Treatment intensity

Studies reported higher doses and rapid titration among SUD and SCD cases, consistent with a prior systematic review linking higher doses with myocarditis and cardiomyopathy. ^{Reference Vickers, Ramineni, Malacova, Eriksson, McMahon and Moudgil42} Whereas British National Formulary (BNF) guidelines advise slow titration in steps of 50–100 mg once or twice weekly to a maximum daily dose of 200–450 mg over 14–21 days, ⁴³ contemporary evidence supports a more individualised approach. The Maudsley Prescribing Guidelines recommend titration tailored to gender and smoking status; slower titration in female non-smokers compared with male smokers, reflecting metabolic differences; and suggest CYP1A2 pharmacogenetic testing to identify patients requiring lower doses. ^{Reference Taylor, Barnes and Young44} Observational data further link aggressive early dose escalation with myocarditis: an Australian case–control study reported a 26% increase in myocarditis risk for each additional 250 mg added in the first 9 days of treatment. ^{Reference Ronaldson, Fitzgerald, Taylor, Topliss, Wolfe and McNeil45} Multiple case reports in this review described accelerated titration rates preceding SCD. ^{Reference Borentain, Millet and Olie19,Reference Chopra and de Leon21,Reference Hagg, Spigset, Bate and Soderstrom24} Collectively, these findings suggest the need for personalised, biomarker-informed, cautious clozapine initiation. What needs to be further established is whether enhanced monitoring is required for those on higher doses.

Concomitant drug interactions

Our finding of valproate co-prescription in two case reports of SCD and myocarditis-related SCD ^{Reference Borentain, Millet and Olie19,Reference Chopra and de Leon21} aligns with meta-analyses identifying increased risk of clozapine-related myocarditis and cardiomyopathy (pooled odds ratio 3.58, 95% CI 1.81–7.06). ^{Reference Vickers, Ramineni, Malacova, Eriksson, McMahon and Moudgil42} One pharmacokinetic study suggested that valproate inhibits clozapine metabolism, with a mean plasma clozapine increase of 22.5%. ^{Reference Flanagan, Obee, Kim and Every-Palmer46} A cohort study found that concomitant valproate during clozapine initiation increased inflammatory adverse events, including myocarditis (odds ratio 3.25, 95% CI 1.03–9.74). ^{Reference Kikuchi, Komatsu, Otsuka, Ito, Yada and Tanifuji47} It is perceived that clozapine-related myocarditis occurs at least partly as a result of an inflammatory reaction, ^{Reference Daniel, Rajaree, Rudy, Jafor, Prasad and Avanthika48} and co-factors such as valproate that raise clozapine exposure or provoke inflammation may increase this risk.

By contrast, our finding of benzodiazepine co-prescription in two case reports of SCD ^{Reference Borentain, Millet and Olie19,Reference Ramaswamy, Boust, Sattar, Bhatia and Petty33} has sparse direct evidence in clozapine cohorts. A retrospective New York cohort of 152 patients co-prescribed clozapine and benzodiazepine reported no cardiac arrests or sudden deaths. ^{Reference Bitter, Demler and Opler49} However, broader cardiology literature describes increased SCD incidence following benzodiazepine initiation (adjusted hazard ratio 2.01, 95% CI 1.42, 2.83), ^{Reference Mo, Wang, Li, Li, Jin and Bai50} suggested as reflecting respiratory depression and secondary arrhythmias. ^{Reference Bounds and Patel51,Reference Kang, Galuska and Ghassemzadeh52}

Sertraline, a selective serotonin reuptake inhibitor (SSRI), was identified in a single case report of SCD, ^{Reference Hoehns, Fouts, Kelly and Tu26} although pooled meta-analysis has not shown an increased SCD risk with SSRIs overall (odds ratio 0.36, 95% CI 0.038–1.600). ^{Reference Prasitlumkum, Cheungpasitporn, Tokavanich, Ding, Kewcharoen and Thongprayoon53} One large pharmacokinetic study reported a negative correlation between sertraline dose and clozapine concentration (rs = −0.535, P = 0.048). ^{Reference Kuzin, Schoretsanitis, Haen, Ridders, Hiemke and Grunder54} However, some studies have shown that SSRIs can have cardiovascular side-effects including fatal arrhythmias, often through QT-interval prolongation. ^{Reference Grace, Medina-Inojosa, Thomas, Krause, Vickers-Douglas and Palmer55–Reference Aronow and Shamliyan57}

Lifestyle and behavioural factors

We identified metabolic comorbidities, including obesity and diabetes mellitus, as potential contributors to SCD and SUD. ^{Reference Es, Melez, Ozgun, Melez, Kertmen and Cavlak22,Reference Khan, Ashraf, Baker, Al-Omary, Savage and Ekmejian29,Reference Modai, Hirschmann, Rava, Kurs, Barak and Lichtenberg32,Reference Ronaldson, Fitzgerald, Taylor, Topliss and McNeil35} Pharmacokinetic studies show that body mass index (BMI) and glycaemic abnormalities affect clozapine metabolism through greater deposition in adipose tissue and influence on hepatic enzyme activity, resulting in higher exposure in some patients. ^{Reference Kuzin, Haen, Hiemke, Bochon, Bochon and Grunder58–Reference Reeves, Bertrand, Obee, Hunter, Howard and Flanagan60} Meta-analyses show that diabetes and obesity independently increased SCD risk (risk ratio 2.02, 95% CI 1.81–2.25 and risk ratio 1.16, 95% CI 1.05–1.28, respectively), ^{Reference Aune, Schlesinger, Norat and Riboli61,Reference Aune, Schlesinger, Norat and Riboli62} whereas clozapine itself causes rapid weight gain and dysglycaemia, further exacerbating cardiac risk, ^{Reference Yuen, Kim, Procyshyn, Panenka, Honer and Barr63} with obesity being associated with increased blood levels of clozapine. ^{Reference Diaz, Josiassen and de Leon64} Accordingly, international guidance advises personalised slower titration for patients with high BMI or other clinical vulnerabilities. ^{Reference de Leon, Schoretsanitis, Smith, Molden, Solismaa and Seppala65}

We identified cigarette smoking as a contributor to SCD in one cohort, ^{Reference Khan, Ashraf, Baker, Al-Omary, Savage and Ekmejian29} appearing to amplify cardiovascular risk. One meta-analysis found that current smokers have a threefold increased risk of SCD than never-smokers (risk ratio 3.06, 95% CI 2.46–3.82). ^{Reference Aune, Schlesinger, Norat and Riboli66} Mechanistically, tobacco smoke induces the liver enzyme CYP1A2, ^{Reference Hukkanen, Jacob, Peng, Dempsey and Benowitz67} producing significantly lower clozapine plasma concentration levels (standard mean difference −0.39, 95% CI −0.55 to −0.22, P < 0.001). ^{Reference Wagner, McMahon, Falkai, Hasan and Siskind68} Similarly, a retrospective cohort study reported clinically important rises in clozapine levels following abrupt smoking cessation. ^{Reference Cormac, Brown, Creasey, Ferriter and Huckstep69} This could be explained by CYP1A2 activity declining following smoking cessation. ^{Reference Faber and Fuhr70} However, this review could not find any association of smoking cessation with SCD or SUD. Closer therapeutic drug monitoring and dose adjustments following smoking cessation warrant further investigation.

We also identified an increased prevalence of illicit drug use (including cannabis, opioids and stimulants) with SCD in one cohort, ^{Reference Khan, Ashraf, Baker, Al-Omary, Savage and Ekmejian29} although the study had insufficient power for statistical analysis. A recent meta-analysis showed that cannabis was associated with increased cardiovascular (risk ratio 2.10, 95% CI 1.29–3.42), ^{Reference Storck, Elbaz, Vindis, Deguilhem, Lapeyre-Mestre and Jouanjus71} chronic opioid exposure with increased CVD risk (pooled odds ratio 1.74, 95% CI 1.12–2.70) ^{Reference Schofield, Conti, Khan and Baldacchino72} and amphetamines with higher cardiovascular mortality (odds ratio 5.12, 95% CI 3.74–7.00). ^{Reference Stockings, Tran, Santo, Peacock, Larney and Santomauro73} Stimulants produce catecholamine surges, coronary vasospasms and direct myocardial toxicity that may precipitate ventricular arrhythmia and promote chronic dilated cardiomyopathy. ^{Reference Dominic, Ahmad, Awwab, Bhuiyan, Kevil and Goeders74,Reference Reddy, Ng, Oh, Moady and Elkayam75}

Monitoring

Two temporal patterns of clozapine-associated SCD were identified: an early-onset inflammatory myocarditis and later-onset cardiomyopathy, consistent with a previous review. ^{Reference Gupta, Sayeed, Singh and Goswami76} In a fatal myocarditis-related SCD case series, Ronaldson et al ^{Reference Ronaldson, Fitzgerald, Taylor, Topliss and McNeil35} reported elevated troponin I (TnI) and T (TnT) in two affected patients. Cardiac troponins are cardiomyocyte-specific regulatory proteins released following myocardial injury, extensively validated as biomarkers for myocardial infarction. ^{Reference Katrukha and Katrukha77,Reference Stark and Sharma78} An Australian case control study reported that 90% of patients with clozapine myocarditis had TnI or TnT concentrations that exceeded twice the upper limit of normal (ULN). ^{Reference Ronaldson, Fitzgerald, Taylor, Topliss and McNeil79} ULN represents the assay-specific maximum threshold, typically defined as the 99th percentile of a healthy reference population. Reported ULNs for TnI and TnT in this study ranged from 0.03 to 0.60 μg/L. ^{Reference Ronaldson, Fitzgerald, Taylor, Topliss and McNeil79} Similarly, data from a London (UK) electronic health registry showed strong predictive performance of troponin-based models in the detection of clozapine-associated myocarditis, with an area under the receiver operating characteristic curve (AUC) of 0.975 in logistic regression analysis. ^{Reference Segev, Iqbal, McDonagh, Casetta, Oloyede and Piper80} Given that AUC values range from 0.5 (no discriminative ability) to 1.0 (perfect discrimination), an AUC of 0.975 reflects excellent accuracy in distinguishing patients with clozapine-induced myocarditis. In broader cardiology cohorts, a large Scottish study found that TnI was more strongly associated with cardiovascular disease (CVD) and CVD mortality, whereas TnT was associated more with non-CVD-related deaths. ^{Reference Welsh, Preiss, Hayward, Shah, McAllister and Briggs81} These findings suggest that TnI may offer better specificity for clozapine-related myocarditis and SCD; however, this requires further validation in dedicated psychiatric cohorts.

Creatinine kinase elevation was also seen with myocarditis-related SCD in the study of Ronaldson et al. ^{Reference Ronaldson, Fitzgerald, Taylor, Topliss and McNeil35} Creatinine kinase is involved in cellular energy metabolism, with its myocardial isoenzyme CK-MB historically used in the diagnosis of myocardial infarction before being replaced by troponins. ^{Reference Aujla, Zubair and Patel82} A post-mortem meta-analysis showed significantly higher CK-MB in SCD patients (standard mean difference 0.63, 95% CI 0.09–1.17, P = 0.02). ^{Reference Xu, Zhang, Zhu and Cao83} A prospective study in Georgia (USA) involving 100 patients (4 developed myocarditis) reported creatinine kinase abnormalities in 30% of first-exposure and 46% of re-initiated patients – considerably higher than troponin I (3.5 and 0%, respectively) or C-reactive protein (CRP) (23 and 39%, respectively), ^{Reference Sandarsh, Bishnoi, Shashank, Miller, Freudenreich and McEvoy84} suggesting that creatinine kinase elevation is common but poorly specific for myocarditis. Rises can occur secondary to non-cardiac factors such as rhabdomyolysis, strenuous exercise or trauma. ^{Reference Kurapati and Zubair85}

CRP, also identified with myocarditis-related SCD, ^{Reference Ronaldson, Fitzgerald, Taylor, Topliss and McNeil35} is an acute-phase inflammatory marker. ^{Reference Singh, Goyal and Patel86} One meta-analysis showed that higher CRP concentrations carried a greater risk of sudden death (hazard ratio 1.19, 95% CI 1.09–1.29). ^{Reference Zhou, Xu, Luan, Wang, Tang and Huang87} Although we did not observe a direct association between CRP and clozapine SCD, a retrospective cohort reported excellent discrimination (AUC 0.876) for clozapine-related myocarditis. ^{Reference Segev, Iqbal, McDonagh, Casetta, Oloyede and Piper80} However, CRP elevation lacks specificity for clozapine-related myocarditis and may be raised in concurrent infection, inflammation, comorbidities and other factors. ^{Reference Gupta, Sayeed, Singh and Goswami76} Despite these limitations, in a proposed monitoring protocol, baseline measurements of troponin, CRP and echocardiography are recommended, with clozapine cessation if troponin exceeds twofold ULN or CRP >100 mg/L. ^{Reference Ronaldson, Fitzgerald, Taylor, Topliss and McNeil79} Similarly, the Maudsley Prescribing Guidelines recommend weekly monitoring of CRP, creatinine kinase and troponin during the first 4 weeks of treatment, or if body temperature exceeds 38°C, with elevated CRP and troponin prompting further assessment using creatinine kinase, B-type natriuretic peptide and echocardiography to confirm or exclude myocarditis. ^{Reference Taylor, Barnes and Young44}

NT-proBNP was raised in one case report of myocarditis-related SCD. ^{Reference Katta, Balla and Aggarwal28} NT-proBNP is released from cardiomyocytes in response to myocardial wall stress, and is validated for heart failure diagnosis. ^{Reference Weber and Hamm88} A cross-sectional study of 38 patients receiving chronic clozapine treatment identified subclinical left ventricular dysfunction in a minority, with NT-proBNP showing strong discriminatory performance (AUC = 0.87). ^{Reference Rostagno, Domenichetti, Pastorelli and Gensini89} Collectively, these data support troponin, CRP and NT-proBNP as useful adjunctive tests, but large prospective validating specific thresholds and efficacy are lacking for clozapine SCD and other cardiovascular side-effects.

Distinct mechanistic pathways of SCD

Evidence from the included studies suggests at least two partially distinct pathways of SCD in clozapine-treated patients, with implications for monitoring and prevention. An inflammatory-structural pathway involves early-onset myocarditis, sometimes progressing to dilated cardiomyopathy, typically within weeks to months of treatment, with highest risk during early exposure, rapid titration and concomitant sodium valproate use. ^{Reference Chopra and de Leon21,Reference Hagg, Spigset, Bate and Soderstrom24,Reference Katta, Balla and Aggarwal28,Reference Kilian, Kerr, Lawrence and Celermajer30,Reference Ronaldson, Fitzgerald, Taylor, Topliss and McNeil35,Reference Vickers, Ramineni, Malacova, Eriksson, McMahon and Moudgil42} By contrast, an electrophysiological pathway may involve QT prolongation, autonomic instability or ventricular arrhythmias leading to SCD without preceding structural disease. ^{Reference Salvo, Pariente, Shakir, Robinson, Arnaud and Thomas3,Reference Hennessy, Bilker, Knauss, Margolis, Kimmel and Reynolds25,Reference Wu, Tsai and Tsai38} These differing mechanisms may suggest the need for different preventive strategies.

Clinical implications

Currently in the UK, the National Institute for Clinical Excellence and BNF advise baseline cardiac assessment before starting clozapine; specialist review if abnormalities are present; discontinuation if myocarditis or cardiomyopathy is suspected; and ongoing cardiovascular risk monitoring, including electrocardiogram review, in patients with cardiovascular risk factors. ^43,90 However, a previous systemic review highlighted the limited high-quality evidence for intensive routine cardiac monitoring, alongside concerns regarding cost-effectiveness, service capacity and barriers to clozapine access. ^{Reference Knoph, Morgan, Palmer, Schak, Owen and Leloux91} In this context, our findings highlight the rationale for considering refinement rather than escalation of monitoring, favouring a risk-stratified and mechanism-informed approach. Specifically, our findings support consideration of more cautious clozapine titration, avoiding rapid dose titration in early treatment and particular care in patients receiving polypharmacy. Monitoring of clozapine plasma levels in individuals at elevated risk may provide contextual pharmacokinetic information; however, supporting evidence is largely limited to clozapine-associated myocarditis, ^{Reference Bellissima, Burns, Helsby, Kingston, Garavan and Tingle92} but evidence linking plasma levels with SCD or SUD remains limited. Early measurement of inflammatory and cardiac biomarkers (CRP and high-sensitivity troponin) in the first few weeks of initiation has been proposed for clozapine-associated myocarditis; however, evidence that such strategies reduce SCD or SUD is lacking, and larger-scale trials are required. Routine cardiovascular risk management, including smoking cessation support, substance misuse management, careful dose adjustment and regular review of high-risk co-medications (particularly valproate and multiple antipsychotics), is important for safe clozapine prescribing but has limited direct evidence for SCD or SUD reduction. Emerging digital and wearable technologies may have a future role in longer-term risk monitoring, although evidence remains preliminary. ^{Reference Krishnadas, Saunders, Donnelly, Okubadejo, Ayadurai and Meiser-Stedman93} Importantly, these risks must be balanced against clozapine’s substantial survival benefits: an important meta-analysis showed a significant reduction in all-cause mortality despite cardiometabolic concerns (risk ratio 0.43, 95% CI 0.34–0.55), and clozapine was the largest protective factor for suicide-related mortality in schizophrenia (risk ratio 0.21, 95% CI 0.15–0.29). ^{Reference Correll, Solmi, Croatto, Schneider, Rohani-Montez and Fairley94} As the only antipsychotic with proven efficacy in treatment-resistant schizophrenia, clozapine is often under-utilised due to safety concerns. Recognition of cardiac risk factors should inform proportionate cardiovascular assessment and tailored monitoring strategies, thereby supporting safe prescribing and improved access, rather than discouraging its use.

Clozapine-associated SCD and SUD appear multifactorial, involving treatment factors, patient comorbidities and pharmacological and lifestyle factors. Future studies should clarify the longer-term effectiveness and cost-effectiveness of these added monitoring strategies needed to address sudden deaths in this vulnerable population.

Strengths and limitations

Strengths include comprehensive database searches, inclusion of broad study designs – from case report to large cohort analyses – and structured synthesis organising risk factors into clinically actionable themes. However, the evidence base is limited by the predominance of case reports and small observational studies, inconsistent definitions of SCD, SUD and myocarditis with variable autopsy rates, heterogenous reporting and an overall high risk of bias, which collectively prevent meta-analysis for most risk factors. Ethnicity was reported in some case reports but was absent from all included cohort studies, limiting assessment of ancestry-related differences in clozapine pharmacokinetics. Evidence suggests that individuals of Asian or Indigenous American ancestry have slower metabolism and increased risk of toxicity, ^{Reference de Leon, Schoretsanitis, Smith, Molden, Solismaa and Seppala65} although this review was unable to evaluate these factors. In addition, the exclusion of grey literature and non-English studies may have led to an under-ascertainment of rare but clinically important adverse effects. Consequently, our findings emphasise pattern recognition and hypothesis generation rather than definitive causal inferences. These limitations highlight the need for well-designed prospective studies and improved pharmacovigilance systems to better characterise clozapine-associated SCD.