
==== Front
Heliyon
Heliyon
Heliyon
2405-8440
Elsevier

S2405-8440(24)05384-2
10.1016/j.heliyon.2024.e29353
e29353
Research Article
Associations of cognitive impairment and longitudinal change in cognitive function with the risk of fatal stroke in middle-aged to older Chinese
Li Jun-xiao ab
Zhong Qiong-qiong b
Zhu Tong c
Jin Ya-li c
Pan Jing c
Yuan Shi-xiang d
Zhu Feng chifengzhu@hotmail.com
ab⁎
a Central Laboratory, Guangzhou Twelfth People's Hospital, Guangzhou, China
b Departments of Public Health and Preventive Medicine, Jinan University, Guangzhou, China
c Department of Science and Education, Guangzhou Twelfth People's Hospital, Guangzhou, China
d Department of Neurosurgery, Guangzhou Twelfth People's Hospital, Guangzhou, China
⁎ Corresponding author. Central Laboratory, Guangzhou Twelfth People's Hospital; No.1 TianqiangLu St., West Huangpu Dadao Rd., Guangzhou, 510620, China. chifengzhu@hotmail.com
15 4 2024
30 4 2024
15 4 2024
10 8 e2935326 10 2023
28 3 2024
5 4 2024
© 2024 The Author(s)
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
It is unclear whether cognitive impairment and the longitudinal change in cognition are associated with the risk of fatal stroke in aging populations. Based on the Guangzhou Biobank Cohort Study data a sum of 26,064 participants at baseline and all deaths caused by stroke in a mean follow-up of 14.3 years (standard deviation = 3.2) were included, and the Cox proportional hazard regression was used in this prospective cohort study. Cognitive impairment was respectively associated with an increased risk of fatal strokes (the adjusted hazard ratio (aHR) = 1.38, 95% CI1.16–1.64, P < 0.001) and fatal ischaemic stroke (aHR = 1.39, 95% CI1.10–1.77, P = 0.007), compared to median cognition; the Delayed Word Recall Test (DWRT) score was associated with a decreasing trend for the risk of fatal strokes in a restricted cubic spline analysis; the longitudinal DWRT score decline was associated with the increased risks of fatal strokes (aHR = 1.42, 95% CI 1.11–1.82, P = 0.006) and fatal haemorrhagic stroke (aHR = 1.75, 95% CI 1.10–2.78, P = 0.02), compared to the longitudinal DWRT score rise. In summary, cognitive impairment and the longitudinal decline in DWRT scores were associated with the increased risk of fatal strokes; early screening of cognitive function should be conducive to predictive intervention in fatal stroke among relatively healthy middle-aged to older populations.

Keywords

Stroke
Cognitive function
Delayed word recall test
Cohort
==== Body
pmc1 Introduction

Stroke in the World Health Organization clinical criteria was defined as rapidly developing clinical signs of (usually focal) disturbance of local or global brain dysfunction, lasting ≥24 h or leading to death [1]. The Global Burden of Disease reported that approximately 101 million people worldwide suffered from stroke in 2019, making it the second cause of death and the third cause of disability [2], and meantime up to 28.76 million Chinese suffered from stroke [3]. It follows that global stroke has become a public health issue. The main cardiovascular risk factors for stroke include hypertension, high total cholesterol, high-density lipoprotein cholesterol, obesity, diabetes mellitus, smoking, short ischaemic attacks, and atrial fibrillation, and the proportion of deaths after the stroke due to pre-existent cardiovascular risk factors before stroke was 27%, this suggests that we must continue to study the factors that cause another 73% of stroke deaths [4]. Cognitive impairment is prevalent and may impact on mortality of the aging population; it is a broad diagnosis of cognitive function, including the full spectrum of mild cognitive impairment (MCI), moderate cognitive impairment, and dementia. MCI is a reversible, daily activity-independent, all-cause mortality-promoting transition state [5,6] between normal cognition and early manifestations of dementia and is a risk factor for dementia [7], wherein pre-stroke dementia linked to the mortality risk of stroke [8]. Additionally, moderate cognitive impairment was related to all-cause mortality [9]; Cognitive impairment is also associated with an increased mortality risk of cardiovascular diseases [10,11]. Previous studies have focused mainly on the relationships between cognitive impairment and those in patients with stroke in a model of retrospective analysis. Among stroke cases with good clinical recovery at three months, the prevalence of any cognitive impairment was 71%, with memory, visual construction and executive functioning most frequently impaired [12]. Multivariate analyses released that the post-stroke prevalence ratio of cognitive impairment increased with older age (2% for each year of age), ethnicity (2.2 fold higher in the black group), and socioeconomic status (42% increased in manual workers) [13]. However, it is not well documented whether cognitive impairment and the longitudinal change in cognitive function are associated with the risk of fatal stroke. Here, we employed a model of prospective analysis to assess the associations of cognitive impairment and the longitudinal change in cognitive function with the risks of fatal stroke (strokes, ischaemic stroke, and haemorrhagic stroke) in relatively healthy middle-aged to older Chinese populations.

2 Methods

2.1 Participants

A prospective cohort study was conducted among all study participants in the Guangzhou Biobank Cohort Study (GBCS) from September 2003 to April 2021. As a tripartite collaborative project between the Guangzhou Twelfth People's Hospital, the University of Hong Kong, China, and the University of Birmingham, UK, GBCS participants at baseline (from September 2003 to January 2008) were recruited from the Guangzhou Health and Happiness Association for Respectable Elders (GHHARE), a large non-social organization with 10 branches in all districts of Guangzhou. The GHHARE membership is open to Guangzhou residents aged ≥50 and a monthly fee of RMB 4 (approximately 50 cents) is required. All surviving participants at baseline were invited to return for the 1st follow-up (from March 2008 to December 2012), and the next follow-ups since then continued every 5 years. The trained nurses at the Guangzhou Twelfth People's Hospital conducted face-to-face interviews to collect information on participants through a computerized questionnaire, including demographic information, socio-economic status, lifestyle, self-reported personal history of illness, and physical examination items; all laboratory tests (fasting blood glucose, lipids, etc.) were completed by the trained staff in the clinical laboratory. This study was reviewed and approved by the Guangzhou Medical Ethics Society, and each subject must read the informed voluntary consent form before receiving the questionnaire and physical examination, and then sign or fingerprint to participate. The reliability of the questionnaires was tested by recalling 200 randomly selected subjects for re-interviewing the satisfied results [14]. Details of GBCS, GHHARE, and prospective studies have been reported previously [15,16].

2.2 Exposure indicators

The Delayed Word Recall Test (DWRT), specifically used in population-based epidemiological studies and screening examinations, was employed in this study, it has been known as a sensitive tool for cognitive function measurement [17,18]. All participants at baseline (2003–2008) and the 1st follow-up (2008–2012) conducted the DWRT examination (a learning task of a 10-word list), wherein the modified 10 words include soy sauce, arm, letter, chairman, ticket, grass, corner, stone, and book; these words adhere to suit Chinese culture and dialectal conventions [19,20]. Every investigator read the words face-to-face, and each word was repeated at 1-s intervals; this process was repeated 3 times; the participants were asked to recall those words as far as possible 5 min later, and a score of 1 was given for each correct word. The maximum DWRT score is 10; a score <4 was defined as cognitive impairment; the mean DWRT score was 5.5 (standard deviation (SD) = 1.85). Subjects of this study were divided into 3 groups: cognitive impairment (the DWRT score <4), median cognitive function (the DWRT score 4–6), and better cognitive function (the DWRT score ≥7) [21].

3 Outcomes

Information on the underlying causes of death was obtained primarily through a record linkage of the Guangzhou Centre for Disease Control and Prevention. The outcome was a fatal stroke (up to April 2021) due to available information on the severity, infarct volume, lesion site, and infectious complications; causes of stroke were coded according to the 10th version of the International Classification of Diseases (ICD) as follows: I60–I69 for strokes, I60.0-I62.9 and I69.0-I69.2 for haemorrhagic, I63.0-I63.9 and I69.3 for ischaemic, and the other codes for unclassified. When the certificates were not issued by a health care provider, causes were verified by the GZCDC (being as a part of the quality assurance program through a cross-checking of past medical history) and verbal autopsies (conducted by 5 senior clinicians from the Guangzhou Twelfth People's Hospital, the Universities of Hong Kong, China, and Birmingham, UK).

3.1 Potential confounders

The risk factors for stroke were chosen as confounders including age, gender, education, smoking, alcohol consumption, physical activity (PA), body mass index (BMI), hypertension, diabetes mellitus, sleep duration, and dyslipidaemia. Model 1 was crude without an adjustment for any confounders. Model 2 was adjusted for 11 confounders (including sex, age, sleep duration, education, smoking (never, ever, and current), alcohol consumption (never, ever, and current), PA, BMI, hypertension, diabetes, and dyslipidaemia). Diabetes was defined as a self-reported diagnosis or taking hypoglycemic medications or fasting blood glucose ≥7.0 mmol/L [22]. Hypertension was defined as a self-reported diagnosis or taking antihypertensive medications or a systolic blood pressure ≥140 mmHg and/or a diastolic blood pressure ≥90 mmHg [23]. Dyslipidaemia was defined as total cholesterol ≥6.2 mmol/L or low-density lipoprotein ≥4.1 mmol/L or high-density lipoprotein <1.0 mmol/L or triglycerides ≥2.3 mmol/L or taking lipid-lowering medications. PA was determined using the International Physical Activity Questionnaire (IPAQ) [24], wherein inactive (not reaching the moderate and active PA levels), moderate (not reaching active but ≥3 days of high-intensity PA and a total of ≥480 MET-min, or ≥5 days of all intensities of PA combined and a total of ≥600 MET-min per week), and active (high-intensity PA ≥3 days and a total of ≥1,500 MET-min, or a total of ≥7 days of all intensities PA and a total of ≥3,000 MET-min of PA per week) were given by corresponding activity intensities and metabolic energy (MET) values.

3.2 Statistical analysis

The Pearson χ2 test and one-way analysis of variance (ANOVA) were used to compare categorical and continuous variables between different DWRT groups. Continuous variables were expressed as mean ± SD; Categorical variables were expressed as frequencies and percentages, respectively. The Restricted Cubic Spline (RCS) analysis was used to depict the trend in the relationship between cognitive score and fatal stroke. Associations of cognitive function and the DWRT Score Change (DWRTc) with stroke were analyzed in the Cox proportional regression model with an unadjusted hazard ratio (HR), adjusted HR (aHR), and 95% confidence intervals (CI), wherein the DWRTc was calculated as the formula: DWRTc = DWRTthe 1st follow-up—DWRTbaseline, and DWRTc rise (the DWRTc level >0), DWRTc stable (the DWRTc level = 0), and DWRTc decline (the DWRTc level <0) were given. All analyses of this study were performed using the IBM SPSS Windows Statistical System (version 26.0, Armonk, NY, USA) and R Studio (version 4.1.3); all P values were two-sided and a statistical significance was defined as P < 0.05.

4 Results

4.1 Basic characteristics at baseline

Among the screened total of 30,430 participants in this study, the numbers of excluded participants were 286 due to a stroke history, 382 due to loss of vital status, 4 due to abnormal sleep duration, and 3,694 due to missing values in variables (sex, hypertension, diabetes, dyslipidaemia, smoking, alcohol consumption, physical activity, BMI, sleep duration, age, and education); eventually, a total of 26,064 subjects without a stroke history at baseline were included, and 793 strokes (393 ischaemic, 252 haemorrhagic, and 148 unclassified) deaths were recorded after a mean follow-up time of 14.3 (SD = 3.2) years or a total person-years of 374,017 years; additionally, a total of 13,126 subjects without a stroke history at the 1st follow-up were included for a longitudinal assessment, wherein 255 strokes (129 ischaemic, 77 haemorrhagic, 49 unclassified) deaths were recorded (Fig. 1).Fig. 1 Flow diagram of the participants selected for analyses in this study.

Fig. 1

Table 1 shows the basic characteristics of the GBCS participants at baseline. Compared to those with DWRT <4, the participants with DWRT ≥7 were younger and more females, had higher education but less sleep duration, shared more current alcohol consumption, active physical activity, and dyslipidaemia but less hypertension, diabetes, BMI ≥28 kg/㎡, former and current smoking, and fatal strokes.Table 1 Baseline characteristics by the DWRT scores in the GBCS participants (n = 26,064).

Table 1Characteristic	the DWRT scores	
cognitive impairment
<4	median cognition
4–6	better cognition
≥7	P value	
Number, n	3,359	15,436	7,269		
Age (mean (SD))	65.15 (7.07)	62.23 (6.90)	59.94 (6.51)	<0.001	
Sleep duration (mean (SD))	6.81 (1.47)	6.92 (1.36)	6.99 (1.28)	<0.001	
Sex, n (%)				<0.001	
 female	2,354 (70.1)	10,837 (70.2)	5,518 (75.9)		
 male	1,005 (29.9)	4,599 (29.8)	1,751 (24.1)		
Body mass index, kg/㎡				0.046	
 <18.5	168 (5.0)	705 (4.6)	276 (3.8)		
 18.5–23.9	1,665 (49.6)	7,688 (49.8)	3,688 (50.7)		
 24–27.9	1,180 (35.1)	5,498 (35.6)	2,611 (35.9)		
 ≥28	346 (10.3)	1,545 (10.0)	694 (9.5)		
Education, n (%)				<0.001	
 primary	2,194 (65.3)	6,837 (44.3)	2,031 (27.9)		
 middle school	1,018 (30.3)	7,294 (47.3)	4,339 (59.7)		
 college	147 (4.4)	1,305 (8.5)	899 (12.4)		
Smoking, n (%)				<0.001	
 never	2,610 (77.7)	12,231 (79.2)	6,128 (84.3)		
 Former	352 (10.5)	1,583 (10.3)	493 (6.8)		
 current	397 (11.8)	1,622 (10.5)	648 (8.9)		
Alcohol consumption, n (%)				<0.001	
 never	2,492 (74.2)	10,856 (70.3)	4,168 (57.3)		
 former	117 (3.5)	388 (2.5)	149 (2.0)		
 current	750 (22.3)	4,192 (27.2)	2,952 (40.6)		
Physical activity, n (%)				<0.001	
 inactive	269 (8.0)	1,355 (8.8)	532 (7.3)		
 moderate	1,491 (44.4)	6,448 (41.8)	2,638 (36.3)		
 active	1,599 (47.6)	7,633 (49.4)	4,099 (56.4)		
Hypertension, n (%)				<0.001	
 no	1,729 (51.5)	8,663 (56.1)	4,423 (60.8)		
 yes	1,630 (48.5)	6,773 (43.9)	2,846 (39.2)		
Dyslipidaemia, n (%)				<0.001	
 no	1,707 (50.8)	7,635 (49.5)	3,422 (47.1)		
 yes	1,652 (49.2)	7,801 (50.5)	3,847 (52.9)		
Diabetes, n (%)				<0.001	
 no	2,815 (83.8)	13,407 (86.9)	6,429 (88.4)		
 yes	544 (16.2)	2,029 (13.1)	840 (11.6)		
Fatal strokes, n (%)	191 (5.7)	482 (3.1)	120 (1.7)	<0.001	
Fatal ischaemic stroke, n (%)	96 (2.9)	245 (1.6)	52 (0.7)	<0.001	
Fatal haemorrhagic stroke, n (%)	53 (1.6)	154 (1.0)	45 (0.6)	<0.001	
Fatal unclassified stroke, n (%)	42 (1.3)	83 (0.5)	23 (0.3)	<0.001	
DWRT: Delayed Word Recall Test; hypertension: systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, and a medication or diagnosis; Diabetes: fasting blood glucose ≥7 and a medication or diagnosis; dyslipidaemia: total cholesterol ≥6.2 mmol/L, triglyceride ≥2.3 mmol/L, low density lipoprotein ≥4.1 mmol/L, high density lipoprotein <1.0 mmol/L, and a medication or diagnosis.

4.2 Cognitive impairment and the risk of fatal stroke occurrence

RCS showed a decreasing trend in relationship between the DWRT score and the risk of fatal strokes, with a cutoff value of the DWRT = 5 (Fig. 2A); participants with cognitive impairment shared increased risks of fatal strokes (aHR = 1.38, 95% CI1.16–1.64, P < 0.001), fatal ischaemic stroke (aHR = 1.39, 95% CI1.10–1.77, P = 0.007), and fatal unclassified stroke (aHR = 1.69, 95% CI1.16–2.46, P = 0.007), respectively, compared to those without cognitive impairment and after adjusting for a series of confounders (Fig. 3).Fig. 2 Associations of the DWRT score (A, n = 26,064) and the DWRT score change (B, n = 13,126) with the risk of fatal strokes in the Restricted Cubic Spline analysis with four knots. The solid line represents the adjusted hazard ratios, with dashed lines showing 95% confidence intervals. A multivariate adjustment was used, including sex, age, diabetes, hypertension, dyslipidaemia, smoking, alcohol consumption, physical activity, body mass index, sleep duration, and education.

Fig. 2

Fig. 3 Associations of DWRT score with the risk of fatal strokes in the GBCS participants (n = 26,064). The graph plots the unadjusted (Model 1) and the adjusted (Model 2) hazard ratios, 95% confidence intervals, and P values for DWRT scores. The adjusted confounders include sex, age, diabetes, hypertension, dyslipidaemia, smoking, alcohol consumption, physical activity, body mass index, sleep duration, and education.

Fig. 3

4.3 The longitudinal change in cognitive function and the risk of fatal stroke occurrence

RCS revealed an increased risk of fatal strokes among those with a DWRT decline but a gradually decreasing trend among those with a DWRT rise (Fig. 2B). According to the basic characteristics at the 1st follow-up (Fig. 1), participants with a DWRT decline were older, had more sleep duration, and fatal strokes, but had less hypertension, diabetes, dyslipidaemia, BMI≥28 kg/㎡, alcohol consumption (former and current), smoking (former and current), and active PA, compared to those with a DWRT rise (Table 2). Participants with a DWRT decline had an increased risk of fatal strokes (aHR = 1.45, 95% CI 1.10–1.91, P = 0.009) and fatal haemorrhagic stroke (aHR = 1.92, 95% CI 1.15–3.22, P = 0.01), compared to those with a DWRT rise and after adjusting for a range of confounders (Fig. 4). Participants with a DWRT rise had faintly decreased risks of fatal strokes (aHR = 0.74, 95% CI 0.52–1.04, P = 0.08) and fatal haemorrhagic stroke (aHR = 0.59, 95% CI 0.32–1.11, P = 0.10), compared to those with a DWRT stable and after adjusting for a range of confounders (Fig. 5).Table 2 Characteristics according to the DWRT changes of the GBCS participants at the 1st follow-up (n = 13,126).

Table 2Characteristics	the DWRT changes	P value	
rise	stable	decline	
>0	= 0	<0	
Number, n	6239	2478	4409		
Age (mean (SD))	64.7 (6.76)	64.6 (7.00)	65.1 (7.19)	0.003	
Sleep duration (mean (SD))	6.97 (1.19)	7.09 (1.25)	7.02 (1.25)	<0.001	
Sex, n (%)				<0.001	
 female	4540 (72.8%)	1807 (72.9%)	3041 (69.0%)		
 male	1699 (27.2%)	671 (27.1%)	1368 (31.0%)		
Body mass index, kg/㎡				0.537	
 <18.5	322 (5.16%)	119 (4.80%)	196 (4.45%)		
 18.5–23.9	3084 (49.4%)	1197 (48.3%)	2165 (49.1%)		
 24–27.9	2204 (35.3%)	891 (36.0%)	1577 (35.8%)		
 ≥ 28	629 (10.1%)	271 (10.9%)	471 (10.7%)		
Education, n (%)				0.506	
 primary	2275 (36.5%)	884 (35.7%)	1634 (37.1%)		
 middle school	3355 (53.8%)	1371 (55.3%)	2347 (53.2%)		
 college	609 (9.76%)	223 (9.00%)	428 (9.71%)		
Smoking, n (%)				0.03	
 never	5222 (83.7%)	2033 (82.0%)	3592 (81.5%)		
 former	506 (8.11%)	209 (8.43%)	396 (8.98%)		
 current	511 (8.19%)	236 (9.52%)	421 (9.55%)		
Alcohol consumption, n (%)				<0.001	
 never	2129 (34.1%)	945 (38.1%)	1732 (39.3%)		
 former	145 (2.32%)	76 (3.07%)	175 (3.97%)		
 current	3965 (63.6%)	1457 (58.8%)	2502 (56.7%)		
Physical activity, n (%)				0.017	
 inactive	79 (1.27%)	24 (0.97%)	43 (0.98%)		
 moderate	1207 (19.3%)	429 (17.3%)	764 (17.3%)		
 active	4953 (79.4%)	2025 (81.7%)	3602 (81.7%)		
Hypertension, n (%)				0.303	
 no	3337 (53.5%)	1290 (52.1%)	2301 (52.2%)		
 yes	2902 (46.5%)	1188 (47.9%)	2108 (47.8%)		
Dyslipidaemia, n (%)				0.892	
 no	2682 (43.0%)	1054 (42.5%)	1901 (43.1%)		
 yes	3557 (57.0%)	1424 (57.5%)	2508 (56.9%)		
Diabetes, n (%)				0.094	
 no	5426 (87.0%)	2151 (86.8%)	3772 (85.6%)		
 yes	813 (13.0%)	327 (13.2%)	637 (14.4%)		
Fatal strokes, n (%)	97 (1.55%)	50 (2.02%)	108 (2.45%)	0.004	
Fatal ischemic stroke, n (%)	56 (0.90%)	23 (0.93%)	50 (1.13%)	0.45	
Fatal hemorrhagic stroke, n (%)	25 (0.40%)	16 (0.65%)	36 (0.82%)	0.02	
Fatal unclassified stroke, n (%)	16 (0.26%)	11 (0.44%)	22 (0.50%)	0.10	
Hypertension: systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, medication or diagnosis; diabetes: fasting blood glucose ≥7, medication or diagnosis; dyslipidaemia: total cholesterol ≥6.2 mmol/L, triglyceride ≥2.3 mmol/L, low density lipoprotein ≥4.1 mmol/L, high density lipoprotein <1.0 mmol/L, and medication or diagnosis.

Fig. 4 Associations of the DWRT score change with the risk of fatal strokes in the GBCS participants (n = 13,126). The graph plots the unadjusted (Model 1) and the adjusted (Model 2) hazard ratios, 95% confidence intervals, and P values for the DWRT score changes. The adjusted confounders include sex, age, diabetes, hypertension, dyslipidaemia, smoking, alcohol consumption, physical activity, body mass index, sleep duration, and education.

Fig. 4

Fig. 5 Associations of the DWRT score change with the risk of fatal strokes in the GBCS participants (n = 13,126). The graph plots the adjusted and unadjusted hazard ratios, 95% confidence intervals, and P values for the DWRT score changes. The adjusted confounders include sex, age, diabetes, hypertension, dyslipidaemia, smoking, alcohol consumption, physical activity, body mass index, sleep duration, and education.

Fig. 5

5 Discussion

In this study, we found that cognitive impairment was associated with increased risks of fatal strokes and fatal ischaemic stroke; a DWRT decline was also associated with increased risks of fatal strokes and haemorrhagic stroke; these results were independent of age, sex, diabetes, hypertension, dyslipidaemia, smoking, alcohol consumption, body mass index, physical activity, sleep duration, and educational attainment. This is the first presentation of the association of longitudinal change in cognitive function with the risk of fatal strokes in aging populations.

The associations of cognitive function with stroke events have been released in recent decades. Patients with ischaemic or haemorrhagic stroke suffer commonly from cognitive impairment [25]; pre- and post-stroke MMSE (Simple Mental State Examination) scores decline in relation to ischaemic stroke [26]; post-stroke cognitive impairment linked to a poor outcome [27]. Such studies focused mainly on cognitive function and post-stroke events.

For pre-stroke cognitive impairment (PCI) on stroke events, patients with MCI shared a higher risk of ischaemic stroke [28], and 13.8% of patients with stroke in France suffered from MCI, wherein pre-existing MCI was associated with 90-day mortality [29]. A number of 30% of patients with ischaemic or haemorrhagic stroke in Norway suffered from PCI [30], wherein PCI was related to cerebrovascular lesions [31]. Severe cognitive impairment was related to a higher risk of stroke in the UK [32]. However, the patients with cognitive impairment and in stroke rehabilitation in the UK had fewer treatments than those without cognitive impairment [33]. PCI has been prevalent worldwide and should be taken its linkage with stroke events into account. Among a series of our findings, we paradoxically showed that cognitive impairment was associated with the increased risks of fatal strokes and fatal ischaemic stroke, and the longitudinal change in cognitive function was associated with the risk of not only fatal strokes but fatal haemorrhagic stroke. Such results suggest that cognitive impairment is a complex indicator to assess the risk of stroke in aging populations, this is possibly due to cognitive impairment being related to a decreased ability to care for oneself in daily activities and memory loss affecting the medications in stroke events [34], and then leads to an increased mortality risk in older populations [35,36].

The main strength of our work is that all analytic data especially the completed assessments of cognitive function were from GBCS. Robust inference of a random linkage of the risk of fatal stroke to pre-stroke cognitive impairment was observed in different outcomes by cognitive impairment and longitudinal changes in cognitive function. However, an unexpected result stated that more subjects (n = 6239, 47%) exhibited better performance in DWRT score changes (Table 2). As cognitive function normally decreases with aging, it's reasonable to assume those participants with slightly smaller year ages and more levels of PA, current alcohol consumption, and never smoking, apart from that the performance of cognitive tests might fluctuate.

There are several limitations to our study. First, the DWRT measurements may not be accurate and comprehensive enough for diagnosis; the Mini-Mental State Examination should be employed for further verification. Second, cognitive impairment includes a broad spectrum of mild- and moderate-cognitive impairment, and dementia although stratified analyses were not conducted because of a lack of data from other neuropsychological tests. Third, the subjects are from South China only and the number of strokes is not sufficient to be further tested and verified, therefore, more cases from GBCS or other national cohorts should be included in the future. Fourth, the longitudinal data from a starting point to follow-ups is lengthy due to the same performance as the baseline; more longitudinal changes in cognitive function should be assessed by linear mixed and group-based trajectory models in subsequent cohort studies. Fifth, the unclassified strokes limited the ability to address fatal strokes, especially ischaemic and haemorrhagic strokes.

6 Conclusion

Cognitive impairment was associated with increased risks of fatal strokes and fatal ischaemic stroke; the longitudinal change in DWRT decline was associated with increased risks of fatal strokes and fatal haemorrhagic stroke. Therefore, early screening of cognitive function should be conducive to predictive intervention in fatal stroke among relatively healthy middle-aged to older populations.

Funding

This work was supported by the 10.13039/501100010256 Guangzhou Municipal Science and Technology Project (2024A03J0427, 201704030132, and 2023A03J0508), the Major Science and Technology Project of Guangzhou Municipal Health Commission (2021A031003), the Guangzhou Municipal Key Medical Discipline (2021–2023), and the 10.13039/501100003785 Guangdong Medical Research Foundation (A2022209). The funders had no role in the study design, data collection or analysis, or preparation of the manuscript.

Data availability statement

Data presented in this study are included in the article/supp. Material/referenced in the article; further inquiries can be directed to the corresponding authors.

Ethics statement

The study was approved by the Guangzhou Medical Ethics Committee of the Chinese Medical Association (No. 2002Z2-E2051). Informed consent was obtained from all the participants and the guardians of dead participants before participation. All methods in this study were performed in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Author details

Jun-xiao Li MSc and Feng Zhu Ph.D.: Central Laboratory at Guangzhou Twelfth People's Hospital and Departments of Public Health and Preventive Medicine, Jinan University, Guangzhou; Qiong-qiong Zhong MSc: Departments of Public Health and Preventive Medicine, Jinan University, Guangzhou; Tong Zhu MSc, Ya-li Jin MSc, and Jing Pan Ph.D.: Department of Science and Education at Guangzhou Twelfth People's Hospital, Guangzhou, China.

CRediT authorship contribution statement

Jun-xiao Li: Writing – original draft, Methodology, Formal analysis, Data curation. Qiong-qiong Zhong: Validation, Methodology. Tong Zhu: Resources, Data curation. Ya-li Jin: Validation, Resources, Methodology, Data curation. Jing Pan: Writing – review & editing, Methodology. Shi-xiang Yuan: Validation, Funding acquisition. Feng Zhu: Writing – review & editing, Supervision, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Abbreviations

aHR adjusted hazard ratio

DWRT Delayed Word Recall Test

MCI mild cognitive impairment

ICD International Classification of Diseases

SD standard deviation

BMI body mass index

CI Confidence interval

MET metabolic energy

GBCS Guangzhou Biobank Cohort Study

GZCDC Guangzhou Centers for Disease Control and Prevention

RCS Restricted Cubic Spline.

Acknowledgments

The Guangzhou Biobank Cohort Study investigators included the Guangzhou Twelfth People's Hospital: Wei-sen Zhang, Min Cao, Tong Zhu, Bin Liu, and Cao-qiang Jiang (Co-PI); the University of Hong Kong: Lin Xu, C.M. Schooling, S.M. McGhee, G.M. Leung, R. Fielding, and Taihing Lam (Co-PI); the University of Birmingham: P. Adab, G. Neil Thomas, and Kar-keung Cheng (Co-PI).
==== Refs
References

1 Coupland A.P. The definition of stroke J. R. Soc. Med. 110 1 2017 9 12 28084167
2 Global, regional, and national burden of stroke and its risk factors, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019 Lancet Neurol. 20 10 2021 795 820 34487721
3 Ma Q. Temporal trend and attributable risk factors of stroke burden in China, 1990-2019: an analysis for the Global Burden of Disease Study 2019 Lancet Public Health 6 12 2021 e897 e906 34838196
4 Portegies M.L. Role of prestroke vascular pathology in long-term prognosis after stroke: the rotterdam study Stroke 47 1 2016 80 87 26604254
5 Paddick S.M. Mortality rates in community-dwelling Tanzanians with dementia and mild cognitive impairment: a 4-year follow-up study Age Ageing 44 4 2015 636 641 25918185
6 Bae J.B. Impact of mild cognitive impairment on mortality and cause of death in the elderly J Alzheimers Dis 64 2 2018 607 616 29914024
7 Domínguez-Chávez C.J. Murrock C.J. Salazar-González B.C. Mild cognitive impairment: a concept analysis Nurs. Forum 54 1 2019 68 76 30261109
8 Hénon H. Influence of prestroke dementia on early and delayed mortality in stroke patients J. Neurol. 250 1 2003 10 16 12527986
9 Campos C.G. Mild and moderate cognitive impairment and mortality among Brazilian older adults in long-term follow-up: the Bambui Health Aging Study Br. J. Psychiatr. 42 6 2020 583 590
10 An J. Cognitive impairment and risk of all-cause and cardiovascular disease mortality over 20-year follow-up: results from the BLSA J. Am. Heart Assoc. 7 15 2018 e008252
11 Li D. Cognition impairment and risk of subclinical cardiovascular disease in older adults: the atherosclerosis risk in communities study Front. Aging Neurosci. 14 2022 889543
12 Jokinen H. Post-stroke cognitive impairment is common even after successful clinical recovery Eur. J. Neurol. 22 9 2015 1288 1294 26040251
13 Douiri A. Rudd A.G. Wolfe C.D. Prevalence of poststroke cognitive impairment: South London stroke register 1995-2010 Stroke 44 1 2013 138 145 23150656
14 Jiang C. Cohort profile: the guangzhou biobank cohort study, a guangzhou-Hong Kong-Birmingham collaboration Int. J. Epidemiol. 35 4 2006 844 852 16844769
15 Huang Y. Mortality in relation to changes in physical activity in middle-aged to older Chinese: an 8-year follow-up of the Guangzhou Biobank Cohort Study J Sport Health Sci 10 4 2021 430 438 32827710
16 Jiang C.Q. Glycemic measures and risk of mortality in older Chinese: the guangzhou biobank cohort study J. Clin. Endocrinol. Metab. 105 3 2020
17 Prince M. Dementia diagnosis in developing countries: a cross-cultural validation study Lancet 361 9361 2003 909 917 12648969
18 Shankle W.R. Methods to improve the detection of mild cognitive impairment Proc. Natl. Acad. Sci. U. S. A. 102 13 2005 4919 4924 15781874
19 Xu L. Sleep duration and memory in the elderly Chinese: longitudinal analysis of the Guangzhou Biobank Cohort Study Sleep 37 11 2014 1737 1744 25364069
20 Xu L. Dose-response relation between physical activity and cognitive function: guangzhou biobank cohort study Ann. Epidemiol. 21 11 2011 857 863 21784658
21 Jin Y.L. Association of hand grip strength with mild cognitive impairment in middle-aged and older people in guangzhou biobank cohort study Int. J. Environ. Res. Publ. Health 19 11 2022
22 Lam K.B. Napping is associated with increased risk of type 2 diabetes: the Guangzhou Biobank Cohort Study Sleep 33 3 2010 402 407 20337199
23 Zhang W.S. Alcohol sensitivity, alcohol use and hypertension in an older Chinese population: the Guangzhou Biobank Cohort Study Hypertens. Res. 32 9 2009 741 747 19557005
24 Deng H.B. Reliability and validity of the IPAQ-Chinese: the guangzhou biobank cohort study Med. Sci. Sports Exerc. 40 2 2008 303 307 18202571
25 Sensenbrenner B. High prevalence of social cognition disorders and mild cognitive impairment long term after stroke Alzheimer Dis. Assoc. Disord. 34 1 2020 72 78 31633560
26 Hong C.T. Poststroke cognitive impairment: a longitudinal follow-up and pre/poststroke mini-mental state examination comparison Curr. Alzheimer Res.19 (10), 2022 716 723
27 Rohde D. The impact of cognitive impairment on poststroke outcomes: a 5-year follow-up J. Geriatr. Psychiatr. Neurol. 32 5 2019 275 281
28 Béjot Y. Influence of preexisting cognitive impairment on clinical severity of ischemic stroke: the dijon stroke registry Stroke 51 6 2020 1667 1673 32397931
29 Graber M. Influence of pre-existing mild cognitive impairment and dementia on post-stroke mortality. The dijon stroke registry Neuroepidemiology 54 6 2020 490 497 31593950
30 Banerjee G. Cognitive impairment before atrial fibrillation-related ischemic events: neuroimaging and prognostic associations J. Am. Heart Assoc. 9 1 2020 e014537
31 Schellhorn T. Pre-stroke cognitive impairment is associated with vascular imaging pathology: a prospective observational study BMC Geriatr. 21 1 2021 362 34126944
32 Stephan B.C.M. Potential value of impaired cognition in stroke prediction: a U.K. Population-based study J. Am. Geriatr. Soc. 65 8 2017 1756 1762 28369710
33 Longley V. Does pre-existing cognitive impairment impact on amount of stroke rehabilitation received? An observational cohort study Clin. Rehabil. 33 9 2019 1492 1502 31020850
34 McLennan S.N. Cognitive impairment predicts functional capacity in dementia-free patients with cardiovascular disease J. Cardiovasc. Nurs. 25 5 2010 390 397 20714236
35 Tangen G.G. Langballe E.M. Strand B.H. Subjective memory impairment, instrumental activities of daily living and longitudinal effect on mortality among older adults in a population-based cohort study: the HUNT Study Scand. J. Publ. Health 48 8 2020 825 831
36 Algarabel S. Recognition memory deficits in mild cognitive impairment Neuropsychol Dev Cogn B Aging Neuropsychol Cogn 19 5 2012 608 619 22247955
