Association of fasting blood glucose level with 90-day unfavorable outcome in acute ischemic stroke patients

Objectives: Fasting blood glucose (FBG) is a risk factor for Acute Ischemic Stroke (AIS). We aimed to systematically assess the association of FBG level and 90-day unfavorable outcome in AIS patients. Methods: FBG levels and related information of the patients were collected at admission. The unfavorable outcome was defined as 90-day mRS 3 – 6. FBG levels were analyzed as continuous variables and tertiles (Q1 – Q3). Odds ratios and 95% confidence intervals were calculated by using multivariate logistic regression analysis. Results: Overall, 677 AIS patients were included. FBG were significantly associated with unfavorable outcome at 90 days (adjusted OR 1.15 [95%Cl, 1.05 – 1.25], P = 0.002). Participants were categorized based on the FBG tertile cut-off points, the Odds ratios was 2.55-fold higher in Q3 than those in Q1 after adjusting (OR 2.55[95% Cl, 1.23 – 5.3], p = 0.012). Threshold effect analysis showed when FBG ≥ 5.5 mmol/L, the correlation between FBG and 90-day unfavorable outcome increased significantly. Subgroup analysis showed that there was no significant interaction between FBG and 90-day unfavorable outcome. Non-diabetic AIS patients with hyper-glycemia (FBG ≥ 7 mmol/L) have a worse prognosis in comparison to those with normal glucose (FBG ˂ 5.6 mmol/L) (OR 8.59 [ 95%Cl, 2.24 – 32.97], p = 0.002). Conclusion: FBG is an independent predictor of 90-day unfavorable outcome after stroke in AIS patients. When FBG ≥ 5.5 mmol/L, the risk of 90-day unfavorable outcome increases significantly.


Introduction
China bears the biggest stroke burden in the world with over two million new cases annually, and acute ischemic stroke (AIS) accounted for 69.6% [1,2].Fasting blood glucose (FBG) is an easy method to evaluate glycemic status which can minimizes the effect of diet, and also has been regarded as an independent predictor for unfavorable outcomes after AIS [3,4].AHA/ASA guidelines recommended maintaining the blood glucose within the range of 7.78-10 mmol/L during the first 24 h after AIS, the quality of the evidence was not high (Level C-LD) [5].The GLIAS study [6] found out that hyperglycemia above 8.61 mmol/L was independently associated with unfavorable outcome.There was also an argument that receiving intensive (4.44-7.22mmol/L) vs standard (4.44-9.94mmol/L) glucose control within 72 h after AIS, but this study did not result in a significant difference in favorable outcome at 90 days in American patients [7].A multicenter cohort study demonstrated that lower and higher FBG levels may lead to an increased risk of stroke recurrence, and the relationship is a J-shaped curve with a nadir value of 4.6 mmol/L [8].The reason of such a big discrepancy was not clear yet, this might be due to the selective inclusion of people with or without diabetes mellitus (DM) cause it is generally now accepted that a worse prognosis in hyperglycemic non-diabetic patients than in diabetic patients [9,10].
Therefore, we investigated the effect of FBG on 90-day unfavorable outcome in the general AIS population and performed further subgroup analysis to assess whether the association between FBG and 90-day unfavorable outcome is robust in distinct population.

Patients
The information of AIS patients from the comprehensive stroke center of Jiangsu Provincial Hospital of Traditional Chinese Medicine from January 1, 2021, to July 31, 2021 was collected continuously for this retrospective observational study.This study was approved by the Ethics Committee (2017NL-012-01).The inclusion criteria were: (1) Patients who met the diagnostic criteria for AIS [11]; (2) visit to the hospital within 24 h of the onset of symptoms.Exclusion criteria were: (1) Patients with severe cognitive impairment or previous stroke resulted in modified Rankin Scale (mRS)≥ 3; (2) with incomplete follow up data; (3) with severe liver, kidney disease or end-stage malignant tumor.Finally, 677 cases were included.The flow chart was shown in Fig. 1.

Clinical data collection
Clinical data, including sex, age, hypertension, hyperlipidemia, diabetes mellitus, atrial fibrillation, hyperlipidemia, coronary heart disease, smoking and drinking, and previous medication history were recorded on admission.Smoking was defined as smoking more than 100 cigarettes in the past.Drinking was defined as drinking at least 12 alcohol drinks a year.The severity of stroke was assessed using the National Institutes of Health Stroke Scale (NIHSS) [12].Stroke were categorized according to the Trial of Organ 10 172 in Acute Stroke Treatment (TOAST).All patients completed diagnostic examinations after admission, including blood biochemical, intravascular and extravascular examinations, and an MRI and/or CT examination within 24-48 h.All data were obtained during the acute phase of ischemic stroke (6-72 h after onset).

Unfavorable outcome and follow-up
The unfavorable outcome was defined as 90-day mRS 3-6 [13].Four specially trained neurologists were responsible for outpatient or telephone visits 90 days later to evaluate the functional outcome of AIS patients.

Laboratory testing
On the first day after admission, blood samples were drawn at approximately 07:30 after fasting for at least 8 h to minimize the effect of diet on blood glucose levels.The samples were sent to the standard serological testing room for testing.

Statistical Analyses
Categorical variables were presented as number (percentage) and continuous variables as median (interquartile range [IQR]).The outcomes were evaluated at 90 days after stroke and divided into favorable outcome group and unfavorable outcome group.The chi-square test for categorical variables and Mann-Whitney U test for continuous variables were applied to compare the two groups.Logistic regression analysis was used to calculate Odds ratios (ORs) and 95% confidence intervals (CIs).Patients were grouped based on the tertiles of the FBG and the lowest tertile was treated as the reference group.
We used a two-piece wise linear regression model to examine the threshold correlation of the FBG and unfavorable outcome.To identify robustness and interactions, we used subgroup analysis and likelihood ratio test.Statistical analyses were performed using the statistical software packages R (http://www.R-project.org,The R Foundation) [14].A P-value of < 0.05 was taken to indicate statistical significance.

Baseline characteristics
A total of 677 AIS patients (444 males, 233 females) were included in the study with the median age of 69 (IQR 60-77) years.The median FBG level was 5.5 (IQR 4.7-7.0)mmol/L.Of these, 65 patients (9.6%) treated with intravenous thrombolysis (IVT) and 41 patients (6.0%) treated with mechanical thrombectomy (MT).Table 1 compares the clinical and biochemical characteristics between the favorable outcome and unfavorable outcome groups.The two groups were significantly different in age, sex, DM, atrial fibrillation (AF), coronary heart disease (CHD), smoking, drinking, TOAST classification, hypoglycemic drugs (HGD), antiplatelet drugs (APD), and levels of FBG and LDL-C.No significant differences in hypertension, hyperlipidemia, HbA1c, IVT, and MT between the two groups.

Threshold Effect Analysis of FBG Level on 90-day unfavorable outcome
According to the median value of the FBG level with a value of 5.5, we further developed a two-piece wise linear regression model to identify the threshold effect.Threshold effect analysis showed that when FBG ≥ 5.5 mmol/L, the risk of 90-day unfavorable outcome increased significantly (OR 1.15 [ 95%Cl, 1.03-1.28],p = 0.032; Table 3).

Subgroups Analyses
Subgroup analyses were performed to assess the impact of FBG on 90-day unfavorable outcome in distinct population.Significant correlations were highly robust in different subgroups (Fig. 2).None of the subgroups considered (age ≥ 65 years or not, sex, IVT, AF, LDL-C ≥ 3.4 mmol/L or not, DM, HGD) have significantly interaction between FBG and 90-day unfavorable outcome.
In addition, the association between FBG and 90-day unfavorable outcome was significant in patients with and without DM.(with DM: OR 1.13 [ 95%Cl, 1.03-1.24],p = 0.014); without DM: OR 1.28 [ 95%Cl, 1.05-1.57],p = 0.015).Diabetic patients and non-diabetic patients were respective divided into three groups according to the cut-points 5.6 mmol/L and 7 mmol/L.These cut-points were chosen based on previous studies [15].The group which FBG ˂ 5.6 mmol/L was used as    the reference.Table 4 showed that non-diabetic AIS patients with hyperglycemia (FBG ≥ 7) have a worse prognosis in comparison to those with normal glucose (FBG ˂ 5.6) (OR 8.59 [ 95%Cl, 2.24-32.97],p = 0.002).

Discussion
In this retrospective study, we found that FBG was independently associated with 90-day unfavorable outcome after AIS.The risk of 90day unfavorable outcome increases significantly when acute phase FBG ≥ 5.5 mmol/L.
Threshold effect analysis further confirmed that When FBG ≥ 5.5 mmol/L, the risk of 90-day unfavorable outcome in AIS patients increased significantly.Subgroup analysis revealed that the independent relationship between FBG and unfavorable outcome was robust in different population, especially in IVT patients (Fig. 2).This is in concordance with the view of Wnuk et al. [16].This might be because the increased blood glucose not only reduce the fibrinolytic activity of r-tPA, inhibiting the dissolution of venous thrombus [17], but also leading to reperfusion injury by increasing oxidative stress and inflammation [18].
Of mention, our subgroup analysis demonstrated that FBG in AIS patients with and without DM were both independently associated with unfavorable outcome (Fig. 2), especially in non-diabetic AIS patients with dysglycemia (FBG ≥ 7 mmol/L).There is evidence that increasing FBG in non-diabetic patients predicts more severe cardiocerebrovascular events in comparison to diabetic patients [9,16,19].This coincides with our results.It has also been shown that abnormal hyperglycemia does not appear to be directly associated with the outcome of acute ischemic stroke when considered stroke severity [20], even the hyperglycemia seems to be a consequence rather than a risk factor [10].For non-diabetic AIS patients with hyperglycemia, further studies are needed to determine whether normalize hyperglycemia is necessary in this type of population.
We divided diabetic patients and non-diabetic patients into three groups according to the cut-points 5.6 mmol/L and 7 mmol/L and using the lowest group as the reference.We found that impaired FBG (between 5.6 and 7 mmol/L) increasing the risk of 90-day unfavorable outcome, although not statistically significant.Multiple studies have shown that prediabetes mellitus may result in cardio-cerebrovascular events and reversion from impaired FBG to normal FBG seemed to be associated with potential benefits [21,22].However, when considering the diagnostic criteria of prediabetes, a cohort study found that prediabetes defined by IFG alone without taking into account the effect of HbA1c was not associated with cardio-cerebrovascular events [23].We might underestimate the effect of impaired FBG in contrast to normal FBG on 90-day unfavorable outcome for regarding low glucose as the reference group.However, previous literature has reported that impaired FBG was still not associated with risk of CVD when normal FBG was separately considered [24].Hence, a larger study including multiple measurements FBG is required to further address this issue.
Our study has several limitations: (1) it was a retrospective study conducted at a single center; (2) the selected center was tertiary-grade A hospital that may have led to a selection bias; (3) our database did not have the information on the use of hypoglycemic drugs after admission and data on the concentration of FBG changes over time.

Conclusion
Our study found that FBG is an independent predictor of 90-day unfavorable outcome in AIS patients.When acute phase FBG ≥ 5.5 mmol/L, the risk of 90-day unfavorable outcome increases significantly.Non-diabetic AIS patients with hyperglycemia have a worse prognosis in comparison to those with normal glucose.

Credit author statement
There is no conflict of interest in the submission of this manuscript, and the manuscript is approved by all authors for submission.I would like to declare on behalf of my co-authors that the work described was original research that has not been published previously, and not under consideration for publication elsewhere, in whole or in part.

Table 4
Further subgroup analysis in diabetic and non-diabetic patients.

Table 1
Clinical and biochemical characteristics of the favorable and unfavorable outcome groups.

Table 2
The relationship between FBG and 90-day unfavorable outcome in AIS patients.

Table 3
Threshold effect analysis of FBG on AIS patients.