Chagas Disease Is an Independent Risk Factor for Stroke

Patients and Methods

We studied consecutive patients in a cardiomyopathy clinic. Patients were enrolled if they had clinical signs suggestive of cardiomyopathy and a transthoracic echocardiogram available within 1 year of study entry. After informed consent, patients underwent a standardized evaluation by a cardiologist, with the following collected prospectively: age, gender, history of diabetes or hypertension, alcohol abuse defined as daily alcohol use, current smoking status, history of coronary artery disease, coronary artery bypass grafting, permanent cardiac pacer use, cardiac arrest, medications currently used, electrocardiogram (ECG), echocardiographic data, and admission blood pressure. Additionally, cardiologists screened for stroke using the Questionnaire for Verifying Stroke-Free Status.⁷ Patients with positive screening underwent an evaluation by a neurologist who confirmed stroke status. Cardiomyopathy was defined by the presence of low ejection fraction (EF) on echocardiogram (<40%) or a borderline-low EF (40 to 49%) plus signs of cardiac dilatation (left ventricle systolic diameter >45 mm and left ventricle diastolic diameter >55 mm). CD was confirmed by appropriate serologic assays. We then compared each variable between patients with or without stroke on enrollment. All patients are being followed in a cohort study, which was approved by the local research ethics committee.

For univariable analyses, Student t test was used for continuous variables, and Fisher exact test for categorical variables, with P<0.05 considered significant. Logistic regression was used for multivariable analysis, including variables with a possible association (P<0.1).

Results

We enrolled 305 patients between February 2002 and February 2003. Mean age was 53±12 years (179 males). CD was the main etiology of cardiopathy (52%). History of hypertension was the most frequent cerebrovascular risk factor (47.7%). Stroke was present in 32 (10.5%) patients, more commonly in CD (15.0%) than in other cardiopathies (6.3%; P=0.015). In 60 patients without evidence of systolic dysfunction (EF≥50%), stroke was present in 6 of 39 (15.4%) patients with CD and only 1 (4.8%) patient with other cardiopathies (P=0.404). Patients with CD had a similar frequency of atrial fibrillation and history of cardioversion as non-CD patients. Apical aneurysm with thrombus was only present in patients with CD.

Variables associated with stroke in univariable analyses (Table 1) were history of diabetes, CD, or cardioversion, and use of amiodarone, antiplatelet agents, and warfarin. Echocardiographic findings did not differ between patients with or without stroke.

In multivariable analysis (Table 2), remaining predictors of stroke were CD (odds ratio [OR], 1.09), cardioversion (OR, 1.07), and diabetes (OR, 1.12). Because medication use was probably consequence and not a cause of stroke in this population, we did not include these in the final analysis. If we excluded patients with concomitant CD and either hypertension or diabetes from the analysis (n=48), CD still showed a trend as a significant predictor of stroke (OR, 1.07; 95% CI, 0.99 to 1.15; P=0.068).

Discussion

CD remains an important cause of cardiopathy in South America. The presence of apical aneurysm and intracardiac thrombus are hallmarks of the disease, making CD a highly embolic condition.⁸

The main finding of our study was an independent association between CD and stroke. Previously, neurological manifestations of the chronic form of CD have been attributed to embolic phenomenon.⁸ However, case series have reported stroke in patients with CD without clinical evidence of cardiopathy.⁶ A case-control study also established CD as an independent risk factor for stroke but did not correct for presence or severity of cardiac disease.⁹ In our patients, stroke was frequently present in patients without evidence of systolic dysfunction by echocardiogram, and the relationship between CD and stroke was independent of cardiac disease severity. When excluding important confounders for stroke risk such as hypertension and diabetes, CD still showed marginal significance (P=0.068) as an independent predictor for stroke, a hypothesis that should be confirmed in future studies.

The reason for an association between CD and stroke is speculative. Associations between chronic infections and stroke have been attributed to activation of inflammatory and coagulation cascade, endothelial dysfunction, and atherogenesis.^4,5 In CD, a chronic activation of the immune system occurs, with persistent lymphomonocytic myocarditis and fibrosis.¹⁰ In experimental CD, microvascular damage,^11,12 endothelial cell changes,¹³ and hyperviscosity¹⁴ have been demonstrated. We hypothesize that such chronic inflammation may explain, at least in part, our findings of stroke in patients without criteria for cardiomyopathy. However, we cannot fully exclude that CD patients with stroke could have intermittent cardiac arrhythmias undetected on clinical examination, history or EKG, or intracardiac thrombus undetected by transthoracic echocardiogram.

There are limitations to our study. First, our population is selected, including only patients with clinical signs of cardiomyopathy. Thus, no patients with latent, arrhythmic or gastrointestinal forms of CD were studied. Second, no data were collected regarding stroke subtypes, which might differ according to the cardiomyopathy etiology. In the present cohort, we are currently collecting data on carotid duplex sonography and neuroimaging, which will allow for a more complete classification.