Life-threatening arrhythmia in patients with suspected acute myocarditis

Patients with suspected acute myocarditis (AM) often undergo prolonged rhythm monitoring due to the risk for life-threatening arrhythmia.

To describe the occurrence, timing and potential early rule-out of life-threatening arrhythmia in patients with AM.

We included consecutive patients with suspected AM admitted to the ICU/IMC for continuous rhythm monitoring into a cohort study. We assessed the incidence and timing of life-threatening arrhythmia (sustained ventricular tachycardia, ventricular fibrillation, cardiac arrest). To rule-out arrhythmia, we evaluated left ventricular ejection fraction (LVEF), maximal cardiac troponin-T (cTnT) levels and a multivariable model.

Among 304 patients with AM (41±16.6 years, 27% female), 13 life-threatening arrhythmias occurred in 10 (3.3%) patients. Of these, 8 occurred within 24h, 2 between 24-48h and 3 after 72h of hospitalization (Figure 1). Patients with life-threatening arrhythmia had substantially higher mortality rates (40% vs. 0.3%, p<0.001). While patients with life-threatening arrhythmia exhibited higher median cTnT levels and lower median LVEF, we did not find a binary cut-off in these two variables to rule-out arrhythmia. We found a substantial overlap in cTnT levels (67% of values overlapped) and in LVEF levels (73% of values overlapped). The temporal relationships of life-threatening arrhythmia occurrence and hs-cTnTpeak are shown in Figure 2: The last life-threatening arrhythmia occurred before the cTnT-peak in 3 (42.9%), simultaneously with the peak in 1 (14.3%), and after the peak in 3 (42.9%) patients. The median times of the last recorded life-threatening arrhythmia before and after the hs-cTnTpeak were 54 and 95 hours, respectively. The final multivariable model included female sex, cTnT, and LVEF and demonstrated an area under the curve of 0.98 (95% CI 0.96-1), with a sensitivity of 99% and specificity of 75% to rule-out life-threatening arrhythmia.

In patients with suspected AM, life-threatening arrhythmias were rare but associated with a 40% mortality rate. Relying solely on the surpassing of the cTnT peak to stop monitoring would have resulted in missing half of life-threatening arrhythmias, despite a clustering of life-threatening arrhythmias around the cTnTpeak and therefore shouldn’t be exclusively used as guidance for earlier discharge from rhythm monitoring. A combined model including 3 clinical variables ruled-out life-threatening arrhythmia with a high sensitivity and may help to guide the indication of rhythm monitoringFigure 1  Figure 2