Metamizole induces voriconazole metabolism and results in subtherapeutic voriconazole concentrations

Abstract

Aims

Voriconazole is extensively metabolized via cytochrome P450 (CYP) enzymes, predominantly CYP2C19 and CYP3A4. Drugs influencing the activity or expression of CYP enzymes can cause clinically relevant changes in the metabolism and voriconazole exposure. Metamizole is known to induce CYP3A4 and CYP2C19. This study aimed to investigate the pharmacokinetic drug–drug interaction between metamizole and voriconazole.

Methods

In this single‐centre retrospective observational cohort study, we compared voriconazole serum trough concentrations before, during and after metamizole treatment.

Results

In the 9 included patients, the median voriconazole trough concentration decreased by 71% during metamizole treatment (P = .028) compared to before start of metamizole. The concentration/dose ratio similarly decreased by 81% during metamizole treatment (P = .018). Additionally, the metabolic ratio (voriconazole‐n‐oxide/voriconazole) increased from 0.9 to 2.4 (P = .028) during metamizole treatment. Subtherapeutic voriconazole trough concentrations were more frequent when combined with metamizole (before 14%, during 70%, after 17%).

Conclusions

Metamizole increases voriconazole metabolism and decreases voriconazole trough concentrations, probably through a CYP3A4 and CYP2C19 inducing effect. It is recommended to avoid concurrent use of metamizole and voriconazole or to closely monitor voriconazole trough concentrations during metamizole treatment and up to 2 weeks after discontinuation of metamizole.

What is already known about this subject

• Metamizole is suspected to induce several cytochrome P450 enzymes including CYP3A4 and CYP2C19.

• Voriconazole is metabolized by CYP3A4 and CYP2C19.

• The drug–drug interaction between metamizole and voriconazole, however, has not been described.

What this study adds

• Metamizole concurrent treatment increases the metabolic ratio of voriconazole (voriconazole‐n‐oxide/voriconazole).

• Metamizole concurrent treatment leads to reduced and even subtherapeutic concentrations of voriconazole.

• Concurrent treatment with metamizole should be avoided in patients treated with voriconazole.

1. INTRODUCTION

Voriconazole is an antifungal drug used as a first‐line treatment for life‐threatening invasive aspergillosis. ¹ Voriconazole is extensively metabolized via cytochrome P450 (CYP) enzymes, predominantly via CYP2C19 and to a lesser extent via CYP2C9 and CYP3A4. ² One of the main voriconazole metabolites through CYP2C19 and CYP3A4 is voriconazole‐n‐oxide. Voriconazole shows autoinhibition via CYP3A4. ² Voriconazole shows nonlinear pharmacokinetics, is bound to plasma proteins for 58% and is renally excreted as an unchanged drug for <2%. ³ Voriconazole has an exposure–response relationship for both efficacy and toxicity. ⁴ Therefore, it is important to attain an adequate voriconazole exposure within the target range. A voriconazole trough concentration >1–1.5 mg/L is associated with improved efficacy and is therefore considered the lower limit of the target range. With trough concentrations above 5–6 mg/L, there is an increased risk of neurotoxicity; this is considered as the upper limit of voriconazole trough concentration. ⁵ , ⁶ , ⁷ , ⁸ , ⁹ There is high intrapatient and interpatient variability in voriconazole pharmacokinetics. Intrapatient variability in voriconazole pharmacokinetics is mostly caused by differences in C‐reactive protein (CRP), since high CRP is associated with reduced voriconazole clearance and higher voriconazole concentrations. ¹⁰ , ¹¹ , ¹² This variability, combined with the narrow therapeutic window, provides a rationale for routine therapeutic drug monitoring (TDM) in patients treated with voriconazole. ⁷ , ⁸ , ⁹

Metamizole is increasingly used in pain management and is generally considered safe. ¹³ In recent years, there has been an increase in the number of publications on the various drug–drug interactions of metamizole. In 2007 it was first shown that metamizole induces CYP3A4 and CYP2B6. ¹⁴ Since then several studies have been published showing reduced drug exposure when metamizole is combined with drugs that are metabolized via CYP3A4 or CYP2B6. ¹⁵ , ¹⁶ , ¹⁷ Additionally, it was demonstrated that metamizole also induces CYP2C19 and CYP2C9 and that this induction is due to activation of the nuclear constitutive androstane receptor (CAR). ¹⁸ Furthermore, in the interaction study between metamizole and midazolam, it was demonstrated that there is a decrease in midazolam concentration as early as 2 days after starting metamizole, with a maximum decrease after 5 days of use of metamizole. ¹⁹ Considering the metabolism of voriconazole through CYP3A4 and CYP2C19, it can be hypothesized that there also is an interaction between metamizole and voriconazole. Voriconazole exposure is known to be reduced by drug–drug interactions with enzyme‐inducing drugs such as rifampicin, ²⁰ phenytoin, ²¹ , ²² and carbamazepine, ²² but also dexamethasone ²³ and flucloxacilline ²⁴ have been shown to reduce voriconazole exposure. Interestingly, since tacrolimus exposure can be reduced by flucloxacillin ²⁵ and metamizole, ¹⁵ this further supports the hypothesis for an interaction between metamizole and voriconazole since the underlying mechanism for flucloxacillin and metamizole reducing the concentration of tacrolimus and voriconazole might overlap. In a study exploring pharmacogenetic influences on voriconazole exposure in children, it has been discussed that metamizole may influence voriconazole exposure in children as it was seen to influence voriconazole trough levels in 1 patient in the study population. ²⁶ A recent case report showed that metamizole was responsible for subtherapeutic voriconazole concentrations in 2 patients. ²⁷ In the present study, we investigated in a patient series whether the combination of metamizole with voriconazole causes a change in voriconazole exposure in patients who received this combination.

2. METHODS

2.1. Study design and study population

A single‐centre retrospective cohort study was performed at the University Medical Center Groningen, The Netherlands. Every patient of 18 years and older treated with the combination of voriconazole and metamizole was eligible for inclusion in this study if at least 1 voriconazole concentration was collected during this drug combination and before, or after the drug combination, in the period between January 2018 and July 2024.

This study was approved by the local ethics committee of University Medical Center Groningen (METc 2023/531).

2.2. Data collection

Eligible patients were identified using a query in the hospital's electronic health records. Patients were selected with at least 1 available voriconazole concentration during at least 24 h of use of concomitant use of metamizole and voriconazole, and either a voriconazole concentration before the start of metamizole, or after discontinuation of metamizole. The primary endpoint was the difference in median voriconazole trough concentration before, during and after discontinuation. There were several secondary endpoints. Since voriconazole dose is adjusted based on trough concentrations during routine TDM, not all patients received the same dose. To determine whether a difference found in the primary endpoint is not due to a difference in dose, the difference in concentration/dose ratio was used as a secondary endpoint. For mechanistical insights in this interaction, the difference in metabolic ratio was determined. The metabolic ratio is the concentration of voriconazole‐n‐oxide divided by the concentration of voriconazole. Lastly, subtherapeutic voriconazole trough concentrations during metamizole treatment were assessed (voriconazole trough concentrations below 1.5 mg/L).

For each patient, the following data were collected from the electronic health records: sex, age, weight, height, blood concentrations (serum or plasma) of voriconazole and voriconazole‐n‐oxide and corresponding doses of voriconazole with the use of metamizole and dose of metamizole, CRP, CYP2C19 and CYP3A4 interacting comedication, and underlying disease.

2.3. Voriconazole concentrations

Voriconazole trough concentrations were used to evaluate the effect of metamizole. Since metamizole is suspected to affect the metabolism and clearance of voriconazole from the blood, voriconazole trough concentrations were considered to be influenced most and were therefore selected as the relevant pharmacokinetic parameter.

Voriconazole and voriconazole‐n‐oxide serum or plasma concentrations were measured as previously described by validated liquid chromatography coupled with tandem mass spectrometry analysis as part of routine care. ²⁸ Samples were stored at −20 °C within 3 h after collection and were analysed within 3 days. The collected voriconazole concentrations were only evaluable as actual trough concentrations if they were collected 12 ± 2 h after the previously administered dose (in twice daily dosing). In case voriconazole was dosed 3 times daily, trough concentrations are defined as samples drawn 8 ± 1 h postdose. If samples were drawn outside the reference time window, the trough concentration for these concentrations was calculated using MwPharm++ (version 2.4, Mediware, Prague, Czech Republic) pharmacokinetic software equipped with the pharmacokinetic model of Purkins et al. using only the concentration to be extrapolated. ²⁹ Concentrations measured during the absorption phase (up until 2 h after administration ³ ) were excluded from this analysis. For patients with more than 1 voriconazole concentration before the start of metamizole, the most recent measurement was selected if this was a representative concentration (comparable CRP, no drug–drug interactions). For the voriconazole concentration during treatment with metamizole, metamizole had to be used for at least 24 h, or metamizole had to be stopped for a maximum of 2 days. In the case of multiple voriconazole concentrations during metamizole concurrent use, the last result was selected. For the voriconazole concentration after discontinuation of metamizole, the blood sample had to be drawn at least 5 days after discontinuation of metamizole. Five days was selected based on the time course of the interaction of metamizole and midazolam. ¹⁹ In the case of multiple voriconazole concentrations, the last result after stopping metamizole, but within 1 year was selected.

2.4. Statistical analysis

Statistical analysis was performed using IBM SPSS Statistics version 28 (IBM Corp, Armonk, NY, USA).

Because of the small sample size, voriconazole trough concentrations have been compared nonparametrically using the Wilcoxon signed rank test for paired data, this is an intraindividual comparison. The voriconazole trough concentrations were compared before start of metamizole vs. during use of metamizole, and during use of metamizole was compared to after discontinuation of metamizole. For all statistical tests, a 2‐tailed significance level of .05 was applied. Numeric data were expressed median with interquartile range (IQR). Categorical data were expressed as percentages.

3. RESULTS

3.1. Baseline characteristics

Nine patients were included with trough levels of voriconazole during the use of metamizole and also a trough level before the start of metamizole or after the use of metamizole. One patient had 2 separate occurrences of the combination of voriconazole and metamizole including trough levels of voriconazole before and after the use of metamizole, making the total number of cases 10. Patient characteristics are shown in Table 1. No patient had relevant interacting comedication or comorbidities and pharmacogenetic status was unknown for all patients. All patients used oral voriconazole and 4 out of 10 patients used metamizole intravenously.

TABLE 1.

Patient characteristics.

Characteristics	Paired data (median [range])
Age (years)	53 (19–65)
Weight (kg)	79.8 (54–130)
Sex
Male	5
Female	4
Voriconazole indication
Possible/probable aspergillosis	8
Candida	1
Metamizole indication
Severe pain and fever	3
Postoperative pain	6
Switch of antifungal treatment	2
Metamizole daily dose	3000 (2000–4000)
Days of use of metamizole before voriconazole trough concentration during	6.7 (2–24)
Number of days metamizole has been discontinued before voriconazole trough concentration after	9.5 (5–44)

3.2. Voriconazole concentrations before, during and after metamizole

In Figure 1 all paired samples are shown. Of all 22 trough concentrations, 12 were calculated using model informed precision dosing. The median voriconazole trough concentration before the start of metamizole was 2.8 (IQR: 1.7–6.1, n = 7), during metamizole treatment 0.8 (IQR: 0.2–2.7, n = 10) and after metamizole 2.7 (IQR: 1.6–3.9, n = 6; before vs. during, P = .028; during vs. after: P = .075). See also Figure 2A. At the point of voriconazole sampling during use of metamizole, voriconazole was used for a median of 8.5 days (IQR: 4–11) and was in steady‐state, and metamizole was used for a median of 4.5 days (IQR: 3.0–6.75, n = 10). The trough concentrations measured after discontinuation of metamizole were a median of 9.5 days after discontinuation (IQR: 5.75–21.5, n = 6). On a descriptive level, the median voriconazole trough concentration decreased by 71% in the group during metamizole treatment compared to the group before the start of metamizole. After discontinuation of metamizole, the median voriconazole trough concentration increased by 238%. There was no difference in CRP between the before, during and after use of metamizole (Figure S1). In the group during use of metamizole 1 out of 10 patients had CRP > 150 g/L; in the group after discontinuation of metamizole 1 out of 6 patients had CRP > 150 g/L; and at the time of all the other voriconazole measurements patients had CRP < 150 g/L.

FIGURE 1.

Voriconazole trough concentrations before, during and after use of metamizole in individual patients. Each line is an individual case, the dashed line is the subtherapeutic threshold at 1.5 mg/L.

FIGURE 2.

Boxplot of voriconazole trough concentrations before, during and after use of metamizole (A). The dashed line indicates the subtherapeutic threshold of 1.5 mg/L. Boxplot of concentration/dose‐ratio of voriconazole before, during and after use of metamizole (B). The voriconazole trough concentration (mg/L) was divided by the daily dose of voriconazole administered (g).

3.3. Voriconazole concentration dose ratio

By correction for the dose of voriconazole, the concentration (mg/L)/dose (g/day) ratio (cd‐ratio) was calculated. In Figure 2B all cd‐ratios are shown. The median cd‐ratio (n = 7) before the start of metamizole was 7.0 (IQR: 2.8–11.3), during the use of metamizole 1.3 (IQR 0.3–5.8) and after discontinuation 4.5 (IQR: 2.2–7.5, n = 6; before vs. during, P = .018; during vs. after, P = .249). On a descriptive level, the median voriconazole cd‐ratio decreased by 81% in the group during metamizole treatment compared to the group before the start of metamizole.

3.4. Voriconazole subtherapeutic concentrations

The frequency of patients with subtherapeutic trough concentrations was compared before, during, and after metamizole. Subtherapeutic concentrations below 1.5 mg/L occurred in 14% before the start of metamizole, 70% during the use of metamizole and 17% after discontinuation of metamizole. Of the 7 patients who had voriconazole concentrations <1.5 mg/L during concomitant use of metamizole, 5 patients had a voriconazole concentration <1.0 mg/L.

3.5. Voriconazole metabolic ratio

Furthermore, the voriconazole metabolic ratios were compared. For 1 patient it was not possible to calculate a metabolic ratio due to a missing voriconazole‐n‐oxide concentration during the use of metamizole. The median metabolic ratio before metamizole was 0.9 (IQR: 0.4–2.9) and during the use of metamizole 2.4 (IQR: 1.0–14.0; before vs. during, P = .028). The metabolic ratio after discontinuation of metamizole was 1.3 (IQR: 0.6–2.2). This does not differ significantly from the metabolic ratio during the use of metamizole (P = .116).

4. DISCUSSION

In this retrospective cohort study, we showed that treatment with metamizole was associated with lower voriconazole trough concentrations, with a median decrease in voriconazole trough concentrations of 71% resulting in 70% of patients with a subtherapeutic voriconazole concentration. Additionally, the difference in cd‐ratio between patients with and without metamizole shows that the different voriconazole trough concentrations are not due to differences in voriconazole doses. As for the metabolic ratio, voriconazole is metabolized through CYP3A4 and CYP2C19 to voriconazole‐n‐oxide. The difference in metabolic ratio shows that the difference in voriconazole trough concentration is likely to be due to increased voriconazole metabolism by metamizole induction of CYP3A4 and CYP2C19. The metamizole‐induced subtherapeutic concentrations of voriconazole can be of clinical relevance in the management of antifungal treatment, as unintentional and unnoticed low voriconazole concentrations represent a risk of undertreatment. ⁴ , ⁵ , ³⁰

To our knowledge, this is the first study to quantify a clinically relevant interaction between metamizole and voriconazole in a series of patients who received this combination. This study confirms the findings described in previous cases. ²⁶ , ²⁷ Other studies have shown that metamizole is an inducer of CYP3A4. Since midazolam is considered a model substrate for CYP3A4, it is relevant to compare our findings with the effect of metamizole on midazolam exposure. Midazolam is mainly metabolized through CYP3A4 and voriconazole by CYP2C19 and by CYP3A4. One study showed a metamizole‐induced decrease in midazolam exposure of 81%. ¹⁹ Another study found a 68% lower midazolam exposure when combined with metamizole. ¹⁸

The results are similar to studies with other drugs metabolized by CYP3A4 and CYP2C19. One study showed a reduced omeprazole exposure of 66%, omeprazole is metabolized by CYP2C19. ¹⁸ Furthermore, sertraline is metabolized by CYP2C19 and CYP3A4 and a retrospective study showed that concurrent treatment with metamizole was associated with 67% lower median sertraline plasma concentrations. ¹⁶ Additionally, this study shows the relevance of this interaction because 40% of the metamizole‐treated patients had sertraline concentrations below the therapeutic reference range, compared to 0% in the control group. This corresponds with the difference in subtherapeutic concentrations seen in our study and a retrospective study investigating the combination of quetiapine and metamizole. Quetiapine is mainly metabolized by CYP3A4 and a 51% lower median quetiapine plasma concentration was found when combined with metamizole. ¹⁷

Moreover, our results are in line with the drug–drug interaction between voriconazole and flucloxacillin. Flucloxacillin is suspected to induce CYP3A4 and a combination of voriconazole and flucloxacillin resulted in subtherapeutic voriconazole trough concentrations for 7 out of 11 patients. ²⁴ Another study showed that 69% of patients with concurrent flucloxacillin had subtherapeutic voriconazole concentrations. ³¹ Based on these studies, it is generally considered the best option to avoid the combination of voriconazole and flucloxacillin.

The underlying mechanism of induction by metamizole is not completely known. It has been found that metamizole does not bind to the Pregnane X Receptor or CAR directly and it is hypothesized that metamizole or its metabolites activate CAR differently. ¹⁴ , ¹⁸ Activation of CAR upregulates messenger RNA of several cytochrome P450 enzymes. ³² However, induction of CYP enzymes appears rather quickly with metamizole, ¹⁹ and induction through upregulation of mRNA generally takes time. Therefore, this might not be the sole explanation for the induction by metamizole. Our study shows a difference in the metabolic ratio of voriconazole before the start of metamizole and during the use of metamizole. This confirms that voriconazole metabolism to voriconazole‐n‐oxide is more active during the use of metamizole than before the start of metamizole. This supports the hypothesis of metamizole‐induced metabolism of voriconazole.

In this study, we showed no significant difference between the voriconazole trough concentration during and after metamizole. This can be explained by a prolonged induction after discontinuation of metamizole. The samples for the trough concentrations after discontinuation of metamizole in our study are drawn at a median of 9.5 days after discontinuation. In CYP induction interactions it can take up to 2–4 weeks for the metabolism by the CYP enzymes to return to normal. ³³ The finding that the voriconazole cd‐ratio in the after metamizole period was numerically lower than the cd‐ratio in the before period suggests that the interaction lasts longer than the anticipated 5 days. This is also indicated by the results of the interaction study of metamizole and midazolam, where it is described that midazolam exposure still was half of the exposure 5 days after discontinuation compared to before the start of metamizole. ¹⁹

This study has some limitations. Firstly, for the primary endpoint, 10 different cases have been included in 9 different patients. Although this is a small sample size, we observe a significant difference in voriconazole trough concentrations. However, the size of the difference might be different in the total population due to this small group. Since the effect of metamizole on voriconazole exposure was convincing, it can be assumed that this is a clinically relevant drug–drug interaction, also in larger populations. However, a larger study is needed to confirm the size of the difference. Furthermore, we only collected retrospective data from the electronic health records. This is not a controlled setting and data are missing. However, even with this method, we found a statistically significant difference in voriconazole trough concentrations before use of metamizole and during use of metamizole. Additionally, our study does not fully elucidate the time course of this interaction. For patients with multiple voriconazole concentrations available within the same period, the most representative concentration was used for analysis as described in the Methods section. This strategy enables a straight analysis of the interaction and an analysis of the influence of time is not possible. Future research could demonstrate what the effect is of a single dose of metamizole and for how long the interaction is relevant.

In clinical practice, switching to another drug for pain management is recommended to avoid concomitant therapy of voriconazole and metamizole. If the antifungal drug is to be switched, it is important to take into account that metamizole may influence the exposure of other antifungals. Isavuconazole is a prodrug that is metabolized by CYP3A4. ³⁴ This means that isavuconazole is suspect to also have an interaction with metamizole and TDM of isavuconazole is recommended in this setting. However, posaconazole is barely metabolized by CYP enzymes ³⁵ and therefore is less likely to have an interaction with metamizole.

In conclusion, we report changes in voriconazole trough concentrations during metamizole treatment leading to metamizole‐induced subtherapeutic voriconazole exposure. We warn physicians and pharmacists to be aware of this drug–drug interaction and recommend avoiding the concurrent use of metamizole and voriconazole. If it is not possible to avoid the combination, close monitoring of voriconazole trough concentrations during metamizole treatment and up to at least 2 weeks after discontinuation of metamizole is strongly recommended.

AUTHOR CONTRIBUTION

Conception and design: all authors. Acquisition of data: S.B. Analysis and interpretation of data: all authors. Drafting of manuscript: S.B. Critical revision and final approval of manuscript: all authors.

CONFLICT OF INTEREST STATEMENT

All authors declare no financial or nonfinancial interest in the subject matter or materials discussed in this manuscript.

Supporting information

FIGURE S1 Boxplot of C‐reactive protein (CRP) before, during and after use of metamizole.

Baan SD, Touw DJ, Lub‐de Hooge MN, Oude Munnink TH. Metamizole induces voriconazole metabolism and results in subtherapeutic voriconazole concentrations. Br J Clin Pharmacol. 2025;91(9):2598‐2604. doi: 10.1002/bcp.70079

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.