Growth hormone (GH), alone or in combination with anabolic steroids, seems to improve the growth rate in Turner syndrome, but to exert a less striking effect on the final height (FH). Reports on the FH usually lack a control group, and the GH effect is determined using the gain in centimeters over projected height. Out of a cohort of 32 Turner syndrome girls under recombinant human GH (rhGH) therapy (0.5 IU/kg/week during the 1st year and 1 IU/kg/week subsequently), 18 (treated for 3-6 years) attained FH. The mean chronological age at the first examination was 9.6 +/- (SD) 2.1 years and at the start of GH therapy 13.0 +/- 2.0 (range 8.8-17.2) years. Eighteen untreated subjects matched for chronological age and karyotype served as control group. The FH as SDS according to Lyon and to unpublished Italian Turner syndrome girl standards was not significantly different as compared with pretreatment. In comparison with Italian cross-sectional Turner syndrome standards (FH 142.5 +/- 7.0 cm), the FH of the control group was quite similar (142.2 +/- 4.9 cm), whereas the rhGH-treated group showed a FH of 147.6 +/- 7.3 cm with a mean increment of about 5 cm. The height gain during therapy (as delta height in SDS either according to Lyon or to Italian SDS standards) was compared for each girl with that of a matched girl of the control group during a comparable observation period. A significantly different delta height was observed in the treated versus control groups: 0.3 +/- 1.1 vs. -1.0 +/- 0.8 according to Lyon (p < 0.001) and 0.8 +/- 0.7 vs -0.3 +/- 0.5 according to Italian standards (p < 0.001). If we compared the FH with the projected height according to Lyon standards, the height gain (as delta height in cm) was significantly higher than in the untreated subjects (-1.1 +/- 4.8 vs. -6.2 +/- 3.9 cm; p < 0.05). It seems worthwhile to undertake GH treatment in Turner syndrome girls who represent a very short stature population, even though the response is less significant than in classic GH deficiency and shows a striking variability, probably due to a sort of peripheral resistance.