The measurement of serum FSH is useful in the diagnostic workup of the infertile male, but fails to predict the presence of sperm in testicular tissue. We investigated whether inhibin B reflects testicular morphology and the presence of sperm more accurately than FSH. Serum inhibin B and gonadotropin levels were determined in 91 infertile men undergoing diagnostic bilateral testicular biopsy. In 52 of the 91 patients multiple samples were taken for testicular sperm extraction (TESE). Inhibin B levels were (mean +/- SEM) 238+/-32 pg/mL in men with normal spermatogenesis (n = 9), 102+/-18 pg/mL in men with spermatogenetic arrest (n = 15), 98+/-16 pg/mL in hypospermatogenesis (n = 23), 41+/-6 pg/mL in focal Sertoli cell-only syndrome (SCO; n = 26), and 27+/-8 pg/mL in complete SCO (n = 18). The percentage of SCO tubuli was more strongly correlated to serum inhibin B (r = -0.58; P<0.01) than to FSH (r = 0.34; P<0.05). Similarly, the percentage of tubules with elongated spermatids was significantly (P<0.05) more strongly correlated to serum inhibin B (r = 0.65; P<0.01) than to FSH (r = -0.4; P<0.01). Thus, inhibin B is slightly more sensitive than FSH as an index of the spermatogenic status. Neither FSH nor inhibin B alone, however, could predict the type of spermatogenetic damage exactly. The combination of FSH and inhibin B had high diagnostic sensitivity (88%) and specificity (83%) for the presence of elongated spermatids in testicular biopsies. Sperm could be retrieved in 34 (65%) of the TESE patients. The combination of inhibin B and FSH measurement showed a sensitivity of 75% and a specificity of 73% when identifying patients in whom sperm could possibly be retrieved by TESE. We conclude that although the measurement of serum inhibin B improves the sensitivity of predictive tests for the presence of sperm in histology or for TESE, this parameter cannot accurately predict TESE outcome.