In-Hospital ECG Findings, Changes in Medical Management, and Cardiovascular Outcomes in Patients With Acute Stroke or Transient Ischemic Attack

Abstract

Background In patients with acute ischemic stroke, little is known regarding the frequency of abnormal ECG findings other than atrial fibrillation and their association with cardiovascular outcomes. We aim to analyze the frequency and type of abnormal ECG findings, subsequent changes in medical treatment, and their association with cardiovascular outcomes in patients with acute ischemic stroke. Methods and Results In the investigator-initiated multicenter MonDAFIS (impact of standardized monitoring for detection of atrial fibrillation in ischemic stroke) study, 3465 patients with acute ischemic stroke or transient ischemic attack and without known atrial fibrillation were randomized 1:1 to receive Holter-ECG for up to 7 days in-hospital with systematic evaluation in a core cardiology laboratory (intervention group) or standard diagnostic care (control group). Outcomes included predefined abnormal ECG findings (eg, pauses, atrial fibrillation, brady-/tachycardias), medical management in the intervention group, and combined vascular end point (recurrent stroke, myocardial infarction, major bleeds, or all-cause death) and mortality at 24 months in both randomization groups. Predefined abnormal ECG findings were detected in 326 of 1693 (19.3%) patients in the intervention group. Twenty of these 326 patients (6.1%) received a pacemaker, and 62 of 326 (19.0%) patients had newly initiated or discontinued β-blocker medication. Discontinuation of β-blockers was associated with a higher death rate in the control group than in the intervention group during 24 months after enrollment (adjusted hazard ratio, 11.0 [95% CI, 2.4-50.4]; P=0.025 for interaction). Conclusions Systematic in-hospital Holter ECG reveals abnormal findings in 1 of 5 patients with acute stroke, and mortality was lower at 24 months in patients with systematic ECG recording in the hospital. Further studies are needed to determine the potential impact of medical management of abnormal ECG findings. Registration URL: https://www.clinicaltrials.gov; Unique identifier: NCT02204267.

Keywords: ECG; beta‐blocker; mortality; stroke; transient‐ischemic attack.

Figure 1. Kaplan–Meier curve for the probability of (A) the CE (ie, mortality, stroke, myocardial infarction, or major bleeding; intervention group, n=1468 and control group, n=1444) and (B) all‐cause death (intervention group, n=1555 and control group, n=1520) within 24 months after the index stroke/transient ischemic attack in patients with different regimens of β‐blocker depicted for each study group separately (patients with missing data at 6 months with regard to β‐blocker status excluded: CE, n=519 and all‐cause death, n=356).

Log‐rank test was used to test group differences. CE indicates composite end point.

Figure 2. Estimated marginal hazard ratios and corresponding 95% CIs are depicted to illustrate the interaction between randomization groups and β‐blocker discontinuation as well as the interaction between randomization groups and heart rate on admission and corresponding 95% CIs.

A, The CE (ie, mortality, stroke, myocardial infarction, or major bleeding) and all‐cause death (B) within 24 months after the index stroke or TIA in multivariable Cox regression models. A time‐varying covariate was used for the discontinuation of β‐blocker during follow‐up discarding patients with unknown β‐blocker status (CE, n=2894, number of events=235; all‐cause death, n=3042, number of events =106, intervention group as reference category) and heart rate on admission (tachycardia >100 bpm and bradycardia <60 bpm, normocardia 60–100 bpm) using separate multivariable Cox regression analyses (CE, n=3367, number of events=477; all‐cause death: n=3354, number of events =174, intervention group as reference category). Cox regression analyses were additionally adjusted for age, sex, stroke severity (National Institutes of Health Stroke Scale score on admission), ischemic stroke or TIA as index event, cardiovascular risk factors at baseline (diabetes, hypertension, coronary artery disease, prior stroke, peripheral artery disease, renal insufficiency, heart failure: diagnosed before or during the hospital stay of the index stroke/TIA). P value for interaction within Cox regression analyses. bpm indicates beats per minute; CE, composite end point; HR, hazard ratio; and TIA, transient ischemic atack.