Prophylaxis, empirical and preemptive treatment of invasive candidiasis.

OBJECTIVE

Invasive candidiasis remains an important infection for ICU patients, associated with poor clinical outcomes. It has been increasingly recognized that the traditional paradigm of culture-directed antifungal treatment is unsatisfactory, and that earlier antifungal intervention strategies, such as prophylaxis, preemptive therapy, and empiric therapy, are required to improve patient outcomes. The purpose of this review is to summarize the recent supportive evidence for such strategies and to highlight the current challenges in their implementation.

RESULTS

Despite new antifungal agents and classes, the mortality from invasive candidiasis remains high. Antifungal prophylaxis remains the best-studied early antifungal intervention strategy; however, unless targeted to patients at highest risk, is inefficient. Recent data suggests that although risk predictive models, using a combination of clinical risk factors and Candida colonization parameters, may be a relatively simple and practical approach to guide prophylaxis or preemptive therapy, further validation of these models is required. A single trial has demonstrated that empiric antifungal therapy is not of benefit when instituted to patients with antibiotic-refractory fever alone.

CONCLUSIONS

On the basis of current knowledge, it is difficult to universally recommend antifungal prophylaxis, apart from patient groups with a known very high risk, such as those with necrotising pancreatitis or recurrent gastrointestinal perforations. Antifungal prophylaxis may also be reasonable where local incidence rates and epidemiology are compelling. Among stable patients with multifocal Candida colonization and/or a multitude of clinical-risk factors, preemptive therapy is currently not indicated, although the development of better risk predictive models may assist with such patients. Among patients with refractory fever despite broad-spectrum antibacterial therapy, empiric antifungal therapy may be reasonable where local incidence rates are high (e.g. >10%); however, a thorough search for alternate causes must be instituted.