Should In Vitro and In Vivo Studies on Antimicrobial Agents during Continuous Renal Replacement Therapy Comply with General Principles of Pharmacokinetics?

LETTER

We carefully read the article on the clearance of isavuconazole during continuous renal replacement therapy (CRRT) (1). However, we are concerned by the methods, the results, and the conclusions.

The comparison of six sessions using AN69 filters to one session using a polyethersulfone filter is questionable owing to major differences regarding adsorption of drugs (2).

In vitro studies are usually designed to assess the clearance provided by the different routes of elimination, which include not only dialysis and filtration but also adsorption (3–6). The authors said that, in in vitro study, values in the effluent samples were below the lower limit of quantification (LLOQ) that precluded any calculation. However, owing to the method of calculation of the extraction coefficient (C_E), it seems unlikely the values of the postfilter concentrations being under the LLOQ, even in the range of tested initial concentrations. Therefore, the authors should have been able to report values of the C_E. There is a recent trend to perform in vitro studies over a period of time as short as 1 h (7). For drugs exhibiting an elimination half-life of about 100 h, such a duration is questionable. Extending results from 1 to 24 h requires the assumption that the C_E is constant. In contrast, in the present study, in patients receiving continuous venovenous hemofiltration (CVVH), there was a ten-fold decrease of the C_E over 24 h, thus evidencing time-dependent kinetics. What about patients receiving continuous venovenous hemodiafiltration (CVVHDF)? The authors acknowledged the limitation related to the incomplete elimination phase sampling data. Conclusions based on the time dependency of the clearance of isavuconazole might have been more appropriate (8). Indeed, time dependency of the values of the sieving coefficient (SC) and C_E means time dependency of filtration and dialysis clearances that are frequently missed (9). Actually, both filtration and dialysis using fresh crystalloids are nonsaturable mechanisms of elimination. Therefore, time-dependency of SC and C_E suggests a masked mechanism of elimination, including adsorption, which is a saturable mechanism defined by a maximal rate of elimination (V_max) and a concentration corresponding to half the V_max (K_m).

The authors reported values of clearance of isavuconazole in the CVVH patient of 13.31 liters/h; meanwhile, the diafiltration flowrate was set at 2 liters/h. Such a discrepancy raises questions about accuracy of determinations. As stated by the authors (1), the clearance induced by filtration is nothing but the filtration flow rate (2 liters/h) multiplied by the SC of isavuconazole (about 0.025). Therefore, the filtration clearance of isavuconazole should have been on the order of 0.05 liters/h. Clarification is needed to explain the discrepancy between clearances measured in patients receiving CVVH compared to those receiving CVVHDF, as the flow rate seems to be the same in both modes of CRRT, 2 liters/h. However, isavuconazole, a small non-ionized molecule easily eliminated by filtration and dialysis, is >90% protein bound, a factor that strongly decreases its elimination by CRRT.

It would be more appropriate to increase data collection on a limited number of parameters rather than drawing conclusions based on a few heterogenous results with limited clinical relevance. As correctly stated by the authors (1), adsorption of isavuconazole by filters remains a pending question.