Background
Stroke is a major cause of serious disability and death [
1], as well as the common cause of acquired cognitive impairment [
2]. Post-stroke cognitive impairment (PSCI) has been reported to be associated with unfavorable prognosis, including major disability, mortality, stroke recurrence and poorer quality of life [
3‐
5]. Thus, novel and reliable predictors are clearly needed for early identification of patients at higher risk of PSCI.
Heart-brain axis has a greater role in the procession of cognitive impairment and dementia [
6‐
8], available population-based evidence reported suboptimal cardiac function or abnormally elevated cardiac biomarkers, such as N-terminal pro-B-type natriuretic peptide (NT-proBNP) or high-sensitivity cardiac Troponin T (hs-cTnT), were associated with worse cognitive performance [
9‐
13], suggesting cardiac biomarkers may be used to identify individuals at higher risk of cognitive impairment. However, whether the predictive roles of cardiac biomarkers persist in the setting of cerebrovascular disease was less consistent [
7]. Soluble suppression of tumorigenesis-2 (sST2), another cardiac stress biomarker of promoting cardiomyocyte hypertrophy and fibrosis, is considered to be an important biomarker of heart failure. Recently, the Framingham Offspring showed that higher sST2 levels were associated with increased risk of incident stroke and subclinical vascular brain injury [
14]. However, the association of sST2 and cognitive impairment in patients with ischemic stroke remains to be addressed.
Therefore, we aimed to prospectively assess the relationship between plasma sST2 levels in the acute phase of ischemic stroke and PSCI at 3 months using the data derived from the China Antihypertensive Trial in Acute Ischemic Stroke (CATIS).
Discussion
This prospective study using data from CATIS trial found that higher plasma sST2 levels were associated with increased risk of PSCI, independently of potential confounders including education, stroke severity and medical history. The significant association remained when further controlled inflammatory biomarker. In addition, adding sST2 levels into the model containing conventional risk factors statistically improved the predictive ability, as evidenced by NRI and IDI statistic. Furthermore, sST2 was one of the promising predictors for PSCI. These findings provided population-based evidence of plasma sST2 as a potential biomarker in predicting PSCI.
Emerging evidence from epidemiological studies support that sST2 is of diagnostic and prognostic value in the setting with various cardiovascular diseases, including heart failure, coronary artery disease, and ischemic stroke [
14,
23‐
26]. For example, the Linz Stroke Unit Study conducted in acute ischemic stroke patients reported a higher level of sST2 in decedents than survivors [
24]. Furthermore, Wolcott et al. demonstrated that sST2 was an independent predictor of short-term mortality, functional outcome and hemorrhagic transformation in patients with ischemic stroke [
25]. Of note, patients with cardiovascular diseases are at higher risk of experiencing cognitive decline. However, clinical studies designed to specifically investigate the association between sST2 and cognitive impairment, especially in the condition of ischemic stroke, are sparse.
A small study of 18 mild cognitive impairment patients and 17 healthy controls showed that serum sST2 levels were significantly higher in patients with mild cognitive impairment than controls [
27]. Furthermore, Andersson et al. using the data from the Framingham Offspring Study reported a cross-sectional association between sST2 concentrations and cognitive impairment, and they found participants in the highest quartile of sST2 had significantly lower brain volumes and poorer delayed performance on the visual reproduction test than those in the lowest quartile [
14]. Similarly, previous studies also suggested significant associations between other cardiac biomarkers with neurological disorders. For example, a cross-sectional analysis of 860 ischemic stroke patients suggested that hs-cTnT was associated with the severity of white matter lesions, which was considered as a predictor of poorer cognitive function [
28,
29]. Recently, the PROSCIS-B (Prospective Cohort With Incident Stroke Berlin) study reported that higher hs-cTnT was associated with higher prevalence of cognitive impairment at baseline and lower Telephone Interview for Cognitive Status-modified during 3-year follow-up in patients with mild-to-moderate ischemic stroke [
20]. The present study, to our knowledge, was the first longitudinal study to directly characterize the relationship of sST2 and PSCI, extending the connection of heart and brain to the patients with ischemic stroke.
The mechanisms underlying the sST2-PSCI association are still unclear, but several potential pathophysiological pathways have been proposed. Cardiac dysfunction was implicated in various pathological conditions, including hemodynamic stress, cerebral hypoperfusion, neuroinflammation, cardiac arrhythmias, and hypercoagulation, and then may further lead to cognitive impairment [
6,
7]. Moreover, prior studies showed that cerebral small vessel disease (CSVD) and brain atrophy had relationship with cognitive dysfunction [
30,
31]. As a serum cardiac marker, sST2 could indicate the load of the CVSD [
7] and elevated sST2 was associated with lower brain volumes [
14], which might affect cognitive function. Furthermore, interleukin 33 was found to be neuroprotective in experimental stroke models [
32], and the administration of interleukin 33 could reduce cognitive decline [
27]. Inflammation might be the potential mechanisms. CRP, a typical inflammatory marker, can induce other proinflammatory factors and was associated with an increased risk of stroke [
33,
34]. However, after additionally adjusting hsCRP in the Model 3, the significant relationships remained. Further studies are required to clarify related mechanism.
Several lines of evidence suggested that age, sex, education attainment, admission BP, stroke severity, medical history, ischemic stroke subtype, and inflammation were associated with cognitive status in the general population or participants with cardiovascular disease [
35‐
39]. In the present study, the sST2-PSCI relationship remained after adjustment for these established risk factors, indicating sST2 independently contributed to the risk of PSCI. This might relate to plasma sST2 reflecting heart and brain injury, and its specificity for cardiac function might distinguish plasma sST2 from other blood-based biomarkers with predictive value for PSCI reflecting other biological processes, such as tHcy, RF, MMP-9 and TIMP-1 [
40‐
42].
In addition, incorporating sST2 into a model with known risk factors statistically improved reclassification for PSCI prediction. Moreover, sST2 was one of the promising predictors for PSCI. Therefore, the evaluation of the association between sST2 and PSCI had important clinical significance given the high prevalence and heavy disease burden of PSCI. These findings, coupled with the evidence that sST2 was of prognosis value in ischemic stroke patients, imply the clinical usefulness of sST2 measurement to identify patients at high risk of PSCI and provide novel therapeutic interventions. Future well-designed clinical trials aimed to test the effect of inhibition of sST2 treatment on cognitive impairment among ischemic stroke patients are warranted.
Our study was based on a subsample of the well-performed CATIS trial with standardized protocol and rigid quality control procedures, enabling us to provide a more comprehensive and valid assessment of the association between plasma sST2 levels with PSCI. However, our study has several limitations. First, patients with BP ≥220/120 mmHg or treatment with intravenous thrombolytic therapy at admission were not included in the CATIS trial. These limited the generalizability of our findings to all acute ischemic strokes. Second, plasma sST2 levels were only measured once at admission, we could not explore its dynamic changes over time and the effect on PSCI. Third, we did not collect the information of pre-stroke cognitive status due to lack of feasibility. However, we included NIHSS score at admission in the multivariate model, which had a subset cognitive dysfunction evaluation and had almost the same diagnostic value as the MMSE (area under the ROC curve values of 0.78 and 0.84, respectively) [
43]. Finally, the data of brain and cardiac imaging, such as the site or type of acute ischemic lesions, left atrial volume or left ventricular dysfunction were also not recorded. Hence, we could not further control these factors.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.