Skip main navigation

High-Serum Brain-Derived Neurotrophic Factor Levels Are Associated With Decreased Risk of Poststroke Cognitive Impairment

Originally publishedhttps://doi.org/10.1161/STROKEAHA.123.044698Stroke. 2024;55:643–650

Abstract

BACKGROUND:

BDNF (brain-derived neurotrophic factor) is widely implicated in the pathophysiological process of stroke, but the effect of BDNF on poststroke cognitive impairment (PSCI) remains unclear. We aimed to investigate the association between baseline serum BDNF and the risk of PSCI at 3 months in a multicenter study based on a preplanned ancillary study of the CATIS trial (China Antihypertensive Trial in Acute Ischemic Stroke).

METHODS:

We examined serum BDNF levels at baseline and used the Mini-Mental State Examination and Montreal Cognitive Assessment to evaluate cognitive function at 3-month follow-up after ischemic stroke. PSCI was defined as Mini-Mental State Examination score <27 or Montreal Cognitive Assessment score <25. Logistic regression analyses were performed to evaluate the association between serum BDNF and the risk of 3-month PSCI.

RESULTS:

In this ancillary study, a total of 660 patients with ischemic stroke with hypertension were included, and 593 patients (mean age, 59.90±10.44 years; 410 males and 183 females) were finally included in this analysis. According to mini-mental state examination score, after adjustment for age, sex, education, baseline National Institutes of Health Stroke Scale score, APOE ɛ4 carriers, and other potential confounders, the odds ratio of PSCI for the highest tertile of BDNF was 0.60 ([95% CI, 0.39–0.94]; P=0.024) compared with the lowest tertile. Multiple-adjusted spline regression model showed a linear association of serum BDNF levels with PSCI at 3 months (P value for linearity=0.010). Adding serum BDNF to conventional prognostic factors slightly improved the risk reclassification of PSCI (net reclassification improvement: 27.46%, P=0.001; integrated discrimination index: 1.02%, P=0.015). Similar significant findings were observed when PSCI was defined by the Montreal Cognitive Assessment score.

CONCLUSIONS:

Elevated serum BDNF levels were associated with a decreased risk of PSCI at 3 months, suggesting that serum BDNF might be a potential predictive biomarker for PSCI among patients with ischemic stroke with hypertension.

Poststroke cognitive impairment (PSCI) is a common neurological sequela following stroke,1 and the prevalence of PSCI among stroke survivors is nearly 81% in China.2 PSCI negatively affects the quality of life and is associated with poor functional outcomes after stroke.3,4 Therefore, identification of novel biomarkers is highly desirable for better refining risk prediction and understanding the pathogenesis of PSCI.

BDNF (brain-derived neurotrophic factor), the most important molecule of neurotrophin family in the central nervous system, plays a critical role in neuronal cell differentiation and survival.5 BDNF levels were decreased with aging,6 and high serum BDNF levels reduced the risk of Alzheimer disease.7 In addition, low BDNF levels were associated with an increased risk of stroke8 and poor functional outcomes after ischemic stroke.9 Nevertheless, there are few epidemiological studies investigating the association of serum BDNF with cognitive decline after stroke, and all these available studies recruited patients with ischemic stroke from a single center and had small sample sizes.10–12 In addition, some important confounders of PSCI (eg, inflammatory biomarkers, APOE ɛ4 status) were not fully taken into account in these previous studies.10–12 Therefore, the significance of serum BDNF on PSCI remains unclear.

Herein, we conducted a secondary retrospective analysis based on the CATIS study (China Antihypertensive Trial in Acute Ischemic Stroke) to investigate the association between serum BDNF levels and the risk of PSCI.

METHODS

Data Availability

The data that support the findings of this study are available from the corresponding author on reasonable request.

Study Design and Population

Our study follows the Strengthening the Reporting of Observational Studies in Epidemiology reporting guideline. The CATIS study was a multicenter, single-blind, blinded end points randomized controlled clinical trial conducted in 26 hospitals across China from August 2009 to May 2013. The CATIS trial was registered late because registration was not required at the time of CATIS design. Detailed methods of CATIS have been described previously.13 Briefly, a total of 4071 computed tomography- or magnetic resonance imaging–confirmed patients with ischemic stroke (aged over 22 years; within 48 hours of symptom onset) with an elevated systolic blood pressure (BP) between 140 and 220 mm Hg were recruited. According to the protocol of CATIS trial, patients were excluded if they had a systolic BP ≥220 mm Hg or diastolic BP ≥120 mm Hg, cerebrovascular stenosis (≥70%), acute myocardial infarction, unstable angina, atrial fibrillation, severe heart failure, resistant hypertension, or aortic dissection, or were in deep coma, or received intravenous thrombolytic therapy.

The present multicenter study was based on a preplanned ancillary study of CATIS (https://clinicaltrials.gov/search?term=NCT01840072), which was designed to evaluate whether early antihypertensive treatment would reduce cognitive impairment at 3 months after ischemic stroke among a random sample of CATIS.14 In this ancillary study, 660 CATIS trial participants were systemically selected before randomization from 7 hospitals for cognitive function assessment at their 3-month follow-up visit. The exclusion criteria for the ancillary study were visual or hearing impairment substantial enough to hinder performance on cognitive testing.14 At the 3-month visit, 15 patients were lost to follow-up, 7 patients were deceased, and a total of 638 patients completed the cognitive function assessment at 3 months. Of these participants, a further 45 participants were excluded because we did not have their blood samples, and 593 patients with ischemic stroke with hypertension were finally included in this analysis.

This study was approved by the institutional review boards at Soochow University in China and Tulane University in the United States. The CATIS trial is registered at clinicaltrials.gov (NCT01840072). Written consent was obtained from all study participants or their immediate family members.

Data Collection

Baseline data on demographic characteristics and clinical features were collected at the time of enrollment. Stroke severity was assessed using the National Institutes of Health Stroke Scale (NIHSS) at admission. Ischemic stroke was classified as large artery atherosclerosis (thrombotic), cardiac embolism (embolic), and small artery occlusion lacunae (lacunar) according to the symptoms and imaging data of the patients. Alcohol drinking was defined as consuming any type of alcoholic beverage at least 12 times in the past year. Three BP measurements were obtained at baseline by trained nurses when the patient was in the supine position. We genotyped APOE rs429358 and rs7412 using SNPscan technology (Catalog: G0104; Genesky Biotechnologies Inc, Shanghai, China) and defined the 3 common APOE alleles (ε2, ε3, and ε4) according to rs429358 and rs7412.15

Serum BDNF Measurement

Fasting blood samples were collected with nonanticoagulated tubes after at least 8 hours of fasting within 24 hours of hospital admission. All collected blood samples were transported at 4 °C to the clinical laboratories of each participating hospital for processing. Serum samples were obtained by centrifuging at 2000g at 4 °C for 10 minutes within 2 hours of blood sampling. All serum samples were immediately frozen at −80 °C until testing. Serum BDNF levels were measured with ELISA using a commercially available ELISA kit (Catalog: DY248; R&D Systems, Minneapolis, MN). Intra-assay and inter-assay coefficients of variation were 2.8% and 7.8%, respectively.

Outcome Assessment

The study outcome was PSCI at 3 months assessed by mini-mental state examination (MMSE)16 and Montreal Cognitive Assessment (MoCA)17 in Chinese. Both MMSE and MoCA have been validated as effective screening tools for cognitive impairment in the Chinese population.16,17 The MMSE is an 11-item scale that tests 5 areas of cognitive function: orientation, registration, attention and calculation, recall, and language, with a maximum score of 30.18 The MoCA evaluates visuospatial/executive functions, naming, memory, attention, language, abstraction, and orientation, and is also scored out of 30.18 One additional point was added to the MoCA score (if <30) for participants with education <12 years to correct for education effects.18 In this study, PSCI was defined as MMSE score <2716 or MoCA score <25.17 PSCI severity was categorized as follows: severe cognitive impairment (MMSE scores, 0–22; MoCA scores, 0–19), mild cognitive impairment (MMSE scores, 23–26; MoCA scores, 20–24), and no cognitive impairment (MMSE scores, 27–30; MoCA scores, 25–30).19,20 In the sensitivity analyses, we also defined PSCI using the following age- and education-adjusted thresholds of MMSE and MoCA: MMSE score <26.5 or MoCA score <19.5 for participants with age ≤75 years and education ≤6 years; MMSE score <22.5 or MoCA score <15.5 for participants with age >75 years and education ≤6 years; MMSE score <28.5 or MoCA score <24.5 for participants with age ≤75 years and education >6 years; and MMSE score <26.5 or MoCA score <24.5 for participants with age >75 years and education >6 years.21

Statistical Analysis

All participants were divided into 3 groups according to the tertiles of serum BDNF levels. The generalized linear regression analysis and the Cochran-Armitage trend χ2 test were used to test for linear trends across BDNF tertiles for continuous variables and categorical variables, respectively. Multivariable logistic regression analyses were performed to calculate odds ratios (ORs) and 95% CIs of PSCI for upper tertiles of BDNF with the lowest tertile as a reference. Potential covariates for PSCI were selected based on prior knowledge.22–24 Covariates included in the multivariable models were age, sex, education, current smoking, alcohol drinking, time from onset to randomization, body mass index, systolic BP, platelet counts, high-sensitivity C-reactive protein, baseline NIHSS score, APOE ɛ4 carriers, 3-month modified Rankin Scale score, medical history (hypertension, diabetes, hyperlipidemia, and family history of stroke), ischemic stroke subtype, and receiving immediate BP reduction. The effect of serum BDNF levels on PSCI severity was assessed using ordinal logistic regression analyses after adjustment for the aforementioned covariates. Sensitivity analyses were conducted to evaluate the robustness of the association between serum BDNF and PSCI using the age- and education-adjusted thresholds of MMSE and MoCA. We also assessed the associations of serum BDNF with the total continuous score and each subitem score of MMSE and MoCA using generalized linear regression analyses.

We used restricted cubic splines to evaluate the pattern of the association between BDNF and PSCI with 4 knots at the 5th, 35th, 65th, and 95th percentiles of BDNF.25 Cubic spline analyses were conducted via SAS LGTPHCURV9 Macro (https://ysph.yale.edu/cmips/research/software/lgtphcurv9_7-3-2011_340182_284_47911_v2.pdf), which fits restricted cubic splines to a logistic regression model to nonparametrically examine the (possibly linear or nonlinear) association between serum BDNF and the OR of PSCI. Subgroup analyses were performed to evaluate the robustness of the association between BDNF and PSCI. The subgroup variables were based on the aforementioned covariates, including age, sex, current cigarette smoking, current alcohol drinking, time from onset to randomization, baseline NIHSS score, APOE ɛ4 carriers, history of hypertension, history of diabetes, and family history of stroke. Interactions between serum BDNF and subgroup variables on PSCI were tested in the models with interaction terms by the likelihood ratio test, adjusting for the aforementioned covariates unless the variable was used as a subgroup variable. Net reclassification improvement (NRI) and integrated discrimination index (IDI) were 2 indexes to assess improvement in model performance accomplished by adding new markers.26 All conventional prognostic factors (aforementioned potential covariates) were directly incorporated into the model when we calculated NRI and IDI. We constructed a conventional model (only including conventional prognostic factors) and a new model (including conventional prognostic factors and serum BDNF) by logistic regression model. To assess the improvement in risk prediction for PSCI by adding serum BDNF to conventional prognostic factors, we calculated NRI and IDI through comparing these 2 models. A 2-tailed P<0.05 was considered to be statistically significant. All analyses were conducted using SAS statistical software (version 9.4, Cary, NC).

RESULTS

Baseline Characteristics

Most baseline characteristics were balanced between patients who were enrolled and those who were excluded in this study, indicating that those assayed are basically representative of the total participants in CATIS (Table 1). There were 593 patients with ischemic stroke with hypertension (mean age, 59.90±10.44 years; 410 males and 183 females) included in this study. Compared with participants with lower BDNF levels, those with higher BDNF levels tended to be younger and drinkers (ie, consuming any type of alcoholic beverage at least 12× in the past year); had higher diastolic BP, body mass index and platelet counts; had higher prevalence of family history of stroke; had lower NIHSS score and high-sensitivity C-reactive protein (Table 2).

Table 1. Baseline Characteristics Between Patients Who Were Enrolled and Those Who Were Excluded

Characteristics*EnrolledExcludedP value
Number of subjects5933478
Demographics
 Age, y59.90±10.4462.31±10.92<0.001
 Male, n (%)410 (69.14)2194 (63.08)0.005
 Current cigarette smoking, n (%)220 (37.10)1265 (36.37)0.734
 Current alcohol drinking, n (%)202 (34.06)1051 (30.22)0.061
Education, n (%)
 Illiteracy47 (7.93)494 (14.20)<0.001
 Primary226 (38.11)1306 (37.55)0.794
 High school282 (47.55)1491 (42.87)0.033
 College or higher38 (6.41)173 (4.97)0.145
Medical history
 Hypertension, n (%)458 (77.23)2751 (79.10)0.305
 Hyperlipidemia, n (%)42 (7.08)235 (6.76)0.771
 Diabetes, n (%)100 (16.86)619 (17.80)0.581
 Coronary heart disease, n (%)64 (10.79)380 (10.93)0.923
 Family history of stroke, n (%)98 (16.53)655 (18.83)0.181
Clinical features
 Time from onset to randomization, h10.75 (5.00–24.00)10.00 (4.50–24.00)0.904
 Baseline systolic BP, mm Hg167.31±16.67165.94±16.930.068
 Baseline diastolic BP, mm Hg98.25±10.0896.41±11.25<0.001
 Body mass index, kg/m224.89±3.0624.97±3.180.576
 Baseline NIHSS score4.00 (3.00–7.00)4.00 (2.00–8.00)0.159
 Three-mo modified Rankin Scale score1.00 (1.00–2.00)1.00 (1.00–3.00)0.050
 Platelet counts, 109/L211.00 (169.00–251.00)208.00 (172.00–248.00)0.548
 High-sensitive C-reactive protein, mg/L1.90 (0.80–4.40)1.90 (0.70–4.80)0.843
 APOE ɛ4 carriers, n (%)81 (15.91)476 (17.02)0.538
Ischemic stroke subtype, n (%)
 Thrombotic381 (64.25)2789 (80.19)<0.001
 Embolic23 (3.88)179 (5.15)0.189
 Lacunar198 (33.39)604 (17.37)<0.001
Receiving immediate BP reduction, n (%)292 (49.24)1746 (50.20)0.666

BP indicates blood pressure; and NIHSS, National Institutes of Health Stroke Scale.

* Continuous variables are expressed as mean±SD or median (interquartile range). Categorical variables are expressed as frequency (%).

Table 2. Baseline Characteristics of the Study Participants According to Serum BDNF Tertiles

Characteristics*BDNF, ng/mLPtrend
<26.4126.41–38.61≥38.61
Number of subjects198197198
Demographics
 Age, y62.97±9.9359.43±10.0057.29±10.61<0.001
 Male, n (%)134 (67.68)140 (71.07)136 (68.69)0.828
 Current cigarette smoking, n (%)70 (35.35)76 (38.58)74 (37.37)0.677
 Current alcohol drinking, n (%)50 (25.25)77 (39.09)75 (37.88)0.008
Education, n (%)
 Illiteracy21 (10.61)14 (7.11)12 (6.06)0.094
 Primary school73 (36.87)81 (41.12)72 (36.36)0.918
 High school87 (43.94)94 (47.72)101 (51.01)0.159
 College or higher17 (8.59)8 (4.06)13 (6.57)0.412
Medical history
 Hypertension, n (%)152 (76.77)155 (78.68)151 (76.26)0.905
 Hyperlipidemia, n (%)12 (6.06)13 (6.60)17 (8.59)0.327
 Diabetes, n (%)39 (19.70)24 (12.18)37 (18.69)0.788
 Coronary heart disease, n (%)24 (12.12)20 (10.15)20 (10.10)0.517
 Family history of stroke, n (%)27 (13.64)24 (12.18)47 (23.74)0.007
Clinical features
 Time from onset to randomization, h10.00 (4.50–24.00)10.00 (4.00–24.00)12.00 (6.00–24.00)0.126
 Baseline systolic BP, mm Hg166.58±15.51166.15±18.40169.18±15.890.121
 Baseline diastolic BP, mm Hg96.23±8.5998.45±11.22100.07±9.94<0.001
 Body mass index, kg/m224.50±2.9524.83±2.8025.31±3.390.008
 Baseline NIHSS score4.00 (3.00–8.00)5.00 (3.00–7.00)4.00 (2.00–6.00)0.007
 Three-mo modified Rankin Scale score1.00 (1.00–2.00)1.00 (1.00–2.00)1.00 (1.00–2.00)0.125
 Platelet counts, 109/L199.00 (165.00–242.00)208.50 (168.00–250.50)223.00 (180.00–257.00)0.001
 High-sensitive C-reactive protein, mg/L2.25 (0.80–5.20)2.20 (0.80–5.10)1.60 (0.70–3.50)0.037
 APOE ɛ4 carriers, n (%)31 (17.61)28 (16.77)22 (13.25)0.273
Ischemic stroke subtype, n (%)
 Thrombotic116 (58.59)141 (71.57)124 (62.63)0.402
 Embolic7 (3.54)10 (5.08)6 (3.03)0.795
 Lacunar77 (38.89)50 (25.38)71 (35.86)0.523
Receiving immediate BP reduction, n (%)100 (50.51)91 (46.19)101 (51.01)0.920

BDNF indicates brain-derived neurotrophic factor; BP, blood pressure; and NIHSS, National Institutes of Health Stroke Scale.

* Continuous variables are expressed as mean±SD or median (interquartile range). Categorical variables are expressed as frequency (%).

Association Between Serum BDNF and PSCI

According to MMSE score at 3 months, 312 participants had cognitive impairment, including 151 patients with mild cognitive impairment and 161 patients with severe cognitive impairment. After adjustment for age, sex, education, and other potential confounders, the OR of PSCI associated with the highest tertile of serum BDNF was 0.60 ([95% CI, 0.39–0.94]; P=0.024; Table 3) compared with the lowest tertile. Multiple-adjusted spline regression analysis indicated an inverse dose-response association of serum BDNF with PSCI at 3 months (P value for linearity=0.010; Figure). Multivariable ordinal logistic regression analysis showed a significant association between serum BDNF and PSCI severity (Ptrend=0.013; Table 3). Similarly, significant findings were observed when cognitive function was measured by MoCA score (Table 3; Figure).

Table 3. ORs and 95% CIs for the Risk of Poststroke Cognitive Impairment According to BDNF Tertiles

BDNF, ng/mLPtrend
<26.4126.41–38.61≥38.61
MMSE<27, n (%)124 (62.63)98 (49.75)90 (45.45)
 Model 11.000.68 (0.45–1.03)0.63 (0.41–0.95)0.030
 Model 21.000.61 (0.39–0.94)0.60 (0.39–0.94)0.024
PSCI severity (MMSE), n (%)
 None (27–30)74 (37.37)99 (50.25)108 (54.55)
 Mild (23–26)58 (29.29)41 (20.81)52 (26.26)
 Severe (0–22)66 (33.34)57 (28.94)38 (19.19)
 Model 11.000.76 (0.52–1.11)0.63 (0.43–0.93)0.019
 Model 21.000.68 (0.46–1.00)0.60 (0.40–0.90)0.013
MoCA<25, n (%)148(74.75)126 (63.96)110 (55.56)
 Model 11.000.70 (0.45–1.10)0.52 (0.34–0.82)0.004
 Model 21.000.61 (0.38–0.98)0.50 (0.31–0.80)0.004
PSCI severity (MoCA), n (%)
 None (25–30)50 (25.25)71 (36.04)88 (44.44)
 Mild (20–24)63 (31.82)60 (30.46)56 (28.29)
 Severe (0–19)85 (42.93)66 (33.50)54 (27.27)
 Model 11.000.74 (0.51–1.07)0.57 (0.39–0.83)0.003
 Model 21.000.65 (0.44–0.96)0.54 (0.36–0.80)0.002

Model 1: adjusted for age, sex, and education. Model 2: adjusted for model 2 and further adjusted for current smoking, alcohol drinking, time from onset to randomization, body mass index, systolic BP, platelet counts, high-sensitivity C-reactive protein, baseline NIHSS score, APOE ɛ4 carriers, 3-mo modified Rankin Scale score, medical history (hypertension, diabetes, hyperlipidemia, and family history of stroke), ischemic stroke subtype, and receiving immediate BP reduction. BDNF indicates brain-derived neurotrophic factor; BP, blood pressure; MMSE, Mini-Mental State Examination; MoCA, Montreal Cognitive Assessment; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio; and PCSI, poststroke cognitive impairment.

Figure.

Figure. Association of serum BDNF (brain-derived neurotrophic factor) with 3-mo poststroke cognitive impairment among patients with ischemic stroke with hypertension. Adjusted odds ratios and 95% CIs derived from restricted cubic spline regression, with knots placed at the 5th, 35th, 65th, and 95th percentiles of serum BDNF levels. Odds ratios were adjusted for the same variables as model 2 in Table 3.

In the sensitivity analyses, the association between high serum BDNF and low risk of PSCI remained significant when PSCI was defined by age- and education-adjusted thresholds of MMSE and MoCA (Ptrend <0.05; Table S1). In addition, serum BDNF levels were significantly associated with the continuum of MMSE scores and MoCA scores (Ptrend<0.05; Table S2). Moreover, the scores of most cognitive subdomains in MMSE and MoCA at 3 months were higher in patients with higher serum BDNF levels (Ptrend<0.05; Tables S3 and S4).

Subgroup Analyses

In the subgroup analyses, we found that high serum BDNF levels were associated with a decreased risk of PSCI in most categories, and there was no interaction between serum BDNF and subgroup factors in the risk of 3-month PSCI (all P value for interaction >0.05; Table S5). Of note, the inverse association between serum BDNF and PSCI remained in patients aged 65 years or older and perhaps was even stronger. However, there were no significant associations between serum BDNF and PSCI among women, drinker, patients with time from onset to randomization exceeding 24 hours, patients with baseline NIHSS score <4, APOE ɛ4 carriers, patients without history of hypertension, patients with history of diabetes, and patients with family history of stroke.

Incremental Predictive Value of BDNF for PSCI

We further assessed the incremental prediction utility of serum BDNF beyond conventional prognostic factors for PSCI. As shown in Table 4, adding serum BDNF to conventional prognostic factors could improve the risk reclassification for MMSE-defined PSCI (NRI, 27.46%; P=0.001; IDI, 1.02%; P=0.015) and MoCA-defined PSCI (NRI, 27.04%; P=0.002; IDI, 1.23%; P=0.014) at 3 months.

Table 4. Reclassification and Discrimination Statistics for 3-Month Poststroke Cognitive Impairment by Serum BDNF Among Patients With Ischemic Stroke With Hypertension

NRIIDI
Estimate (95% CI)P valueEstimate (95% CI)P value
Cognitive impairment: MMSE
 Conventional modelReferenceReference
 Conventional model+BDNF (tertiles)27.46 (12.48–42.44)0.0011.02 (0.20–1.83)0.015
Cognitive impairment: MoCA
 Conventional modelReferenceReference
 Conventional model+BDNF (tertiles)27.04 (11.24–42.85)0.0021.23 (0.25–2.20)0.014

Conventional model included age, sex, education, current smoking, alcohol drinking, time from onset to randomization, body mass index, systolic BP, platelet counts, high-sensitivity C-reactive protein, baseline NIHSS score, APOE ɛ4 carriers, 3-mo modified Rankin Scale score, medical history (hypertension, diabetes, hyperlipidemia, and family history of stroke), ischemic stroke subtype, and receiving immediate BP reduction. BDNF indicates brain-derived neurotrophic factor; BP, blood pressure; IDI, integrated discrimination index; MMSE, Mini-Mental State Examination; MoCA, Montreal Cognitive Assessment; NIHSS, National Institutes of Health; and NRI, net reclassification improvement.

DISCUSSION

In the present study, patients with ischemic stroke in the highest tertile of serum BDNF had a 40% decreased risk of 3-month PSCI compared with those in the lowest tertile. Moreover, serum BDNF could slightly improve the risk reclassification of PSCI beyond the conventional prognostic factors, suggesting that serum BDNF might be a potential predictive biomarker for PSCI. Further prospective studies are needed to confirm our findings.

Compared with other blood-based biomarkers, BDNF is a key component in the maintenance of synaptic plasticity and synaptogenesis, which is the cellular biological basis of memory acquisition and consolidation.27 Decreased BDNF levels could induce neuronal degeneration, and subsequently led to hippocampal atrophy and cognitive decline.28 BDNF expression is decreased with aging, and low BDNF levels are associated with age-related hippocampal volume loss and decreased spatial memory ability.6 In addition, serum BDNF levels are related to various neurodegenerative diseases with cognitive impairment, including Alzheimer disease and Parkinson disease.7

In this multicenter study of 593 patients with ischemic stroke with hypertension, we observed a significant association of increased serum BDNF levels with a lower risk of PSCI at 3 months. We also found that serum BDNF levels were inversely associated with age, NIHSS score, and high-sensitivity C-reactive protein. Given that BDNF was associated with age, we adjusted age in the multivariable models and performed subgroup analyses stratified by age to eliminate the confounding effects of age. We found that the significant association of serum BDNF with PSCI was independent of age. In addition, there were no modified effects of age on the association between serum BDNF and PSCI, and the association between high BDNF and low risk of PSCI remained in different age subgroups. Of note, it seems counterintuitive that BDNF levels are higher in alcohol drinkers, and high BDNF levels are associated with higher diastolic BP, body mass index, and platelets in the univariate analysis. However, we found no associations of BDNF with alcohol drinking, diastolic BP, body mass index, and platelets in the age- and sex-adjusted analyses. Therefore, further studies are needed to confirm these.

From findings of our study, patients with low BDNF levels at admission may be at high risk of PSCI. Many vascular risk factors (eg, hypertension, diabetes, and hyperlipidemia) were reported to be associated with increased risk of PSCI,22 so it is critical to monitor these vascular risk factors for patients with ischemic stroke with decreased BDNF levels to prevent subsequent cognitive decline. Further studies with test-treatment outcome, health economic and implementation science methods are needed before making recommendations around using serum BDNF at scale.

Several biological mechanisms may underlie the association of BDNF with PSCI. BDNF can regulate neuroregeneration and neuroprotection by activating TrkB (tropomyosin-related kinase receptor B).5 BDNF/TrkB activation could alleviate the extent of ischemic injury and promote angiogenesis, suggesting a protective role of BDNF in preventing brain damage after ischemia.29,30 Hippocampus-dependent learning in the Morris water maze, contextual fear, and passive avoidance tests are associated with a rapid and transient increase in BDNF mRNA expression in the hippocampus.31,32 In addition, treatment with anti-BDNF antibodies causes impairment of memory in the water maze and passive avoidance tests.33,34 Further studies are needed to explain the mechanisms underlying the association between serum BDNF and PSCI.

This study has several strengths. First, this is the first multicenter study to examine the relationships between serum BDNF and cognitive impairment after ischemic stroke. Second, this study is based on the CATIS, a randomized clinical trial with standardized protocols and rigid quality control procedures in data collection and outcome assessment, which enables us to provide a more valid appraisal of the association between BDNF and PSCI. However, our study has some limitations. First, the participants included in this ancillary study of CATIS were relatively young with high follow-up rate and high survival rate, so there might be selection bias in this study. Second, we do not have data on mood disorders and sensory impairment in this study, so further studies are needed to confirm our findings with adjustment of these important moderators of cognition. Third, given the difficulty of testing cognitive function in acute phase of ischemic stroke, data on cognitive function of patients at baseline was unavailable, which may influence the observed associations in this study. However, the NIHSS has a subset of cognitive function assessment,35 and APOE genotypes are well-known genetic factors for cognitive function. We had adjusted baseline NIHSS score and APOE ɛ4 carriers in the multivariable models, so the effect of baseline cognitive function on our results was minimal. Fourth, poststroke cognition is dynamic in the early phase and a single direct test can only provide a snapshot of cognition. Further studies are warranted to assess the association between serum BDNF and PSCI at 1 year. Finally, the mean age of patients with stroke in this study was lower than that in China (59.9 years versus 66.4 years),36 so our findings might not be generalizable to older patients with stroke. However, our subgroup analyses found that the significant association between low serum BDNF levels and high risk of PSCI remained among patients with ischemic stroke aged over 65. Further studies conducted among older patients are needed to verify our findings.

CONCLUSIONS

Elevated serum BDNF levels were associated with a decreased risk of 3-month PSCI, suggesting that BDNF might be a potential predictive biomarker for PSCI among patients with ischemic stroke with hypertension. Further prospective studies are warranted to investigate whether there is a causal association between BDNF and PSCI or whether BDNF assays could have clinical utility.

ARTICLE INFORMATION

Acknowledgments

We thank the study participants and their relatives and the clinical staff at all participating hospitals for their support and contribution to this project.

Supplemental Material

STROBE checklist

Tables S1–S5

Nonstandard Abbreviations and Acronyms

BDNF

brain-derived neurotrophic factor

BP

blood pressure

CATIS

China Antihypertensive Trial in Acute Ischemic Stroke

IDI

integrated discrimination index

MMSE

Mini-Mental State Examination

MoCA

Montreal Cognitive Assessment

NIHSS

National Institutes of Health Stroke Scale

NRI

net reclassification improvement

OR

odds ratio

PSCI

poststroke cognitive impairment

TrkB

tropomyosin-related kinase receptor B

Disclosures None.

Footnotes

*X. Chang and J. You contributed equally.

The podcast and transcript are available at https://www.ahajournals.org/str/podcast.

For Sources of Funding and Disclosures, see page 650.

Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/STROKEAHA.123.044698.

Correspondence to: Zhengbao Zhu, MD, PhD, Department of Epidemiology, School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, Suzhou Medical College of Soochow University, 199 Renai Rd, Industrial Park District, Suzhou, Jiangsu Province 215123, China. Email

REFERENCES

  • 1. Kim KY, Shin KY, Chang KA. Potential biomarkers for post-stroke cognitive impairment: a systematic review and meta-analysis.Int J Mol Sci. 2022; 23:602. doi: 10.3390/ijms23020602CrossrefGoogle Scholar
  • 2. Qu Y, Zhuo L, Li N, Hu Y, Chen W, Zhou Y, Wang J, Tao Q, Hu J, Nie X, et al. Prevalence of post-stroke cognitive impairment in China: a community-based, cross-sectional study.PLoS One. 2015; 10:e0122864. doi: 10.1371/journal.pone.0122864CrossrefMedlineGoogle Scholar
  • 3. Yuan M, Guo YS, Han Y, Gao ZK, Shen XY, Bi X. Effectiveness and mechanisms of enriched environment in post-stroke cognitive impairment.Behav Brain Res. 2021; 410:113357. doi: 10.1016/j.bbr.2021.113357CrossrefGoogle Scholar
  • 4. Liao X, Zuo L, Dong Y, Pan Y, Yan H, Meng X, Li H, Zhao X, Wang Y, Shi J, et al. Persisting cognitive impairment predicts functional dependence at 1 year after stroke and transient ischemic attack: a longitudinal, cohort study.BMC Geriatr. 2022; 22:1009. doi: 10.1186/s12877-022-03609-zCrossrefGoogle Scholar
  • 5. Li Y, Li F, Qin D, Chen H, Wang J, Wang J, Song S, Wang C, Wang Y, Liu S, et al. The role of brain derived neurotrophic factor in central nervous system.Front Aging Neurosci. 2022; 14:986443. doi: 10.3389/fnagi.2022.986443CrossrefGoogle Scholar
  • 6. Erickson KI, Prakash RS, Voss MW, Chaddock L, Heo S, McLaren M, Pence BD, Martin SA, Vieira VJ, Woods JA, et al. Brain-derived neurotrophic factor is associated with age-related decline in hippocampal volume.J Neurosci. 2010; 30:5368–5375. doi: 10.1523/JNEUROSCI.6251-09.2010CrossrefGoogle Scholar
  • 7. Gao L, Zhang Y, Sterling K, Song W. Brain-derived neurotrophic factor in Alzheimer’s disease and its pharmaceutical potential.Transl Neurodegener. 2022; 11:4. doi: 10.1186/s40035-022-00279-0CrossrefGoogle Scholar
  • 8. Kaess BM, Preis SR, Lieb W, Beiser AS, Yang Q, Chen TC, Hengstenberg C, Erdmann J, Schunkert H, Seshadri S, et al; CARDIoGRAM. Circulating brain-derived neurotrophic factor concentrations and the risk of cardiovascular disease in the community.J Am Heart Assoc. 2015; 4:e001544. doi: 10.1161/JAHA.114.001544LinkGoogle Scholar
  • 9. Zhu Y, Sun L, Huang T, Jia Y, Yang P, Zhang Q, Fang C, Shi M, Guo D, Peng Y, et al. High serum brain-derived neurotrophic factor is associated with decreased risks of poor prognosis after ischemic stroke.Stroke. 2023; 54:1789–1797. doi: 10.1161/STROKEAHA.122.042362LinkGoogle Scholar
  • 10. Hassan TM, Yarube IU. Peripheral brain-derived neurotrophic factor is reduced in stroke survivors with cognitive impairment.Pathophysiology. 2018; 25:405–410. doi: 10.1016/j.pathophys.2018.08.003CrossrefGoogle Scholar
  • 11. Chen HG, Wang M, Jiao AH, Tang GT, Zhu W, Zou P, Li T, Cui GQ, Gong PY. Research on changes in cognitive function, β-amyloid peptide and neurotrophic factor in stroke patients.Eur Rev Med Pharmacol Sci. 2018; 22:6448–6455. doi: 10.26355/eurrev_201810_16057CrossrefGoogle Scholar
  • 12. Baccaro A, Wang YP, Candido M, Conforto AB, Brunoni AR, Leite CDC, Busatto Filho G, Lotufo PA, Benseñor IM, Goulart AC. Post-stroke depression and cognitive impairment: study design and preliminary findings in a Brazilian prospective stroke cohort (EMMA study).J Affect Disord. 2019; 245:72–81. doi: 10.1016/j.jad.2018.10.003CrossrefGoogle Scholar
  • 13. He J, Zhang Y, Xu T, Zhao Q, Wang D, Chen CS, Tong W, Liu C, Xu T, Ju Z, et al; CATIS Investigators. Effects of immediate blood pressure reduction on death and major disability in patients with acute ischemic stroke: the CATIS randomized clinical trial.JAMA. 2014; 311:479–489. doi: 10.1001/jama.2013.282543CrossrefMedlineGoogle Scholar
  • 14. Bu X, Zhang Y, Bazzano LA, Xu T, Guo L, Wang X, Zhang J, Cui Y, Li D, Zhang F, et al. Effects of early blood pressure reduction on cognitive function in patients with acute ischemic stroke.Int J Stroke. 2016; 11:1009–1019. doi: 10.1177/1747493016660094CrossrefMedlineGoogle Scholar
  • 15. Caselli RJ, Dueck AC, Osborne D, Sabbagh MN, Connor DJ, Ahern GL, Baxter LC, Rapcsak SZ, Shi J, Woodruff BK, et al. Longitudinal modeling of age-related memory decline and the APOE epsilon4 effect.N Engl J Med. 2009; 361:255–263. doi: 10.1056/NEJMoa0809437CrossrefGoogle Scholar
  • 16. Zhu Y, Zhao S, Fan Z, Li Z, He F, Lin C, Topatana W, Yan Y, Liu Z, Chen Y, et al. Evaluation of the mini-mental state examination and the montreal cognitive assessment for predicting post-stroke cognitive impairment during the acute phase in Chinese minor stroke patients.Front Aging Neurosci. 2020; 12:236. doi: 10.3389/fnagi.2020.00236CrossrefGoogle Scholar
  • 17. Sun R, Ge B, Wu S, Li H, Lin L. Optimal cut-off MoCA score for screening for mild cognitive impairment in elderly individuals in China: a systematic review and meta-analysis.Asian J Psychiatr. 2023; 87:103691. doi: 10.1016/j.ajp.2023.103691CrossrefGoogle Scholar
  • 18. Nasreddine ZS, Phillips NA, Bédirian V, Charbonneau S, Whitehead V, Collin I, Cummings JL, Chertkow H. The montreal cognitive assessment, MoCA: a brief screening tool for mild cognitive impairment.J Am Geriatr Soc. 2005; 53:695–699. doi: 10.1111/j.1532-5415.2005.53221.xCrossrefMedlineGoogle Scholar
  • 19. Delavaran H, Jönsson AC, Lövkvist H, Iwarsson S, Elmståhl S, Norrving B, Lindgren A. Cognitive function in stroke survivors: a 10-year follow-up study.Acta Neurol Scand. 2017; 136:187–194. doi: 10.1111/ane.12709CrossrefMedlineGoogle Scholar
  • 20. Webb AJ, Pendlebury ST, Li L, Simoni M, Lovett N, Mehta Z, Rothwell PM. Validation of the Montreal cognitive assessment versus mini-mental state examination against hypertension and hypertensive arteriopathy after transient ischemic attack or minor stroke.Stroke. 2014; 45:3337–3342. doi: 10.1161/STROKEAHA.114.006309LinkGoogle Scholar
  • 21. Zhang S, Qiu Q, Qian S, Lin X, Yan F, Sun L, Xiao S, Wang J, Fang Y, Li X. Determining appropriate screening tools and cutoffs for cognitive impairment in the Chinese elderly.Front Psychiatry. 2021; 12:773281. doi: 10.3389/fpsyt.2021.773281CrossrefGoogle Scholar
  • 22. Rost NS, Brodtmann A, Pase MP, van Veluw SJ, Biffi A, Duering M, Hinman JD, Dichgans M. Post-stroke cognitive impairment and dementia.Circ Res. 2022; 130:1252–1271. doi: 10.1161/CIRCRESAHA.122.319951LinkGoogle Scholar
  • 23. Rosenblat JD, Brietzke E, Mansur RB, Maruschak NA, Lee Y, McIntyre RS. Inflammation as a neurobiological substrate of cognitive impairment in bipolar disorder: evidence, pathophysiology and treatment implications.J Affect Disord. 2015; 188:149–159. doi: 10.1016/j.jad.2015.08.058CrossrefGoogle Scholar
  • 24. Tajeddinn W, Fereshtehnejad SM, Seed Ahmed M, Yoshitake T, Kehr J, Shahnaz T, Milovanovic M, Behbahani H, Höglund K, Winblad B, et al. Association of platelet serotonin levels in Alzheimer’s disease with clinical and cerebrospinal fluid markers.J Alzheimers Dis. 2016; 53:621–630. doi: 10.3233/JAD-160022CrossrefGoogle Scholar
  • 25. Durrleman S, Simon R. Flexible regression models with cubic splines.Stat Med. 1989; 8:551–561. doi: 10.1002/sim.4780080504CrossrefMedlineGoogle Scholar
  • 26. Pencina MJ, D’Agostino RB, D’Agostino RB, Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond.Stat Med. 2008; 27:157–172. doi: 10.1002/sim.2929CrossrefMedlineGoogle Scholar
  • 27. Lu B, Nagappan G, Guan X, Nathan PJ, Wren P. BDNF-based synaptic repair as a disease-modifying strategy for neurodegenerative diseases.Nat Rev Neurosci. 2013; 14:401–416. doi: 10.1038/nrn3505CrossrefMedlineGoogle Scholar
  • 28. Kapogiannis D, Mattson MP. Disrupted energy metabolism and neuronal circuit dysfunction in cognitive impairment and Alzheimer’s disease.Lancet Neurol. 2011; 10:187–198. doi: 10.1016/S1474-4422(10)70277-5CrossrefGoogle Scholar
  • 29. Xiong LL, Chen J, Du RL, Liu J, Chen YJ, Hawwas MA, Zhou XF, Wang TH, Yang SJ, Bai X. Brain-derived neurotrophic factor and its related enzymes and receptors play important roles after hypoxic-ischemic brain damage.Neural Regen Res. 2021; 16:1453–1459. doi: 10.4103/1673-5374.303033CrossrefGoogle Scholar
  • 30. Hang P, Zhao J, Cai B, Tian S, Huang W, Guo J, Sun C, Li Y, Du Z. Brain-derived neurotrophic factor regulates TRPC3/6 channels and protects against myocardial infarction in rodents.Int J Biol Sci. 2015; 11:536–545. doi: 10.7150/ijbs.10754CrossrefMedlineGoogle Scholar
  • 31. Yamada K, Mizuno M, Nabeshima T. Role for brain-derived neurotrophic factor in learning and memory.Life Sci. 2002; 70:735–744. doi: 10.1016/s0024-3205(01)01461-8CrossrefGoogle Scholar
  • 32. Hall J, Thomas KL, Everitt BJ. Rapid and selective induction of BDNF expression in the hippocampus during contextual learning.Nat Neurosci. 2000; 3:533–535. doi: 10.1038/75698CrossrefGoogle Scholar
  • 33. Mu JS, Li WP, Yao ZB, Zhou XF. Deprivation of endogenous brain-derived neurotrophic factor results in impairment of spatial learning and memory in adult rats.Brain Res. 1999; 835:259–265. doi: 10.1016/s0006-8993(99)01592-9CrossrefGoogle Scholar
  • 34. Alonso M, Vianna MR, Depino AM, Mello e Souza T, Pereira P, Szapiro G, Viola H, Pitossi F, Izquierdo I, Medina JH. BDNF-triggered events in the rat hippocampus are required for both short- and long-term memory formation.Hippocampus. 2002; 12:551–560. doi: 10.1002/hipo.10035CrossrefGoogle Scholar
  • 35. Cumming TB, Blomstrand C, Bernhardt J, Linden T. The NIH stroke scale can establish cognitive function after stroke.Cerebrovasc Dis. 2010; 30:7–14. doi: 10.1159/000313438CrossrefMedlineGoogle Scholar
  • 36. Wang W, Jiang B, Sun H, Ru X, Sun D, Wang L, Wang L, Jiang Y, Li Y, Wang Y, et al; NESS-China Investigators. Prevalence, incidence, and mortality of stroke in China: results from a nationwide population-based survey of 480 687 adults.Circulation. 2017; 135:759–771. doi: 10.1161/CIRCULATIONAHA.116.025250LinkGoogle Scholar

eLetters(0)

eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.