Pulse frequency analysis of LH, PRL, testosterone, and cortisol was carried out with the Cluster Analysis Program in eight male cocaine abusers and eight aged-matched normal men. Four of the eight cocaine abusers had hyperprolactinemia (range, 22.08-44.65 micrograms/L). Cocaine users as a group had significantly higher mean peak height (P less than 0.02) than control subjects. Cocaine users with hyperprolactinemia had higher mean peak height than control subjects or cocaine users with normal PRL levels (P less than 0.01). Cocaine users with hyperprolactinemia also had higher mean amplitude increments than control subjects (P less than 0.02). Cocaine users with hyperprolactinemia had a higher mean valley than controls (P less than 0.01) and cocaine users with normal PRL levels (P less than 0.03). However, there were no significant differences in PRL peak frequency, peak duration, or interpulse intervals between cocaine users with or without hyperprolactinemia and control subjects. There were minimal differences between cocaine users and control subjects in pulse frequency analysis of LH parameters; the small differences in mean LH levels and average interpulse interval were not in the abnormal range and were probably not biologically significant. No differences between cocaine users and controls were detected for pulse frequency analysis of testosterone or cortisol. Cocaine-induced hyperprolactinemia may contribute to disorders of sexual and reproductive function in men who abuse the drug, and recent reports that PRL modulates immune function suggest that cocaine-induced derangements of PRL secretion may also contribute to cocaine-related comorbidity in infectious disease. Since cocaine users with hyperprolactinemia had a higher mean valley as well as a higher peak pulse PRL height than control subjects, but did not have greater PRL pulse frequencies, we conclude that hyperprolactinemia in these men may be due to a cocaine-induced derangement of dopaminergic inhibition of basal PRL secretion.