The Euro-Growth Study: Why, who, and how

Background: Assessment of growth is important in health management of infan ts and children. Evaluation of growth performance requires anthropometric m easurements, with proper interpretation depending on the use of appropriate references. Europe-specific growth references have not been available. The purpose of the present study was to generate such references for infants a nd children from birth to 3 years of age. The study was further intended to assess the influence of nutrition and lifestyle factors on growth. Methods: The Euro-Growth Study was designed as a multicenter longitudinal c ohort study. This report describes the study design, the sample (cohort), a nd the methods used. Quality control measures included standardized measure ment techniques with ongoing cross-sectional and longitudinal consistency c hecks. Selectivity in participation and discontinuation and reproducibility of data over time were evaluated. Results: Of 2245 infants who were enrolled in the study at 22 study sites i n 11 countries, 1746 (78%) provided longitudinal data until 12 months of ag e, 1205 (57%) until 24 months of age, and 1071 (48%) until 36 months of age . Anthropometric measurements were performed on 21,773 occasions. As a resu lt of cross-sectional and longitudinal data checks, 209 data points (0.09% of all data points) were rejected. Comparison of participants's demographic data with those of nonparticipating local subjects indicated that the coho rt was selective in maternal age (higher), maternal education level (higher ), household location, and family structure. The withdrawal rate during the study was low (24%), and withdrawals occurred at random. Analysis of repro ducibility over time resulted in the elimination of the data from one study site. Conclusion: The Euro-Growth Study provided longitudinal growth data from a large cohort of normal European children. Because of the chosen sampling me thod, the study cohort ws somewhat selective relative to the local backgrou nd population. Internal validity was satisfactory in that the reproducibili ty of anthropometric measurements was high in 21 sites and the withdrawal r ate was low and random. (C) 2000 Lippincott Williams & Wilkins, Inc.