Is Cardiac Troponin Useful in Detecting In-Hospital Afib After Ischemic Stroke? | Journal Scan
What parameters, available on admission, predict in-hospital paroxysmal atrial fibrillation (pAF) in ischemic stroke patients?
This work was part of the TRELAS (Troponin Elevation in Acute Ischemic Stroke) study, which looked at the relationship between highly sensitive cardiac troponin T (hs-cTnT) and ischemic stroke. In a Berlin stroke unit, consecutive patients with ischemic stroke from February 2011 to December 2013 had hs-cTnT levels measured. Patients with atrial fibrillation, pacemakers, or ST-elevation myocardial infarction were excluded. The authors collected demographic and clinical information, including stroke severity. CHADS2 and CHA2DS2-VASc scores were calculated, excluding the index stroke event. The distribution of hs-cTnT was skewed and was, therefore, log transferred for analysis. Multivariable logistic regression was used to predict the diagnosis of pAF. A sensitivity analysis was done to account for variation in the duration of electrocardiogram (ECG) monitoring among patients.
There were 1,228 patients included in the analysis and 114 (9.3%) were found to have pAF while hospitalized. Patients had continuous ECG monitoring in the hospital for a median of 3 days. On univariate analysis, age (p < 0.001), stroke severity (p < 0.001), duration of monitoring (p < 0.001), CHADS2 score (p < 0.001), CHA2DS2-VASc score (p < 0.001), congestive heart failure (p = 0.006), hypertension (p < 0.001), smoking (p = 0.023), log hs-cTnT values (p < 0.001), low glomerular filtration rate (p < 0.001), and insular cortex involvement (p < 0.001) were associated with pAF. The incidence of new atrial fibrillation did not differ between left and right insular cortex involvement (p = 0.44). The presence of multiple lesions in ≥1 vascular territory was not associated with pAF (p = 0.35). In the multivariate analysis, the following were associated with pAF: age 65-74 years (odds ratio [OR], 2.69; 95% confidence interval [CI], 1.19-6.09), age, ≥75 (OR, 4.50; 95% CI, 1.98-10.25), hypertension (OR, 2.28; 95% CI, 1.12-7.44), log hs-cTnT (OR, 1.81; 95% CI, 1.03-3.20), hs-cTnT value >17 ng/L (OR, 1.65; 95% CI, 1.05-2.59), insular cortex involvement (OR, 3.06; 95% CI, 1.93-4.87), and duration of monitoring (OR, 1.10; 95% CI, 1.03-1.19). The findings from a sensitivity analysis that accounted for variations in the duration of ECG monitoring were consistent with the primary results. In regression models that included hs-cTnT and insular cortex involvement along with CHADS2 or CHA2DS2-VASc score, the c-statistic had modest improvement. When duration of ECG monitoring was included in the model, only the combination of hs-cTnT and insular cortex involvement was significant.
In this study, predictors of in-hospital atrial fibrillation after ischemic stroke include: age, insular cortex involvement, hypertension, hs-cTnT level, and duration of monitoring.
Atrial fibrillation is an important risk factor for ischemic stroke, but since it is often paroxysmal, detecting it can be challenging. Identifying predictors of pAF after ischemic stroke would be useful to help determine the intensity of monitoring for pAF. This study expands on the roles of hs-cTnT and lesions of the insular cortex in predicting pAF. Of note, insular cortex lesions could be the result of pAF, but these lesions are also associated with arrhythmias, and therefore, could also be a cause of pAF. Further research is needed in this area. Somewhat surprisingly, lesions in multiple vascular distributions were not predictive of pAF, although this finding has been noted in other studies as well. Echocardiography findings, such as left atrial size, were not included in the models to predict pAF, which may limit the results.
Clinical Topics: Arrhythmias and Clinical EP, Heart Failure and Cardiomyopathies, Prevention, Implantable Devices, SCD/Ventricular Arrhythmias, Atrial Fibrillation/Supraventricular Arrhythmias, Acute Heart Failure, Hypertension, Smoking
Keywords: Arrhythmias, Cardiac, Atrial Fibrillation, Cerebral Cortex, Electrocardiography, Glomerular Filtration Rate, Heart Failure, Hypertension, Logistic Models, Multivariate Analysis, Risk Factors, Smoking, Stroke, Troponin, Troponin T
< Back to Listings