Potential Factors for Inadequate Voriconazole Plasma Concentrations in Intensive Care Unit Patients and Patients with Hematological Malignancies

Abstract

Voriconazole plasma concentrations (VPCs) vary widely, and concentrations outside the therapeutic range are associated with either worse outcome in invasive aspergillosis (IA) or increased toxicity. The primary goal of this cohort study conducted in a real-life setting was to identify potential factors associated with inadequate VPCs in ICU patients and patients with hematological malignancies. Within a period of 12 months, trough VPCs were obtained and analyzed with high-performance liquid chromatography, and the adequate range was defined as 1.5 to 5.5 mg/liter. VPCs of <1.5 mg/liter were defined as low, whereas VPCs of >5.5 mg/liter were defined as potentially toxic. A total of 221 trough VPCs were obtained in 61 patients receiving voriconazole, and 124/221 VPCs (56%) were found to be low. Multivariate analysis revealed that low VPCs were significantly associated with clinical failure of voriconazole, prophylactic use, younger age, underlying hematological malignancy, concomitant proton pump inhibitor (PPI) (pantoprazole was used in 88% of the patients), and absence of side effects. Low VPCs remained an independent predictor of clinical failure of voriconazole. The defined adequate range was reached in 79/221 (36%) VPCs. In 18 samples (8%), potentially toxic levels were measured. Multivariate analysis revealed higher body mass index (BMI), absence of hematological malignancy, therapeutic application, and diarrhea as factors associated with potentially toxic VPCs. Neurotoxic adverse events occurred in six patients and were mostly associated with VPCs in the upper quartile of our defined adequate range. In conclusion, potential factors like younger age, prophylaxis, underlying hematological malignancy, BMI, and concomitant PPI should be considered within the algorithm of voriconazole treatment.

INTRODUCTION

Invasive fungal infections (IFI), in particular invasive aspergillosis (IA), are an important cause of mortality among patients with hematological malignancies and patients in intensive care units (ICUs) (1–4). Voriconazole has broad-spectrum antifungal activity and is currently considered a gold standard in therapy of IA (5). The drug may, however, be associated with adverse events (AEs), including visual disturbance, encephalopathy, and hepatic enzyme elevation (6). Voriconazole has nonlinear pharmacokinetics and undergoes extensive hepatic metabolism by the cytochrome P450 system (mainly CYP2C19 and CYP3A4) that depends on age, genetic factors, and interactions with other drugs (7). Thus, significant interpatient variability is observed after administration of the same dose. Studies have shown that not only interpatient, but also intrapatient variability of voriconazole plasma concentrations (VPCs) is significant (8–10). Plasma concentrations below 1.5 mg/liter have been associated with a worse outcome in IA and VPCs above 5.5 mg/liter with increased toxicity (6, 8–11). In a recent randomized, assessor-blinded, controlled, single-center trial that included 110 adult patients (mostly patients with hematological malignancies), Park and colleagues showed that routine therapeutic drug monitoring (TDM) of voriconazole may reduce drug discontinuation due to adverse events and improve the treatment response in invasive fungal infections (12). Various studies have recommended TDM, not only in the adult, but also in the pediatric patient cohort (13, 14). While most studies have concentrated on patients with hematological malignancies, comparably limited data exist about TDM of voriconazole in ICU patients (11). As the therapeutic range of voriconazole is narrow, knowledge of potential factors associated with inadequate VPCs is crucial.

We conducted a monocentric prospective study to identify potential factors associated with inadequate trough VPCs in patients with hematological malignancies and ICU patients.

(Original data from the study were presented at ID Week 2012 in San Diego, CA, and at ECCMID 2013 in Berlin, Germany.)

MATERIALS AND METHODS

The cohort study was conducted from August 2011 to September 2012 at the Medical University Hospital of Graz, Graz, Austria. During this time, all trough VPCs that were obtained from patients with underlying hematological diseases and patients admitted to ICUs were included. The objectives were to analyze trough VPCs in this real-life setting and to identify potential factors associated with inadequate VPCs.

Patients receiving voriconazole were identified and screened by clinical rounds. Patients' medical records were reviewed individually by using a standardized data collection template in order to collect demographic information and clinical data on outcomes of therapy and adverse events, mycological laboratory test results, and voriconazole dosing information and concomitant medications taken during voriconazole therapy (15). Voriconazole dosing records for each patient were used to verify the time of voriconazole concentration sampling in relation to the dose. The standard dosages (all divided into two doses) were a 12-mg/kg of body weight loading dose followed by 8-mg/kg maintenance for intravenous therapy (with no dosage adjustment in cases of renal impairment) and an 800-mg loading dose followed by 400-mg maintenance for oral therapy. In our clinical setting, trough VPCs were measured 12 h after administration of the last voriconazole dose, with the initial VPC being measured on day 4 of voriconazole prophylaxis/therapy and then repeated once or twice a week in the case of adequate VPCs or up to four times a week in the case of inadequate VPCs, toxicity, or treatment failure. In cases of voriconazole dosage adjustment, it was recommended to wait at least 24 h before measuring the next VPC. Therefore, VPCs were included if measured 96 h or more after receiving the initial dose (the first 2 doses were always loading doses) or 24 h after dosage adjustment. Only patients above 18 years of age with underlying hematological malignancies or admitted to the ICU were included. Patients with underlying hematological malignancies, including those with hematopoietic stem cell transplantation (HSCT), who were admitted to the ICU at the time of testing were allocated to the hematological-malignancy group. One case represented a single patient during hospitalization and was completed at the patient's discharge. Patients receiving long-term voriconazole prophylaxis were counted as a single case regardless of the number of times they were readmitted. Response to therapy was defined as resolution/major improvement of symptoms and signs of IA (including radiographic changes on chest X rays or computed tomography scans, as well as negative serum galactomannan results). Response to prophylaxis was defined as absence of breakthrough IFI. Failure was defined as deterioration or lack of significant improvement of the same parameters (including death of the patient or drug withdrawal with evidence of infection still present) or occurrence of breakthrough IFI.

IFI was defined according to consensus definitions of the European Organization for Research and Treatment of Cancer Invasive Fungal Infections Cooperative Group (EORTC) and the Mycoses Study Group of the National Institute of Allergy and Infectious Disease (MSG), which are, however, not specifically designed for the ICU (16, 17).

Trough VPCs were measured by employing the CE-IVD-marked Chromsystems Voriconazole Reagent Kit (Chromsystems GmbH, Munich, Germany) based on high-performance liquid chromatography (HPLC) with fluorescence detection, with a lower limit of quantitation of 0.3 mg/liter. Therapeutic targets for voriconazole have not yet been defined; however, tentative recommendations suggest targeting a trough level between 1.5 mg/liter and 5.5 mg/liter for both treatment and prophylaxis (9, 10). Therefore, concentrations within the range of 1.5 mg/liter to 5.5 mg/liter were defined as adequate; inadequate concentrations were either below (low VPC) or above (potentially toxic VPC) this range.

At our center, general practice in response to low VPCs was to increase the dosage up to a maximum of 12 mg/kg/day divided into two doses. The voriconazole dosage was decreased in response to potentially toxic VPCs. Consultations by our infectious disease service recommending dosage adjustment were regularly performed for patients included in the analysis.

The study adhered to Declaration of Helsinki, 1996, good clinical practice, and the study protocol was approved by the local ethics committee, Medical University of Graz, Graz, Austria.

Statistical analysis was performed using SPSS, version 19 (SPSS Inc., Chicago, IL, USA). Continuous data (i.e., VPCs) are presented as medians (interquartile ranges [IQR]) and categorical data as proportions. Proportions were compared using the chi-squared or Fisher's exact test as appropriate. Analyses of continuous data were performed using the Mann-Whitney U test or Wilcoxon signed-rank test as appropriate due to the nonnormality of voriconazole concentrations. The P values of the Mann-Whitney U tests were not corrected for multiple comparisons and are therefore only descriptive. Univariate and multivariate logistic regression analyses were used to identify potential factors that contribute to the variability in VPCs, and odds ratios (OR) are shown. A P value of less than 0.05 was considered statistically significant.

RESULTS

A total of 221 trough VPCs (median, 1.20 mg/liter; IQR, <0.3 to 3.4) were obtained in 61 patients receiving voriconazole (144 VPCs in 40 hematological malignancy patients; 77 VPCs in 21 ICU patients). Twenty hematological patients received voriconazole for prophylactic (HSCT, n = 8; induction chemotherapy-related prolonged neutropenia, n = 12); all other patients received voriconazole for therapeutic reasons. Demographic and clinical data for the patients are depicted in Table 1.

Table 1.

Demographic and clinical data for patients with underlying hematological malignancies admitted to the ICU

Parametera	Value for patients:		P value (if significant)
	Hematological malignancies	ICU
N	40	21
No. (%) male/female	25 (63)/15 (38)	11 (52)/10 (48)
Age (yr) [median (range)]	51.55 (18.73–81.27)	64.31 (41.83–78.64)	0.002
No. (%) with primary underlying disease/condition
AML	17 (43)	0
MDS	8 (20)	0
NHL	7 (18)	0
MM	5 (13)	0
ALL	3 (8)	0
SOTb	0	6 (29)
Cardiac surgeryc	0	3 (14)
COPD (GOLD 3–4)	0	2 (10)
Solid tumor	0	2 (10)
Systemic connective tissue diseases	0	2 (10)
Trauma	0	2 (10)
Neurological disorder/surgery	0	2 (10)
Abdominal surgeryc	0	1 (5)
AIDS	0	1 (5)
IFI classification [no. (%)]
No IFI (prophylaxis group)	20 (50)	0
Possible IFI	6 (15)	10 (48)
Probable IFI	13 (33)	8 (38)
Proven IFI	1 (3)	3 (14)
No. (%) with prophylaxis/therapy	20 (50)/20 (50)	0/21 (100)	<0.001
Response to therapy/prophylaxis [no. (%)]	33 (83)	14 (67)
6-week overall mortality [no. (%)]	7 (18)	3 (14)
12-week overall mortality [no. (%)]	11 (28)	4 (19)
AE to voriconazole [no. (%)]	3 (8)	8 (38)	0.005

^aALL, acute lymphocytic leukemia; AML, acute myeloid leukemia; COPD, chronic obstructive pulmonary disease; MDS, myelodysplastic syndrome; MM, multiple myeloma; NHL, non-Hodgkin's lymphoma; SOT, solid-organ transplantation.

^bThree patients each with renal and liver transplants.

^cAll associated with a prolonged stay in the ICU (>30 days).

Initial VPCs on day 4 of treatment were low or within the range defined as adequate in 27 patients each, while in 7 patients, potentially toxic VPCs were measured. The median oral weight-based dosage was 3.29 mg/kg every 12 h (q12h) (IQR, 2.74 to 3.83) and was therefore significantly lower than the median intravenous dosage (P < 0.001).

Overall, 124 of 221 VPCs (56%) were found to be low. No significant difference was found between hematological and ICU patients receiving voriconazole therapeutically. Seventy-nine of 221 (36%) VPCs obtained were found to be within the range defined as adequate. In 18 samples, potentially toxic VPCs were obtained. Neurotoxic AEs occurred in six patients (hallucinations in three patients, visual disturbances in two patients, and encephalopathy in one patient, associated with VPCs of 3.0, 4.5, 4.7, 4.9, 5.1, and 5.9 mg/liter) and cholestatic hepatopathy in five patients. VPCs were significantly higher in patients who experienced an AE (P < 0.001; median, 4.7 mg/liter; IQR, 4.2 to 5.1) than in those who did not (median, 1.1; IQR, <0.3 to 2.95). While reduction of the voriconazole dosage was sufficient in four patients experiencing an AE related to voriconazole, the drug had to be discontinued in seven patients due to the AE. Further, voriconazole had to be discontinued in six other cases due to constant low VPCs despite escalation of the dosage to 12 mg/kg body weight. In three more cases, the switch from oral to intravenous administration led to VPCs within the defined adequate range.

According to the Mann-Whitney U test (P values were not corrected for multiple comparisons and are therefore only descriptive), significantly higher VPCs were observed in female patients (P = 0.021), older patients (P < 0.001), patients who received therapeutic (P < 0.001) and intravenous (P = 0.018) voriconazole, patients without concomitant proton pump inhibitors (PPI) (P = 0.015), patients who experienced an AE (P < 0.001), and patients who died within 6 weeks (P = 0.03). Patients with low VPCs had significantly lower body mass indexes (BMI) (P = 0.006) and were significantly younger (P < 0.001). Low VPCs were also significantly associated with male sex (P = 0.019), prophylactic (P < 0.001) and oral (P = 0.029) voriconazole, concomitant PPI treatment (P = 0.004), patients who survived at week 6 (P = 0.008), and patients who did not experience a voriconazole-associated AE (P = 0.001). VPCs within the defined adequate range were significantly associated with older age (P = 0.002), female sex (P = 0.002), ICU patients (P = 0.028), therapeutic application (P < 0.001), absence of concomitant PPI (P = 0.005), absence of diarrhea (P = 0.012), and patients who experienced an AE (P = 0.019). Potentially toxic VPCs were significantly associated with older age (P = 0.027), higher BMI (P < 0.001), therapeutic (P = 0.015) and intravenous (P = 0.014) voriconazole, diarrhea (P = 0.015), and patients who died within 6 weeks (P = 0.02). Details of VPCs obtained in patients with hematological malignancies and ICU patients are depicted in Table 2.

Table 2.

VPCs of patients with underlying hematological malignancies admitted to the ICU

Parameter	Value for patientsa:		P valueb (if significant)
	Hematological malignancies	ICU
N	144	77
No. of VPCs per patient	3 (1–8)	3 (1–8)
Male/female [no. (%)]	86 (60)/58 (40)	45 (58)/32 (42)
Age (yr)	49.17 (45.22–52.15)	64.71 (57.23–71.72)	<0.001
Weight (kg)	69.5 (61–79)	80 (61–95)	0.008
BMI	24.68 (19.99–27.47)	27.90 (23.44–30.70)	<0.001
Prophylaxis/therapy [no. (%)]	94 (65)/50 (35)	0/77 (100)	<0.001
Administration [no. (%)]
Intravenous	62 (43)	61 (79)	<0.001
Oral	82 (57)	16 (21)	<0.001
Concomitant PPI [no. (%)]	113 (78)	75 (97)	<0.001
Concomitant systemic cortiocosteroids [no. (%)]	41 (28)	23 (30)
Concomitant rifampin [no. (%)]	2 (1)	0
Diarrhea [no. (%)]	39 (51)	4 (5)	<0.001
Response to therapy/prophylaxis [no. (%)]	123 (85)	56 (72)	0.022
6-week overall mortality [no. (%)]	21 (15)	15 (19)
12-week overall mortality [no. (%)]	31 (22)	17 (22)
VPCs (mg/liter)	0.95 (0.2–2.5)	1.5 (0.45–3.95)
VPCs within targeted range [no. (%)]	44 (31)	35 (45)	0.028
Low VPCs [no. (%)]	87 (60)	37 (48)
Potentially toxic VPCs [no. (%)]	14 (10)	5 (6)

^aMedian (IQR) for continuous data unless otherwise indicated.

^bP values were calculated with Fisher's exact/chi-squared or Mann-Whitney U test.

Low VPCs remained an independent predictor of clinical failure of voriconazole treatment/prophylaxis (P = 0.041; OR, 2.92; 95% confidence interval [CI], 1.05 to 8.14) in multivariate analysis. The results of univariate and multivariate analyses of factors associated with low VPCs (<1.5 mg/liter) and potentially toxic VPCs (>5.5 mg/liter) are depicted in Table 3. Those associated with adequate VPCs are shown in Table 4.

Table 3.

Univariate and multivariate analyses of factors associated with low and potentially toxic VPCs

Risk factor	Insufficient VPCs (<1.5 mg/liter)						Potentially toxic VPCs (>5.5 mg/liter)
	Univariate analysis			Multivariate analysis			Univariate analysis			Multivariate analysis
	OR	95% CI	P valuea	OR	95% CI	P valuea	OR	95% CI	P valuea	OR	95% CI	P valuea
Age (per yr)	0.95	0.92–0.99	0.009	0.96	0.93–0.99	0.003	1.06	0.99–1.12
Male sex	7.31	2.03–26.37	0.002				29.92	1.06–848	0.046
Wt (kg)	0.91	0.85–0.97	0.003				0.89	0.78–1.02
BMI	1.27	1.05–1.53	0.013				1.61	1.06–2.44	0.026	1.14	1.02–1.28	0.024
Hematological malignancy	3.51	1.13–10.89	0.030	3.85	1.53–9.67	0.004	0.75	0.01–0.55	0.01	0.2	0.05–0.74	0.016
Therapeutic application	0.08	0.03–0.25	<0.001	0.09	0.04–0.24	<0.001	9.92	1–98.44		14.65	3.2–67.07	0.001
Intravenous formulation	0.62	0.26–1.50					2.01	0.38–10.73
Concomitant systemic corticosteroids	2.27	0.97–5.42					0.74	0.17–3.27
Concomitant PPI	6.67	1.93–23.04	0.003	5.31	2–14.09	0.001	1.23	0.2–7.7
Concomitant rifampin	0.43	0.02–11.97					124.88	1.2–12967	0.042
Diarrhea	0.58	0.19–1.73					4.02	0.61–26.34		5.82	1.7–19.96	0.005
Overall mortality
6 wk	0.05	0.01–0.88	0.041
12 wk	8.00	0.56–115.47
Clinical response to voriconazole	0.11	0.03–0.46	0.002	0.40	0.17–0.93	0.033	4.75	0.49–45.89
AE due to voriconazole	0.37	0.11–1.18		0.08	0.09–0.68	0.021	6.63	0.37–119.43

^aP values are shown if significant.

Table 4.

Univariate and multivariate analyses of factors associated with VPCs within the defined adequate range (≥1.5 to ≤5.5 mg/liter)

Risk factor	Univariate analysis			Multivariate analysis
	OR	95% CI	P valuea	OR	95% CI	P valuea
Age (per yr)	1.04	1.01–1.08	0.027	1.03	1.01–1.05	0.024
Male sex	0.15	0.05–0.5	0.002
Wt (kg)	1.09	1.03–1.15	0.004
BMI	0.78	0.65–0.93	0.005
Hematological malignancy	0.81	0.29–2.28
Therapeutic application	4.16	1.51–11.44	0.006	6.43	2.91–14.25	<0.001
Intravenous formulation	1.05	0.45–2.44
Concomitant systemic corticosteroids	0.59	0.26–1.34
Concomitant PPI	0.2	0.07–0.59	0.004	0.17	0.07–0.42	<0.001
Diarrhea	0.45	0.15–1.41
Overall mortality
6 wk	8.69	0.52–146.31
12 wk	0.26	0.01–3.47
Clinical response to voriconazole	5.85	1.43–23.97	0.014	2.7	1.2–6.1	0.017
AE due to voriconazole	2.26	0.50–10.17

^aP values are shown if significant.

DISCUSSION

We conducted a prospective study to evaluate VPCs among ICU patients and patients with underlying hematological malignancies. Three main findings are evident. First, VPCs were frequently low in both patient groups, and multivariate analysis identified younger age, prophylaxis, underlying hematological malignancy, and concomitant PPI as potential associated factors. Further, low levels remained a significant predictor of failure of voriconazole treatment and prophylaxis. Second, higher BMI, absence of hematological malignancy, therapeutic voriconazole, and diarrhea remained significant predictors for potentially toxic VPCs. Third, most neurotoxic AEs occurred at VPCs between 4.5 and 5.1 mg/liter, which were still within our defined adequate range.

We found that 48% of VPCs in the ICU group, and as many as 60% in the hematological-malignancy group, were low. These low VPCs were significantly more frequent in patients with oral/prophylactic application than in those with intravenous/therapeutic application. A possible reason may have been that we used fixed standard dosages for oral application in contrast to the weight-based intravenous dosage. The standard 200 mg twice a day (b.i.d.) dosage may indeed be insufficient to reach adequate VPCs in many patients. Some studies have suggested weight-based dosage for oral application as well, while others have suggested higher oral dosages (9, 18, 19). Our dose adjustment strategy generally worked well, although voriconazole had to be discontinued in about 10% of the patients, who did not reach adequate VPCs despite escalation to the maximum dosage of 12 mg/kg/day. Studies evaluating higher voriconazole dosages than 12 mg/kg in patients with low trough levels are therefore needed.

Multivariate analyses also identified younger age, underlying hematological malignancy, concomitant PPI, clinical failure of voriconazole, and absence of AEs as potential factors associated with low VPCs. On the other hand, clinical response to voriconazole therapy, older age, therapeutic voriconazole, and absence of concomitant PPI were significant predictors of VPCs within our defined adequate range, while higher BMI, absence of hematological malignancy, therapeutic voriconazole, and diarrhea remained significant predictors for potentially toxic VPCs. With regard to clinical response, age, and therapeutic application, comparable results were also reported in other studies (8–10, 19–21). Interestingly, there was a significant negative correlation between PPI treatment and high VPCs. PPI treatment, in fact, remained a significant predictor of low VPCs in multivariate analysis. These findings are in contrast to previous findings reporting increased VPCs in cases of concomitant PPI use (8, 22, 23). Other studies, however, found no effect of PPI treatment on the pharmacokinetics of voriconazole (24). The reason for the differing results may have been that the vast majority of patients, especially those in the ICU group, had received PPI treatment in our study. Another reason for discrepant findings may rely on the fact that metabolization of drugs using the CYP2C19 pathway (including omeprazol, clopidogrel, sertaline, and azoles) can be affected by genetic polymorphisms, resulting either in poor metabolizers when a loss-of-function allele (*2 or *3) is present or in extensive metabolizers with a gain-of-function allele (*17); in addition, which kind of PPI was used seems to be important, as the use of pantoprazol (which had been used in 88% of the patients in this study) seems less influential than the use of omeprazol (25). We also found an interesting correlation of diarrhea with potentially toxic VPCs. A possible explanation may be that plasma volume is decreased in such patients, resulting in higher VPCs after intravenous administration.

Although the results were not significant, Lee and colleagues found that even a single initial VPC may correlate with the clinical outcome in hematological-malignancy patients (26). The fact that serial testing is essential, however, has also been observed in our study collective, as trough VPCs frequently varied markedly over time in the same patients. This intrapersonal variability, along with nonlinear saturable pharmacokinetics of voriconazole in adults, may raise the concern that the adjustment of the voriconazole dosage based on TDM at a single time point may result in suboptimal VPCs at a later time (27, 28).

TDM may help to maximize the efficacy and minimize the risk of toxicity. Nevertheless, AEs to voriconazole occurred in nearly 20% of patients and led to discontinuation of the drug in nearly 10% of the patients included. The fact that voriconazole was continued at a lower dosage in four of the patients experiencing an AE may be attributable directly to TDM, however, as similar findings were reported in Park's randomized TDM study (12).

In a very interesting recent study, Koselke and colleagues reported a strong association between potentially toxic concentrations and morbidly obese patients with weight-based dosage (29). Higher BMI also remained a significant predictor of potentially toxic VPCs in multivariate analysis in our study. Dosing voriconazole based on an ideal body weight or adjusted body weight may therefore be appropriate for morbidly obese patients, but nevertheless, TDM should still be performed.

We found that experiencing a voriconazole-related AE was associated with significantly higher VPCs. Other studies have also found significantly higher VPCs among patients experiencing an AE than among patients who did not (6, 20). In particular, neurological AEs have been associated with VPCs of >5.0 mg/liter (8, 10). In our study, neurological AEs occurred in six patients; associated VPCs were found in the upper quartile of our defined adequate range in most cases (four patients had VPCs between 4.5 and 5.1 mg/liter), and only one patient exhibited VPCs above 5.5 mg/liter. These results may suggest that lowering the upper VPC cutoff from 5.5 to 4.5 mg/liter would be meaningful. Similar results were recently reported in a study by Pascual et al., where the optimal upper cutoff was identified as 4.5 mg/liter by logistic multivariate regression analysis. That cutoff was associated with a probability of neurotoxicity below 15% in that study (9). In another study exclusively evaluating prophylaxis among lung transplant recipients, a cutoff of 4.0 mg/liter was even suggested (11).

Our study has several limitations, including its observational design. We also did not evaluate CYP2C19 status, where poor metabolizers have been shown to have lower VPCs (12, 30). The frequency of poor CYP2C19 metabolizers was reported to be significantly higher in Asians (15% to 20%) than in Caucasians (2% to 3%), and in this study, none of the included patients was of Asian origin (31). Further, some previously described factors, such as concomitant PPI treatment, were present in the vast majority of included patients, while other factors, such as concomitant phenytoin or rifampin medication, were present in very few. Nevertheless, these results suggest that low VPCs are frequently found in ICU patients and in patients with hematological malignancies. Higher and/or bodyweight-adjusted oral dosages may increase low VPCs, which are found frequently and in particular among patients receiving oral/prophylactic regimens. On the other hand, potentially toxic VPCs are mostly found in patients with very high BMIs and intravenous bodyweight-adjusted dosages. Dosing voriconazole based on an ideal body weight or adjusted body weight may therefore be appropriate for morbidly obese patients. Finally, with regard to neurotoxicity, the results suggest that lowering the upper VPC cutoff from 5.5 to 4.5 mg/liter may be beneficial.