Clozapine is an atypical antipsychotic drug mostly used in patients with treatment-resistant schizophrenia. Over the years, its use has increased in many countries, with the prevalence of clozapine consumption in 2014 ranging from 0.9 to 173.2 per 100,000 persons, depending on the country (Bachmann et al., 2017; Remington et al., 2017). Safety issues associated with clozapine are common and include agranulocytosis (occurring in approximately 1.0% of patients) and neutropenia (occurring in approximately 3.0% of patients) (Rajagopal, 2005). Due to these safety concerns, clozapine is usually reserved for treatment-resistant schizophrenia despite its demonstrated efficacy in managing positive, negative, and overall symptoms and relapse rates in schizophrenia, compared to first-generation antipsychotics and pooled first-/second-generation antipsychotics (Wagner et al., 2021).

Infections have also been reported with clozapine, potentially leading to the elevation of serum levels of the drug. A systematic review identified 40 cases of infections with demonstrated elevated clozapine levels (Clark et al., 2018). In a Chinese cohort of patients undergoing therapeutic drug monitoring during infection and non-infection periods, the median levels of clozapine were significantly higher in the infection period compared to those in the non-infection period (n = 42; p < 0.001) (Zhang et al., 2021). Elevation of clozapine levels with infection may be due to downregulation of metabolizing enzymes such as cytochrome P450 (Clark et al., 2018). Additionally, a retrospective study from the UK found an increased risk of COVID-19 infection in clozapine-treated patients compared to those on other antipsychotic drugs (adjusted hazard ratio = 1.76; 95% CI 1.14–2.72) (Govind et al., 2021).

Clozapine might be more strongly associated with pneumonia than with other infections. A study on a Taiwanese registry with 33,024 inpatients with schizophrenia found that the current use of clozapine was associated with a dose-dependent increase in the risk of pneumonia (adjusted risk ratio = 3.18; 95% CI: 2.62–3.86) (Kuo et al., 2013). A study using VigiBase^®, the WHO global safety report database, supports the prominent role of pneumonia in mortality associated with clozapine adverse drug reactions (De Leon et al., 2020).

Patients with schizophrenia often receive polypharmacotherapy, exposing them to potential drug–drug interactions (DDIs). A cross-sectional observational study in a psychiatric hospital found a high prevalence (88.7%) of potential DDI in this population (Bačar Bole et al., 2023). Clozapine is metabolized by various CYP450 enzymes, including CYP1A2, CYP3A4, CYP2D6, CYP2C9, and CYP2C19 (Chetty and Murray, 2007; Pardiñas et al., 2019). Therefore, the association of clozapine with CYP450 inhibitors could lead to clozapine overdose and increase the risk of infection.

Sex-related differences in clozapine tolerability have also been described in 147 treatment-resistant patients treated with clozapine (Martini et al., 2021). Age onset incidence of schizophrenia also differs by sex (Jauhar et al., 2022). As sex-specific differences in the susceptibility to infections and immune response to infection have been reported (McClelland and Smith, 2011), it is also possible that sex differences in clozapine-associated infections exist.

However, until now, studies about clozapine-associated infections remain scarce and are poorly described. Except for clozapine-induced pneumonia, which has been widely described (De Leon et al., 2020), it is unclear what kind of other infection clozapine exposure might lead to. The mechanism behind these infections also needs to be understood as it could be hypothesized that clozapine-induced infections are caused by clozapine-associated neutropenia. Dose-dependency, patient-associated risk factors, and potential DDIs leading to infections also need to be explored to improve preventive measures. Thus, for a better understanding of clozapine-associated infections, a study was conducted using VigiBase^®. The objectives were to assess if clozapine was associated with an over-reporting of infections, to characterize these infections, to evaluate dose-dependency and sex differences, and to investigate potential DDIs associated with infections. The study also explored if clozapine-associated infections were more associated with neutropenia.

In our study, conducted under real-life conditions using the largest pharmacovigilance database worldwide, we identified a significant association between clozapine and reported infections. This signal was consistent over time (Figure 1). Various types of infections were linked to clozapine use in our study, with 94.3% of them considered serious and 9.8% resulting in death at the time of reporting. Although serious adverse effects tend to be more reported than non-serious effects, our findings imply that people treated with clozapine should be carefully followed in the case of symptoms of infection.

The most frequently reported infections were respiratory and gastrointestinal in nature. Patients and families should be educated about the risk of infections, particularly respiratory and gastrointestinal infections.

Clozapine may be particularly associated with pneumonia compared to other infections. A Taiwanese registry study involving a nationwide cohort of 33,024 inpatients with schizophrenia found that the current use of clozapine (adjusted RR = 3.18; 95% CI: 2.62–3.86, p < 0.001) was linked to a dose-dependent increase in the risk of pneumonia. On the other hand, quetiapine, olanzapine, zotepine, and risperidone showed a lesser extent of association with an increased risk of pneumonia. Unlike clozapine, no clear dose-dependent relationship was observed for these antipsychotics. The risk of pneumonia with clozapine was found to be strongest within the first 30 days of treatment (Kuo et al., 2013).

A retrospective study based on clinical records from the UK also reported an increased risk of COVID-19 infection in clozapine-treated patients compared to those on other antipsychotic medications (adjusted HR = 1.76; 95% CI 1.14–2.72) (Govind et al., 2021). However, a register-based cohort study from the Stockholm region did not find an association between clozapine treatment and severe COVID-19 infection. The adjusted HR for the exposed group compared to the unexposed group was 0.96 (95% CI: 0.54 and 1.70) for inpatient care, 1.69 (0.48 and 5.93) for care in an intensive care unit (ICU), and 0.86 (0.26 and 2.80) for death (Ohlis et al., 2022). This study might have lacked statistical power due to the low number of clozapine-treated patients suffering from COVID-19 during the study period. In another English retrospective 1-year cohort study, the incidence of infection was compared between 64 patients starting clozapine and 120 patients starting paliperidone palmitate long-acting injection (PPLAI). The incidence of infection was greater in clozapine starters than in PPLAI starters (28% vs. 6%; p = 0.001; adjusted odds ratio 5.82 (95% CI = 2.15–15.76). Similar to our findings, infectious episodes in clozapine patients were not statistically related to changes in neutrophil counts. According to the authors’ classification, the most commonly reported infection in the clozapine group was respiratory infection; however, the majority of infections were non-respiratory-related (Mace et al., 2022).

Regarding the risk of gastrointestinal infection, particularly appendicitis, a retrospective study including 465 patients, of whom 65 were on clozapine, showed that the clozapine exposure group exhibited a higher incidence of appendicitis during the observation period than the non-exposure group (863 cases vs. 124 cases per 100,000 person-years) (Kawakita et al., 2022). Additionally, a case series of six patients with perforated appendicitis during clozapine treatment reported a 20-fold increase in appendicitis incidence compared to the general population in male subjects of the same age group (Steinert and Jans, 2021). The authors also suspected a dose-dependency as they observed high clozapine serum levels in three of those patients during the course of appendicitis.

Possible indirect mechanisms of clozapine predisposition to infection, particularly aspiration pneumonia, include sialorrhea and impairment of swallowing function with esophageal dilatation and hypomotility (Abdelmawla and Ahmed, 2009). A systematic review during the first month of clozapine treatment indicated that up to 50% of patients develop fever and flu-like symptoms, seemingly driven by increased cytokines (Røge et al., 2012). A recent study comparing the levels of secondary antibodies of clozapine users with those of the users of antipsychotics other than clozapine (control group) noted lower secondary antibodies among clozapine users. Total serum immunoglobulins [immunoglobulin (Ig)G, IgA, and IgM] and specific immunoglobulin antibodies to Haemophilus influenzae and pneumococcus were decreased (Ponsford et al., 2019). Lower immunoglobulin levels might contribute to the onset of infections.

In our study, neutropenia was less associated with clozapine-associated reports of infections compared to other clozapine-associated reports (p-value < 0.005), indicating that most of these infections were probably not linked to clozapine-associated neutropenia. However, agranulocytosis was more associated with clozapine-associated reports of infections compared to other clozapine-associated reports (p-value < 0.005), which is expected as infection is very common in agranulocytosis patients. Indeed, a retrospective Finn study on 163 patients with clozapine-induced agranulocytosis found that 78.6% of the patients presented with an infection (Lahdelma and Appelberg, 2012). The median age of the patients in this study was 49 years, which is very close to the median age in our study (50 years), but the sex ratio was lower in their study (1.09 vs. 1.57 in our study). However, since agranulocytosis was reported in only 3.33% of the reports of clozapine-associated infections, it is not the main cause of the signal of infections observed in our study.

Although men and women have a similar prevalence of schizophrenia (Li et al., 2016), research indicates that men often experience the onset of severe schizophrenia symptoms at an earlier age compared to women. Specifically, men tend to encounter the peak period of initial pronounced psychotic symptoms between 20 and 24 years of age, whereas women tend to experience these symptoms 5 or more years later (Kahn et al., 2015). This might be the reason why there are important gender differences in clozapine prescription. Indeed, several studies, including ours, have reported that men represented between 63.1% and 78.6% of patients treated with clozapine (Taylor, 2004; Silveira et al., 2015). Consequently, sex is probably not a risk factor for clozapine-induced infections.

The median dose in our study was 250 mg (IQR = 100–400). This aligns with the typical dose found in the clozapine-treated population as Caucasians are usually prescribed 300–600 mg/day to reach the therapeutic range, while in Asian countries, average clozapine doses are lower than 300 mg/day (de Leon et al., 2020). However, it is essential to consider that dose reporting was unavailable in more than 98% of the reports in the database. The results of the tests performed in this study to explore a dose-dependency effect do not allow us to draw a conclusion.

The results of the disproportionality analysis for drug–drug interactions (DDI) between clozapine and CYP450 inhibitors on the reporting of infections are more challenging to interpret. Except for ciclosporin (an inhibitor of CYP3A4), a trend toward a DDI was found with all CYP inhibitors, for which a disproportionality analysis could be computed. However, it is unclear which specific CYP450 enzymes may be involved. The associations observed with certain CYP inhibitors, such as clarithromycin and metronidazole, may be influenced by an indication bias since these drugs are commonly prescribed in the context of infections. VPA, one of the most co-reported drugs in the DDI analysis, has been associated with an increase in oxidative stress. Given the potential for clozapine to induce oxidative stress as well, the combination of VPA and clozapine may potentially favor the onset of certain types of infections (Gai et al., 2020; Salimi et al., 2022).

Some authors suspect that the association of clozapine with proton pump inhibitors (PPI) might increase the formation of reactive metabolites and contribute to the increase in hematological adverse drug reactions (ADRs) (Wiciński et al., 2017). However, the oxidative stress hypothesis is inconclusive with PPIs as these drugs usually alleviate oxidative stress (Gao et al., 2019; Gandhi et al., 2021). PPIs have been associated with respiratory infections in different retrospective studies, suggesting a potential pharmacodynamic interaction (Ho et al., 2017; Wang et al., 2019). In a retrospective chart review, aiming to explore the potential effect of polypharmacy on the hematologic profiles of clozapine patients, 24 out of 26 (96%) of the subset of patients who were prescribed a PPI or ranitidine concomitantly with clozapine experienced hematological ADRs (Shuman et al., 2014). A case of infection during PPI treatment with elevated plasma clozapine levels was reported in a 51-year-old woman and could potentially be linked to the switch from omeprazole to esomeprazole (Wagner et al., 2011). The authors speculated that this might be due to the removal of induction of clozapine metabolism by omeprazole. However, the delay was not in favor of this hypothesis. Omeprazole and lansoprazole, in addition to being CYP2C19 inhibitors, are also CYP1A2 inducers. In a case series involving two patients, the prescription of omeprazole was associated with a reduction in clozapine plasma concentrations of 41.9% and 44.7% (Frick et al., 2003). Another retrospective study in 13 psychiatric patients found that the switch from omeprazole to pantoprazole led to an increase in clozapine levels in non-smoking patients and to a decrease in clozapine levels in smoking patients. This was probably caused by the discontinuation of enzyme induction in the cytochrome P450 enzyme 1A2 by omeprazole in non-smokers, whereas CYP1A2 remained induced in smokers (Mookhoek and Loonen, 2004). As the prevalence of cigarette smoking is high in schizophrenia patients (around 70%–80%) and probably even higher in treatment-resistant schizophrenia patients (Ding and Hu, 2021), we believe that the DDI observed in our study between PPIs and clozapine is probably more linked to the CYP2C19 inhibition by PPIs.

Regardless of the cause of infection, several reports showed that infection leads to an increase in the toxicity levels of clozapine and its metabolites in the serum (Darling and Huthwaite, 2011; Espnes et al., 2012; Zhang et al., 2021; Chengappa et al., 2022). This is likely to be mediated by cytokine suppression of cytochrome P450 1A2 (CYP1A2), the main hepatic microsomal system involved in clozapine metabolism, which is also involved in the metabolism of several antibiotics commonly used to treat infections. This further enhances the potential for clozapine toxicity (Røge et al., 2012). However, despite sex being a factor of cytochrome P450 expression (Zanger and Schwab, 2013), we did not find any sex-related significant differences in clozapine-associated infections.

This study has revealed a significant safety signal concerning the association between clozapine and reported infections. Respiratory infections, as well as gastrointestinal infections, including appendicitis, were the most commonly reported infections. The co-administration of clozapine with valproic acid (VPA) or proton pump inhibitors (PPIs) may potentially increase the risk of infection. However, as this study was based on pharmacovigilance data, a definitive causal relationship between clozapine exposure and the occurrence of infections cannot be established with certainty. Nevertheless, in clinical practice, psychiatrists should remain vigilant for signs of infections when prescribing clozapine and ensure that mandatory vaccines, including pneumococcal vaccination, have been administered. Conducting a study to evaluate the relevance of therapeutic drug monitoring of clozapine when the patient’s treatment regimen is altered (especially with VPA or PPIs) or when an infection occurs would be of considerable interest.