Fluconazole Underexposure in Critically Ill Patients: a Matter of Using the Right Targets?

Abstract

We read with great interest the paper entitled “Optimization of Fluconazole Dosing for the Prevention and Treatment of Invasive Candidiasis Based on the Pharmacokinetics of Fluconazole in Critically Ill Patients” recently published by J. M. Boonstra and colleagues (1). We applaud the authors for investigating fluconazole exposure in critically ill patients. Despite the fact that fluconazole is a frequently used triazole in the setting of candidemia and invasive candidiasis, its pharmacokinetics in critically ill patients are still not well documented (2–4) and the ideal dosing regimen is a matter of debate (5). The authors documented a large variability in exposure and developed a population pharmacokinetics model based on 561 concentrations measured in 33 patients in an attempt to support personalized fluconazole dosing. They found that almost half of their studied population was underexposed, with organ transplant recipients and patients on renal replacement therapy (RRT) having the highest risk, when taking into account a target area under the concentration-time curve from 2 to 24 h (AUC0–24) of 400 mg · h/liter. Based on a simulation analysis, higher fixed (a 1,200-mg loading dose [LD] followed by 1,000 mg daily) or weight-based (a 13-mg LD/kg of body weight followed by 8 mg/kg daily) doses would be necessary to achieve the above-mentioned target in >95% of patients. However, they also mention that overexposure, potentially leading to hepatotoxicity, might be encountered in an important number of patients when using these high doses. Therefore, they recommend guiding fluconazole dosing based on therapeutic drug monitoring (TDM), targeting trough concentrations (the minimum concentration of a drug in serum [Cmin]) of 15 mg/liter, which was confirmed, also in earlier reports, to correlate very strongly with an AUC0–24 of 400 mg · h/liter (1, 6). Although the study is nicely designed, we question the interpretation of the exposure results and final conclusion. As mentioned in their Materials and Methods, the authors aimed to reach an AUC for the free, unbound fraction of a the drug (fAUC)/MIC of 100, taking into account protein binding of 12%, which is indeed the internationally recommended pharmacokinetic/pharmacodynamic target (7, 8). The target AUC0–24 is typically calculated by taking into account the non-species-related breakpoint for susceptibility as a denominator, which is defined as 2 mg/liter by EUCAST and corresponds to the epidemiological cutoff of Candida albicans and Candida tropicalis (7). This, hence, yields a target AUC0–24 of 200 mg · h/liter. If this target were used, more patients would have reached the target, and potentially more well-defined subpopulations would have been derived as being underexposed. Besides, the simulation analysis would have yielded lower fixed or weight-based dosing, leading to target attainment, presumably more in line with the standard 800-mg LD followed by 400 mg daily, or 12-mg/kg LD followed by 6 mg/kg daily. Finally, when this target is used, the recommendation for TDM might be questioned, as a Cmin of 7.5 mg/liter would be reached in the majority of patients. Based on this target, we believe that future efforts should focus on dose increases for well-defined subpopulations at risk for underdosing, e.g., for patients with RRT, augmented renal clearance, or obesity. This would avoid the need for TDM, which is often difficult to implement, as assays with short turn-around times are not available.