In-hospital mortality of HIV-associated cryptococcal disease in patients treated with amphotericin B versus fluconazole
Poswa, Xoliswa Pennley
Introduction Cryptococcal disease (CD) is the most common cause of morbidity and mortality among patients living with HIV I AIDS in many parts of sub-Saharan Africa. Globally the highest number of HIV -associated CD cases occur in sub-Saharan Africa (720 000/957 900) and mortality rates among patients on antifungal treatment remain unacceptably high. This study aimed to estimate and compare in-hospital mortality of HIV-associated CD among patients who were treated with amphotericin B versus fluconazole versus mixed treatment with amphotericin B and fluconazole. Materials and Methods We performed an analytical, cross-sectional analysis of data from a national laboratory-based surveillance programme through a network of public sector laboratories in South Africa. The study period was 1 January 2005 to 31 December 2006. The analysis used a subset of data from laboratory confirmed cases with completed case report forms and available data on outcome. The exposure was measured in 3 levels defined as treatment during the induction phase of therapy for at least 7 days with either amphotericin B or fluconazole or mixed treatment (initiation of treatment with one regimen and switching on to the other within 7 days of treatment). Outcome was defined as: patients who died between 7 and 30 days in hospital. Chi-squared test was used to compare characteristics among the treatment groups and multiple logistic regression models were constructed to identify risk factors for in-hospital death. Results Sixty two percent of the patients (1,363/2,211) were treated for ?:.7 days and 38% (848/2,211) for <7 days. In the group treated for ?:.7 days, mortality was 359/1,363 (26%) and the median time to death was 12 days (IQR 9-18). There was no significant difference in case fatality among patients treated with: amphotericin B (29%), fluconazole (27%) or mixed treatment (24%), (p-value = 0.28). On multivariate analysis, factors significantly associated with in-hospital mortality were: patients between the age group of 40-59 years, province in which the patients resided and altered mental status. In the group treated for <7 days patients treated with fluconazole were 82% less likely to die than patients treated with amphotericin B. Discusslon and Conclusion In-hospital mortality was high and similar among all treatment regimens in the ?:.7 days group. However a significant reduction in mortality was noted in the <7 days group treated with fluconazole. The reason for the later findings is unclear. It may be that amphotericin B alone is not superior but equivalent to fluconazole in the first 7 days of treatment and 5-flucytocine may be the intervention required to improve outcome in the first 7 days. Or it may be that patients are reaching care too late for a significant impact in disease outcome to be observed and prevention of CD is required in the form of ART and primary prophylaxis. Selection bias in that patients in the fluconazole group are less sick than in the amphotericin B-group is the most likely reason for lower mortality. The study findings call for more standardization of optimum treatment as well as advocacy for the availability of 5-flucytocine. Factors associated with high mortality require further investigations for interventions to improve patient outcomes.
Thesis (M.Sc. (Med.) (Epidemiology and Biostatistics))--University of the Witwatersrand, Faculty of Health Sciences, 2012.