Infection with Cryptococcus neoformans is an increasing problem in immunocompromised patients, particularly those with acquired immune deficiency syndrome (AIDS). Amphotericin B and fluconazole are currently acceptable therapies for cryptococcal meningitis; however, their effects remain suboptimal and recurrence or treatment failure is still a problem. Antifungal susceptibility testing may be an important tool for guiding therapy, but for C. neoformans, a reliable method is still not available. This retrospective study evaluated minimal inhibitory concentration (MIC) for amphotericin B and fluconazole, and minimal fungicidal concentration (MFC) and timed-kill curves for amphotericin B against 16 clinical isolates of C. neoformans obtained from AIDS patients with cryptococcal meningitis. No correlation between clinical outcome and MIC was observed for amphotericin B. In selected cases, the MFC seemed to be a better predictor of outcome than MIC. In this study, amphotericin B timed-kill curves appeared to show a correlation with clinical outcome of the 16 patients with AIDS-associated cryptococcal meningitis. These in vitro tests must be further evaluated in prospective studies to confirm their potential usefulness for guiding cryptococcal meningitis therapy.