Introduction
Immune checkpoint inhibitors (ICI) are an effective treatment in modern oncology and have become a cornerstone for the treatment of many cancers. ICI activate the immune system to target cancer cells via blockade of co-inhibitory molecules present on T lymphocytes which regulate the immune system. ICI therapy has resulted in a significant reduction in cancer morbidity and mortality in a range of solid tumours, including melanoma, renal cancer, urological malignancies, head and neck cancers, and small cell and non-small cell lung cancer [
1]. Activation of the immune system by ICIs is recognised to cause a range of inflammatory immunotoxicities, including myocarditis and pericarditis which have been associated with high morbidity and mortality, as reported in an observational, retrospective, pharmacovigilance study by Salem et al. in 2018 [
2]. However, the full range of cardiovascular complications related to the inhibition of immune checkpoints is not fully understood.
Our cardio-oncology service has provided specialist cardiology care to patients from our partner oncology centre, where ICI have been used over the last ten years. Here we describe our real-world experience of diagnosing and managing a range of cardiovascular complications observed in cancer patients receiving ICI therapy.
Methods
This study is a retrospective analysis of all patients referred to the cardio-oncology service receiving ICI treatment with potential cardiovascular complications. All patients underwent assessment with cardiovascular investigations including a 12-lead electrocardiogram (ECG), measurement of cardiac troponin I, natriuretic peptides (BNP and NTproBNP) and transthoracic echocardiography. Selected patients had cardiac magnetic resonance (CMR) including cine imaging for volumetric analysis, oedema assessment with STIR-T2 imaging and T1 and T2 parametric mapping when available, and late gadolinium enhancement imaging, cardiac positron emission tomography-computed tomography (PET-CT), Holter ECG monitoring, myocardial perfusion scan, CT or invasive coronary angiography, and endomyocardial biopsy, when indicated. Each case was reviewed by the cardio-oncology team and the referring oncology team.
Inclusion criteria for this study were the following:
Exclusion criteria were defined as follows:
-
Patients who were found not to have any cardiac complications after assessment in the cardio-oncology service.
-
Patients who were lost to follow up before completing cardiac investigations.
-
Patients referred for assessment of a pre-existing cardiovascular disease who did not develop a new event after starting therapy with immune check-point inhibitors.
-
Patients referred for assessment of an intracardiac metastasis to monitor response to ICI therapy.
Definitions
Immune checkpoint inhibitors related cardiovascular adverse events are defined as all new cardiovascular diseases developed during ICI therapy (within 90 days of a dose of ICI), which are not explained by another cause. The diagnosis of ICI-related myocarditis was based on the definition and classification proposed by Bonaca et al. and classified as possible, probable and definite based on a combination of pathology, imaging, clinical and biomarkers findings [
3]. Non-inflammatory left ventricular dysfunction (NILVD) was defined as a new diagnosis of asymptomatic reduction of left ventricular ejection fraction (LVEF) to a value < 50% confirmed by echocardiography or CMR or symptomatic heart failure with LVEF 50–53% with a reduced global longitudinal strain and/or natriuretic peptide elevation. The diagnosis of NILVD requires the exclusion of other causes of acute cardiac dysfunction and the absence of active inflammation on CMR and of a new cardiac troponin elevation.
Immune checkpoint inhibitor-related supraventricular tachyarrhythmias were subdivided into two groups: primary, in which there was no other cardiovascular or systemic ICI-related toxicity, and secondary, where another ICI-related toxicity likely contributed to the risk of the arrhythmia, e.g., thyrotoxicosis, myocarditis, or cytokine release syndrome.
Statistical analysis
Data were collected prospectively and analysed with the SPSS statistical package (SPSS Inc., Chicago, IL, USA) and R studio Version 1.4.1717. Normally distributed data are presented as mean ± standard deviation. Non-parametric data are presented as median and interquartile ranges (IQR), and categorical data as percentages. Results from myocarditis and non-inflammatory left ventricular dysfunction cohorts were compared using Chi-squared analyses and Fisher’s exact test for categorical variables, two-sided t-tests for continuous variables with Gaussian distribution and Mann-Whitney test for non-parametric variables and a multivariable logistic regression analysis was performed to assess for independent risk factors of myocarditis or NILVD. A P‐value of < 0.05 was considered statistically significant.
Discussion
Immune checkpoint inhibitors have a clinically significant impact on cancer survival proven in randomised controlled trials and are licensed for the treatment of various cancers. This therapy has had a particularly dramatic effect on the outcomes for patients with cancers previously associated with high mortality, such as metastatic melanoma, metastatic renal carcinoma and advanced lung carcinoma [
9]. Adjuvant and neoadjuvant indications are also now indicated for some cancers and are being assessed across a range of malignancies. Immunotoxicities are a common side effect of ICI, with as many as 70% of patients who may develop immune related adverse events and 40% of them needing to have their therapy interrupted [
10,
11]. The importance of understanding the range of cardiovascular risks associated to the use of ICI therapies increases with the expanding number of licensed indications for this treatment, as does an understanding of the risk:benefit in specific cohorts of cancer patients where the absolute benefit is lower (adjuvant and neoadjuvant indications). In the last six years, the focus has been on myocarditis as the most serious cardiovascular adverse event caused by ICI therapy. This paper describes a widening spectrum of ICI-related cardiac complications including arrhythmias, NILVD, pericarditis, and acute coronary syndrome. While not intending to imply causation, given this is a descriptive study of our real-life experience and not a randomised control trial, we believe the reported findings are of high relevance to the cardio-oncology community.
Myocarditis was the most frequent ICI-related cardiovascular complication in our cohort. There were rare case reports prior to 2016, but awareness increased after Johnson et al. reported two cases of fulminant myocarditis with the use of combined therapy, Ipilimumab and Nivolumab. The authors reported that myocarditis was a rare complication (0.09% cases treated with ICI) but with high mortality (33%) [
11]. Since this first publication, further cohorts have been published with increasing numbers of patients and events [
4,
12‐
15]. Diagnostic criteria have evolved and in the latest definition from the International Cardio-oncology Society, myocarditis is confirmed either by histopathology (endomyocardial biopsy) or the combination of a new cardiac troponin rise and either a diagnostic CMR or at least 2 minor criteria [
16]. In this cohort, among the 17 patients characterised as definite myocarditis, 11 (65%) fulfilled this new definition. Four patients (44%) with probable myocarditis met the criteria as well as 3 out of the 7 patients with possible myocarditis. Non-invasive imaging techniques (echocardiography, CMR), ECG, ECG telemetry and cardiac biomarker measurement are recognised as critical investigations [
17] in this context and should be performed rapidly in cancer patients receiving ICI therapy who have suspected myocarditis [
3]. In the cohort presented in this article the use of all these diagnostic tools, together with prompt treatment with high dose steroids at first instance, led to outstanding clinical outcomes with no cardiovascular deaths after > 200 days of median follow-up time.
The question of routine surveillance for cardiovascular complications and its utility in patients under ICI treatment has been raised before [
18]. This is a cohort of patients which includes cases as early as from 2014 when no formal recommendations were available. Cardiovascular investigations were mainly driven by signs or symptoms of cardiac disease.
and decided by the treating team. Our current recommendation is to follow the recently published cardio-oncology guidelines [
19] where ECG, NP and cTn measurements are recommended in all patients before starting ICI therapy and before doses 2, 3 and 4 to detect subclinical ICI-related CV toxicity.
In clinical practice some cases of myocarditis are more challenging to diagnose. One example is that of the cancer patients who have received steroid treatment for another immune related toxicity, with myocarditis suspected subsequently. Partially treated ICI-related myocarditis may not meet the formal criteria, e.g. CMR evidence of prior myocarditis on late gadolinium enhancement but without active inflammation. In cases where diagnostic uncertainty exists, we recommend cardiac PET-CT for clinically stable patients, as suggested by Boughdad et al. who demonstrated that this method is highly sensitive in detecting myocardial inflammation [
20], although further research is still needed. In cases of clinical instability and uncertain diagnosis, endomyocardial biopsy is the diagnostic tool of choice.
The frequency of ICI-related myocarditis in this study, is substantially higher than initially recognised, when the reported rate was 0.09%, and more similar to, although slightly higher than, that reported in 2018 by Mahmood et al., 1.14% [
21]. Our study presents a cancer population treated with ICI therapy of 2647 patients with a rate of myocarditis of 1.25% (33/2647). In this cohort, there were no cardiovascular deaths reported among the patients with ICI-related myocarditis. This is in stark contrast to the 17–27% mortality in previously published cohorts [
21,
22] and there are several factors which we believe may explain the difference. Firstly, the mentioned publications included patients until 2017. Our study has included cases until December 2020 and 56% of the patients had been referred in the last 2 years. This is very likely to have had an impact on the outcomes since, over several years, both oncologists and cardiologists in the two centres have developed a high index of suspicion to consider the possibility of, and investigate for, myocarditis. This has led to a reduction in the threshold for referral of cancer patients with suspected ICI-related myocarditis by the oncologists and rapid assessment by the cardio-oncology service, with the consequence of early diagnosis and treatment. Most patients diagnosed with myocarditis, or at high clinical suspicion and where clinically unstable, were started on high-dose steroids and ICI treatment interrupted. The current protocol used in our service is the following: intravenous methylprednisolone 500–1000 mg once daily for three days minimum and continuing until troponin stabilises at < 80 ng/L and any clinical complications (heart failure, ventricular arrhythmias) have settled, then switching to oral prednisolone 1 mg/kg with weaning scheme whilst monitoring cardiac troponin. Early treatment of ICI-myocarditis with high dose steroids is believed to reduce MACE and cardiovascular mortality [
23], and this management pathway is probably another reason for such low mortality rates.
This paper also raises awareness that there are other ICI-related cardiovascular complications besides myocarditis. Tachyarrhythmias, such as atrial fibrillation, were among the most frequent. Pericarditis is also observed, either with myocarditis (peri-myocarditis), or as a separate diagnosis. Less frequent, but of high clinical importance is ICI-related ischaemic heart disease, including acute myocardial infarction with the need of urgent percutaneous coronary intervention; bradyarrhythmia leading to pacemaker implantation; and pulmonary artery hypertension.
ICI treatment is usually interrupted in cancer patients who develop ICI-related cardiovascular complications until the diagnosis is established. This is critical as myocarditis is a relative contraindication to further ICI treatment. In contrast, it may be restarted in patients with other ICI-related cardiac events, including AF, pericarditis and NILVD, after this has been treated. The current approach is to follow general cardiology guidelines for these other ICI-related cardiac toxicities, except that ICI-related pericarditis may require steroid therapy in addition to colchicine.
One of the most important aspects of our study is the description of NILVD as a new cardiovascular complication associated with the use of immune check-point inhibitors. This event was observed in 0.6% of the referred patients (15/2647) and happened in those who remained on ICI treatment for longer courses, the median time to presentation being 26 weeks after starting therapy. Several other characteristics distinguish myocarditis from NILVD including the increase in cardiac troponin and the presence of late gadolinium enhancement on CMR, which are both higher in myocarditis, given the inflammatory nature of this complication, and the requirement of these tests to be abnormal for the confirmation of the diagnosis of ICI-related myocarditis. Left ventricular systolic function was lower in the NILVD cohort. We also observed that concomitant right ventricular dysfunction was a frequent finding in these patients, suggesting that longer-term ICI treatment may lead to myocardial impairment in a subgroup of patients, affecting both ventricles, via a non-inflammatory mechanism. We have initiated guideline-based heart failure treatment, temporarily interrupted ICI treatment in severe cases, and then restarted ICI treatment after recovery of ventricular function. On the other hand, LV dysfunction was only present in 27% of myocarditis patients, stressing the point that myocarditis should not be ruled out on the basis of a normal LV function, this is aligned with previously published observations [
21]. Finally, several risk factors were associated with myocarditis but not NILVD such as the presence of other immune related toxicity and the use of combined therapy (Ipilimumab and Nivolumab). Therefore, it appears that immune checkpoint inhibitors can trigger two different types of myocardial injury: myocarditis via active inflammation with T lymphocyte infiltration of the myocardium and a non-inflammatory type of ventricular dysfunction. The pathological mechanisms leading to NILVD are unknown, although in the year 2001 Nishimura et al. showed that PD1 knock-out mice develop a form of dilated myocardiopathy, implying that this receptor is crucial to maintaining myocardial vitality and blocking it might lead to myocardial dysfunction [
24] Further research in the field is required to better understand the pathophysiology, diagnosis, and treatment of both ICI-related myocarditis and ICI-related NILVD.
Limitations
There are several limitations to this study that the authors would like to acknowledge. First, this is a single-centre study limiting the cases to a relatively low number compared to previous pharmacovigilance publications [
2]. However, we believe that data proceeding from real world experience during routine clinical practice are increasingly valuable as a resource and significantly complement the information coming from these other sources. Secondly, the patients included were referred by the oncology and haematology teams which might have led to referral bias. There might be other cases which were not referred due to non-cardiac causes and therefore not included, or cases in which cardiovascular complications may not have been recognised or considered to be associated with ICI therapy. This limits the possible conclusions regarding the incidence of cardiovascular events in the wider population of patients treated with ICI. Furthermore, a matched untreated control cohort was not available to compare the cardiac event rates spontaneously occurring in this group of patients and therefore conclusions regarding causality cannot be drawn and the number of endomyocardial biopsies was limited to 2, not allowing to perform any statistically valid comparison between the histological and clinical findings. Finally, specific data points, such as the date of ICI treatment commencement or the number of cycles received prior to developing the adverse events, were not available in all cases.
Declarations
The remaining authors have nothing to disclose.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.