OBJECTIVE: To study the effects of GH treatment for up to 42 months on bone mineral density (BMD) and bone turnover. DESIGN AND METHODS: BMD with dual energy X-ray absorptiometry, serum type I procollagen carboxy-terminal propeptide (PICP), serum type I collagen carboxy-terminal telopeptide (ICTP) and serum IGF-I were assessed in 71 adults with GH deficiency. There were 44 men and 27 women, aged 20 to 59 (median 43) years. Thirty-two patients completed 36 months and 20 patients 42 months of treatment. RESULTS: The BMD increased for up to 30-36 months and plateaued thereafter. In the whole study group, the maximum increase of BMD was 5.0% in the lumbar spine (P<0. 001), 5.9% (P<0.01) in the femoral neck, 4.9% (NS, P>0.05) in the Ward's triangle and 8.2% (P<0.001) in the trochanter area. The serum concentrations of PICP (202.6+/-11.5 vs 116.3+/-5.4 microg/l; mean+/-s.e.m.) and ICTP (10.5+/-0.6 vs 4.4+/-0.3 microg/l) doubled (P<0.001) during the first 6 months of GH treatment but returned to baseline by the end of the study (130.0+/-10.4 and 5.6+/-0.7 microg/l respectively), despite constantly elevated serum IGF-I levels (39. 6+/-4.1 nmol/l at 42 months vs 11.9+/-0.9 nmol/l at baseline; P<0.001). The responses to GH treatment of serum IGF-I, PICP, ICTP (P<0.001 for all; ANOVA) and of the BMD in the lumbar spine (P<0.05), in the femoral neck and the trochanter (P<0.001 for both) were more marked in men than in women. At the end of the study the BMD had increased at the four measurement sites by 5.7-10.6% (P<0.01-0.001) in patients with at least osteopenia at baseline and by 0.1-5.3% (NS P<0.05) in those with normal bone status (P<0.001 for differences between groups; ANOVA). Among patients who completed 36-42 months of treatment, the number of those with at least osteopenia was reduced to more than a half. The response of BMD to GH treatment was more marked in young than in old patients at three measurement sites (P<0. 05-<0.001; ANOVA). In the multiple regression analysis the gender and the pretreatment bone mass appeared to be independent predictors of three measurement sites, whereas the age independently determined only the vertebral BMD. CONCLUSIONS: GH treatment in GH-deficient adults increased BMD for up to 30-36 months, with a plateau thereafter. Concurrently with the plateau in BMD the bone turnover rate normalized. From the skeletal point of view GH-deficient patients exhibiting osteopenia or osteoporosis should be considered as candidates for GH supplementation of at least 3-4 years.