|Home | About | Journals | Submit | Contact Us | Français|
Voriconazole is increasingly used in allogeneic hematopoietic stem cell transplantation for prophylaxis and treatment of fungal infections. Hepatic dysfunction is common in hematopoietic stem cell transplant patients and may impact the clinical decision to institute voriconazole.
Retrospective review was conducted of all adult and pediatric hematopoietic stem cell transplant recipients who received >1 day of voriconazole from January 2005 through July 2007. Clinical hepatotoxicity was defined as the subjective attribution of liver enzyme elevations, even mild, as hepatotoxicity due to voriconazole by the treating physician and leading to discontinuation of voriconazole. Biochemical hepatotoxicity was defined as elevations in one or more liver enzymes to > 3 times upper limit of normal or > 3 times baseline if abnormal at baseline. Liver enzymes included: aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase, and total bilirubin. Simple and multiple logistic regressions were used to define risks for hepatic dysfunction. Wilcoxon signed rank tests were used to assess the difference between liver function test values before, during, and after use of voriconazole.
Sixty-eight (34%) out of 200 patients developed hepatotoxicity on voriconazole. Median duration of administration was 72 days (range: 1–804 days). Biochemical hepatotoxicity occurred in 51 (75%) and clinical hepatotoxicity in 17 (25%) of patients. Thirty-five (51%) of patients with hepatotoxicity required discontinuation of therapy. In simple logistic regression, acute graft-versus-host disease was a risk factor for hepatotoxicity. Receipt of a T-cell depleted allograft was protective. In multiple logistic regression, acute GVHD (p=0.002) remained significant. There were no cases of liver failure or death attributed to voriconazole.
In this cohort of allogeneic hematopoietic stem cell transplant patients, the rate of hepatotoxicity while on voriconazole was 34%. In general, the hepatic dysfunction was mild and reversible. Voriconazole therapy, with monitoring, appears reasonably safe in hematopoietic stem cell transplant recipients at high risk for invasive fungal infections.
Voriconazole is a highly orally bioavailable azole with broad activity against Candida and Aspergillus . Triazole antifungal agents have been reported to cause both cholestatic and hepatocellular injury [2,3]. Rarely, fulminant hepatitis with hepatic necrosis is reported in patients receiving triazoles. Severe hepatitis can be enhanced upon re-challenge and may be fatal [1,4–12]. In clinical trials of voriconazole for the treatment of invasive aspergillosis, transaminase elevations were observed in up to 19% of patients with 4% being serious hepatic adverse events. In an observational study of patients with hematologic malignancies, up to 69% of patients developed transaminase elevations. However, only 7% of patients were thought to have clinically significant hepatotoxicity (HT) and required discontinuation of voriconazole . The incidence of liver enzyme elevations and clinically significant hepatotoxicity in hematopoietic stem cell transplant (HSCT) recipients treated with voriconazole is not known. Because of its broad antifungal spectrum, relatively low rate of toxicities, and convenience of administration, voriconazole is increasingly used for prophylaxis and treatment of invasive fungal infections in HSCT recipients who are already at risk for hepatotoxicity [3,14].
Liver dysfunction in allogeneic HSCT recipients may be due to a variety of factors including toxicity from the preparative regimen and other medications, infection, veno-occlusive disease (VOD), and acute and chronic graft-versus-host disease (GVHD) of the liver. These conditions may increase the risk for hepatotoxicity associated with voriconazole in HSCT recipients. Over the last 2.5 years at our institution, voriconazole has been used for prophylaxis and treatment of fungal infections in HSCT. We report our experience on the hepatic safety of voriconazole in this population.
We conducted a retrospective review to determine the frequency of HT in HSCT patients who received voriconazole and to establish if any additional transplant-related factors increased the risk for HT.
The study was approved by the Memorial Sloan-Kettering Cancer Center (MSKCC) Institutional Review Board. Adult and pediatric patients who received their first allogeneic HSCT at MSKCC from January 2005 through July 2007 were included in the study. Patients who received >2 consecutive doses of voriconazole from 1/1/05–2/1/08 were identified from pharmacy records. Patients were included in the study if they had liver enzymes measured within 2 weeks prior and during voriconazole therapy.
Biochemical hepatotoxicity (BIO-HT) was defined as elevation of aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (Alk Phos), and/or total bilirubin (TBil) to >3 times ULN or >3 times baseline if abnormal at baseline, regardless of whether the abnormalities led to discontinuation or interruption of voriconazole. Clinical hepatotoxicity (CL-HT) was based on the subjective assessment of the treating physician (CL-HT) was defined as mild elevations in LFTs (1.5–2.9 times above the upper limit of normal or above baseline if abnormal) attributed to voriconazole and leading treatment discontinuation.
Indication was considered prophylaxis if voriconazole was initiated prior to engraftment for the prevention of fungal infections. For this study, all other indications for voriconazole including secondary prophylaxis while receiving intensive immunosuppression for GVHD and treatment of fungal infections were defined collectively as “all other indications”. These conditions are frequently associated with additional potential reasons for HT. Duration of therapy was defined as the number of days between first and last dose of voriconazole. If the therapy was interrupted for more than 5 days, resumption of therapy was defined as a second course.
Prophylaxis against opportunistic infections was instituted according to the standards of care at MSKCC. To prevent herpes simplex and varicella zoster virus reactivation acyclovir was given form beginning of conditioning until immune reconstitution.
Cytomegalovirus (CMV) prophylaxis was not routinely used. Patients at risk for CMV disease were monitored and treated if they had evidence of CMV reactivation. For Pneumocystis jiroveci prophylaxis, trimethoprim/sulfamethoxazole (TMP/SMX) or pentamidine (in cases of sulfa allergy) were given from day-7 to day-3 followed by oral TMP/SMX or inhaled pentamidine from day+30 and continued until immune reconstitution. Penicillin VK or equivalent was used for prophylaxis against invasive disease due to Streptococcus pneumoniae in patients with chronic GVHD or splenectomy.
Recipients of conventional grafts received standard prophylaxis for GVHD. All patients on cyclosporine, tacrolimus, or sirolimus had routine therapeutic drug monitoring of the immunosuppressants. An elevated trough level was defined >600 ng/ml for cyclosporine and >15 ng/ml for tacrolimus and sirolimus during administration of voriconazole.
For the prevention of invasive fungal infections, low risk patients received fluconazole from the day of admission for stem cell transplant until at least day+75. Patients with any risk factors for invasive mold infection (prior possible, probable, or proven invasive mold infection, GVHD, CMV infection, HLA-mismatched or unrelated donor, prior or current extensive corticosteroid use, or age >50) received intravenous voriconazole 6 mg/kg every 12 hours for 2 doses, then 4 mg/kg every 12 hours starting early post-transplant, followed by oral voriconazole 200 mg every 12 hours until at least day+75 or cessation of intensive immunosuppression. Patients with GVHD or treatment with corticosteroids also received voriconazole until discontinuation of immunosuppression. Patients intolerant to voriconazole received intravenous micafungin. For treatment of mold infections, voriconazole was administered at the same dose as for prophylaxis. There was no routine therapeutic drug monitoring for voriconazole during the study period. Voriconazole levels were checked at the discretion of the treating physician.
Patients who received >2 consecutive doses of voriconazole were evaluated for HT. Patient demographics, primary underlying disease, indication for voriconazole, prior concomitant medications, concurrent medical events and liver function test (LFT) results were obtained from the electronic medical record. All LFT values from 2 weeks prior to the start of therapy, during treatment, and 2 to 4 weeks after the end of voriconazole were recorded. The maximum value was recorded for each of the three time periods. Concomitant medications were defined as any therapy taken while on voriconazole. Timing and reasons for discontinuation of voriconazole, as documented in the medical record, were recorded.
Descriptive statistics were used to summarize patient characteristics and safety outcomes. Simple logistic regression was used for univariate analysis of age, sex, duration of voriconazole treatment (greater or less than 30 days), underlying disease (leukemia, myelodysplastic syndrome, non-Hodgkin’s or Hodgkin’s lymphoma, non-malignant, or other) donor type (matched related, matched unrelated, mismatched related, or mismatched unrelated), stem cell source (peripheral blood, bone marrow, or cord blood), graft manipulation (T-cell depletion or unmodified), conditioning intensity (ablative or non-ablative, conditioning regimen (total body irradiation (TBI)-containing, busulfan-containing, or other), acute GVHD (grades 0–1 or grades 2–4), and indication for voriconazole (prophylaxis or all other indications). The Fisher’s exact test was used to determine the significance of the association between elevated levels of voriconazole or immunosuppressants (cyclosporine, tacrolimus, and sirolimus) and HT. Wilcoxon signed rank tests were used to assess the difference between maximum LFT values at baseline, during treatment, and post-treatment. P-values ≤ 0.05 were considered significant. All analyses were performed using Stata version 8.
A total of 334 adult and pediatric patients underwent allogeneic HSCT at MSKCC during the study period. Two hundred patients who received >2 consecutive doses of voriconazole and had liver enzyme monitoring prior to and during therapy were included in the study. One hundred seventeen patients (58.5%) received voriconazole for post-transplant prophylaxis, and 83 (41.5%) for all other indications. The median overall duration of voriconazole therapy was 72 days (range: 1–804). The median duration for patients with HT was 63 days (range: 3–530) compared to 77 days (range: 1–804) in those without HT.
Overall 68 (34%) patients developed HT on voriconazole. HT developed after a median of 26 days (range: 0–341). Fifty-one (75%) patients met criteria for BIO-HT. The remaining 17 (25%) patients had CL-HT. Figure 1 shows the rates of HT for the study patients. Voriconazole therapy was discontinued in 35 (51%) patients with HT. Six patients were re-challenged with a second course of voriconazole. Two of the six patients discontinued the second course due to recurrent HT. The second course of treatment was not included in any of the following analyses.
In 22 (32%) of the 68 patients, HT occurred contemporaneously with major diagnoses associated with HT including: GVHD of the liver or gut (N=11), VOD of the liver (N=3), sepsis with multisystem organ failure (N=4), acute infectious hepatitis (N=1), major bleed leading to TBil elevations (N=1), cholecystitis (N=2). While we cannot exclude the contribution of voriconazole to the HT, the presentation and temporal association of HT in these patients was more consistent with alternative causes.
For patients that experienced biochemical hepatotoxicity we compared the maximum value of the liver enzyme at baseline, during treatment, and post-discontinuation of voriconazole to assess the magnitude and reversibility of hepatotoxicity. Figure 2 shows the box plots of hepatic function tests for the 3 time-points. For the majority of the patients the maximum elevations in AST, ALT, TBil and Alk Phos during treatment were moderate. Post-treatment values of AST, ALT, and Alk Phos were significantly lower than the values during treatment and were similar to baseline values. Post-treatment TBil was similar to values during treatment and significantly higher compared to baseline (p=0.02). Ten patients had TBil elevations and followup monitoring after discontinuation of voriconazole. In 6 (60%) patients with TBil elevations, HT was attributable to other causes. In 4 other patients Tbil elevations returned to normal values.
Fifteen patients died prior or during week 2 after the discontinuation of voriconazole and did not have post treatment LFT. Voriconazole HT was not thought to contribute to death in any of the 15 patients. Causes of death included septic shock with multisystem organ failure (N=7), relapsed underlying malignancy (N=4), alveolar hemorrhage (N=2), and VOD (N=2).
Table 1 shows the characteristics of the 68 patients who developed HT compared to the 132 patients who did not. We conducted simple logistic regression to identify risk factors for hepatotoxicity. All variables listed in Table 1 were included in the analysis as putative risk factors for HT. Table 2 shows significant variables. Acute GVHD was a risk factor for HT (OR 2.98, CI 1.52–5.85, p=0.001). This was supported by an association between the receipt of T-cell depleted graft and less HT (OR 0.51, CI 0.28–0.93, p=0.027). In multivariate analyses, HT was associated with acute GVHD grades 2–4 (OR 3.17, CI 1.50–6.68, p=0.002).
The majority of patients received multiple medications known to cause HT alone or synergistically with voriconazole. The most commonly used concomitant medications with potential for HT are shown in Table 3. Overall, there was no substantial difference in the utilization of each major category of concomitant medications between patients who developed HT while on voriconazole compared to those who did not. Ten (15%) patients with HT had received concomitant agents which were temporally related and likely contributed to HT including: combination of antifungal agents (N=5), oral contraceptives (N=3), deferasirox (N=1), and anabolic steroid (N=1).
During the study period, voriconazole level monitoring was performed at the discretion of the clinician. Voriconazole steady state trough levels were checked in 36 patients who developed HT and 29 patients who did not develop HT. Three (8.3%) with HT and 4 (13.8%) without HT had at least one through level >6 mg/L (p=NS). All patients on cyclosporine, tacrolimus, or sirolimus had routine therapeutic drug monitoring. Elevated trough level was noted for 26 of 63 patients on cyclosporine, 21 of 38 patients on tacrolimus and 3 of 16 patients on sirolimus. The proportion of patients with elevated trough levels was not significantly different between patients with and without HT: cyclosporine 44% vs 39% (p=0.97), tacrolimus 73% vs 44% (p=0.35), sirolimus 50% vs 14.3% (p=0.69).
Evaluating hepatic safety of medications in HSCT recipients is challenging because of frequent confounding conditions and medications associated with hepatic abnormalities [15–17]. Our observational study is the largest-to-date assessment of the hepatic safety of voriconazole in allogeneic HSCT outside the context of a randomized, controlled clinical trial. We analyzed hepatotoxicity based on the measurements of serum transaminase levels. The assessment of clinical hepatotoxicity was in part subjective.
The exposure to voriconazole was long with a median of 72 days. Hepatotoxicity (HT) occurred at a median of 26 days of therapy. Approximately a third of the patients (34%) developed HT during voriconazole therapy. This is higher compared to the rate of 19% reported in randomized clinical trials of voriconazole (19%) . In our cohort, determining the causality of hepatotoxicity in relation to voriconazole was extremely difficult. Almost a half of our patients had other concomitant conditions or medication exposures causing or contributing to HT. Some patients with mild transaminase changes continued voriconazole therapy with their transaminases improving or remaining only mildly elevated.
Since the majority of the allogeneic HSCT patients at MSKCC receive voriconazole prophylaxis, identifying a comparator cohort of patients who did not receive voriconazole and had similar characteristics to our study cohort was not feasible. We therefore compared the group of patients who developed HT on voriconazole to patients who did not to identify risk factors for hepatotoxicity. In multivariate analyses, only acute GVHD grades 2–4 remained a significant risk factor for HT.
The extensive exposure to other potentially hepatotoxic medications posed an additional challenge in the assessment of competing causes for HT. Cyclosporine A, tacrolimus, warfarin, and calcium channel blockers can cause HT and have a known pharmacokinetic interaction with voriconazole. When we compared the utilization of immunosuppressants and other hepatotoxic drugs between patients who developed HT versus those who did not, we found no substantial differences between the two groups (all p>0.1). The proportion of patients with elevated levels of calcineurin inhibitors (cyclosporine and tacrolimus) or mTOR inhibitor (sirolimus) was not significantly different in the HT group. Interestingly, in univariate analysis recipients of T-cell depleted grafts who do not require additional exogenous immunosuppression for GVHD prophylaxis, were less likely to develop hepatotoxicity compared to recipients of unmodified grafts. Lack of exogenous immunosuppression or lower rates of GVHD among the recipient of T-cell depleted grafts may account for this finding.
Voriconazole is extensively metabolized by the liver. Cytochrome CYP2C19, the major enzyme responsible for voriconazole metabolism, exhibits genetic polymorphisms which result in variability of voriconazole metabolism and exposure. Poor metabolizers may experience up to a 4-fold increase in drug exposure, and such phenotype is found in up to 20% of non-Indian Asians but only 3–5% of Caucasians or Blacks . In our population, only 9 (4%) patients were Asian. The association between elevated voriconazole levels with HT is not well defined. An increase in voriconazole trough levels has been associated with HT, and this is certainly a concern in HSCT patients who are already at risk for liver dysfunction [19–23]. Since therapeutic drug monitoring was at the discretion of clinicians in our study, testing bias is plausible. Patients with hepatic function abnormalities or those requiring high doses of voriconazole would be more likely to have levels checked. Nonetheless we did not observe any significant differences in the number of patients with voriconazole trough levels >6 mg/L between the groups.
When we assessed the trend of transaminases with time, there was a significant decrease after withdrawal of therapy in all enzymes except TBil. The elevations in Tbil could be explained by alternative conditions and comorbidities in the majority of patients. Our data strongly suggests that HT was largely reversible after discontinuation of voriconazole. Although there was a temporal association of rise in transaminases with treatment and a fall with discontinuation of voriconazole for many patients with HT, a causal relationship is less clear. When we reviewed the medical records of patients who developed HT and died while on voriconazole therapy, voriconazole therapy was not thought to contribute to death in any of these patients.
In summary, in this large cohort of high risk HSCT patients, 34% of patients experienced mild to moderate elevations in liver enzymes. The abnormalities were reversible after stopping voriconazole with the exception of total bilirubin. Acute GVHD was a significant risk factor for HT. Approximately 50% of the patients had concomitant diagnoses or medications associated with hepatotoxicity. In approximately 50% of the patients, liver function abnormalities, normalized or remained mildly elevated, and did not require discontinuation of voriconazole therapy. Voriconazole therapy, with monitoring, appears reasonably safe in hematopoietic stem cell transplant recipients at high risk for invasive fungal infections.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.