Immune checkpoint inhibitor (ICI) therapy is an important option for a number of malignancies, but have been increasingly associated with immune-related adverse events, including cardiovascular toxicity. ICI-related cardiotoxicity can range from asymptomatic troponin elevation to fulminant myocarditis. Specifically, after receiving ICI therapy, atrial arrhythmias (AA) incidence ranges between 1.1% and 8.2%, and ICI-related arrhythmic events were found to be severe with death occurring in 26.1% on arrhythmia cases. ,,, While previous studies have noted the occurrence of AA during ICI therapy, there is a lack of data regarding which patients are at highest risk and how the development of these arrhythmias independently influences long-term survival. This study was specifically designed to address two questions: (1) What clinical variables are associated with new-onset AA in a large ICI cohort? (2) Does the development of AA serve as an independent marker for increased all-cause mortality? We hypothesize that AA is not merely a bystander of ICI therapy but a significant prognostic indicator of poor clinical outcomes.
We retrospectively reviewed adult patients treated with ICI therapy at Mayo Clinic Rochester, Mayo Clinic Florida, or Mayo Clinic Arizona between December 2011 and March 2020. ICI therapy included one or more of the following drugs: ipilimumab, nivolumab, pembrolizumab, avelumab, atezolizumab, cemiplimab, and durvalumab. Patients who were diagnosed with an AA prior to initiation of ICI therapy were excluded from this study. Of the 2292 patients started on ICI therapy, 1944 patients met the inclusion criteria and were reviewed for further analysis. Baseline clinical characteristics, malignancy type, and cardiovascular comorbidities were obtained through electronic medical record review. Patients who developed AA (atrial fibrillation or atrial flutter) after ICI therapy were identified, and management decisions along with outcomes were recorded. For this study, major adverse cardiovascular event (MACE) was characterized as either myocardial infarction, heart failure hospitalization, stroke/transient ischemia attack, or cardiovascular death. Cox regression was used to evaluate clinical predictors of AA and to examine the association of AA as a time-dependent predictor of all-cause mortality.
The median follow-up duration was 15.25 months, with an interquartile range from 5.06 to 51.84 months. New onset AA occurred in 138 patients (7.1%), corresponding to a cumulative incidence of 4.1% at 1 year and 6.9% and 5 years ( Figure 1 ). Patients who developed AA were more likely to be male (65.9% vs 56.0%), older (68.39 vs 61.42 years old), and had higher incidences of hypertension (60.9% vs 41.9%), hyperlipidemia (57.2% vs 35.5%), diabetes (21.7% vs 12.5%), and heart failure (14.5% vs 4.3%). Although patients with AA trended to have a higher incidence of vascular disease (31.2% vs 24.1%), this difference was not statistically significant. Among patients without AA, the most common malignancies were melanoma (31.1%), lung (28.4%), and other cancers (33.2%). These distributions did not differ significantly from patients who developed AA, where melanoma accounted for 24.6% of patients, lung for 34.1% of patients, and other cancers for 36.2% of patients.
Time from the start of ICI therapy administration to the onset of atrial arrhythmia. This survivorship curve demonstrates the cumulative incidence of AA after starting ICI therapy. The cumulative incidence of AA was 4.1% at 1 year, 5.1% at 2 years, and 6.9% at 5 years.
AA typically occurred spontaneously (60.1%), but was also noted in the context of infection (10.9%) or lung/respiratory disease (15.2%). Management of AA primarily involved rate control (72.5%) and anticoagulation (51.4%). Despite treatment, outcomes were poor: 34% of patients died within 6 weeks of AA onset, and 16.7% of patients experienced a MACE. In a separate time-dependent model, AA was independently associated with all-cause mortality (HR 2.00; p < 0.001) after adjustment for age and sex ( Table 1 ).
Table 1
. Cox regression, proportional hazards, univariable and multivariable modeling for new onset AA
| Characteristic | Univariable prediction of AA | Multivariable prediction of AA | ||||
|---|---|---|---|---|---|---|
| HR | 95% CI | p value | HR | 95% CI | p value | |
| Age | 1.05 | 1.03, 1.06 | <0.001 | 1.04 | 1.03, 1.06 | <0.001 |
| Sex | 1.57 | 1.10, 2.23 | 0.012 | 1.47 | 1.03, 2.09 | 0.035 |
| Hypertension | 2.05 | 1.45, 2.88 | <0.001 | 1.33 | 0.92, 1.93 | 0.13 |
| Diabetes | 2.07 | 1.38, 3.10 | <0.001 | 1.45 | 0.94, 2.22 | 0.094 |
| Stroke/TIA | 2.84 | 1.61, 5.04 | <0.001 | 1.71 | 0.94, 3.11 | 0.078 |
| Vascular disease (PAD/CAD) | 1.28 | 0.89, 1.84 | 0.18 | 0.65 | 0.44, 0.96 | 0.031 |
| Heart failure | 3.80 | 2.37, 6.11 | <0.001 | 2.76 | 1.64, 4.63 | <0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
