Establishing the proportion of severe/moderately severe vs mild cases of progressive disseminated histoplasmosis in patients with HIV

BACKGROUND: Progressive disseminated histoplasmosis remains a major but neglected cause of death among patients with advanced HIV. Recently, aiming to reduce avoidable deaths, the Pan American Health Organization issued the first diagnosis and treatment guidelines for HIV-associated histoplasmosis. But what proportion of progressive disseminated histoplasmosis in HIV-infected patients is severe is currently not known. Because this proportion influences treatment needs, we aimed to estimate this in a cohort of 416 patients in French Guiana. METHODS: We used the definition in the recent PAHO/WHO guidelines for severity. We used regression modelling to predict the impact of CD4 count on the proportion of severe cases. In a territory where treatment cost is not a limiting factor and where histoplasmosis is well known, we assumed that clinicians’ initial treatment reflected their perception about the severity of the case and therefore, the needs for different treatments. RESULTS: Using these definitions, since the beginning, there were 274 (65.9%) severe/moderately severe cases and 142 (34.1%) mild cases. In practice 186 cases were treated with deoxycholate or liposomal amphotericin B (44.7%) and 230 (55.3%) cases treated with itraconazole as first line treatment. The Kappa concordance measure between the guideline definition and the actual treatment given was 0.22. There was a 9% risk difference for death within 30 days of antifungal treatment initiation between severe/moderately severe and mild cases. Over threequarters (77%) of early deaths were attributed to severe/moderately severe cases. CONCLUSIONS: This is the only rigorous estimate of the proportion of severe/moderately severe cases of progressive disseminated histoplasmosis in symptomatic HIV patients on the largest published cohort. These numbers may help defend budget needs for rapid diagnostic tests and liposomal amphotericin B.