From March 2003 to January 2006 a total number of 29 critically ill patients (17 male, 12 female) needed a treatment with intravenous voriconazole as a reserve antifungal at the 10-bed general intensive care unit of the Mainz University Hospital. 15 patients (51.7 %) required intermittent dialysis due to acute (14 patients) or chronic (1 patient) renal failure. In four patients suffering from renal insufficiency the trough plasma levels of voriconazole and sulphobutylether beta cyclodextrin sodium were measured. Blood samples were collected in Monovette plastic tubes and immediately centrifuged at 2000g at 4°C. This investigation was approved by the Institutional Review Board (Rheinland-Pfalz, Germany). The board waived the need for informed consent. Additionally, monitoring for possible toxic effects of sulphobutylether beta cyclodextrin sodium regarding level of consciousness, hemodynamic stability, dermal reactions, and liver function tests was performed.

Voriconazole was determined according to a previously published method 9. In brief, plasma samples were spiked with internal standard and borate buffer was added. The samples were loaded onto solid phase extraction (SPE) columns, washed with buffer and methanol/water, and eluted with methanol/acetic acid mixture. After drying, the extracts were reconstituted with mobile HPLC phase and analyzed by HPLC/UV. The limit of quantification was 0.2 μg/ml. The precision of determination (coefficient of variation) was 7.2, 7.9, 3.2, and 3.7% at quality control concentrations of 0.2, 0.4, 3.4, and 6.9 μg/ml, respectively. Sulphobutylether beta cyclodextrin sodium was determined according to a previously published method 8: Phosphate buffer was added to plasma samples before loading onto SPE columns (Chromabond, cyclohexyl). The loaded SPE columns were washed with buffer and eluted with methanol/water (30/70) mixture. After drying, the extracts were reconstituted in methanol/water (10/90) and analyzed by HPLC/Fluorescence. The limit of quantification was 4.00 μg/ml. The precision of determination (coefficient of variation) was 5.7, 6.7, 7.0, and 6.9% at quality control concentrations of 4.00, 14.9, 65.2, and 126 μg/ml, respectively. The analytical methods were validated according to FDA validation guidelines and fulfilled the respective quality assurance for accuracy and precision.

A 75-year-old woman submitted with acute myeloid leukemia showed a maximal sulphobutylether beta cyclodextrin sodium plasma level of 145 μg/ml in the initial phase of treatment with intravenous voriconazole (Figure 1, Table 1, patient no. 1). After a few days renal function recovered and the plasma levels came down to less than 20 μg/ml after 10 days of treatment with intravenous voriconazole.

In contrast to the patient with a recovery of renal function the remaining three patients showed renal failure during the complete period of intravenous treatment with voriconazole. In these patients an accumulation of sulphobutylether beta cyclodextrin sodium plasma levels was determined with a maximum of 523 μg/ml in a 18-year-old man submitted with myocarditis (Table 1, patient no. 2), 409 μg/ml in a 57-year-old man submitted with multiple myeloma (Table 1, patient no. 3), and 581 μg/ml in a 47-year-old man submitted with intestinal ischemia (Table 1, patient no. 4).

Besides sulphobutylether beta cyclodextrin sodium voriconazole plasma levels were determined which were most commonly less than 1.5 μg/ml and showed no evidence for accumulation (Table 1). In patient no. 1 the levels were, in part, not detectable. In patient no. 3 all voriconazole measurements were distinctly elevated as compared to the remaining patients which was due to a concomittant hepatic insufficiency.

Hemodialysis in the four patients was performed using a low flux AM-BIO 750 cartridge (ASAHI KASEI MEDICAL EUROPE GmbH, Frankfurt, Germany). The average frequency of the dialysis sessions was every 24 to 48 hours, the duration was 4 hours, blood flow 200 to 220 ml/min, and fluid extraction 2 to 3 liters to maintain proper fluid balance. Clinical and biochemical data to assess the toxicity of accumulation of sulphobutylether beta cyclodextrin sodium were determined but in these critically ill patients with complicated disease it was difficult to trace back possible effects. Patient no. 3 was sedated, intubated and mechanically ventilated during the whole course of intravenous voriconazole treatment. Patients no. 2 and 4 were awake but tracheotomized and not fully orientated. Patient no. 1 was extubated but she was somnolent and could not answer questions. None of the patients showed an obvious deterioration of the level of consciousness during the treatment with voriconazole. A more detailed assessment like testing for the frequently described side effect of visual disturbances under intravenous voriconazole could not be performed in these patients. Dermal reactions were not observed under intravenous voriconazole treatment. Patients no. 1, 2, and 4 were hemodynamically stable without catecholamines. Patient no. 3 needed treatment with noradrenaline which required adjustment to individual situations like during high fever or during dialysis. Liver function test results at the beginning of intravenous voriconazole treatment compared to the end of treatment showed the hepatic impairment of patient no. 3 (Table 2).

The present data indicate an accumulation of sulphobutylether beta cyclodextrin sodium in patients with renal failure and intermittent dialysis therapy given intravenous voriconazole in whom restoration of renal function was not achieved during the treatment course. In contrast, in the patient in whom restoration of renal function was obtained early during the treatment with intravenous voriconazole, accumulation of sulphobutylether beta cyclodextrin sodium was not observed. Fortunately, there was no evidence for toxic effects related to the concentrations of sulphobutylether beta cyclodextrin sodium measured in these patients. Assuming no abnormalities in the volume distribution (the steady-state volume of distribution of sulphobutylether beta cyclodextrin sodium is approximately 0.2 L/kg, which is similar to extracellular fluid volume in humans) the maximal concentration values of sulphobutylether beta cyclodextrin sodium would be translate back to a total amount of about 100 mg/kg accumulated sulphobutylether beta cyclodextrin sodium on board. This extrapolated dose value was about 10-fold lower than the usual doses used in animal models for toxicity assessment. In animal experiments the minimal single lethal dose has been determined to be 2000 mg/kg, while daily doses of 1000 mg/kg for one month or 600 mg/kg for six months did not produce functional renal changes 4. Assuming a half-life of sulphobutylether beta cyclodextrin sodium in healthy animals of 1.6 hours the single doses of 1000 or 600 mg/kg would have been eliminated within a few hours (animals receiving 1000 mg/kg would have less than 100 mg/kg on board within about 3.5 hours). In contrast, the 100 mg/kg in our patients with renal failure would represent a continuous baseline level. Therefore, the range of the presently determined sulphobutylether beta cyclodextrin sodium concentrations might well be considered to be toxic from clinical perspective. Nevertheless, for comparisons with animal data a word of caution is suitable regarding the presently investigated patients with renal insufficiency. The sick kidney may be protected from the toxic action of sulphobutylether beta cyclodextrin sodium, which is nephrotoxic through its reuptake from glomerular filtrate by the kidney tubules. This phenomenon has been shown for aminoglykoside antibiotics, which share this mechanism of toxicity. However, accumulation of drug delivery vehicles bare many potential risks to these critically ill patients and should be prevented 1011. Alternative antifungals like liposomal amphotericin B or caspofungin should be considered or intravenous voriconazole therapy should be switched to oral administration as soon as possible. The possible accumulation of sulphobutylether beta cyclodextrin sodium in patients under renal replacement therapy should be studied in prospective clinical studies in more detail. Furthermore, different renal replacement techniques like dialysis with high-flux membranes or continuous veno-venous hemofiltration need to be evaluated regarding the clearance of sulphobutylether beta cyclodextrin sodium.

Regarding the voriconazole plasma levels published data in patients with renal insufficiency is rare. A case report in a patient with chronic renal insufficiency treated by continuous veno-venous hemodiafiltration, which focussed solely on voriconazole levels, demonstrated that no dosage adjustment was required 12. The voriconazole plasma levels presently found in patients no. 1, 2, and 4 were in line with published data on healthy subjects which received 3 mg kg-1 twice daily. By day 12 this previous study showed a maximal concentration of voriconazole of 2.4 μg/ml and a half-life of 6.4 hours 13. The trough levels of voriconazole presently determined were at several points lower than the minimum inhibitory concentrations for aspergillus spp., candida spp., and for most emerging fungal pathogens (minimal concentration > 0.8 μg/ml). Nevertheless, voriconazole treatment levels < 0.2 μg/ml have frequently been observed in successfully treated patients and the manufacturer does not suggest a dosage adjustment 7. The higher levels of voriconazole in patient no. 3 were in line with our previous observations that patients with liver insufficiency frequently show levels around 5 μg/ml (our valid range for voriconazole plasma levels is from 0.2 to 10 μg/ml). Previous studies showed that liver function test abnormalities were statistically significant but weakly associated with elevated plasma voriconazole concentrations. To investigate potential threshold effects, patients with plasma voriconazole concentrations above and below various threshold values (1, 2, 3, 4, 5, 6 μg/ml) were analyzed. These investigations did not identify threshold plasma concentrations above which the risk of a liver function test abnormality was higher compared with plasma concentrations below the threshold 4.