Clindamycin/primaquine as prophylaxis for Pneumocystis carinii pneumonia.

The records of 206 patients with advanced infection due to human immunodeficiency virus type 1 who were receiving prophylaxis with clindamycin/primaquine (C/P), trimethoprim-sulfamethoxazole (TMP-SMZ), or dapsone to prevent Pneumocystis carinii pneumonia (PCP) were retrospectively examined. Two hundred sixty-two patient-years of prophylaxis were accrued (176.2 of TMP-SMZ, 63.4 of dapsone, and 22.8 of C/P). The rates of PCP in the TMP-SMZ, dapsone, and C/P groups were 3.4, 11.0, and 30.7 per 100 patient-years, respectively. Pairwise comparisons showed C/P to be less effective than TMP-SMZ (relative risk [RR], 9.02; 95% confidence interval [CI], 3.03-26.83). A similar trend was apparent for C/P vs. dapsone (RR, 2.78; 95% CI, 0.98-7.93). When only those receiving primary prophylaxis were analyzed, C/P recipients remained at greater risk than TMP-SMZ recipients (RR, 13.19; 95% CI, 3.54-49.12) and dapsone recipients (RR, 3.85; 95% CI, 1.12-13.31). Failure of C/P prophylaxis could be due, at least in part, to underdosing (clindamycin, 300 mg/d; primaquine, 15 mg/d). C/P recipients had more nonspecific diarrhea than did TMP-SMZ recipients (RR, 2.99; 95% CI, 1.61-5.55).