Voriconazole hepatotoxicity as a result of steroid withdrawal in a patient with allergic bronchopulmonary aspergillosis

Voriconazole is a triazole antifungal agent usually used for the treatment of allergic bronchopulmonary aspergillosis (ABPA) in combination with glucocorticoids 1, 2. This drug is extensively metabolized to inactive metabolites by hepatic cytochrome P450. 3, 4. Several studies have demonstrated that trough concentrations >1 μg ml⁻¹ and < 6 μg ml⁻¹ are associated with treatment success and minimized toxicity 5, 6, 7, 8. In this context, therapeutic drug monitoring of voriconazole is of particular interest. It has been demonstrated that some commonly used drugs can affect the plasma concentration of voriconazole. However, there is limited data regarding voriconazole interaction with glucocorticoids. Due to the wide administration of both voriconazole and glucocorticoids in patients with ABPA, it is necessary to understand the onset and severity of this interaction. Here, we describe a case report of voriconazole hepatotoxicity as a result of steroid withdrawal in a patient with allergic bronchopulmonary aspergillosis (ABPA).

An 82‐year‐old woman diagnosed with ABPA started long‐term treatment with 0.5 mg kg⁻¹ prednisone for 28 days and voriconazole (200 mg every 12 h for 16 weeks). On day +14, the patient progressed satisfactorily and the plasma concentration of voriconazole was in therapeutic range (2.58 μg ml⁻¹). On day +28, a clinical decision was made to remove the prednisone but voriconazole was maintained. One month after prednisone discontinuation, the patient presented photophobia, proximal myalgia and asthenia. In the physical exam, the patient presented icteric coloration. A blood test showed elevated bilirubin (6.8 mg dl⁻¹, direct bilirubin 5.9 mg dl⁻¹) and high liver enzymes: AST 869 UI/L, ALT 1085 UI/L, GGT 1283 UI/L, AP 291 UI/L and moderate hyponatremia, 129 mmol l⁻¹. Plasma concentration of voriconazole was determined resulting in 14.47 μg ml⁻¹. All medications received by the patient were checked to find probable interactions, which could explain the high concentration. The only change in the medication between the first determination of voriconazole and the second one had been the suspension of prednisone. DIPS Algorithm (Drug Interaction Probability Scale) 9 was used to evaluate the causation of a potential drug interaction between prednisone and voriconazole (Table 1). The score was 7 points, suggesting this drug interaction as the probable cause of the patient's hepatotoxicity. In addition, due to a transaminases rise of 19 times above the upper limit of normal with a hepatocellular pattern of liver injury (R factor = 8.5) 10, voriconazole met the criteria necessary to be qualified as DILI (Drug‐Induced Liver Injury). Thus, treatment with voriconazole was removed and progressive decreases in plasma concentrations were observed. Photophobia and muscular weakness disappeared few days after voriconazole discontinuation. However, the recovery of liver parameters was very slow, taking 2 months until complete normalization.

Table 1.

DIPS Algorithm (Drug Interaction Probability Scale); the interaction is considered highly probable if the score is 8 or higher, probable if 5 to 8, possible if 2 to 4, and doubtful if 2 or less

Question	Yes	No	Do not know
Are there previous credible reports of this interaction in humans?	+1	−1	0
Is the observed interaction consistent with the known interactive properties of precipitant drug?	+1	−1	0
Is the observed interaction consistent with the known interactive properties of object drug?	+1	−1	0
Is the event consistent with the known or reasonable time course of the interaction (onset and/or offset)?	+1	−1	0
Did the interaction remit upon dechallenge of the precipitant drug with no change in the object drug?	+1	−2	0
Did the interaction reappear when the precipitant drug was readministered in the presence of continued use of object drug?	+2	−1	0
Are there reasonable alternative causes for the event?	−1	+1	0
Was the object drug detected in the blood or other fluids in concentrations consistent with the proposed interaction?	+1	0	0
Was the drug interaction confirmed by any objective evidence consistent with the effects on the object drug (other than drug concentrations from previous question)? (serum concentration of ALT, AST)	+1	0	0
Was the interaction greater when the precipitant drug dose was increased or less when the precipitant drug dose was decreased?	+1	−1	0
Total score	7

Numbers in bold are the puntuation in the patient

Voriconazole has the potential to interact with multiple medications. The main pathway for these interactions is the hepatic enzyme induction or inhibition involving CYP2C19, CYP3A4 and/or CYP2C9 4. An in vitro study has identified glucocorticoid receptor binding sites in the promoter region of the CYP2C19 gene and demonstrated up‐regulation of CYP2C19, supporting an inductive effect of glucocorticoids on CYP2C19 11, 12. There are limited data regarding the interaction between voriconazole and corticosteroids in clinical practice 13, 14, 15, 16, 17. Dolton et al. found that concomitant dexamethasone use was associated with a 3.75‐fold decrease in voriconazole concentrations 13. Similarly, Cojutti et al. found that coadministration of methylprednisolone and dexamethasone was associated with a 1.65 and 1.93‐fold decrease in voriconazole concentration, respectively 14. Finally, Wallace et al. observed a two‐fold increase in the plasma concentration of voriconazole when the dexamethasone dose was reduced gradually in one patient with a fungal brain abscess treated with both drugs 16.

Our patient presented plasma concentrations of voriconazole in a therapeutic range during treatment with prednisone. Upon discontinuation of prednisone, an increase in voriconazole serum concentrations was observed along with toxicity, suggesting that prednisone had been inducing metabolism of voriconazole. On the other hand, it should be noted that genetic polymorphisms in CYP2C19, as a primary elimination pathway of voriconazole, is another factor influencing trough plasma concentration. Poor metabolizers have an increased risk of achieving higher plasma concentration compared with other genotypes 18. We observed that our patient was an extensive metabolizer (CYP2C19*1*1); therefore, the high concentrations of voriconazole were not attributed to CYP2C19 polymorphism. This is the first clinical case describing probable voriconazole hepatotoxicity as a result of steroid withdrawal. Nevertheless, we need further studies investigating the effect of corticosteroid dosing on voriconazole metabolism as well as the effect of different glucocorticoids. In addition, the effect on voriconazole concentration caused by prednisone in the present case was higher to the observed in previous studies (a 5.60‐fold increase in voriconazole concentration when prednisone was discontinued); therefore, other causes may have contributed to decrease the elimination of the drug from the blood such as the inflammatory state of the patient 19, 20, 21.

Considering that the concomitant treatment of glucocorticoids and voriconazole is relatively frequent, we must take into account this possible pharmacokinetic interaction and its possible clinical repercussion in the future. A proper knowledge of this interaction will allow healthcare professionals to improve the treatment of patients. Future pharmacokinetic, pharmacodynamic and pharmacogenetic studies are needed to better understand these complex interactions. In the meantime, routine therapeutic drug monitoring of voriconazole is a useful tool that can help us to optimize the management of these patients. We must pay special attention to changes in the treatments, such as the withdrawal of glucocorticoids, and consider a voriconazole dose adjustment if necessary in order to avoid toxicity.

Considering the clinical relevance of the adverse reaction it was reported to the Spanish Pharmacovigilance System, and the sample for the genetic determination was collected after obtaining the informed consent of the patient. The study of pharmacokinetic‐pharmacogenetic of voriconazole was approved by the local ethics committee.

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY 22, and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 23.

Competing Interests

There are no competing interests to declare.

Blanco‐Dorado, S. , Marques Afonso, A. T. , Bandín‐Vilar, E. J. , Novo‐Veleiro, I. , Ferrón Vidán, F. , Latorre‐Pellicer, A. , Maroñas Amigo, O. , Fernández‐Ferreiro, A. , and Lamas, M. J. (2019) Voriconazole hepatotoxicity as a result of steroid withdrawal in a patient with allergic bronchopulmonary aspergillosis. Br J Clin Pharmacol, 85: 460–462. 10.1111/bcp.13819.