normality of CDC-defined height SDS distributions. Frequency of individuals shorter than the CDC-1st centile and shorter than an empirically defined 1st centile. RESULTS In Australia, means of CDC-defined height SDS distributions are always greater than 0 and the CDC-1st centile identifies only the shortest 0·5% of children. Means may vary with age and occasionally between genders in contemporary populations. Normality and SDs of 1 are retained. CONCLUSIONS The secular trend has resulted in an underestimate of the number of Australian children eligible for GH treatment using the CDC-1st centile cut-off. Contemporary, local data should be used to construct standards. Using the 2nd CDC centile would approximate the 1st local centile until new standards are constructed. The secular trend does not account for the gender bias in GH therapy.