Abstract
A survey of psychiatric and cardiac medical staff from the The Prince Charles Hospital Health Service District in late 2002 indicated that nine of the 94 patients initiated on clozapine over the preceding 3 years developed cardiovascular complications and had been diagnosed with clozapine-related myocarditis or cardiomyopathy. During this same time-period there were no staff reports of the more publicized clozapinerelated adverse effect of agranulocytosis. Staff reported that serious psychotic relapse, increased hospitalization rates and diminished quality of life appeared to be associated with clozapine discontinuation. To facilitate the aim of offering clozapine treatment to all potential candidates and particularly given our unique co-location with a specialist cardiothoracic facility, a review of identified cases of myocarditis and cardiomyopathy was undertaken to inform decisions regarding clozapine treatment in this service
Literature review
Post-marketing drug surveillance is imperative in detecting new risks, quantifying old ones and modifying prescribing behaviour accordingly [1].
Clozapine-related myocarditis and cardiomyopathy
Incidence
The World Health Organization (WHO) program for international drug monitoring has shown a significant association between clozapine and both myocarditis and cardiomyopathy. This association is much stronger than for all other antipsychotic medications combined. As of early 2001, WHO had reports of at least 231 cases of clozapine-related myocarditis or cardiomyopathy [2]. It has been suggested that if diagnostic definitions of myocarditis and cardiomyopathy had been standardized earlier and had more advanced computing technology been available, this association could have been detectable by 1994 [3].
Reports of regional incidence are variable, with clozapine-related myocarditis reported as occurring inbetween 0.03% and 0.19% of initiations [3–6] and one estimate for clozapine-related cardiomyopathy being 51.5 cases per 100 000 treatment years [6]. The reasons underlying this apparent variability have not been fully delineated, but regional differences in diagnostic classification, adverse drug reaction reporting mechanisms and other unknown environmental factors have been implicated. With respect to regional diagnostic classification, not only do definitions of myocarditis and cardiomyopathy vary, but there are a number of published case reports involving diagnoses other than myocarditis and cardiomyopathy that document similar clinical details in presentation, management and outcomes. These diagnoses include pericarditis [7–10], acute cardiomyopathy [11], [12], cardiac dysfunction or dysrhythmia [13–15], congestive heart failure [16] and fatal myocardial failure [17].
Postulated mechanism(s)
The exact mechanism(s) underlying clozapine-related myocardial disease have not been determined, but the possibilities of a type 1 hypersensitivity reaction or direct toxic effect have been raised [6]. Research has also suggested a common pathophysiological process underlying psychosis and neuromuscular dysfunction and a possible link between this, clozapine-induced blockade of cell membrane calcium-dependent ion channels and cardiac dysfunction [14],[17–22].
Clozapine also appears to have complex variable effects on multiple cell and humoral-mediated immunological pathways [23], [24]. The balance of these effects can change over time and may interact with other immune system challenges such as exposure to infectious agents to produce symptomatology [25].
Time to onset
In published case reports of clozapine-related myocarditis, the time to onset of cardiovascular symptoms from initiation of clozapine varies from between 4 days [8], [9], [26] and five-and-a-half months [27], with occasional reports of emergence up to 7 years from initiation [10]. Reports involving national cohorts indicate that 75% of myocarditis cases occur within the first month [3] and 85% occur within the first 2 months of treatment [28].
The time to onset of clozapine-related cardiomyopathy has been variably reported as between 5 months [29] and 4 years [30]. National cohort studies indicate that most cases emerge within 6–9 months of initiation [4], [6].
Clinical presentation
Adverse drug reaction bulletins published by both independent and pharmaceutical company sources have alerted clinicians to consider clozapine-related myocarditis in patients who develop tachycardia or hypotension accompanied by other features of heart failure (such as chest pain, dyspnea, arrhythmias or raised jugular venous pressure) or other non-specific symptoms (such as fatigue, flu-like symptoms or fever) [28], [31].
Published case reports of clozapine-related myocarditis indicate that it is common for non-specific flu-like symptoms such as malaise, myalgia, dry cough, pleuritic chest pain and low-grade fevers to precede more overt signs of heart failure [3], [6], [11], [26],[32–34]. Rashes are sometimes also evident at this stage and are suggestive of hypersensitivity [15], [17].
Transient tachycardia or hypotension are frequently observed during clozapine initiation and can usually be conservatively managed. As demonstrated by published case reports however, it is significant if tachycardia or hypotension are worsening over time, or are associated with the emergence of new symptoms (such as dizziness, palpitations, dyspnea or chest pain) or other signs of heart failure (such as arrhythmias, raised jugular venous pressure or oedema) [3], [4], [6], [7], [27].
The clinical presentation of clozapine-related cardiomyopathy is usually that of insidious, progressive heart failure [4], [29], although sudden deterioration can always supervene [30].
Investigations
The results of blood tests and cardiac imaging investigations are not diagnostic of clozapine-related myocarditis, but they can objectively demonstrate suspected abnormalities of cardiac function, suggest the presence of an allergic or inflammatory process and aid detection of any other non-drug-related comorbidity such as infection.
Blood test results suggestive of an allergic or inflammatory process include peripheral eosinophilia and raised erythrocyte sedimentation rate or C-reactive protein. Raised creatine kinase levels indicate that muscle tissue has generally been affected by these allergic or inflammatory processes, while detectable troponins specifically indicate myocardial damage. Combinations of these abnormalities are documented in most published case reports.
Cardiac imaging investigations routinely performed in published case reports include chest X-ray, electrocardiogram (ECG) and echocardiogram. Chest X-ray features consistent with heart failure include demonstration of cardiomegaly, pulmonary congestion or pleural effusions [8], [10], [31]. ECG abnormalities are often nonspecific and can include low voltages, ST segment changes and arrhythmias [9], [11], [13], [26]. Echocardiogram features consistent with impaired myocardial function include demonstration of decreased ejection fraction with regional or global wall dyskinesias, increased wall thickness, chamber dilatation or pericardial effusion [3], [7], [27], [34].
Right ventricular endomyocardial biopsy has been held as the reference standard for the diagnosis of myocarditis. This is challenged by data including its low sensitivity of only about 35% compared with the clinical standard of improved left ventricular function over time [35]. Additionally, the results of the Myocarditis Treatment Trial demonstrated only approximately 10% of suspected myocarditis cases had diagnostic biopsy findings [36]. In the face of negative biopsy results clinicians continue to need to titrate treatment against clinical presentation and progress, leading many clinicians to question the necessity of a specific histological diagnosis. Endomyocardial biopsy is only rarely referred to in published case reports of clozapine-related myocarditis and cardiomyopathy [4].
Blood tests are less useful in the investigation of suspected clozapine-related cardiomyopathy. Cardiac imaging however, can objectively demonstrate features consistent with compromised myocardial functioning [4], [31].
Treatment
The general management of myocarditis requires avoidance of agents that exacerbate myocarditis or depress myocardial function and conventional therapy for heart failure, arrhythmias and any other underlying contributory illnesses. The Myocarditis Treatment Trial demonstrated no advantage with immunosuppression treatment [37]. The management strategies outlined in published case reports of clozapine-related myocarditis are generally consistent with these guidelines. General supportive care and possible treatment with diuretics, ACE inhibitors, digoxin or beta-blockers are strategies that are frequently referred to [35], [38]. Clozapine discontinuation is reported in the majority of patients [11–13],[16], [26], but there are several reports of symptomatic improvement despite clozapine continuation or rechallenge [6], [9], [34].
Conventional treatment for cardiomyopathy is similar to that for myocarditis. The management strategies outlined in published case reports of clozapine-related cardiomyopathy are generally consistent with this and include clozapine discontinuation in the majority of patients [12], [29].
Clinical outcome
Up to 25% of cases of acute clozapine-related myocarditis run a fulminant course resulting in death [6], [17], [28], [39]. However, the outcomes in most published case reports indicate symptom resolution within several days [11], [12], [16] to several weeks [3] of discontinuing clozapine. After acute symptom resolution there is often continued improvement of cardiac function as reflected in imaging results [29]. Some case reports suggest that continuing or rechallenging clozapine during the acute episode is associated with more prolonged symptoms [7], [8], [10]. There is however, at least one report of resolution of pericarditis/myocarditis-associated symptoms and imaging abnormalities, despite the continuation of clozapine [9].
The usual course of cardiomyopathy is insidious and progressive, but there is a heightened risk of sudden death, usually from dysrhythmia [4], [12], [28], [30]. There are several case reports of progressive symptoms of heart failure with continued clozapine treatment. While there are no specific guidelines for the management of clozapinerelated cardiomyopathy there are single case reports of cardiac dysfunction being treated and monitored along usual cardiology guidelines while clozapine is continued and also of cardiac dysfunction improving with clozapine discontinuation [6], [29].
Clozapine rechallenge
The mechanism(s) underlying successful clozapine rechallenge are unknown. There is at least one case report of clozapine rechallenge following an episode of myocarditis, prompted by the patient's continued mental state deterioration following the original discontinuation of clozapine. This decision involved the participation of the patient, his family and the cardiologist who had made the original diagnosis. Rechallenge was undertaken without the re-emergence of cardiac side-effects, and the patient experienced a significant improvement in his symptoms and functioning. The mechanism(s) underlying the first episode of myocarditis as well as the subsequent successful rechallenge were not clear to the authors [34].
Method
Nine patients previously diagnosed with clozapine-related myocarditis or cardiomyopathy at the hospital were identified through group and individual interviews with psychiatric and cardiac medical staff. During these semistructured interviews, clinicians were invited to recount their experiences with clozapine treatment, including adverse drug effects and emergent cardiovascular symptomatology. All cases identified by this method had occurred between 1999 and 2002 and all had used the same brand of clozapine. A number of staff have worked in this hospital for over 10 years, and given the repetition with which identified cases were reported by various staff, we were confident of identifying all diagnosed cases. The hospital's medical records department could not generate any other cases by diagnostic search. Only Clozapine Patient Monitoring System (CPMS) haematology results were made available by the pharmaceutical company.
In Australia, clozapine is licensed for use as a third line agent in the management of treatment-resistant schizophrenia and as such, patients initiated on clozapine are more likely to suffer with chronic, severe forms of the illness. In this way it was considered that the group of identified patients from this hospital would be representative of the overall population of patients who would be candidates for clozapine treatment in Australia. Moreover it is generally accepted that patients suffering with severe psychotic illnesses experience a high rate of physical comorbidity and that this may affect assessment of any emergent physical symptomatology. Because of this, a review of medical history and baseline physical state was included in the reevaluation of cases in this study.
This study was designed as a retrospective review involving comprehensive chart review and multidiscipline consensus panel re-evaluation.
Hospital charts for each of the nine identified cases were reviewed and information relating to psychiatric and physical presentation, treatments, progress and outcomes were organized to demonstrate the dynamic interplay between these domains over time. Additionally, all related pathology investigations were validated with the hospital's pathology laboratory database, all X-rays were reviewed with a consultant radiologist and a cardiac specialist reviewed all cardiac investigations.
The multidiscipline consensus panel was made up of three cardiologists, a general physician, a microbiologist, a chemical pathologist and the psychiatrists involved in the care of identified patients. The primary aim of this panel's standardized re-evaluation of cases was to determine the likelihood of the original diagnoses having been true. In an attempt to quantify this diagnostic likelihood for each case, a process of stratification of likelihood was employed. There is a lack of published guidelines for this process and so for this study, guidelines were based on internationally accepted diagnostic criteria for myocarditis and cardiomyopathy, taking into account published reports of clozapine related cardiovascular adverse effects and articles regarding the establishment of causality in adverse drug effects.
Table 1 displays the agreed upon criteria of defining clinical features that were to be applied to each case, stratifying the likelihood of the original diagnosis having been true to one of five groups: definite; highly probable; probable; possible; or unlikely. Secondary aims of the panel meeting were to consider issues relating to clozapine continuation during an acute episode of myocarditis and rechallenge following such an episode to facilitate the future rational management of clozapine treatment in this service.
Minimum clinical features defining stratification of diagnostic likelihood of clozapine-related myocarditis or cardiomyopathy
Results
Clozapine-related myocarditis
Overview
All patients in this study suffered with severe treatment-resistant schizophrenia or schizo-affective disorder and had failed trials of multiple other antipsychotic medications before the decision to trial clozapine was made. Baseline cardio-respiratory function was within normal limits in all cases.
Dosage titrations were consistent with CPMS-recommended initiation protocols with most patients taking 200–300 mg/day at the time of diagnosis. There were no cases of direct clozapine toxicity as reflected by plasma clozapine levels at the time of diagnosis. All plasma levels were within the accepted therapeutic range with the highest being 750 μ g/L.
The presentation and clinical course of patients in this study were strikingly similar to both each other and to other internationally reported cases, which was most obvious in the highly probable cases. There were no pre-existing medical conditions, concomitant illnesses or co-prescribed medications that better accounted for these clinical presentations.
Incidence
Consensus panel re-evaluation of the eight cases of clozapine-related myocarditis resulted in their stratification into the following categories of diagnostic likelihood: three highly probable cases (38-year-old male, 42-year-old female, 46-year-old female); three probable cases (26-year-old male, 23-year-old female, 37-year-old female); and two possible cases (27-year-old male, 47-year-old male).
Onset and clinical presentation
There were clear temporal relationships between the introduction of clozapine and the onset of physical symptoms between day 11 and day 18 (mean = 14.4), usually marked by complaints of feeling generally unwell and ‘flu-ey’ followed by dyspnea and/or chest pain. The patient's clinical presentations tended to continue deteriorating over the next day or two until myocarditis was diagnosed between day 13 and day 19 (mean = 15.8). At the time of diagnosis, patients were usually febrile to 37.3–39.4 ° C (mean = 38.4 C), tachycardic to 100–136 bpm (mean = 115.5), hypotensive to 85–140/50–80 mmHg (mean = 110.3/70.3) and tachypneic to 14–30 breaths/min (mean = 20.9). There was clinical evidence of heart failure in the highly probable cases, including elevated jugular venous pressure, added heart sounds, lung crepitations and peripheral cyanosis.
Investigations: imaging
Chest X-ray findings at the time of diagnosis were consistent with compromised cardiac functioning in the highly probable cases and included increased pulmonary markings, small pleural effusions and relatively increased cardiothoracic ratios.
ECG findings at the time of diagnosis were abnormal in seven of the cases. These abnormalities tended to be non-specifc and included right axis deviation, poor R-wave progression and ST segment and T-wave abnormalities.
Echocardiography at the time of diagnosis in the highly probable cases demonstrated clearly impaired cardiac function with ventricular dilatation and reduced ejection fraction. Probable cases demonstrated a tendency to high-normal wall thicknesses and low-normal ejection fraction at the time of diagnosis that improved over time. Possible cases were within normal limits at the time of diagnosis, but there was improvement in the ejection fraction over time in one of these two cases.
Endomyocardial biopsy and coronary angiography was undertaken in one of the highly probable cases but was negative.
Investigations: blood tests
There was evidence of an emerging inflammatory or allergic process at the time of diagnosis in all cases as reflected in elevated ESR to 40–140 mm/h (mean = 93.3, reference range < 12) and elevations in C-reactive protein to 19–286 mg/L (mean = 147.1, reference range <5). Also demonstrated at the time of diagnosis in all cases were 1–4.3-fold increases in eosinophil counts from baseline (mean = 2.8-fold increase) and 1.8–3.2-fold decreases in lymphocyte counts from baseline (mean = 2.3-fold decrease). At the time of diagnosis, total creatine kinase was increased in four cases up to 869 U/L (mean of patients with elevations = 476, mean of all patients = 288.4, reference range < 200). Troponin-I was measured at the time of diagnosis in seven cases and was detectable in all of them from 0.2–0.8 μ g/L (mean = 0.43, reference range < 0.2). Troponin-T was measured in the other case and was detectable at 0.14 μ g/L (reference range < 0.05).
Treatment and clinical outcomes
All patients demonstrated early positive mental state response to clozapine and subsequent rapid deterioration on its discontinuation. This deterioration necessitated further trials of alternative neuroleptic medication and prolonged psychiatric hospitalization in seven cases.
There were clear temporal relationships in seven cases between clozapine discontinuation at the time of diagnosis and physical recovery 4–21 days later (mean = 9). Physical recovery was facilitated by general supportive care in all cases. This was augmented by short-term treatment with various combinations of ACE inhibitors, beta-blockers, diuretics and digoxin in all of the highly probable cases and in one of the probable cases.
Clozapine was continued after myocarditis was diagnosed in one of the possible cases, with continued improvement in clinical and investigative abnormalities over the next 9 days.
Rechallenge
The highly probable cases had all been seriously physically compromised during the episode of myocarditis and the cardiologists did not consider that clozapine rechallenge was viable for these patients, as it was seen that the immediate physical risks outweighed any future potential psychiatric benefits.
The probable cases had not been as seriously physically compromised and the cardiologists considered that it would be possible to undertake clozapine rechallenge in these patients. Clozapine rechallenge was in fact successfully undertaken in one of these probable cases 10 months after the acute episode of myocarditis.
In one of the possible cases, clozapine was continued throughout the acute episode and clinical and investigative abnormalities resolved within 9 days. Clozapine was discontinued at the time of diagnosis in the other possible case but the cardiologists considered that this patient was a potential candidate for rechallenge.
Clozapine-related cardiomyopathy
There was one case of clozapine-related cardiomyopathy that was reevaluated by the panel and the diagnosis was maintained as at least possible. This patient (46-year-old female) had been hospitalized throughout her clozapine treatment and died suddenly 161 days after initiation. Staff had been given no cause for concern regarding new onset cardiac problems prior to this time and her vital signs and physical examination had been stable throughout the initiation period. ECG, echo and sestambi stress tests were normal at baseline. There were increased eosinophils over days 20-48 (2.5-fold increase from baseline), and on day 160, creatine kinase rose to 550. The coroner reported that the heart valves and chambers appeared enlarged with microscopic evidence of hypertrophy and foci of fibrosis and reported the cause of death as dilated cardiomyopathy and obesity. The hospital's histopathologist reviewed available histology specimens and while disagreeing with the assessment of myocyte hypertrophy, did generally agree that an indeterminate cardiac event leading to acute pulmonary oedema was the likely cause of death.
Discussion
The results of this study indicate that our service has had three highly probable, three probable and two possible cases of clozapine-related myocarditis during 1999–2002. Rates of incidence depend on which cases are included in calculations. When considering all 235 clozapine initiations undertaken since clozapine's release at this hospital in 1993, calculated rates of incidence are: 1.3% if only highly probable cases are included; 2.5% if probable and highly probable cases are included; and 3.4% if all cases are included. These calculated rates of incidence are increased by a factor of 2.5 if only initiations undertaken during 1999-2002 are considered. This would indicate that our local incidence of clozapinerelated myocarditis has been atypically high, for which there is no clear explanation. The one possible case of clozapine-related cardiomyopathy may represent an incidence of up to 0.4% in all clozapine initiations undertaken at this hospital since 1993, which is comparable with published rates of incidence.
This study involved retrospective case reviews, so only information gathered at the time of the original episode could be examined. The lack of endomyocardial biopsy in the majority of patients resulted in no patients being able to be assigned a definite diagnosis on that basis. It was accepted that even after having excluded other causative processes as much as possible, we were dealing with probabilities only. The influence of possible batch-related and/or community health or other environmental factors could not be determined.
It could be argued that heightened vigilance as a function of our co-location with cardiac services and the sensitization of staff following the first cases in this series may have led to the false-positive identification of cases. However, the striking similarities between cases in this study to both each other and to other published case reports, along with the maintenance of diagnoses as at least possible on panel re-evaluation, supports the alternate view that there was increased sensitivity in detecting true cases.
The onset of new physical symptoms and/or deterioration in physical signs after early toleration of clozapine suggests another process supervening on the more usual pattern of adaptation to clozapine titration, as was the pattern in all eight cases of clozapine-related myocarditis in this study. Establishing clozapine as a causal factor in a suspected adverse drug effect is supported by the demonstration of clear temporal relationships between initiation and clinical/investigative abnormalities and between discontinuation and clinical/investigative normalization, as well as the exclusion of other potential explanatory factors, including pre-existing/concomitant illnesses and co-prescribed medications.
The consensus panel agreed that modifying monitoring procedures at this hospital during clozapine initiation and maintenance treatment could assist detection and management of emergent cardiovascular symptomatology. In addition to CPMS-recommended investigations, this could include baseline echocardiography, inflammatory markers and storage of serum to aid interpretation of any future investigations and maintaining a low threshold for repeating ECG and troponin studies, especially during the first 8 weeks of treatment.
The exact mechanism(s) underlying clozapine-related myocarditis and cardiomyopathy remain unknown. As well, the mechanism(s) underlying cases of both successful clozapine continuation during an acute episode of myocarditis and successful clozapine rechallenge following such an episode are not known. It is also not clear whether successful continuation and rechallenge reflect similar underlying process. These processes may prove to be multi-factorial and involve interaction between the patient's baseline physical state, the environment and clozapine effects on metabolic processes, cell membrane ion channel function and immune reactivity. This multifactorial interaction may be dynamic over time and exert its influence on symptom development as more of a state than trait phenomenon. In this way the absolute risk of clozapine exposure may vary between individuals and the relative risk within an individual may vary depending on the time and place.
There are a number of case reports of successful clozapine rechallenge including at least one from this study. The decision to proceed with clozapine rechallenge in an individual patient should follow consideration of the balance of physical risks (both the short-term risk of repeated myocarditis and the longer-term risk of cardiomyopathy) to potential psychiatric benefits. If the patient was not so seriously physically compromised during the acute episode of myocarditis that re-exposure to clozapine would be considered as categorically too dangerous, it may be reasonable to proceed under close supervision, informed by the initial pattern of presentation but aware of the possibility of sensitization. The decision to rechallenge should be contingent on the co-operation and consent of the patient and on close liaison between psychiatric and cardiac medical staff. It would be prudent for rechallenge to be initiated in hospital with slow dose titration and close clinical and investigative monitoring. The return of symptoms during rechallenge would require prompt clozapine discontinuation and medical management and would preclude further exposure to clozapine.
It is only with the reporting and consideration of all suspected drug-related adverse effects that our knowledge and rational management of them can improve.
Footnotes
Acknowledgements
The authors acknowledge the staff that supported this project and participated in the consensus panel meeting: John Bowles, Deborah Meyers, Elizabeth Whiting, Chris Coulter, George Javorsky, Robert AhHoon, Chris Parker, Lyn Waller and Gloria Carpenter.