Effects of a 12-week progressive strength training period on serum concentrations of testosterone, cortisol and sex-hormone-binding globulin (SHBG) as well as on strength development of the leg extensor muscles were investigated in nine middle-aged males (M50; range 44-57 years) and in nine middle-aged females (F50; range 43-54 years) as well as in 10 elderly males (M70; range 64-73 years) and in 11 elderly females (F70; range 66-73 years). Substantial increases took place in maximal isometric strength during the 12-week training period both in M50 (from 2834 +/- 452 to 3941 +/- 772 N; P < 0.001) and in F50 (from 2627 +/- 725 to 3488 +/- 1017 N; P < 0.001) as well as in M70 (from 2591 +/- 736 to 3075 +/- 845 N; P < 0.01) and in F70 (from 1816 +/- 427 to 2483 +/- 408 N; P < 0.001). The relative increases in strength during the 12-week training period did not differ significantly between the groups. However, during the last 4 weeks of the training none of the groups demonstrated further increases in strength but it actually decreased in F50 (P < 0.05), M70 (P < 0.01) and in F70 (P < 0.05). No systematic changes were observed during the training in the mean concentrations of serum total testosterone, free testosterone, cortisol, and SHBG, nor in testosterone/cortisol and testosterone/SHBG ratios. However, the individual levels of serum testosterone and testosterone/cortisol ratio and the individual changes in strength during the last four most intensive training weeks of the 12-week period were in significant positive linear correlation in F70 (r = 0.57; P < 0.05) and in M70 (r = 0.61; P < 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)