Is Acute Ischemic Stroke Really Associated with Left Ventricular Systolic Dysfunction? A Case-Control Study
Keywords: Acute ischemic stroke, atrial fibrillation, systolic dysfunction
Heart failure and acute ischemic stroke (AIS) are frequently associated, due to similar risk factors and intertwined pathophysiologic mechanisms, and both result in a high mortality rate. In the general population, there is 2–12% prevalence of asymptomatic left ventricular (LV) dysfunction which seems to be associated with a higher risk of AIS. Our hypothesis was that LV systolic dysfunction (LVSD) is associated with AIS.
The present study was a prospective, matched, case-control study on patients with AIS. From January 1, 2016 until August 31, 2017, out of 460 AIS patients admitted in Stroke Unit, 110 patients with AIS were enrolled in our study. Inclusion criteria were: (1) admission for AIS within 24 h of onset and (2) absence of any major concurrent illness, including malignancies, severe renal failure (stage 4 or above), acute, or severe (i.e. respiratory failure-inducing) pulmonary disease. The patients in the control group (CG) without history of AIS and the same inclusion criteria were selected from an Internal Medicine ward and were matched for age, gender, and atrial fibrillation (AF) prevalence.
AIS diagnosis and severity. The etiology of AIS was determined by a neurologist according to TOAST classification capitalizing on the information provided by (1) cervical ultrasonography, transcranial Doppler (and, when necessary, CT angiography) for carotid and vertebral atherosclerosis and dissection, (2) continuous 5 days ECG monitoring for new onset AF, and (3) transthoracic echocardiography searching for putative cardiac sources of thromboembolism, respectively. AIS severity was graded according to National Institutes of Health Stroke Scale (NIHSS) in three categories: mild (NIHSS ≤8), moderate (NIHSS between 9 and 15), and severe (NIHSS ≥16).
Echocardiographic evaluation. All AIS patients underwent transthoracic echocardiography within 5 days of AIS onset and were compared with the CG. An ecocardiography board-certified specialist performed and interpreted transthoracic two-dimensional standard parasternal, apical, and subcostal echocardiography. The analysis of the echocardiographic data was blinded to the outcome of the study. LV ejection fraction, the parameter used for assessing systolic dysfunction, was determined by biplane Simpson's method. Based on the ejection fraction, the systolic function was classified as normal (EF ≥55%), mild dysfunction (EF 55–45%), moderate dysfunction (EF 45–30%), and severe dysfunction (EF <30%).
The risk factors taken into account were: arterial hypertension, dyslipidemia, diabetes mellitus, chronic alcohol consumption, obesity, and smoking. Arterial hypertension was defined as blood pressure >130/90 mmHg or normal blood pressure under antihypertensive treatment associated with a positive history of hypertension. Dyslipidemia was considered to be present if serum level of total cholesterol >200 mg/dL and/or of tryglicerides >150 mg/dL or if cholesterol and/or tryglicerides levels were normal under lipid lowering treatment in association with a positive history of dyslipidemia. Diabetes mellitus was defined as fasting glucose level >126 mg/dL or fasting glucose level was normal under oral antidiabetics or insulin treatment associated with a positive history of diabetes. An ingestion of more than one drink per day for women and more than two drinks per day for men was considered chronic alcohol consumption. A body mass index of 30.0 or higher indicated obesity. A patient was deemed to be current smoker if he/she has smoked > 100 cigarettes in his/her lifetime and has smoked in the last month.
Fisher's exact test was used for statistical analysis. The statistical calculations were performed using the R language and environment for statistical computing and graphics. The results were considered statistically significant if the P value was below the generally accepted threshold of 0.05.
Informed consent. The study was designed in compliance with the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the Ethics Committee of Carol Davila University of Medicine and Pharmacy. All participants in this study gave their informed consent either directly or through a surrogate when appropriate (patients with aphasia, altered state of conscience).
The mean age of our patients was 67.2 years in both groups. New onset AF (paroxysmal AF or persistent AF) was present in 35 of 110 (31.81%) AIS patients. Persistent AF was present in 32 of 110 (29.09%) patients in CG. All the demographics and risk factors in both groups are shown in [Table 1]. There were no significant differences between the two groups.
The etiology of AIS was classified based on the TOAST (Trial of Org 10172 in Acute Stroke Treatment) criteria as: cardioembolic (in 29.09% of the patients), large artery atherosclerosis (in 50.90%), small-vessel lacunar (in 3.64%), other causes (in 1.81%), and undetermined/cryptogenic (in 14.54%).
Systolic dysfunction. LVSD of any degree was not statistically more or less frequent in AIS patients than in CG ones. The same was true if only the patients with AF were considered (the prevalence of LVSD was not significantly different in patients with AIS and AF than in controls with AF). However, among the patients without AF, the proportion of patients with normal systolic function was significantly higher in the controls than in the AIS group (P = 0.036) [Table 2].
Analyzing only AIS patients, there was no significant difference regarding the prevalence of systolic dysfunction between those with AF and those without AF. Nonetheless, if only the controls were assessed, those with AF had systolic dysfunction in a significantly higher proportion than those without AF [Table 3].
The comparison between the patients with cardioembolic AIS and those with noncardioembolic AIS demonstrated no significant difference regarding the prevalence of systolic dysfunction [see [Table 4]].
In our study, the prevalence of LVSD was the same in AIS patients and controls matched for age, sex, and (crucially) AF prevalence, although the prevalence of LVSD was indeed higher in AIS patients without AF than in controls without AF. The results of our study may seem contradictory to those of two previous studies that arrived at the conclusion that LVSD was more frequent in AIS patients than in controls, both when analyzed globally (for all degrees of LVSD),, and for subsets such as moderate/severe and mild LVSD. The explanation lies probably with the different choice of CG: apparently healthy individuals from the community in the previous studies versus stroke-free patients in the present study. This difference in the CGs is also reflected by the difference in the prevalence of the various AIS risk factors among the controls: hypertension, coronary artery disease, AF, and diabetes were all significantly more prevalent in the controls of both previous studies (and also hypercholesterolemia and smoking in the study of Kelly et al.) but not in the present study. However, in the study of Hays et al., the association between AIS and LVSD persisted even after adjustment for LVSD risk factors, but there are still other LVSD risk factors [both known, such as inflammation, or as yet unknown] and it is virtually impossible to take into account the whole lot. Interestingly, the controls in the present study were matched with the AIS patients only for age, sex, and prevalence of AF, and not for the prevalence of any other risk factor. Nonetheless, the two groups (AIS and CG) turned out to be matched also for other risk factors (including hypertension, obesity, diabetes, dyslipidemia, smoking, chronic alcohol intake, heart failure, previous myocardial infarction), as if AF and its risk factors are by far the main risk factors for LVSD or at least are accounting for all the main risk factors for LVSD. This conclusion is logical when bearing in mind the pathophysiological chain leading from LVSD, through increased LV diastolic and left atrial pressures to increased left atrial wall stress and/or volume, and finally to AF.,
Nevertheless, there might still be another reason for the higher prevalence of LVSD in AIS patients without AF than in controls without AF: a transient cardiac dysfunction induced by AIS. Indeed, there are studies pointing out a transient LV dysfunction (possibly as a consequence of a burst in sympathetic activity) that may accompany AIS without necessarily preceding or causing it. A follow-up echo study performed several weeks after the AIS onset demonstrating the remission of LVSD could have been interpreted as a proof for such a transient AIS-induced LVSD. Unfortunately, the design of the present study did not include such a follow-up echocardiogram (mainly for logistical reasons).
Regarding the fact that AIS did not mark out a subset of AF patients more prone to LVSD [which agrees with the results obtained by Kelly et al.], but it did so in patients without AF, a sensible explanation might be that AF by itself is a strong marker of heart disease,, and particularly of LV dysfunction because both systolic and diastolic dysfunction result in higher LA pressure and wall stress, and consequently increased risk of AF (as stated above). By contrast, the absence of AF leaves the ground open for other markers/comorbidities of cardiovascular disease – prominent among these is AIS. In the absence of both markers of cardiovascular disease (AF and AIS), there is no clear reason for the CG patients without AF to have an increased prevalence of LVSD, as long as no other risk factors are thought about.
The novelty of this study is that the CG was matched not only for age and sex, as in previous studies, but also for AF prevalence – this last matching criterion appeared to be essential as the results of this study contradicted those yielded by previous studies, suggesting that the crucial mediator of the AIS-LVSD association is AF. Our study concluded that the prevalence of LVSD was the same in AIS patients and controls matched for age, sex, and AF prevalence, although the prevalence of LVSD was indeed higher in AIS patients without AF than in controls without AF.
As cardioembolic is the subtype of AIS with the highest in-hospital mortality and the short-term prognosis of patients with cardioembolic stroke is poor compared to other ischemic stroke subtypes, a comparison between patients with cardioembolic and noncardioembolic AIS was mandatory. However, no significant difference regarding the prevalence and severity of LVSD between these two main etiological categories of AIS was found, a result which is to be expected in a study in which there was no such difference between AF and non-AF patients with AIS, considering the very strong association between cardioembolic AIS and AF. Indeed, in all but three of the AF patients cardioembolism was considered the cause of AIS. In the remaining three, athersosclerosis with either severe stenosis or symptomatic arterial occlusion was considered the most probable cause.
There are several directions for future research in this area. As the main conclusion of our study is contrary to the conclusions of earlier studies, the most obvious direction for the future is the undertaking of another study with a much larger sample size in order to convincingly settle this controversy. Another would be a prospective study aimed at establishing the proportion of AIS patients with LVSD but without AF that subsequently develop AF over time. Given the fact that women differ from men in the distribution of stroke risk factors, subtype, severity, and outcome, an indispensable line of research in the future would be the assessment of whether LVSD shows gender differences in patients with AIS.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
The approval from the ethics committee was obtained: 29-02-2016.
[Table 1], [Table 2], [Table 3], [Table 4]