Prediction of relapse after treatment of coccidioidomycosis.

Relapse after apparently successful treatment of coccidioidomycosis has been a problem with both amphotericin B and the azoles. We conducted a retrospective cohort study of 34 patients who required therapy for coccidioidomycosis between 1973 and 1993; 10 relapsed and 25 (one patient received two courses of therapy) did not relapse during follow-up. The mean time to relapse after completion of therapy was 7.3 months (range, 1-21 months). All 34 patients responded clinically to therapy. A fourfold or greater decrease in titers of antibody, as determined by complement fixation (CF), during therapy was seen in seven (78%) of nine patients who relapsed and 17 (85%) of 20 patients who did not relapse (P = .956). There was no significant difference between relapsers and nonrelapsers in terms of the lowest CF titer during therapy, the CF titer at the end of therapy, or the peak CF titer. The risk of relapse was increased among those with a peak CF titer of > or = 1:256 (relative risk [RR] = 4.7; 95% confidence interval [CI] = 1.4-16.1), as compared with patients who did not mount such a high antibody response. Similarly, the risk of relapse was higher among those with serially negative coccidioidin skin tests (CSTs) than those with serially positive CSTs (RR = 4.8; 95% CI = 1.2-19.5). We conclude that clinical response, lowest CF titer, end-of-therapy CF titer, and decrease in the CF titer of at least fourfold are not predictive of relapse in patients with coccidioidomycosis. Negative serial coccidioidin skin tests and a peak CF antibody titer of > or = 1:256 are independently associated with increased risk of relapse.