Bioactive FSH and immunoreactive FSH were determined in 193 infertile men and in 23 men with proven fertility using the Sertoli cell aromatase bioassay for bioactive FSH measurement and a two-site fluoroimmunoassay for immunoreactive FSH measurement. Overall bioactive and immunoreactive FSH levels correlated well (r = 0.74, p < 0.001) but were significantly different from fertile men (bioactive FSH: 6.2 \ m=+-\ 0.3 U/l; immunoreactive FSH: 4.1 \ m=+-\ 0.4 U/l) in patients with Klinefelter's syndrome (24.1 \ m=+-\ 6.1; 26.9 \ m=+-\ 3.0), non-obstructive azoospermia (25.1 \ m=+-\ 4.3; 22.2 \ m=+-\ 4.0), maldescended testes (12.5 \ m=+-\ 4.6; 14.6 \ m=+-\1.6), and patients with severe oligozoospermia (11.9 \ m=+-\1.2; 11.2 \ m=+-\1.0). Infertile men with moderate oligozoospermia (8.9 \m=+-\1.5; 8.0 \m=+-\1.1) and normal sperm counts (9.6 \ m=+-\1.1; 7.6 \ m=+-\ 1.0) had insignificantly elevated bioactive FSH and immunoreactive FSH levels. Bioactive to immunoreactive FSH ratios were significantly reduced in all patient groups except for patients with normal sperm counts when compared with fertile men. A considerable number of patients exhibited elevated immunoreactive FSH concomitant with normal bioactive FSH levels. We conclude that 1. determination of immunoreactive FSH suffices for classification of patients; 2. bioactive to immunoreactive FSH ratios are reduced in infertile men; 3. some men might secrete immunoreactive FSH with reduced bioactivity. Elevated serum FSH levels in infertile men are re¬ garded as an indication of primary testicular failure resulting in insufficient negative testicular feedback. Such failure is due to irreversible de¬ struction of the germinal epithelium (1). The cor¬ rect determination of serum FSH is, therefore, of prime importance for differential diagnosis and assignment to treatment protocols (1,2). However, determination of FSH by RIA (immunoFSH) does not always reflect the biologically active hormone (bioFSH) (3, 4). Differences between bioFSH and immunoFSH are found in various endocrine situ¬ ations. These are a consequence of variations in the glycosylation of FSH isoforms or are due either to the secretion of an immunoreactive but biologi¬ cally inactive abnormal form of the hormone or to the hypersécrétion of isolated subunits with no bioactivity (5, 6). Therefore the determination of bioFSH in infertile men would be of greater im¬ portance than immunoFSH. This, however, has not been possible because suitable assay proce¬ dures for the determination of bioactive FSH in serum were not available.Recently two in vitro bioassays based upon the FSH-dependent aromatase activity of either rat granulosa cells (7) or immature rat Sertoli cells (8) were validated for human serum FSH. The two as¬ says yielded different results when applied to measure FSH in serum from women during nor¬ mal menstrual cycles (8). Using the granulosa cell aromatase bioassay, Wang et al. (9) reported a de¬ crease in the ratios of bioFSH to immunoFSH (B/I) in infertile men.In the present study we have evalu...