ObjectiveWe wanted to examine the relationship between initial growth response to recombinant human Growth Hormone (rhGH) treatment and body composition in children with growth hormone deficiency (GHD). Design and methodsForty-two patients (21 boys and 21 girls) aged between 5.7–15.5years (mean age: 10.8±2.6years) with isolated GHD. The auxological and laboratory data (GH and IGF-I levels) and results of bioelectrical impedance analyses were evaluated. Children with GHD were followed up for 12months and categorized according to growth response to rhGH into good and poor responders (change in height of >0.7 SDS or <0.7 SDS over one year respectively). Mean doses of rhGH per kg of fat free mass (FFM) were calculated. ResultsForty-eight percent of patients showed a good growth response to rhGH therapy. At study entry, mean age, height SDS, weight SDS, serum IGF-1 SDS, IGFBP-3 SDS, growth velocity prior to rhGH therapy, GH after clonidine and l-dopa were similar in the two groups. At baseline, BMI SDS and waist–hip ratio were significantly higher in good responders (p=0.02 and p=0.006, respectively). Good responders had lower percentages of FFM (73.4±8.9 vs. 83.1±5.9) and total body water (TBW) (56.5±5.3 vs. 63.1±4.4), compared to poor responders (p<0.05). There were significant correlations between changes in height SDS over one year and baseline body composition in children with GHD on rhGH treatment (r=−0.617 for percentage of FFM, r=−0.629 for percentage of TBW, p<0.001). A correlation between BMI SDS, waist–hip ratio, mean rhGH dose per FFM and growth response was observed only in prepubertal subjects. ConclusionBaseline body composition data in children with GHD can be used to predict the growth response to rhGH treatment. A management strategy that involves titrating rhGH dose according to FFM as a means of optimizing the growth response to intervention requires further study.