Is Itraconazole Superior to Voriconazole for Treatment of Histoplasmosis?

摘要

Methods: This is a retrospective cohort study of all adult cases of proven or probable histoplasmosis treated with either itraconazole or voriconazole at a single center between 2002 and 2017. Participants were allowed to have had an initial course of amphotericin but those receiving initial treatment with other azoles were excluded. The investigators examined many potential confounders and evaluated the association between treatment choice and cumulative incidence of death by 180 days. To compare mortality outcomes, they used Cox regression to control for potential confounders in the cumulative incidence analysis up to 180 days with an additional component called a “Heaviside function” that accounts for violation of the proportional hazards assumption and divided the analysis into 2 periods (early vs late). Results: Of 512 patients evaluated, 194 were evaluable; 175 (90%) received itraconazole and 19 (10%) received voriconazole. These were adult patients, most were immunocompromised (60%), and around half had disseminated histoplasmosis. The choice of voriconazole was not significantly associated with the presence of disseminated disease or other putative confounders. In the final multivariable analysis, which included disseminated disease as a (significant) variable (all other variables tested were not associated with outcome), patients who received voriconazole had a significantly higher incidence of death during the first 42 days (adjusted hazard ratio [HR] = 4.3; 95% confidence interval [CI] 1.3-13; P = .015). Mortality was not different during the 43-180 day period (HR = 0.0; 95% CI 0.0->99), but there were very few events. Comments: This study suffers from the usual problem of any observational study addressing causality—“Why did these patients get voriconazole when standard therapy at their institution is clearly itraconazole?” Was there something special about them that might also increase their risk of death or treatment failure? The authors did try to adjust for suspected confounders, including immunocompromise, age, sex, ICU status, and others (all were not significant), but no available approach can adjust for unrecognized confounders. There is biological plausibility for the difference. Fluconazole appears to be inferior to itraconazole for treatment of disseminated histoplasmosis in patients with HIV, perhaps because of induction of fluconazole resistance mutations (eg, 14a-demethylase mutants); these mutations also confer voriconazole resistance, but it is not known whether voriconazole induces them. However, this study did not specifically look at cause of death and did not evaluate control of infection (eg, serial Histoplasma blood cultures or serum antigens), so it is not clear if this is the mechanism for the outcome discrepancy. The investigators also do not report therapeutic drug monitoring results, so difficulty with attaining fungicidal levels might also contribute to the difference if it is truly related to treatment failure. The absence of children in this study raises the question of whether the findings are applicable in pediatrics—poor absorption and tolerability of itraconazole might be worse in children and might decrease any efficacy advantage (but probably does not, especially with therapeutic drug monitoring). Conclusions: These data, in combination with biological plausibility and an absence of high-quality supportive data, suggest that we should avoid voriconazole for the treatment of histoplasmosis in adults and probably also in children. Based on this study and extensive clinical experience, the drug of choice remains itraconazole.