Nephrotoxicity of continuous amphotericin B in critically ill patients with abdominal sepsis: a retrospective analysis with propensity score matching—authors’ response

We thank the respondents for their interest in our study on the nephrotoxicity of continuous amphotericin B in critically ill patients with abdominal sepsis.1 We agree with the respondents that continuous conventional amphotericin B (CCAB) is not described in the guidelines for patients with abdominal infection and significant risk factors for candidiasis.2 The respondents state that the presence of alternatives in these guidelines makes the concept of CCAB unsound. We strongly disagree with this statement. The reason that conventional amphotericin B (c-AmB) is not mentioned in guidelines is because only literature about intermittent amphotericin B administration was considered, showing nephrotoxicity.3 c-AmB as first-line treatment for invasive fungal infections was substituted for safety reasons, not for lack of effectiveness. Demonstrating that not all dosage regimens of amphotericin B cause nephrotoxicity is valuable additional information that has not previously been included in drug policy considerations. Amphotericin B has a broader antifungal spectrum compared with fluconazole, and azoles carry higher risks of resistance.4,5 We agree with the respondents that fluconazole can play a role in treating invasive fungal disease (IFD), however, in clinical practice in Dutch Hospitals, fluconazole is only used if the yeast or fungus is known to be susceptible to fluconazole. In addition, the use of fluconazole has several disadvantages in ICU settings, because of multiple drug–drug interactions, adverse effects in critically ill patients and its relatively large infusion volume. The volume of infusion of fluconazole is 200–400 mL per day, which, for some patients, can contribute to volume overload in combination with other infusions. Compared with c-AmB, the volume of infusion of fluconazole is around four times higher. Based on our findings, CCAB can be considered as a safe, first-line antifungal in critically ill patients with suspected or documented invasive candidiasis, pending confirmation or exclusion of the diagnosis and identification of the organism.6 Further, the respondents state that we ignored the pharmacodynamic data in relation to efficacy. Although our study did not focus on efficacy, we thoroughly addressed the pharmacokinetic and pharmacodynamic properties of amphotericin B in our report. Some studies conclude that amphotericin B has both time- and concentration-dependent activity,7 while several animal models indicate that optimal activity of amphotericin B may be achieved by optimizing the Cmax/MIC.8 Clinical trials confirm the idea that CCAB is not only safe but also effective in several patient groups.3,9–11 Further studies to explore the clinical efficacy of CCAB in ICU patients with abdominal sepsis in comparison with other antifungal agents are needed. In our study, the focus of therapeutic drug monitoring (TDM) was to prevent amphotericin B accumulation and possible nephrotoxicity.6 The upper limit of 1.0 mg/L was pragmatically chosen, as concentrations >0.5 mg/L result in fungicidal concentrations in the peritoneum.12 More information on the usefulness of TDM for amphotericin B is required, also in relation to the efficacy of CCAB. The respondents state that c-AmB has shown infusion-related adverse events, anaemia and electrolyte disorders. These potential side effects were not seen in previous studies on CCAB. The infusion reactions seen with c-AmB were caused by the high rate of administration of c-AmB, and are not expected with CCAB because of the slow continuous administration of amphotericin B. In the study cited by the respondents, the electrolyte disturbances were described as clinical manifestations of renal dysfunction.13 Possible electrolyte disturbances do not cause clinically relevant problems in an ICU setting, due to the continuous monitoring and the presence of rescue medication. Overall, we found the added value too limited to study possible side effects other than renal dysfunction in CCAB. We agree with the respondents that the retrospective design may have led to potential biases. However, in the comment it is stated that selection bias is increased by the applied propensity matching, which is not the case. The selection bias was actually limited by performing propensity score matching by age, baseline serum creatinine and APACHE III score as possible predictors. We chose to exclude patients taking amphotericin B for <3 days because the nephrotoxicity previously seen with c-AmB is dependent on the cumulative dose. Therefore, nephrotoxicity was not expected in the 3 days of treatment. Also, the long elimination half-life of amphotericin B of 15 days supports our hypothesis that a possible nephrotoxic effect will show up after a longer period of treatment. Finally, we agree with the respondents that the cost-effectiveness is not fully studied. The costs for a CCAB patient are €14.74 per day, which is low compared with other antifungals.6 Costs related to efficacy are not known because there are no studies that compare CCAB with other antifungal drugs. There is no need to compare CCAB with fluconazole because patients who are not resistant to fluconazole switch to this drug. In addition, if follow-up studies show that TDM is not of benefit (or is only beneficial in a limited group of CCAB patients) the costs of the blood concentration determination decrease. Overall, CCAB may have economic benefit, even if it is not clear what the exact amount would be. We conclude that CCAB with TDM as pre-emptive antifungal therapy in ICU patients with abdominal sepsis did not show renal impairment and can be safely used for this group of patients in clinical practice. This study was supported by internal funding. None to declare.