OBJECTIVE An inverse relationship has been shown between body mass index (BMI) and the peak growth hormone (GH) response to stimulation in adults and in children with short stature. This relation is observed even within a normal range of BMI. The aim of this study was to investigate the effect of BMI on the GH response to clonidine in a large number of children with short stature. METHODS We conducted a retrospective study on the GH response to clonidine in a single centre. METHODS We studied 202 children with short stature (135 M and 67 F) who underwent clonidine testing from 2007 to 2009. RESULTS One hundred and twenty-eight patients had a GH peak >10 μg/l. In univariate regression analysis, the peak GH after clonidine was negatively correlated with BMI-standard deviation score (BMI-SDS) and positively correlated with height velocity-SDS and IGF-I-SDS. Only the relationship between peak GH and BMI-SDS remained significant in children with a BMI-SDS from -2 to +2. In the multivariate stepwise regression analysis, BMI-SDS and IGF-I-SDS were the only significant variables in the entire cohort, explaining 19·5% of the variance in peak GH. When only subjects with BMI-SDS between -2·0 and +2·0 were included in the analysis (n = 173), BMI-SDS alone explained 21·4% of the variability in peak GH. The number of patients who failed the clonidine test increased with increasing BMI-SDS. CONCLUSIONS BMI affects the GH response to clonidine in children with short stature and should be considered when interpreting the results to the stimulation test.