Abstract
Objectives
Glucose fluctuations are more harmful than persistent hyperglycemia for chronic complications of diabetes. However, the relationship between cognition and glucose fluctuations in patients with acute ischemic stroke (AIS) complicated by type 2 diabetes mellitus (T2DM) remains unclear. We aimed to evaluate the association between short-term glucose fluctuations and cognition in patients with AIS complicated by T2DM.
Design
A cohort study with a 2-year follow-up.
Setting and participants
We included 554 patients with mild AIS (mean age: 62 years; 170 females and 384 males).
Measurements
Glucose variability (GV) was evaluated using glycated hemoglobin (HbA1c), stress hyperglycemia (SHR), standard deviation of blood glucose (SDBG), mean postprandial blood glucose (MPBG), mean amplitude of glycemic excursion (MAGE), and time in range (TIR). We evaluated the relationship between GV, fasting blood glucose (FBG) and cognition during the acute phase using linear regression analysis. We evaluated the relationship between GV, FBG and the occurrence of post-stroke cognitive impairment (PSCI) using a logistic regression model. Mediation analyses were fitted to explore whether the relationships of HbA1c with cognition were mediated by cerebral small vessel disease (CSVD).
Results
A clear pattern of age-related GV was observed. Higher SHR in middle-aged participants; higher HbA1c, and lower TIR in older participants; and higher MAGE, MPBG, and SDBG across a broad age range (50–80 years) were associated with cognitive impairment in the acute phase of AIS. Higher SHR and SDBG together with lower TIR in middle-aged participants, higher HbA1c in older participants, and higher FBG, MPBG, and MAGE across a broad age range (50–80 years) were associated with the occurrence of PSCI. The association between HbA1c and cognition was partially mediated (proportion: 7–16%) by CSVD.
Conclusions
Short-term glucose fluctuations are associated with cognition and a higher risk of PSCI in patients with AIS complicated by T2DM. CSVD might play an important role in the relationship between short-term glucose fluctuations and cognition.
Keywords: Acute ischemic stroke, Glucose variability, Cognitive impairment
1. Introduction
Stroke is the second leading cause of death worldwide and a major cause of disability [1]. Post-stroke cognitive impairment (PSCI), characterized by cognitive impairment that occurs after a stroke event and persists for up to 6 months [2], is becoming more prominent. Type 2 Diabetes Mellitus (T2DM) is prevalent among older adults, and accurate measurement of glycemic control is crucial. Glycated hemoglobin (HbA1c) is the gold standard for assessing glycemic control in patients with diabetes [3]; however, it does not capture acute fluctuations in blood glucose levels [4]. With the development of glucose monitoring technology, glucose variability (GV) has become a sensitive indicator for describing the pattern of glycemic control, considering all blood glucose concentrations over a given period to better account for glucose fluctuations [5].
In recent years, studies have indicated that a larger GV in Alzheimer’s disease (AD) is associated with neurological deterioration [6] and that diabetes is associated with an increased risk of dementia [7]. Furthermore, previous studies have demonstrated that long-term GV is associated with small atherosclerotic arteries in Chinese adults [8]. This suggests that the degree of GV may be closely related to the development of cognitive impairment in patients with acute ischemic stroke (AIS) combined with T2DM, possibly because glycemic fluctuations may contribute to cognitive impairment by exacerbating microvascular diseases of the central nervous system (CNS). Glycemic control involves several indicators, and middle and elderly patterns are always heterogeneous, suggesting that glycemic management should be evaluated more accurately. Because the diagnosis of diabetes alone cannot fully reflect changes in blood glucose levels, more glycemic indicators need to be explored in a large-sample study. More importantly, few studies have explored the association between GV and the development of cognitive impairment in patients with AIS. If the relationship between glycemic control and cognitive dysfunction in patients with AIS and T2DM is understood, it may be easier to identify high-risk groups and determine disease severity and prognosis. Therefore, this study aims to (i) investigate the correlation between GV and cognition in patients with AIS and T2DM; (ii) test whether the effects of GV on cognition are mediated by cerebral small vessel disease (CSVD) burden.
2. Material and methods
2.1. Study design and participants
We included 554 patients with mild AIS who were admitted to the Department of Neurology at the First Hospital of Jilin University between April 2019 and December 2022. (Registration number: 2018YFC1312301; Registration date: April 11, 2019). A flow diagram of the study is shown in Fig. 1. The inclusion criteria were as follows: (i) meeting the World Health Organization (WHO) diagnostic criteria for acute ischemic stroke [9] and being aged 50–80 years; (ii) presenting symptoms of focal neurological deficit and new infarcted lesions on brain magnetic resonance (MR) diffusion-weighted imaging (DWI); (iii) having a National Institutes of Health Stroke Scale (NIHSS) score ≤6. The exclusion criteria were as follows: (i) cognitive impairment due to a combination of medical or surgical systemic disease, infectious disease, toxic metabolic disease, or psychiatric disease; (ii) history of memory loss, diagnosis of cognitive impairment before stroke, or use of medication affecting cognition in the previous 2 weeks; (iii) suffering from severe schizophrenia, depression, and anxiety; and (iv) incomplete clinical baseline information. Participants who met the WHO diagnostic criteria [10] for diabetes or were undergoing hypoglycemic therapy for a history of hyperglycemia were included in the diabetic group (n = 252), and the rest were included in the non-diabetic group (n = 302). In this study, patients were followed up at 6 months ± 14 days and 1 year ± 14 days, and subjects were required to be visited and discharged from the group if a new ischemic event occurred. Upon discharge, a specialized neurologist and a nurse collaborate to deliver routine health education to both the patient and their family members, in accordance with the guidelines for primary care of ischemic stroke [11]. General information, laboratory data, neuropsychological scale scores, and imaging data were recorded during the follow-up period. The study protocol was approved by the Ethics Committee of the First Hospital of Jilin University and was conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants.
Fig. 1.
Flow chart illustrating the categorization of study participants.
Abbreviations: AIS, Acute ischemic stroke; PSNCI, Post-Stroke non-cognitive impairment; PSCI, Post-Stroke cognitive impairment.
2.2. Data acquisition
We collected clinical data from general patients during hospitalization, including age, sex, and education level, as well as vascular risk factors including smoking, drinking, history of stroke, hypertension, diabetes, dyslipidemia, atrial fibrillation, and coronary artery disease. Laboratory findings included fasting blood glucose (FBG) and glycated hemoglobin (HbA1c). Within 24 h of AIS onset, patients underwent a 12 -h overnight fast, after which blood samples were drawn from the antecubital vein in the morning. These samples were analyzed within 4 h of collection. FBG levels were determined using the glucose oxidase method with a Hitachi LABOSPECT 008As automatic biochemistry analyzer (Hitachi Limited, Tokyo, Japan). HbA1c levels were measured through high performance liquid chromatography using an ADAMSA1cHA-8180 fully automated HbA1c analyzer (ARKRAY Corporation, Kyoto, Japan).
2.3. GV evaluation
A total of 209 patients underwent regular fingertip 7-point blood glucose monitoring with pre-prandial, postprandial, and bedtime blood glucose monitoring, including blood glucose monitoring half an hour before and 2 h after three meals and 1 h before bedtime. A professional physician or nurse performed the entire procedure. The patient’s fingertip blood glucose level was monitored daily using a blood glucose meter (Huayi Jingdian Biotechnology Co, Beijing, China). The concentration of glucose in the participants’ fingertip capillaries was determined in vitro using blood glucose test strips manufactured by the company in conjunction with the glucose oxidase method. GV was assessed, including stress hyperglycemia (SHR), standard deviation of blood glucose (SDBG), mean postprandial blood glucose (MPBG), mean amplitude of glycemic excursion (MAGE), and time in range (TIR). SHR was defined as the admission glucose divided by the estimated average glucose derived from HbA1c, aiming to capture the relative increase in glucose attributed to the inflammatory and neurohormonal derangements that occur during a major illness. The formula for SHR [12] is as follows: SHR = FBG (admission FBG)/(1.59 * HbA1c - 2.59). SDBG was defined as the maximum value of the standard deviation of blood glucose at multiple points within a day during blood glucose monitoring. MAGE was determined by including for calculation only swings whose size was >1 standard deviation of the mean glycemic values obtained during the study period, and the average of the 3 days with the largest fluctuations was taken as the final MAGE. TIR was defined as the percentage of time during which blood glucose was within the target range (3.9–10.0 mmol/L) for 24 h, and the average TIR of the period was taken as the final. The average number of days of fingertip glucose monitoring was 5.52 days (range 2–15 days).
2.4. Imaging evaluation
All participants underwent a 3.0 T cranial MR with sequences including T1-weighted imaging, T2-weighted imaging, DWI, fluid-attenuated inversion recovery sequences, and magnetic susceptibility-weighted imaging. The patients were graded according to the overall CSVD imaging burden score proposed by Staals [13]: (i) presence of lacunar infarctions (LIs) (1 point); (ii) presence of cerebral microbleeds (CMB) (1 point); (iii) presence of enlarged perivascular spaces (EPVS)was counted if there were moderate to severe (grade 2–4) EPVS in the basal ganglia (1 point); (iv) presence of white matter hyperintensities (WMH) was defined as either (early) confluent deep WMH (Fazekas score 2 or 3) or irregular periventricular WMH extending into the deep white matter (Fazekas score 3) (1 point).
2.5. Cognition evaluation
According to the Expert Consensus on the Management of Cognitive Impairment after Stroke 2021, the section on neuropsychological assessment in the acute phase of stroke [14], cognitive psychological assessments were conducted on all included patients on day 7 following AIS. This time frame was chosen when patients’ cerebral infarction symptoms had stabilized. Assessments were performed by psychometrically qualified assessors who were blinded to blood glucose levels. The assessments took place in a relatively quiet environment, ensuring optimal concentration, with no other patients present simultaneously. There was no increase in NIHSS scores after admission of these patients with mild stroke (Those patients with increased scores have been excluded). Additionally, patients with HAMD and HAMA scores ≥7 were excluded to minimize the potential impact of post-stroke anxiety and depressive states on cognitive function assessment. To ensure the validity and reliability of the assessment process, specialized attending neurologists and neuropsychological evaluators collaboratively licensed the patients to access the above assessment. The overall cognitive function was assessed using the Montreal Cognitive Assessment (MoCA), with a total score of 30. The MoCA is strongly correlated with educational level. Studies have revealed that the corrected MoCA score increases by 1 when the duration of education is <12 years [15]. We classified the study subjects into PSCI group (n = 85, MoCA <24 for more than 12 years of education, MoCA <22 for 7–12 years of education, MoCA <19 for no more than 6 years of education) and post-stroke non-cognitive impairment (PSNCI) group (n = 111, MoCA ≥24 for more than 12 years of education, MoCA ≥22 for 7–12 years of education, MoCA ≥19 for no more than 6 years of education) based on the follow-up MoCA score. The PSNCI group was divided into a cognitive function decline group (n = 35) and a cognitive function non-decline group (n = 50) based on whether there was a decline in the MoCA during follow-up. A cut-off of 65 years was proposed to determine whether participants were categorized as middle-aged (<65 years: n = 313) or older (≥65 years: n = 241), following the WHO’s definition of older people as those aged ≥65 years. This population is characterized by a gradual decline in the function of all organs of the body and an increased prevalence of various underlying diseases [16].
2.6. Statistical analysis
Statistical analyses were performed using Statistical Package for the Social Sciences (SPSS) software, version 26.0 (IBM, New York, NY, USA), and demographic characteristics were evaluated using the chi-square test (for categorical variables) and the Mann-Whitney U test or t-test (for continuous variables). The correlation between GV and cognition was further validated in a multiple linear regression (MLR) model after adjusting for covariates (including age, sex, years of education, disease history, and NIHSS) using the non-diabetic group as the reference. This involved exploring each of the six GV variables (HbA1c, SHR, SDBG, MPBG, MAGE, and TIR), and FBG with cognition. Logistic regression analyses were used to verify the correlation between GV and the occurrence of PSCI and whether MoCA decreased in the PSNCI group. A two-tailed P < 0.05 was considered statistically significant. To explore whether the relationship between HbA1c and cognition was mediated by CSVD, mediation analyses were performed using the method proposed by Baron and Kenny [17]. Mediation effects were established if the following criteria were simultaneously met: (i) HbA1c was significantly associated with CSVD; (ii) HbA1c was significantly associated with cognition; (iii) CSVD was significantly associated with cognition; and (iv) the associations of HbA1c with cognition were attenuated when CSVD (the mediator) were added in the regression model. The model was corrected for age, sex, years of education, and stroke history.
3. Results
3.1. Baseline characteristics of participants
Table 1 presents the characteristics of the participants. Among the 554 patients, 252 were in the diabetic group, and 302 were in the non-diabetic group. The mean age of the participants was 62 years (SD = 7.51), and 170 (30.69%) were female. Comparisons of age, sex, vascular risk factors (history of hypertension, hyperlipidemia, coronary artery disease, atrial fibrillation, history of smoking, and alcohol consumption), NIHSS scores, and Mini-Mental State Examination scores between the two groups did not show any statistically significant differences (P > 0.05). Compared to the non-diabetic group, patients with diabetes demonstrated a better level of education and were more likely to have hyperlipidemia (P < 0.001). MoCA scores were lower in the diabetes group (P < 0.001).
Table 1.
Baseline characteristics of different groups.
| Variable | Diabetes | Non-diabetes | Total | P |
|---|---|---|---|---|
| Demographic characteristics | ||||
| N | 252 | 302 | 554 | |
| Age, mean (SD) | 62.63 (7.30) | 62.7 (7.70) | 62.7 (7.51) | 0.907 |
| Sex, (male/female) | 178/74 | 206/96 | 384/170 | 0.538 |
| Education level, year mean (SD) | 10.88 (4.05) | 9.54 (4.20) | 10.51 (5.18) | <0.001 |
| Vascular risk factors | ||||
| Hypertension, n (%) | 85 (33.73) | 126 (41.72) | 211 (38.09) | 0.091 |
| Hyperlipemia, n (%) | 78 (30.95) | 48 (15.89) | 126 (22.74) | <0.001 |
| Previous stroke, n (%) | 90 (35.71) | 81 (26.82) | 171 (30.87) | 0.552 |
| Coronary artery disease, n (%) | 27 (10.71) | 21 (6.95) | 48 (8.66) | 0.208 |
| Atrial fibrillation, n (%) | 5 (1.98) | 7 (2.32) | 12 (2.17) | 0.564 |
| Current smoker, n (%) | 111 (44.04) | 156 (51.66) | 267 (48.19) | 0.063 |
| Regular alcohol user, n (%) | 134 (53.17) | 177 (58.61) | 311 (5.61) | 0.224 |
| Clinical assessment | ||||
| NIHSS (IQR) | 2.4 (1−3) | 2.2 (1−3) | 2.3 (1−3) | 0.247 |
| MMSE (IQR) | 25.39 (22.75−28) | 24.87 (22−28) | 25.11 (23−28) | 0.248 |
| MoCA (IQR) | 17.96 (13−21) | 19.32 (15−24) | 18.52 (16−24) | <0.001 |
| FBG (IQR) | 7.99 (6.26−8.83) | 5.02 (4.75−5.49) | 6.37 (4.99−6.86) | <0.001 |
| HbA1c (IQR) | 7.6 (6.5−8.4) | 5.7 (5.5−6.0) | 6.6 (5.6−7.2) | <0.001 |
Abbreviations: SD = Standard deviation; NIHSS = National Institutes of Health Stroke Scale; IQR = Interquartile range; MMSE = Mini-Mental State Examination; MoCA = Montreal Cognitive Assessment; FBG = fasting blood glucose; HbA1c = glycated hemoglobin.
3.2. Relationship between blood glucose and cognitive function in the acute phase of AIS
Among all participants, cognitive levels were lower in the diabetes group (Fig. 2A, P < 0.001), and in the subgroups stratified by age, the differences remained significant only in middle-aged participants (Fig. 2B, P = 0.030). In the diabetes group, higher HbA1c was associated with cognitive decline (β = −0.462, P = 0.036), and in the subgroups stratified by age, it remained significant only in older participants (β = −0.740, P = 0.043). No association was observed between FBG levels and cognitive function. In the non-diabetic group, higher HbA1c was associated with cognitive decline (β = −0.824, P = 0.046). In the age-stratified subgroup, it remained significant only in middle-aged participants (β = −0.134, P = 0.031). The non-diabetic group was divided into lower glycemia (n = 130) and higher glycemia (n = 172) groups according to the mean FBG (5.02 mmol/L). In the lower glycemia group, lower FBG was associated with cognitive decline (β = 0.973, P = 0.046), which remained significant only in older participants (β = 0.740, P = 0.043). No association was found between the FBG levels and cognitive function in the higher glycemia group. Additionally, these differences remained significant after adjusting for various covariates (P < 0.05). (Table 2)
Fig. 2.
Comparison of MoCA between diabetes and non-diabetes groups (A–B).
Abbreviations: MoCA = Montreal Cognitive Assessment.
Table 2.
Relationship between blood glucose and cognitive function in the acute phase of AIS.
| Total |
Middle-aged participants |
Older participants |
|||||
|---|---|---|---|---|---|---|---|
| Variable | β | P | β | P | β | P | |
| Diabetes | |||||||
| FBG | 0.017 | 0.881 | 0.056 | 0.703 | 0.013 | 0.947 | |
| HbA1c | −0.462 | 0.036 | −0.291 | 0.288 | −0.740 | 0.043 | |
| Non-diabetes | |||||||
| lower glycemia | FBG | 0.973 | 0.046 | 0.155 | 0.791 | 0.74 | 0.043 |
| higher glycemia | FBG | −1.604 | 0.309 | 0.522 | 0.787 | −0.511 | 0.067 |
| HbA1c | −0.824 | 0.046 | −0.134 | 0.031 | −0.134 | 0.575 | |
Abbreviations: FBG = fasting blood glucose; HbA1c = glycated hemoglobin.
A higher HbA1c level was associated with an increased CSVD burden (a = 0.099, P = 0.003), which remained significant in the subgroups stratified by age for both middle-aged (a = 0.151, P = 0.002) and older participants (a = 0.050, P = 0.024). HbA1c appears to be a better indicator of long-term control of blood glucose. Therefore, it was used to perform the subsequent mediation analyses. The results showed that the association between HbA1c and cognitive impairment was partly mediated by the burden of CSVD, with a mediation proportion of approximately 13.3%, which was significant for both middle-aged (6.2%) and older (15.7%) participants. (Fig. 3A–C)
Fig. 3.
Mediation effects of CSVD on cognition (A–C).
Abbreviations: MoCA = Montreal Cognitive Assessment; HbA1c = glycated hemoglobin; CSVD = cerebral small vessel disease.
3.3. Relationship between GV and cognitive function in the acute phase of AIS
The above results indicate that diabetes and poor glycemic control are important risk factors for cognitive impairment. Therefore, we investigated the relationship between the control pattern of blood glucose level and cognition. Higher SHR was associated with cognitive decline (β = −0.157, P = 0.014), which remained significant only in middle-aged participants (β = −0.16, P = 0.044). Lower TIR was associated with cognitive decline (β = 0.217, P = 0.019), which remained significant only in older participants (β = 0.138, P = 0.024). Higher MPBG was associated with cognitive decline (β = −0.430, P < 0.001), which remained significant both in middle-aged participants (β = −0.330, P < 0.001) and older participants (β = −0.499, P < 0.001). Higher MAGE was associated with cognitive decline (β = −0.762, P < 0.001), which remained significant for both middle-aged (β = −0.723, P < 0.001) and older participants (β = −0.782, P = 0.004). Higher SDBG was associated with cognitive decline (β = −0.575, P < 0.001), which remained significant for both middle-aged (β = −0.505, P < 0.001) and older participants (β = −0.350, P < 0.001). All the differences remained significant after adjusting for various covariates. (P < 0.05), respectively (Table 3, Fig. 4A–F)
Table 3.
Associations between GV and cognition in the acute phase of AIS.
| Variable | Total |
Middle-aged participants |
Older participants |
|||
|---|---|---|---|---|---|---|
| β | P | β | P | β | P | |
| SHR | −0.157 | 0.014 | −0.16 | 0.044 | −0.076 | 0.466 |
| TIR | 0.217 | 0.019 | −0.217 | 0.47 | 0.138 | 0.024 |
| MPBG | −0.43 | <0.001 | −0.33 | <0.001 | −0.499 | <0.001 |
| MAGE | −0.762 | <0.001 | −0.723 | <0.001 | −0.782 | 0.004 |
| SDBG | −0.575 | <0.001 | −0.505 | <0.001 | −0.35 | <0.001 |
Abbreviations: SHR = stress hyperglycemia; SDBG = standard deviation of blood glucose; MPBG = mean postprandial blood glucose; LAGE = largest amplitude of glucose; MAGE = mean amplitude of glycemic excursion; TIR = time in range.
Fig. 4.
Associations between GV (SHR, TIR, MPBG, MAGE, SDBG) and cognition in the acute phase of AIS (A–F).
Abbreviations: SHR = stress hyperglycemia; MPBG = mean postprandial blood glucose; LAGE = largest amplitude of glucose; MAGE = mean amplitude of glycemic excursion; TIR = time in range; MoCA = Montreal Cognitive Assessment.
3.4. Relationship between GV and the occurrence of PSCI
The above results suggest that GV is an important factor in cognitive impairment. Therefore, we investigated the relationship between the occurrence of PSCI and GV during the recovery period. Table 4 shows the characteristics of participants in the PSCI and PSNCI groups. There were 92 patients (30.46%) with PSCI in the non-diabetic group, and 35 were female (38.04%), with a mean age of 62.8. There were 111 patients (56.63%) with PSCI in the diabetic group, and 39 were female (35.14%), with a mean age of 63.6. In the non-diabetic group, no association was found between the occurrence of PSCI and HbA1c levels. The non-diabetic group was further divided into lower glycemia and higher glycemia groups based on the mean FBG (5.02 mmol/L). In the lower glycemia group, lower FBG was associated with an increased incidence of PSCI (OR = 1.470, P = 0.014, 95%CI: 0.170−0.297), which remained significant only in older participants (OR = 1.627, P = 0.033, 95%CI: 0.241−0.631). No association was found between FBG levels and the occurrence of PSCI in the higher glycemia group. In the diabetic group, higher FBG was associated with an increased incidence of PSCI (OR = 1.313, P < 0.001, 95%CI: 1.158–1.490). This finding was significant for both middle-aged (OR = 1.377, P < 0.001, 95%CI: 1.152–1.647) and older participants (OR = 1.246, P = 0.045, 95%CI: 1.044–1.487). Higher HbA1c levels were associated with an increased incidence of PSCI (OR = 1.048, P = 0.047), which was significant only in older participants (OR = 1.919, P = 0.049). Higher SHR was associated with an increased incidence of PSCI (OR = 53.692, P < 0.001), which was significant only in middle-aged participants (OR = 824.245, P < 0.001). Lower TIR was associated with an increased incidence of PSCI (OR = 0.628, P = 0.039), which was significant only in middle-aged participants (OR = 0.472, P = 0.018). Higher MPBG was associated with an increased incidence of PSCI (OR = 1.658, P < 0.001), which remained significant for middle-aged (OR = 1.547, P < 0.001) and older participants (OR = 1.818, P < 0.001). Higher MAGE was associated with an increased incidence of PSCI (OR = 2.697, P < 0.001), which was significant for both middle-aged (OR = 3.455, P < 0.001) and older participants (OR = 2.489, P < 0.001). SDBG was associated with an increased incidence of PSCI only in middle-aged participants (OR = 11.525, P < 0.001). All the differences remained significant after adjusting for various covariates. (P < 0.05) (Table 5).
Table 4.
Follow-up characteristics between the PSCI and PSNCI groups.
| Variable | PSNCI | PSCI | total | P |
|---|---|---|---|---|
| Demographic characteristics | ||||
| N | 85 | 111 | 196 | – |
| Age, mean (SD) | 60.9 (7.13) | 63.6 (7.21) | 62.4 (7.19) | 0.012 |
| Sex, (male/female) | 66/19 | 72/39 | 138/58 | 0.052 |
| Education, year mean (SD) | 10.88 (4.06) | 9.54 (4.20) | 11.38 (3.33) | <0.001 |
| Vascular risk factors | ||||
| Hypertension, n (%) | 54 (63.53) | 71 (63.96) | 125 (63.78) | 0.508 |
| Previous stroke, n (%) | 20 (23.53) | 53 (47.75) | 73 (37.24) | 0.001 |
| Hyperlipemia, n (%) | 26 (30.59) | 31 (27.93) | 57 (29.08) | 0.514 |
| Coronary artery disease, n (%) | 13 (15.29) | 10 (9.00) | 23 (11.73) | 0.389 |
| Atrial fibrillation, n (%) | 2 (2.35) | 4 (3.6) | 6 (3.06) | 0.361 |
| Current smoker, n (%) | 44 (51.76) | 49 (44.14) | 93 (47.45) | 0.317 |
| Regular alcohol user, n (%) | 40 (47.06) | 55 (49.55) | 95 (48.47) | 0.721 |
| Clinical assessment | ||||
| NIHSS (IQR) | 1.9 (1−3) | 2.5 (1−3) | 2.2 (1−3) | 0.007 |
| MMSE (IQR) | 27.74 (27−29) | 24.96 (20−27) | 25.24 (23−28) | 0.067 |
Abbreviations: SD = Standard deviation; NIHSS = National Institutes of Health Stroke Scale; IQR = Interquartile range; MMSE = Mini-mental State Examination; PSNCI = Post-Stroke non-cognitive impairment; PSCI = Post-Stroke cognitive impairment.
Table 5.
Associations between GV and the occurrence of PSCI.
| Variable | Total |
Middle-aged participants |
Older participants |
||||||
|---|---|---|---|---|---|---|---|---|---|
| OR | P | 95% CI | OR | P | 95% CI | OR | P | 95% CI | |
| HbA1c | 1.048 | 0.047 | 1.208−1.818 | 0.966 | 0.715 | 0.802−1.163 | 1.919 | 0.049 | 1.168−1.724 |
| SHR | 53.629 | <0.001 | 10.680−269.302 | 824.245 | <0.001 | 51.408−13215.425 | 5.341 | 0.078 | 0.827−34.499 |
| TIR | 0.628 | 0.039 | 0.372−0.975 | 0.472 | 0.018 | 0.253−2.333 | 1.300 | 0.615 | 0.467−3.615 |
| MPBG | 1.658 | <0.001 | 1.388−1.982 | 1.547 | <0.001 | 1.242−1.928 | 1.818 | <0.001 | 1.331−2.483 |
| MAGE | 2.697 | <0.001 | 2.047−3.555 | 3.455 | <0.001 | 2.250−5.305 | 2.489 | <0.001 | 1.628−3.806 |
| SDBG | 1.582 | 0.497 | 0.421−5.937 | 11.525 | <0.001 | 4.823−27.541 | 1.149 | 0.919 | 0.078−17.028 |
Abbreviations: HbA1c = glycated hemoglobin; SHR = stress hyperglycemia; SDBG = standard deviation of blood glucose; MPBG = mean postprandial blood glucose; LAGE = largest amplitude of glucose; MAGE = mean amplitude of glycemic excursion; TIR = time in range; OR = Odds ratio; CI = Confidence interval.
3.5. Relationship between GV and the occurrence of cognitive decline in the PSNCI group
The above results indicate that GV is an important risk factor for the occurrence of PSCI. Therefore, we aimed to determine the relationship between GV and the occurrence of cognitive decline in patients with PSNCI during the recovery period from stroke. In the non-diabetic group, no relationship was found between HbA1c or FBG levels and the occurrence of cognitive decline in the PSNCI group. In the diabetic group, higher HbA1c levels were associated with an increased incidence of cognitive decline in the PSNCI group (OR = 1.615, P = 0.004). This finding was significant in both middle-aged (OR = 1.494, P = 0.046) and older participants (OR = 1.885, P = 0.029). A higher MAGE score was associated with an increased incidence of cognitive decline in the PSNCI group (OR = 1.764, P = 0.003) and was only significant in middle-aged participants (OR = 2.243, P = 0.003). No associations were found between the FBG, TIR, MPBG, SHR, or SDBG levels and the incidence of cognitive decline in the PSNCI group. (Supplementary Table S1).
4. Discussion
This large-scale study systematically explored the relationship between GV, CSVD, and cognitive impairment in patients with mild AIS. Our results clearly show that glycemic control is associated not only with the cognitive level of patients during the acute phase of AIS but also with the probability of developing PSCI and cognitive decline during the recovery period for PSNCI. Our findings suggest that CSVD partially mediates the effects of HbA1c on cognitive function, indicating an association between GV and CSVD. Notably, these associations were significantly dependent on age and glycemic control patterns.
First, by comparing the results of different GV indices, we clearly delineated the age-related patterns of GV with cognition and CSVD burden in patients with acute AIS. In the non-diabetic group, higher HbA1c levels (especially in middle-aged participants) and lower FBG levels (especially in older participants with below-average FBG levels) were risk factors for cognitive impairment. In the diabetic group, acute glycemic fluctuation after AIS was a risk factor for cognitive impairment. Specifically, higher GV, especially higher MAGE, SDBG, and MPBG, are risk factors for cognitive impairment, showing a strong association with cognitive impairment across a broad age range (50–80 years). Higher SHR was also identified as a risk factor for cognitive impairment, although this association became less significant in older participants. Higher HbA1c and lower TIR were risk factors for cognitive impairment, but this association became less significant in middle-aged participants. Additionally, our study estimated that CSVD might partially mediate the effect of HbA1c on cognition by approximately 13%. Blood glucose fluctuations may contribute to cognitive dysfunction by exacerbating the CSVD burden. Subsequently, we investigated the age-related patterns of GV and the occurrence of PSCI during the patient's recovery period. In the non-diabetic population, we found that a lower FBG level (especially in older participants with below-average FBG levels) was a risk factor for the occurrence of PSCI. Conversely, in the diabetic group, higher SHR and SDBG together with lower TIR were identified as risk factors for the development of PSCI in middle-aged participants. Moreover, higher HbA1c was a risk factor for the development of PSCI in older participants. Additionally, higher MPBG, MAGE, and FBG showed a strong association with the occurrence of PSCI across a wide age range. Furthermore, we observed that higher MAGE and HbA1c levels were also risk factors for cognitive decline during recovery from AIS combined with diabetes in middle-aged participants with PSNCI. Overall, our findings suggest that long-term unsatisfactory glycemic control, stress hyperglycemia, acute elevation of postprandial glucose, substantial glucose fluctuations, and excessively low FBG levels in older participants are all risk factors for an unfavorable prognosis in patients recovering from AIS.
Currently, the evaluation of blood glucose fluctuation indicators is complex and difficult to implement in clinical practice, and there is no uniform international standard. MAGE is regarded as the "gold standard" for the detection of GV, as it accurately reflects blood glucose fluctuations rather than merely being a statistically dispersed characteristic after removing all minor blood glucose fluctuations with amplitude below a certain threshold value. The development of continuous glucose monitoring (CGM) has increased the accuracy of MAGE calculations. However, CGM is expensive, relatively complex to perform, and not universally available among the diabetic population, especially after discharge from the hospital. Among individuals with diabetes, seven-finger prick glucose monitoring using a glucometer remains more common. Previous studies [18] have shown that the error in data from seven glucose monitoring sessions used to calculate MAGE is within acceptable limits and has a good correlation with the CGM [19]. In this present study, MAGE, compared to SDBG, could better predict acute intraday blood glucose fluctuations in patients and was related to cognition in the acute phase and PSCI. The significant correlation between MAGE and cognition in patients in the acute phase, occurrence of PSCI, and cognitive decline in PSNCI further supports the notion that vascular damage is more pronounced with major glucose fluctuations associated with meals. In our study, 87.3% (220/252) of patients with AIS combined with T2DM had above-normal HbA1c levels, indicating that most patients do not have optimal glycemic control and that cognitive and microvascular benefits can be obtained by implementing a more finely stratified management program in this patient population. This study suggests that glycemic control should be more stringent in middle-aged participants, as SHR, MPBG, and MAGE were associated with cognitive impairment in both the acute and recovery phases. This indicates that acute fluctuations in blood glucose with postprandial hyperglycemia are more detrimental to cognitive function in middle-aged participants after an acute episode of AIS. In older participants, glycemic control can be adjusted appropriately [20]. HbA1c, MPBG, and MAGE were associated with cognitive impairment in both the acute and recovery phases, suggesting that postprandial hyperglycemia management, particularly acute elevations, and maintaining mean glucose levels over time should be prioritized for older patients. Previous studies [21] have also shown that postprandial plasma glucose excursions and HbA1c levels are associated with impairment in global, executive, and attention functioning in older patients with T2DM. Large cohort studies have further supported that the coefficient of variation of HbA1c and postprandial glucose levels increases the risk of AD in patients with diabetes [22]. In addition to the exploration of short-term glucose fluctuations for AIS complicated by T2DM, our study has also found that for non-diabetic patients, managing HbA1c levels and preventing hypoglycemia in older participants should be emphasized. This has been confirmed by studies in which even a single episode of severe hypoglycemia or a milder degree of hypoglycemia has been found to be independently associated with a worse prognosis in AIS.
Fluctuating blood glucose levels are more detrimental to cerebral function and neurological inflammation compared to sustained high or low blood glucose levels [23] and are associated with cognition [24]. While the specific mechanisms through which acute fluctuations in blood glucose levels affect cognition remain unclear, several hypotheses have been proposed. First, blood glucose fluctuations can induce oxidative stress [25]. The acute transition from normoglycemia to hyperglycemia triggers oxidative stress in microglia, which induces HSP70 and HO-1 expression to resist stress injury, as observed in animal experiments using a diabetic mouse model [26]. Second, fluctuations in blood glucose levels can damage the vascular endothelium [27]. An experimental study on the effects of glucose fluctuations on endothelial cell function in T2DM rats confirmed that high glucose fluctuations promote the production of nitrotyrosine and 8-hydroxydeoxyguanosine (8-OHdG) [28]. Through the poly ADP ribose polymerase pathway, glucose fluctuations generate and promote the synthesis of ROS in the mitochondrial respiratory chain, enhance oxidative stress, and induce endothelial cell apoptosis. This, in patients with diabetes, acute glucose fluctuations induce more severe oxidative stress and endothelial dysfunction than steady high glucose levels. Third, fluctuations in blood glucose levels activate the inflammatory response [29]. The shift from normal to high glucose levels induces inflammatory stress and activates microglia. Excessive ROS induced by dysglycemia can activate redox-sensitive proinflammatory transcription factors, such as NF-κB, activator protein-1, and early growth response-1. The genes regulated by these transcription factors are inducible, as evidenced by the increased expression of various proinflammatory cytokines and chemokines (e.g., TNF-α, IL-6, and MCP-1) at the mRNA level in both monocytes and plasma. Thus, diabetic inflammation may occur [26]. In addition, the breakdown of the blood-brain barrier may lead to the infiltration of neurotoxic blood products into the parenchyma, thereby exacerbating a range of pathological processes. However, it is important to acknowledge some limitations of our study. First, this study was a single center study, which may limit the generalizability of the findings. To ensure the robustness of our results, future research should involve larger sample sizes and multiple centers. Second, our investigation was limited to PSCI. To gain a more comprehensive understanding of the relationship between GV and cognitive function, future studies should explore diverse types of dementia, such as AD, frontotemporal dementia, and dementia with Lewy bodies, and others. This broader approach will provide insights into the specific nuances of GV across different forms of cognitive impairment.
5. Conclusions
The findings of this study are relevant to the clinical management of AIS combined with T2DM, suggesting that assessing GV in addition to blood glucose levels can be beneficial. These findings provide new insights into glycemic control and PSCI prevention in diabetes care. In summary: (i) screening for risk factors of cognitive impairment may go beyond the diagnosis of diabetes or reliance on a single glycemic indicator. A broader range of indicators related to blood glucose fluctuations and the stress response following acute CNS events could be considered. If it is difficult to achieve blood glucose control before and after all three meals, this can be achieved by developing different blood glucose management priorities for different subgroups stratified by age. This will hopefully reduce the incidence of PSCI and complications of diabetes. (ii) Similarly, the monitoring indicators for glucose-lowering therapy in cognitively impaired patients should extend beyond the focus on lowering FBG and PBG. Attention should be given to fluctuating blood glucose levels, and management should be stratified based on age. Controlling daily blood glucose fluctuations is crucial for achieving HbA1C standards and preventing diabetic complications; (iii) there may be different mechanisms by which different indicators of GV affect cognitive impairment. Further studies are required to elucidate these mechanisms.
Author contributions
Z-RL, W-CX, Z-M and S-L conceived the protocol. Z-RL and W-CX contributed to the analysis and interpretation of data. Z-RL and W-CX grafted the manuscript. Z-RL, W-CX, Z-M and S-L critically revised the manuscript. All authors agree to be fully accountable for ensuring the integrity and accuracy of the work and read and approved the final manuscript.
Fundings
This study was supported by the General Program of the National Natural Science Foundation of China (No. 82071442), the Jilin Provincial Department of Finance (JLSWSRCZX2021-004), the Major Chronic Disease Program of the Ministry of Science and Technology of China (No. 2018YFC1312301), and STI2030-Major Projects (No. 2021ZD0201802).
Declaration
The experiments comply with the current laws of our country. The study protocol was approved by the Ethics Committee of the First Hospital of Jilin University and was conducted according to the principles of the Declaration of Helsinki.
Conflict of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest
Acknowledgements
We would like to thank Editage (www.editage.cn) for English language editing.
Footnotes
Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.jnha.2024.100252.
Appendix A. Supplementary data
The following is Supplementary data to this article:
References
- 1.Regenhardt R.W., Das A.S., Stapleton C.J., Chandra R.V., Rabinov J.D., Patel A.B., et al. Blood pressure and penumbral sustenance in stroke from large vessel occlusion. Front Neurol. 2017;8:317. doi: 10.3389/fneur.2017.00317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Dong Q.W.K. Expert consensus on post-stroke cognitive impairment. Chin J Stroke. 2021;16(4):376–389. doi: 10.3969/j.issn.1673-5765.2021.04.011. [DOI] [Google Scholar]
- 3.American Diabetes Association Professional Practice C 2. Classification and diagnosis of diabetes: standards of medical care in diabetes-2022. Diabetes care. 2022;45(Suppl 1):S17. doi: 10.2337/dc22-S002. [DOI] [PubMed] [Google Scholar]
- 4.Kovatchev B., Cobelli C. Glucose variability: timing, risk analysis, and relationship to hypoglycemia in diabetes. Diabetes Care. 2016;39(4):502–510. doi: 10.2337/dc15-2035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sparks J.R., Kishman E.E., Sarzynski M.A., Davis J.M., Grandjean P.W., Durstine J.L., et al. Glycemic variability: importance, relationship with physical activity, and the influence of exercise. Sports Med Health Sci. 2021;3(4):183–193. doi: 10.1016/j.smhs.2021.09.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Gordon Weston R., Salamo Russell M., Behera A., Chibnall J., Alshekhlee A., Callison Richard C., et al. Association of blood glucose and clinical outcome after mechanical thrombectomy for acute ischemic stroke. Intervent Neurol. 2018;7(3-4):182–188. doi: 10.1159/000486456. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Zhao Q., Zhang Y., Liao X., Wang W. Executive function and diabetes: a clinical neuropsychology perspective. Front Psychol. 2020;11:2112. doi: 10.3389/fpsyg.2020.02112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Zhang Y., Wu S., Li M., Wang T., Xu M., Lu J., et al. Long-term glycemic variability is associated with arterial stiffness in chinese adults. Front Endocrinol. 2021;12 doi: 10.3389/fendo.2021.711540. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Stroke--1989, Recommendations on stroke prevention, diagnosis, and therapy: report of the who task force on stroke and other cerebrovascular disorders. Stroke. 1989;20(10):1407–1431. doi: 10.1161/01.STR.20.10.1407. [DOI] [PubMed] [Google Scholar]
- 10.Alberti K.G.M.M., Zimmet P.Z., Consultation W.H.O. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus. Provisional report of a who consultation. Diabet Med. 1998;15(7):539–553. doi: 10.1002/(SICI)1096-9136(199807)15:7<539::AID-DIA668>3.0.CO;2-S. doi: 10.1002/(SICI)1096-9136(199807)15:7<539::AID-DIA668>3.0.CO;2-S. [DOI] [PubMed] [Google Scholar]
- 11.Association CM, House CMJP, Practice CSoG, Association EBoCJoGPoCM, Disease EGoGfPCoNS Guideline for primary care of ischemic stroke (2021) J Clin Endocrinol Metab. 2021;(09):927–946. [Google Scholar]
- 12.Roberts G.W., Quinn S.J., Valentine N., Alhawassi T., O’Dea H., Stranks S.N., et al. Relative hyperglycemia, a marker of critical illness: introducing the stress hyperglycemia ratio. J Clin Endocrinol Metab. 2015;100(12):4490–4497. doi: 10.1210/jc.2015-2660. [DOI] [PubMed] [Google Scholar]
- 13.Staals J., Makin S.D.J., Doubal F.N., Dennis M.S., Wardlaw J.M. Stroke subtype, vascular risk factors, and total mri brain small-vessel disease burden. Neurology. 2014;83(14):1228–1234. doi: 10.1212/WNL.0000000000000848. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Dong K.W.Q. Experts consensus on post-stroke cognitive lmpairment management 2021. Chin J Behav Med Brain Sci. 2021;16(No.4):376–389. doi: 10.3969/j.issn.1673-5765.2021.04.011. [DOI] [Google Scholar]
- 15.Yangjuan J., Ning H., Meirong W., Yanli J., Jingru Z., Peiyuan L., et al. Application of moca and mmse in screening for cognitive impairment in acute ischemic stroke. Chin J Behav Med Brain Sci. 2017;(1):46–50. [Google Scholar]
- 16.Beard J.R., Officer A.M., Cassels A.K. The world report on ageing and health. Gerontologist. 2016;56(Suppl 2):S163–S166. doi: 10.1093/geront/gnw037. [DOI] [PubMed] [Google Scholar]
- 17.Baron R.M., Kenny D.A. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986;51(6):1173–1182. doi: 10.1037/0022-3514.51.6.1173. [DOI] [PubMed] [Google Scholar]
- 18.Service F.J., O’Brien P.C., Rizza R.A. Measurements of glucose control. Diabetes Care. 1987;10(2):225–237. doi: 10.2337/diacare.10.2.225. [DOI] [PubMed] [Google Scholar]
- 19.Inchiostro S., Candido R., Cavalot F. How can we monitor glycaemic variability in the clinical setting? Diabetes Obesity Metab. 2013;15(s2):13–16. doi: 10.1111/dom.12142. [DOI] [PubMed] [Google Scholar]
- 20.Geriatric Endocrinology and Metabolism Specialised Committee of the Geriatrics Society of the Chinese Gerontological Society ECWGoGDDaTM Expert consensus on diagnostic and treatment measures for diabetes mellitus in the elderly (2013 edition) Chin J Intern MedC. 2014;(3):243–251. [Google Scholar]
- 21.Abbatecola A.M., Rizzo M.R., Barbieri M., Grella R., Arciello A., Laieta M.T., et al. Postprandial plasma glucose excursions and cognitive functioning in aged type 2 diabetics. Neurology. 2006;67(2):235–240. doi: 10.1212/01.wnl.0000224760.22802.e8. [DOI] [PubMed] [Google Scholar]
- 22.Li T.C., Yang C.P., Tseng S.T., Li C.I., Liu C.S., Lin W.Y., et al. Visit-to-visit variations in fasting plasma glucose and hba(1c) associated with an increased risk of alzheimer disease: Taiwan diabetes study. Diabetes Care. 2017;40(9):1210–1217. doi: 10.2337/dc16-2238. (1935-5548 (Electronic)) [DOI] [PubMed] [Google Scholar]
- 23.Ceriello A., Colagiuri S. International diabetes federation guideline for management of postmeal glucose: a review of recommendations. Diabetic Med. 2008;25(10):1151–1156. doi: 10.1111/j.1464-5491.2008.02565.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ceriello A., Esposito K., Piconi L., Ihnat M.A., Thorpe J.E., Testa R., et al. Oscillating glucose is more deleterious to endothelial function and oxidative stress than mean glucose in normal and type 2 diabetic patients. Diabetes (New York, NY). 2008;57(5):1349–1354. doi: 10.2337/db08-0063. [DOI] [PubMed] [Google Scholar]
- 25.Quagliaro L., Piconi L., Assaloni R., Martinelli L., Motz E., Ceriello A. Intermittent high glucose enhances apoptosis related to oxidative stress in human umbilical vein endothelial cells: the role of protein kinase c and nad(p)h-oxidase activation. Diabetes (New York, NY). 2003;52(11):2795–2804. doi: 10.2337/diabetes.52.11.2795. [DOI] [PubMed] [Google Scholar]
- 26.Hsieh C.-F., Liu C.-K., Lee C.-T., Yu L.-E., Wang J.-Y. Acute glucose fluctuation impacts microglial activity, leading to inflammatory activation or self-degradation. Sci Rep. 2019;9(1):840. doi: 10.1038/s41598-018-37215-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Toya S.P., Malik A.B. Role of endothelial injury in disease mechanisms and contribution of progenitor cells in mediating endothelial repair. Immunobiology (1979) 2012;217(5):569–580. doi: 10.1016/j.imbio.2011.03.006. [DOI] [PubMed] [Google Scholar]
- 28.Jing-shang W.A.N.G., YIN Hui-jun G.C.-Y., Ye H.U.A.N.G., Cheng-dong X.I.A., Qian L.I.U. Influence of high blood glucose fluctuation on endothelial function of type 2 diabetes mellitus rats and effects of panax quinquefolius saponin of stem and leaf. Chin J Integrative Med. 2013;19(3):217–222. doi: 10.1007/s11655-012-1093-z. [DOI] [PubMed] [Google Scholar]
- 29.Bémeur C., Ste-Marie L., Desjardins P., Vachon L., Butterworth R.F., Hazell A.S., et al. Dehydroascorbic acid normalizes several markers of oxidative stress and inflammation in acute hyperglycemic focal cerebral ischemia in the rat. Neurochem Int. 2005;46(5):399–407. doi: 10.1016/j.neuint.2004.11.007. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.