Objective: To determine whether antifungal agents given prophylactically or empirically decrease morbidity and mortality in patients with cancer complicated by neutropenia. Design: Meta-analysis of randomised trials of amphotericin B, various lipid soluble formulations of amphotericin B (for example, AmBisome), fluconazole, ketoconazole, miconazole, or itraconazole compared with placebo or no treatment. Setting: Trials conducted anywhere in the world. Subjects: Patients with cancer complicated by neutropenia. Main outcome measures: Mortality, invasive fungal infection (defined as positive blood culture, oesophageal candidiasis, or lung or deep tissue infection), and colonisation. Results: 24 trials with 2158 randomised patients were reviewed; the total number of deaths was 434. Prophylactic or empirical treatment with antifungals as a group had no effect on mortality (odds ratio 0.92; 95% confidence interval 0.74 to 1.14). Amphotericin B decreased mortality significantly (0.58; 0.37 to 0.93) but the studies were small and the difference in number of deaths was only 15. Antifungal treatment decreased the incidence of invasive fungal infection (0.47; 0.35 to 0.64) and fungal colonisation (0.45; 0.30 to 0.69). For every 73 patients treated (95% confidence interval 48 to 158) one case of fungal invasion was prevented in surviving patients. Conclusions: There seems to be no survival benefit of antifungal agents given prophylactically or empirically to patients with cancer complicated by neutropenia. These agents should be restricted to patients with proved infection and those in randomised trials. A large, definitive placebo controlled trial of amphotericin B is needed.
