[Growth curves as a screening instrument--possibilities and limits]

The application of growth charts as screening tools for children is a commonplace. Growth charts published during the past 15 years in Germany are now under debate with respect to their validity. An analysis of the problems related to growth curves gave the following results: A clear, sharp boundary between normal and non-normal permitting reliable grouping of children does not exist. Due to the heterogeneity of the population of children coming for screening, the call for an allround optimal and representative growth-curve remains wishful thinking. The growth charts so far published, show marked differences in the course of the percentile lines. The discrepancies can be explained by different sample sizes and algorithms of calculation. In the example of the longitudinal Zurich growth study on normal children it can be demonstrated, that the change between percentile ranges during the initial 4 years of life often considered to be pathological occurs in almost 50% of the probands. By conservative interpretation all growth charts can be used, because only with rare exceptions the diagnosis of growth disorders will be solely based on the measurement of tallness and its comparison to percentile values. As a rule: The grouping of a child as having non-normal growth is the result of multiparametric medical considerations.