Special Issue Article
HbA1c is associated with increased all-cause mortality in the first year after acute ischemic stroke Shuolin Wu1, Chunxue Wang1, Qian Jia1, Gaifen Liu1, Kolin Hoff2, Xianwei Wang1, Anxin Wang1, Chunjuan Wang1, Xingquan Zhao1, Yilong Wang1, Liping Liu1, Yongjun Wang1, On Behalf of the Investigators for the Survey on Abnormal Glucose Regulation in Patients with Acute Stroke across China (ACROSS) 1
Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, China, 2Division of Endocrinology, Diabetes and Metabolism Department, Perelman School of Medicine, the University of Pennsylvania, Philadelphia, PA, USA
Objectives: To assess the association between baseline HbA1c and the poor outcomes within 1 year after acute ischemic stroke. Methods: Acute ischemic stroke patients with HbA1c values at baseline (n 5 2186) were selected from the abnormal glucose regulation in patients with acute stroke across China study (ACROSS). Logistic regressions were performed to assess the association between HbA1c quartiles (,5.5% [37 mmol/mol], 5.5 to ,6.1% [37 to ,43 mmol/mol], 6.1 to ,7.2% [43 to ,55 mmol/mol], and 7.2% [ 55 mmol/mol]) and the poor outcomes within 1 year. Poor outcomes were defined as all-cause mortality (modified Rankin scale [mRS] 5 6) and poor functional outcome (mRS [2–6]). Results: The risk for all-cause mortality was significantly increased in HbA1c level .5.5% [.37 mmol/mol] when compared to HbA1c quartile ,5.5% [,37 mmol/mol] and dramatically increased to two to three times higher in the highest HbA1c quartile 7.2% [.55 mmol/mol] (1-year all-cause mortality model, odds ratios [ORs] were 1.07, 1.01, and 2.45, P for trend 0.009). After the further analysis with previous diabetes mellitus (DM) and post-stroke insulin use stratified, the risk of mortality was increased across the HbA1c levels (P for trend 0.020) and dramatically augmented in HbA1c 7.2% [.55 mmol/mol] in patients without a history of DM and without post-stroke insulin use. Discussion: Elevated HbA1c (from 5.5% [37 mmol/mol]) presenting pre-stroke glycemia status has a significant trend in increasing the risk of 1-year all-cause mortality. HbA1c 7.2% (.55 mmol/mol) is an independent risk predictor for 1-year all-cause mortality after acute first-ever ischemic stroke. Such an association might be altered by glycometabolism status. Keywords: Diabetes mellitus, Ischemic stroke, Mortality, Risk factors
Introduction Diabetes mellitus (DM) is a well-known risk factor for macrovascular disease including stroke. A recent study has demonstrated that stroke patients with diabetes had a significantly higher incidence of death or dependency at 6 months after stroke.1 Hyperglycemia on admission has been linked to poor outcomes and increased mortality in patients admitted for ischemic stroke.2 HbA1c at admission is a widely accepted parameter for describing pre-stroke glycemic status (PSGS).3 Kamouchi’s group has reported that pre-stroke glycemic control status4 (measured as HbA1c) is associated Correspondence to: Chunxue Wang, Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, No.6 Tiantanxili, Dongcheng District, Beijing 100050, China. Email: [email protected]
444
ß W. S. Maney & Son Ltd 2014 DOI 10.1179/1743132814Y.0000000355
with poor outcomes within 24-hour of hospitalization for acute ischemic stroke.5 It is unclear how PSGS is related to long-term outcomes. We investigated the association between PSGS (measured as HbA1c) and all-cause mortality as well as poor functional outcome in the first year after acute ischemic stroke.
Methods Subjects ‘ACROSS’ was a large prospective cohort study conducted in China from 2008 to 2009 that investigated on abnormal glucose regulation in acute strokes. The baseline characteristics and the recruiting criteria have been published in Stroke.6 A total of 3450 patients with new onset ischemic stroke, intracerebral
Neurological Research
2014
VOL .
36
NO .
5
Wu et al.
hemorrhage, or subarachnoid hemorrhage were sequentially recruited in ACROSS. The present study is a sub-analysis (only ischemic stroke patients included) of the ACROSS. Out of 2639 patients with ischemic stroke, 2186 patients (82.8%) had HbA1c values collected on admission. Among these patients, 1948 patients had follow-up at 3 months, 1810 patients were available for follow-up at 12 months. Among patients with 1-year mRS scores, 1779 (81.3%) patients were analyzed for death analysis, and 13 patients from this group were not included in the analysis for 1-year poor functional outcome due to lack of accurate mRS scores (Fig. 1).
Baseline clinical data Acute stroke was diagnosed according to World Health Organization (WHO) criteria7 and the neuroimaging information was obtained from brain CT or MRI. Stroke subtypes were determined based on the Trial of ORG 10172 in Acute Stroke Treatment (TOAST).8 Baseline HbA1c value as well as other blood biochemical variables were obtained following an 8–12 fasting state within the first 24 hours after admission. All the variables were analyzed in the certified central laboratory. HbA1c was measured using ‘high performance liquid chromatographic analysis (HPLC)’ with a Bio-Rad Variant II analyzer (Bio-Rad Laboratories, Hercules, CA, USA) with a reference value of 4.1–6.5%, which is aligned with the Diabetes Control and Complications Trial and National Glycohemoglobin Standardization Program (NGSP) standards.9 The intra-assay coefficient of variation (CV) was 2.5%, and the inter-assay CV was ,4.0%, both within NGSP acceptable limits.10 HbA1c was classified into four groups according to quartiles (,5.5% [,37 mmol/mol], 5.5 to ,6.1% [37 to 43 mmol/mol], 6.1 to ,7.2% [43 to ,55 mmol/mol], 7.2% [ 55 mmol/mol]). Data on age, gender, education status, alcohol intake, tobacco use, body mass index (BMI), systolic blood pressure (SBP), diastolic blood pressure (DBP), National Institutes of Health Stroke scale (NIHSS),11 and past medical history were collected at admission. Diabetes diagnosis was based on a self-reported history of diabetes confirmed by the medical records or using an oral hypoglycemic agent for at least 6 months.
Ethics The Ethics Committees of Beijing Tiantan Hospital at all participating centers approved the procedures. Written informed consent was obtained from all patients or from the designated family member when the patient was unable to complete it.
Clinical assessment during hospitalization and follow-up Blood biochemical samples were obtained within 24 hours after admission. The values of the homeostasis
HbA1c and all-cause mortality after ischemic stroke
model assessment of insulin resistance (HOMA2-IR)12 were calculated by HOMA2-IR model available from www.ocdem.ox.ac.uk. Telephone interviews were conducted for 3- and 12-month follow-up. Death was defined as all-cause mortality and confirmed through telephone interviewing by two different representatives. All interviewers were centrally trained on the usage of the interview protocol. Poor functional outcome was defined as mRS (2–6); Mortality was indicated by mRS 5 6.
Statistical analysis The clinical characteristics among HbA1c quartiles were compared using one-way ANOVA or Kruskal– Wallis H test for continuous variables. Categorical variables were analyzed by chi-square test. Univariate analysis was conducted followed by multivariate logistic regression analysis for poor outcomes. Risk factors were first compared among the four HbA1c quartiles, and then the significant ones were selected based on the initial assessment and further for comparison of the HbA1c level. The interaction effects analysis between HbA1c and the related risk factors were also assessed in the multivariate analysis. The association between the risk factors and the poor outcome were assessed by univariate logistic regression. Multivariate models were established for evaluating the association between HbA1c and outcomes. The association between HbA1c and 1-year all-cause mortality was further analyzed stratified by insulin use and history of DM, since insulin use is associated with death risk13 and both DM and insulin use are closely related to post-stroke glycemic control. The variables include traditional confounding factors, newly identified risk factors including HOMA2-IR, white blood cell count, insulin therapy,14,15 complications during hospitalization as well as factors that may alter HbA1c levels or outcomes including red blood cell count, hemoglobin, and medications (model 1 for 1-year allcause mortality, Model 2 and 3 for 3-month and 1-year poor functional outcome). Statistical analyses were conducted using the SPSS 17.0 software (SPSS Inc., Chicago, IL, USA).All P values are two-sided and considered significant when P , 0.05.
Results Demographics of the patients Table 1 shows clinical characteristics of patients according to HbA1c quartiles. The patients in the highest HbA1c quartile was much older, had more moderate alcohol intake, much higher BMI but not much severer ischemic strokes at baseline, and received more insulin therapy. Patients with the other three higher quartiles had much higher level of FG, triglyceride, SBP, DBP at baseline, and received more
Neurological Research
2014
VOL .
36
NO .
5
445
Wu et al.
HbA1c and all-cause mortality after ischemic stroke
Figure 1 Flow chart of patient selection mRS, modified Rankin scale.
446
Neurological Research
2014
VOL .
36
NO .
5
Wu et al.
therapy of anti-hypertension and lipid-lowering agents. The significant variables in Table 1 were further categorized into different levels for comparison of HbA1c level (Table 2). Female had a higher level of HbA1c than male. Most cardiovascular risk factors were significant associated with a higher level of HbA1c except moderate to severe drinking and current smoking in our data. Table 3 shows the association of risk factors with poor clinical outcome. Red cell counts and hemoglobin level were protective factors for all poor outcomes. While age of over 45, increasing white cell counts (per 16109/l), previous DM, previous hypertension, stroke severity at admission (NIHSS 9), and insulin use in hospital were all significant with poor clinical outcome.
Association between HbA1c and 1-year all-cause mortality Multivariate analysis indicated that the increased risk of all-cause mortality among higher HbA1c quartiles had a significant trend (P 5 0.009), particularly in HbA1c quartile 7.2% [ 55 mmol/mol]. In Model 1 (Fig. 3),
HbA1c and all-cause mortality after ischemic stroke
the increased risk in HbA1c 7.2% [ 55 mmol/mol] is two-fold higher than the lowest quartile (OR 5 2.45, P 5 0.006). When further stratified by history of DM and insulin use in hospital with the same variables as those adjusted in Model 1 (Table 4), in patients without previous DM and without insulin use in hospital, HbA1c levels were still significantly associated with 1-year all-cause mortality (P for trend 0.020), and particularly in HbA1c quartile 7.2% [ 55 mmol/mol] (P 5 0.004). However, among those with previous DM, whether insulin use in hospital or not, and those with insulin use but without previous DM, the all-cause mortality risk of patients in higher HbA1c levels was not significantly higher than that in lower HbA1c levels (all P . 0.05). All the interaction effect between HbA1c and risk factors significant in the univariate analysis (Table 1 and 2) including age levels, gender, tobacco use, drinking, BMI, stroke severity, past medical history of hypertension, dyslipidemia, coronary heart disease and DM, triglyceride, HOMA-IR, and insulin use in hospital were assessed in the above multivariate
Table 1 Clinical characteristics according to HbA1c quartiles HbA1c quartiles Baseline characteristics Age* (years) Male (%)* Tobacco use (%)* Never Quit Current Alcohol intake (%)* Never Moderate Severe BMI* NIHSS*, median(Q1–Q3) Past medical history Hypertension (%)* Hyperlipidemia (%)* Diabetes (%)* Coronary heart disease (%)* HbA1c (%)* Fasting glucose (mmol/l) * Triglyceride (mmol/l) * HOMA2-IR* Ischemic stroke subtypes (%) Cardioembolism Non-Cardioembolism Large artery atherothrombosis Small artery occlusion Other Medication in hospital Insulin therapy (%)* Anti-thrombotic agent (%) Anti-hypertension agent (%) Lipid-lowering agent (%)
,5.5% (n 5 540)
5.5 to ,6.1% (n 5 507)
6.1 to ,7.2% (n 5 579)
7.2% (n 5 560)
60 ¡ 13 367 (68.0)
63 ¡ 12 331 (65.3)
65 ¡ 12 339 (58.5)
63 ¡ 11 332 (59.3)
316 (58.5) 44 (8.1) 180 (33.3)
261 (51.5) 54 (10.7) 192 (37.9)
352 (60.8) 59 (10.2) 168 (29.0)
332 (59.3) 45 (8.0) 183 (32.7)
365 (67.6) 88 (16.3) 87 (16.1) 24.3 ¡ 3.2 5 (2–9)
327 (64.5) 87 (17.2) 93 (18.3) 24.6 ¡ 3.8 4 (4–8)
410 (70.8) 95 (16.4) 74 (12.8) 25.0 ¡ 3.6 4 (2–8)
318 (58.9) 38 (7.0) 34 (6.3) 48 (8.9) 4.8 ¡ 0.5 5.3 ¡ 1.3 1.56 ¡ 0.97 1.11 ¡ 0.71
296 (58.4) 66 (13.0) 38 (7.5) 55 (10.8) 5.8 ¡ 0.2 5.3 ¡ 1.1 1.72 ¡ 1.00 1.29 ¡ 0.80
355 (61.3) 81 (14.0) 134 (23.1) 99 (17.1) 6.5 ¡ 0.3 6.1 ¡ 1.8 1.77 ¡ 1.00 1.27 ¡ 0.79
377 (67.3) 104 (18.6) 79 (14.1) 25.3 ¡ 4.0 4 (2–7) 372 (66.4) 86 (24.4) 334 (59.6) 84 (15.0) 9.1 ¡ 1.7 9.6 ¡ 3.5 2.14 ¡ 1.61 1.39 ¡ 0.83
21 519 341 143 35
(7.6) (96.1) (63.1) (26.5) (6.5)
36 471 313 124 34
(7.1) (92.9) (61.7) (24.5) (6.7)
40 539 341 161 37
(6.9) (93.1) (58.9) (27.8) (6.4)
23 537 362 147 28
(4.1) (95.9) (64.6) (26.3) (5.0)
22 505 208 358
(4.1) (93.5) (38.5) (66.3)
16 468 227 367
(3.2) (92.3) (44.8) (72.4)
38 550 286 436
(6.6) (95.0) (49.4) (75.3)
235 527 262 426
(42.0) (94.1) (46.8) (76.1)
* Indicates P , 0.05 when comparing between four groups. HbA1c: glycated hemoglobin; mRS: modified Rankin scale; NIHSS: the National Institutes of Health Stroke scale; HOMA2-IR: the correctly solved computer model for homeostasis model assessment of insulin resistance; OTHER: other causes and stroke of undetermined etiology.
Neurological Research
2014
VOL .
36
NO .
5
447
Wu et al.
HbA1c and all-cause mortality after ischemic stroke
Table 2 Comparison of HbA1c level according to clinical characteristics Clinical characteristics
HbA1c (%)
Age (years)
Gender Smoking Drinking Body mass index (kg/m2) Stroke severity (NIHSS) Past medical histories Hypertension Dyslipidemia Diabetes mellitus Coronary heart disease Triglyceride (mmol/l) HOMA-IR Insulin therapy in hospital
,45 45–65 .65 Male Female Current No Never Moderate Severe 25 ,25 9 ,9
6.0 6.7 6.6 6.5 6.7 6.6 6.6 6.6 6.7 6.4 6.8 6.5 6.5 6.6
¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡
1.6 2.0 1.7 1.8 1.9 1.8 1.9 1.8 2.0 1.8 1.9 1.8 1.9 1.8
Yes No Yes No Yes No Yes No 1.7 ,1.7 .1 1 Yes No
6.7 6.4 7.0 6.5 8.1 6.1 6.8 6.6 6.9 6.4 6.8 6.2 8.7 6.3
¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡
1.9 1.8 1.9 1.8 2.0 1.5 1.8 1.8 2.0 1.7 2.0 1.5 2.3 1.5
P ,0.001
0.005 0.617 0.06
,0.001 0.011
0.004 ,0.001 ,0.001 0.015 ,0.001 ,0.001 ,0.001
analysis, and were not significant (all P values were .0.05).
Association between HbA1c and 3-month and 1-year poor functional outcome Figure 2 shows mRS ratios according to HbA1c quartiles. Higher HbA1c quartiles occurred more mRS from 2–6. The 3-month and 1-year poor functional outcome multivariate analysis (Model 2 and 3 in Fig. 3) indicated that compared to the lowest level of HbA1c ,5.5% [,37 mmol/mol], HbA1c in other three higher quartiles had no significant trend in increasing risk of poor functional outcome (Model 2 for 3-month and Model 3 for 1-year, P for trend were 0.165 and 0.718). All the interaction effect between HbA1c and the related risk factors were all not significant (all P . 0.05).
Other significant risk factor in multivariate analysis Age of over 65 was a risk factor of 1-year all-cause mortality and 3-month and 1-year poor functional outcome (OR with 95% CI, 1.12 (1.04–3.28), 1.06 (1.01– 4.13), and 1.13 (1.05–3.42), respectively). Previous hypertension and SBP were associated with 1-year allcause mortality (OR with 95% CI, 1.31 (1.02–3.58) for previous hypertension and 1.16 (1.08–4.22).
Discussion Given the annual morbidity and mortality rates caused by stroke have exceeded cardiovascular disease to become the leading cause of death and disability16 and
448
Neurological Research
2014
VOL .
36
NO .
5
the increased prevalence of abnormal glycometabolism in China,17 it is critical to have thorough understanding on how PSGS may influence prognosis of stroke. HbA1c is an accurate and stable parameter18 that reflects long-term glycemic control.5,19 Pre-stroke glycemic status (measured as HbA1c) was first proposed by the Japan Diabetes Society.4 Our current study indicated that even patients with a good PSGS (baseline HbA1c of 5.5% [ 37 mmol/mol]) was significantly associated with an increased risk of 1year all-cause mortality after acute first-ever ischemic stroke onset. In the highest HbA1c quartile ( 7.2% [ 55 mmol/mol]) the risk of 1-year all-cause mortality is dramatically increased. Such an association is prominent in acute ischemic stroke without a history of DM and without post-stroke insulin use.
HbA1c and 1-year all-cause mortality Our finding indicated HbA1c was significantly associated with increased risk for the all-cause mortality. Such an association started from HbA1c 5 5.5% [37 mmol/mol] and was sharply enlarged when HbA1c achieved 7.2% [55 mmol/mol] or higher. This is consistent with a previous study suggesting high ‘normal’ HbA1c is associated with all-cause mortality in all cardiovascular diseases.20 However, the association of HbA1c and death in the present study was somewhat different with that from other studies. It was indicated in the present study that an increased risk of all-cause mortality in the HbA1c level 7.2% [ 55 mmol/mol] which is somewhat inconsistent with a study from Action in Diabetes and Vascular disease: PreterAx and DiamicroN-MR Controlled Evaluation (ADVANCE). In their study, HbA1c over 7.0% [53 mmol/mol] was regarded as a significant risk for death from macrovascular diseases21. No association was observed in Kamouchi’s study between HbA1c and mortality during hospitalization5 while it is not the case when long-term outcomes are assessed in our studies. Monami M’s study with a mean 5.7 ¡ 3.5 year follow-up found HbA1c of 6.5–7.4% [48–57 mmol/mol] had the lowest death risk. They also stated that glycemic target should be individualized according to the characteristics of each type 2 diabetics because insulin treatment significantly increased the death risk associated with HbA1c of ,6.5% [48 mmol/ mol] in the study.13 Several reasons may account for the discrepancy. Our study focused on ischemic stroke (separate from cardiovascular diseases) while the ADVANCE trial included macro-and micro-vascular diseases; Shortterm poor outcomes were studied in Kamouchi’s study while much longer ones are assessed in ours; Not like Monami M’s study, ours enrolled both diabetics and non-diabetics and had a shorter followup period.
Wu et al.
HbA1c and all-cause mortality after ischemic stroke
Table 3 Association of clinical variables with poor clinical outcome 1-year all-cause 3-month poor 1-year poor mortality OR (95% CI) outcome OR (95% CI) outcome OR (95% CI)
Clinical characteristics Age (years)
Gender Smoking Drinking
Past medical histories Hypertension Dyslipidemia Diabetes mellitus Coronary heart disease Body mass index (kg/m2) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) NIHSS Red cell counts, per 161012/l White cell counts, per 16109/l Hemoglobin, per 1 g/l High density lipoprotein (mmol/l) Low density lipoprotein (mmol/l) Triglyceride (mmol/l) HOMA-IR In-hospital therapy Antihypertensive Lipid-lowering Antithrombotic Insulin therapy
.65 45–65 ,45 Female Male Current No Moderate to severe Mild Never
4.65 2.80 Ref 1.08 Ref 0.82 Ref 1.78
(1.86–11.60) (1.12–7.04)
1.51 (0.57–4.05) Ref
0.54 (0.41–0.70) Ref
0.77 (0.59–1.00) Ref
Yes No Yes No Yes No Yes No 25 ,25 140 ,140 90 ,90 9 ,9
1.55 ,1.55 3.37 ,3.37 1.7 ,1.7 .1 1
1.32 Ref 0.63 Ref 1.48 Ref 1.32 Ref 0.65 Ref 1.00 Ref 1.00 Ref 2.05 Ref 0.76 1.10 0.99 1.02 Ref 0.87 Ref 0.99 Ref 0.95 Ref
1.22 Ref 0.92 Ref 1.26 Ref 1.27 Ref 0.98 Ref 1.26 Ref 1.04 Ref 8.60 Ref 0.67 1.10 0.99 1.12 Ref 0.94 Ref 0.85 Ref 1.05 Ref
1.17 Ref 0.69 Ref 1.54 Ref 1.02 Ref 0.82 Ref 1.16 Ref 0.98 Ref 5.16 Ref 0.65 1.09 0.99 0.87 Ref 0.87 Ref 0.83 Ref 1.07 Ref
No Yes No Yes No Yes No Yes
1.00 Ref 1.50 Ref 1.86 Ref 0.53 Ref
(0.81–1.47) (0.60–1.10) (0.70–4.49)
(1.01–1.81) (0.37–1.06) (1.08–2.02) (0.88–1.99) (0.48–0.88) (0.73–1.39) (0.74–1.36) (1.50–2.81) (0.59–0.99) (1.05–1.16) (0.98–1.00) (0.64–1.63) (0.63–1.19) (0.73–1.33) (0.69–1.29)
(0.75–1.35) (1.09–2.05) (1.08–3.22) (0.37–0.75)
3.30 1.92 Ref 1.80 Ref 0.63 Ref 0.75
1.06 Ref 1.29 Ref 1.04 Ref 0.67 Ref
(2.15–5.05) (1.25–2.95) (1.48–2.17) (0.52–0.76) (0.58–0.97)
(1.00–1.48) (0.69–1.23) (1.02–1.55) (0.97–1.66) (0.81–1.19) (0.92–1.54) (0.86–1.25) (6.72–11.02) (0.57–0.79) (1.05–1.14) (0.98–1.00) (0.83–1.51) (0.77–1.15) (0.70–1.03) (0.88–1.27)
(0.88–1.28) (1.05–1.60) (0.68–1.59) (0.52–0.87)
4.36 2.40 Ref 1.64 Ref 0.62 Ref 0.72
0.88 Ref 1.20 Ref 1.06 Ref 0.56 Ref
(2.82–6.77) (1.55–3.71) (1.35–1.99) (0.51–0.76) (0.55–0.94)
(1.02–1.42) (0.52–1.03) (1.25–1.91) (0.78–1.35) (0.68–0.99) (0.94–1.43) (0.81–1.20) (4.62–6.62) (0.55–0.77) (1.04–1.13) (0.98–0.99) (0.64–1.18) (0.71–1.06) (0.68–1.00) (0.88–1.31)
(0.73–1.07) (0.96–1.48) (0.69–1.64) (0.43–0.74)
OR indicates odds ratio, CI indicates confidence interval. NIHSS indicates the National Institutes of Health Stroke scale; HOMA-IR indicates the correctly solved computer model for homeostasis model assessment of insulin resistance.
In the further stratified analysis, the association between baseline HbA1c (implying PSGS) and 1-year all-cause mortality was only observed in patients without a history of DM and without post-stroke insulin use in hospital. On the contrary, in those with a history of DM or accepting insulin therapy in hospital, higher HbA1c levels were not detrimental to 1-year all-cause mortality. Those without a history of DM but accepting insulin use in hospital were in a large possibility of newly-onset DM.22 Thus, we conclude that baseline HbA1c, indicating PSGS, is very likely associated with risk of mortality only in ischemic stroke patients without DM. The more prominent association between glycemia and poor outcome in non-diabetes than diabetes was partly in
line with Zsuga, et al.23 The difference was that poststroke hyperglycemia was also significant in DM group. The index was hyperglycemia post stroke in their research but PSGS in our study and the different following-up period might lead to the discordance. In DM, post stroke glycemic control was not only necessarily needed but also significant in the whole course of treatment and affecting later prognosis, however, it is not necessarily applied in each nondiabetic acute ischemic stroke patient, thus the effect of glycemia on mortality might be largely attributed to PSGS or acute hyperglycemia response22 but not the post-stroke glycemic control level in those without DM. As fasting plasma glucose, partly reflecting
Neurological Research
2014
VOL .
36
NO .
5
449
Wu et al.
HbA1c and all-cause mortality after ischemic stroke
Figure 2 Description for the 3-month and 1-year mRS ratios in HbA1c quartiles mRS, modified Rankin scale.
acute hyperglycemia response, was included in the multivariate analysis and it showed not significant, we believed that PSGS is significantly detrimental to 1-year all-cause mortality in non-diabetic ischemic stroke patients independently of acute hyperglycemia response. We also learned from our study that PSGS was not significant with 1-year all-cause mortality in acute ischemic stroke patients with DM. Concerning the consequence of post-stroke glycemic control in the treatment of DM and the following prognosis, we postulated that post-stroke glycemic control level has much more effect on the poor outcome than PSGS in acute ischemic stroke with DM. However, we still cannot well compare the effect of PSGS and poststroke glycemic control on 1-year all-cause mortality, since the longitude change of post-stroke glycemic level was not recorded in our study. Thus, further investigation is needed for comparison between poststroke glycemic control and PSGS in the association with post-stroke outcome in acute stroke patients.
HbA1c and poor functional outcome within 1 year following acute ischemic stroke onset A previous study from Kamouchi’s group has shown that increased HbA1c is associated with worsening of outcomes during hospitalization after acute ischemic
Circles in black stand for OR; diamond in black stand for 95% CI.; q2: HbA1c quartile 5.5 to ,6.1%;; q3: HbA1c quartile 6.1 to ,7.2%;; q4: HbA1c quartile 7.2%;; OR: odds ratio;; Model 1 is on 1-year mortality adjusted for age, gender, education status, tobacco use, alcohol intake, medical histories of hypertension and coronary heart disease, body mass index (BMI), systolic (SBP) and diastolic blood pressure (DBP), fasting glucose, high and low density lipoprotein, white blood cell counts, red blood cell counts, hemoglobin level, NIHSS, TOAST, and OCSP stroke subtypes, HOMA2-IR, triglyceride, the usages of insulin, anti-hypertension agent, lipid-lowering agent, and anti-thrombotic agent during hospitalization, urinary infection and pneumonia during hospitalization.; Model 2 is on 3-month poor functional outcomes analysis adjusted for age, gender, education status, alcohol intake, tobacco use, medical histories of hypertension and coronary heart disease, BMI, SBP and DBP, NIHSS, fasting glucose, high and low density lipoprotein, HOMA2-IR, triglyceride, white blood cell counts, red blood cell counts, hemoglobin level, OCSP, and TOAST subtypes, the usages of anti-hypertension agent, lipid-lowering agent, and anti-thrombotic agent during hospitalization, urinary infection, and pneumonia during hospitalization.; Model 3 is on 1-year poor functional outcomes analysis adjusted for variables in model 2 plus the usages of 3-month anti-hypertension agent, lipid-lowering agent, and antithrombotic agent, and 3-month stroke recurrence. Figure 3 HbA1c quartiles versus risk of all-cause mortality and poor functional outcome compared to HbA1c ,5.5%
stroke.5 These results indicated that poor functional outcome was higher in patients with poor PSGS when HbA1c is over 8.4% [68 mmol/mol]. That was not in line with our findings. Our finding did not suggest higher levels of HbA1c had a higher risk of poor functional outcome within 1 year after ischemic stroke. The longer follow up period in our study might partly explain the paradox.
Table 4 Association between HbA1c levels and 1-year all-cause mortality stratified by diabetes mellitus (DM) and insulin use* Diabetes mellitus HbA1c levels Q1 Q2 Q3 Q4 P for trend
(OR with 95% CI)
Insulin use
P Without insulin use
Ref 3.00 (0.25–16.32) 0.52 (0.04–6.58) 1.28 (0.15–10.88)
Ref 0.388 0.46 (0.04–5.41) 0.612 1.17 (0.23–5.81) 0.823 2.45 (0.54–11.15) 0.117
Without diabetes mellitus
P
0.534 0.851 0.245 0.377
OR indicates odds ratio, CI indicates confidence interval. * Adjusted for the same variables in Model 1. Q1: ,5.5%; Q2: 5.5 to ,6.1%; Q3: 6.1 to ,7.2%; Q4: 7.2%.
450
Neurological Research
2014
VOL .
36
NO .
5
Insulin use Ref 0.88 (0.06–12.97 2.62 (0.30–15.00) 0.97 (0.18–5.23)
(OR with 95% CI)
P
0.923 0.383 0.970 0.678
Without insulin use Ref 1.09 (0.60–1.97) 1.40 (0.79–2.45) 2.54 (1.34–4.82)
P
0.780 0.248 0.004 0.020
Wu et al.
HbA1c in ischemic stroke needs to be highlighted in clinician The present study showed that even a non-diabetic HbA1c value (,6.5%) has a trend for an increased risk of 1-year all-cause mortality in acute ischemic stroke, especially in patients without diabetes. HbA1c is not only a tool for assessing glycemic control but also meaningful to those without diabetes in acute ischemic stroke. Since its significance for secondary stroke prevention24 and poor clinical outcome, baseline HbA1c is recommended as a routine test in all acute ischemic stroke patients regardless of diabetes diagnosis. For patients whose HbA1c are in the high ‘normal’ level ( 6.5%), clinicians should be aware that the 1-year mortality risk might be increased 7% or so in this population. Pre-stroke glycemic status indicated by HbA1c at admission after acute ischemic stroke onset should be a vital notion to clinicians as the HbA1c level of over 7.2% predicts a two-fold or more increase in the risk for 1-year mortality. Since the association between PSGS and all-cause mortality differs in glycometabolism diagnosis and post stroke glycemic control status (use insulin or not), further investigation is needed for comparison between PSGS and post-stroke glycemic control in the association with poor clinical outcomes in acute stroke patients stratified by diagnosis of glycometabolism, including DM, prediabetes, and normoglycemia.
Other variables significant for poor clinical outcome Age was a traditional risk factor of poor outcome which was also observed in our study. Hypertension25 or SBP26 at admission were also related to higher risk of stroke and its poor clinical outcome which was also confirmed by others.
Limitations and strengths There were some limitations in our study. In total, 452 patients (17.2%) with ischemic stroke were excluded due to lack of HbA1c values, although this is slightly lower than that of a similar study from Japan (22.2%),5 a selective bias could not be avoided; Non-responder and non-mRS ratios were not as high as other clinical studies concerning HbA1c and CVD,27 which may potentially affect OR. Certain biochemical and physiological data, including HbA1c, SBP, and LDL, could not be collected before acute stroke or during the following-up period due to limitations of the current study’s design. Despite these limitations, we provided detailed and quantified evaluation for long-term prognosis of patients with ischemic stroke. Still many physiologic and biochemical parameters and traditional and newly established risk factors were accounted for multivariate analysis, such as blood pressure,28 insulin therapy, HOMA2IR, and complications during hospitalization.
HbA1c and all-cause mortality after ischemic stroke
Conclusions Elevated baseline HbA1c (from 5.5% [37 mmol/mol]) presenting pre-stroke glycemia status has a significant trend in increasing the risk of 1-year all-cause mortality. HbA1c (from 5.5% [37 mmol/mol]) is an independent risk factor for 1-year all-cause mortality after acute first-ever ischemic stroke onset. Such an association might be altered by glycometabolism status.
Disclaimer Statements Contributors SW drafted the manuscript; CW and KH helped to draft the manuscript; QJ helped to collect the data; GL, XW, AW and CW performed the statistical analysis; XZ participated in the design of the study and explanation of the data; LL, YW and YW participated in the design of the study. All authors read and approved the final manuscript. Funding This study was funded jointly by the Beijing Science and Technology Committee (grant no. 7102050, www.bjkw.gov.cn) and the National Science Foundation (grant no. 81071115, www.nsfc.gov.cn) Conflicts of interest None. Ethics approval The Ethics Committees of Beijing Tiantan Hospital at all participating centers approved the procedures. Written informed consent was obtained from all patients or from the designated family member when the patient was unable to complete it.
Acknowledgements We thank all participating hospitals, colleagues, patients, and caregivers involved in the study. This study was funded jointly by the Beijing Science and Technology Committee (grant no. 7102050, www.bjkw.gov.cn) and the National Science Foundation (grant no. 81071115, www.nsfc.gov.cn).
References 1 Jia Q, Zhao X, Wang C, Wang Y, Yan Y, Li H, et al. Diabetes and poor outcomes within 6 months after acute ischemic stroke: the China National Stroke Registry. Stroke. 2011;42:2758–62. 2 Fuentes B, Castillo J, San Jose´ B, Leira R, Serena J, Vivancos J, et al. The prognostic value of capillary glucose levels in acute stroke: the GLycemia in Acute Stroke (GLIAS) study. Stroke. 2009;40:562–8. 3 Chmielewska B, Hasiec T, Belniak-Legiec´ E, Stelmasiak Z. Usefulness of current and retrospective markers of glycemia for early prognostication of ischemic stroke in diabetics and nondiabetics. Ann Univ Mariae Curie Sklodowska Med. 1999;54:261–7. 4 Japan Diabetes Society. Treatment objectives and control indicators. In: Japan Diabetes Society, editor. Treatment guide for diabetes 2007. Tokyo: Bunkodo; 2007. p. 1–84. 5 Kamouchi M, Matsuki T, Hata J, Kuwashiro T, Ago T, Sambongi Y, et al. Prestroke glycemic control is associated with the functional outcome in acute ischemic stroke: the Fukuoka Stroke Registry. Stroke. 2011;42:2788–94. 6 Jia Q, Zheng H, Zhao X, Wang C, Liu G, Wang Y, et al. Abnormal glucose regulation in patients with acute stroke across China: prevalence and baseline patient characteristics. Stroke. 2012;43:650–7. 7 WHO Task Force on Stroke and other Cerebrovascular Disorders: Stroke-1989. Recommendations on stroke prevention, diagnosis, and therapy. Stroke. 1989;20:1407–31.
Neurological Research
2014
VOL .
36
NO .
5
451
Wu et al.
HbA1c and all-cause mortality after ischemic stroke
8 Adams HP Jr, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, et al. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke. 1993;24:35–41. 9 National Glycohemoglobin Standardization Program (NGSP). List of NGSP certified methods. Available from: http:// www.ngsp.org/certified.asp (accessed 2012 May 12). 10 Schwartz KL, Monsur JC, Bartoces MG, West PA, Neale AV. Correlation of same-visit HbA1c test with laboratory-based measurements: a MetroNet study. BMC Fam Pract. 2005;6:28. 11 Brott T, Adams HP Jr, Olinger CP, Marler JR, Barsan WG, Biller J, et al. Measurements of acute cerebral infarction: a clinical exami Renee Holleran, Robert Eberle, Vicki Hertzberg, Marvin nation scale. Stroke. 1989;20:864–70. 12 Levy JC, Matthews DR, Hermans MP. Correct homeostasis model assessment (HOMA) evaluation uses the computer program. Diabetes Care. 1998;21:2191–2. 13 Monami M, Vitale V, Lamanna C, Bartoli N, Martelli D, Zannoni S, et al. HbA1c levels and all-cause mortality in type 2 diabetic patients: epidemiological evidence of the need for personalized therapeutic targets. Nutr Metab Cardiovasc Dis. 2013;23:300–6. 14 Conway BN, May ME, Blot WJ. Mortality among low-income African Americans and whites with diabetes. Diabetes Care. 2012;35:2293–9. 15 Hayashi T, Kawashima S, Nomura H, Itoh H, Watanabe H, Ohrui T, et al. Age, gender, insulin and blood glucose control status alter the risk of ischemic heart disease and stroke among elderly diabetic patients. Cardiovasc Diabetol. 2011;10:86. 16 Liu L, Wang D, Wong KS, Wang Y. Stroke and stroke care in china: huge burden, significant workload, and a national priority. Stroke. 2011;42:3651–4. 17 Yang W, Lu J, Weng J, Jia W, Ji L, Xiao J, et al. Prevalence of diabetes among men and women in China. N Engl J Med. 2010;362:1090–101. 18 Derr R, Garrett E, Stacy GA, Saudek CD. Is HbA (1c) affected by glycemic instability? Diabetes Care. 2003;26:2728–33.
452
Neurological Research
2014
VOL .
36
NO .
5
19 Selvin E, Coresh J, Golden SH, Boland LL, Brancati FL, Steffes MW, et al. Glycemic control, atherosclerosis, and risk factors for cardiovascular disease in individuals with diabetes: the atherosclerosis risk in communities study. Diabetes Care. 2005;28:1965–73. 20 Smith SA. Higher ‘normal’ glycated hemoglobin levels were associated with increased risk for diabetes, CVD, stroke, and mortality in adults. Ann Intern Med. 2010;153:JC1–13. 21 Zoungas S, Chalmers J, Ninomiya T, Li Q, Cooper ME, Colagiuri S, et al. Association of HbA1c levels with vascular complications and death in patients with type 2 diabetes: evidence of glycaemic thresholds. Diabetologia. 2012;55: 636–43. 22 Jia Q, Zheng H, Liu L, Zhao X, Wang C, Jing J, et al. Persistence and predictors of abnormal glucose metabolisms in patients after acute stroke. Neurol Res. 2010;32:359–65. 23 Zsuga J, Gesztelyi R, Kemeny-Beke A, Fekete K, Mihalka L, Adrienn SM, et al. Different effect of hyperglycemia on stroke outcome in non-diabetic and diabetic patients – a cohort study. Neurol Res. 2012;34:72–9. 24 Wu S, Shi Y, Wang C, Jia Q, Zhang N, Zhao X, et al. Glycated hemoglobin independently predicts stroke recurrence within one year after acute first-ever non-cardioembolic strokes onset in A Chinese cohort study. PLoS One. 2013;8:e80690. 25 Yong H, Foody J, Linong J, Dong Z, Wang Y, Ma L, et al. A systematic literature review of risk factors for stroke in China. Cardiol Rev. 2013;21:77–93. 26 Rodrı´guez-Ya´n˜ez M, Castellanos M, Blanco M, Garcı´a MM, Nombela F, Serena J, et al. New-onset hypertension and inflammatory response/poor outcome in acute ischemic stroke. Neurology. 2006;67:1973–8. 27 Adams RJ, Appleton SL, Hill CL, Wilson DH, Taylor AW, Chittleborough CR, et al. Independent association of HbA1c and Incident cardiovascular disease in people without diabetes. Obesity (Silver Spring). 2009;17:559–63. 28 Jia Z, Wu S, Zhou Y, Wang W, Liu X, Wang L, et al. Metabolic syndrome and its components as predictors of stroke in middle-aged and elderly Chinese people. Neurol Res. 2011;33:453–9.
HbA1c is associated with increased all-cause mortality in the first year after acute ischemic stroke Shuolin Wu1, Chunxue Wang1, Qian Jia1, Gaifen Liu1, Kolin Hoff2, Xianwei Wang1, Anxin Wang1, Chunjuan Wang1, Xingquan Zhao1, Yilong Wang1, Liping Liu1, Yongjun Wang1, On Behalf of the Investigators for the Survey on Abnormal Glucose Regulation in Patients with Acute Stroke across China (ACROSS) 1
Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, China, 2Division of Endocrinology, Diabetes and Metabolism Department, Perelman School of Medicine, the University of Pennsylvania, Philadelphia, PA, USA
Objectives: To assess the association between baseline HbA1c and the poor outcomes within 1 year after acute ischemic stroke. Methods: Acute ischemic stroke patients with HbA1c values at baseline (n 5 2186) were selected from the abnormal glucose regulation in patients with acute stroke across China study (ACROSS). Logistic regressions were performed to assess the association between HbA1c quartiles (,5.5% [37 mmol/mol], 5.5 to ,6.1% [37 to ,43 mmol/mol], 6.1 to ,7.2% [43 to ,55 mmol/mol], and 7.2% [ 55 mmol/mol]) and the poor outcomes within 1 year. Poor outcomes were defined as all-cause mortality (modified Rankin scale [mRS] 5 6) and poor functional outcome (mRS [2–6]). Results: The risk for all-cause mortality was significantly increased in HbA1c level .5.5% [.37 mmol/mol] when compared to HbA1c quartile ,5.5% [,37 mmol/mol] and dramatically increased to two to three times higher in the highest HbA1c quartile 7.2% [.55 mmol/mol] (1-year all-cause mortality model, odds ratios [ORs] were 1.07, 1.01, and 2.45, P for trend 0.009). After the further analysis with previous diabetes mellitus (DM) and post-stroke insulin use stratified, the risk of mortality was increased across the HbA1c levels (P for trend 0.020) and dramatically augmented in HbA1c 7.2% [.55 mmol/mol] in patients without a history of DM and without post-stroke insulin use. Discussion: Elevated HbA1c (from 5.5% [37 mmol/mol]) presenting pre-stroke glycemia status has a significant trend in increasing the risk of 1-year all-cause mortality. HbA1c 7.2% (.55 mmol/mol) is an independent risk predictor for 1-year all-cause mortality after acute first-ever ischemic stroke. Such an association might be altered by glycometabolism status. Keywords: Diabetes mellitus, Ischemic stroke, Mortality, Risk factors
Introduction Diabetes mellitus (DM) is a well-known risk factor for macrovascular disease including stroke. A recent study has demonstrated that stroke patients with diabetes had a significantly higher incidence of death or dependency at 6 months after stroke.1 Hyperglycemia on admission has been linked to poor outcomes and increased mortality in patients admitted for ischemic stroke.2 HbA1c at admission is a widely accepted parameter for describing pre-stroke glycemic status (PSGS).3 Kamouchi’s group has reported that pre-stroke glycemic control status4 (measured as HbA1c) is associated Correspondence to: Chunxue Wang, Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, No.6 Tiantanxili, Dongcheng District, Beijing 100050, China. Email: [email protected]
444
ß W. S. Maney & Son Ltd 2014 DOI 10.1179/1743132814Y.0000000355
with poor outcomes within 24-hour of hospitalization for acute ischemic stroke.5 It is unclear how PSGS is related to long-term outcomes. We investigated the association between PSGS (measured as HbA1c) and all-cause mortality as well as poor functional outcome in the first year after acute ischemic stroke.
Methods Subjects ‘ACROSS’ was a large prospective cohort study conducted in China from 2008 to 2009 that investigated on abnormal glucose regulation in acute strokes. The baseline characteristics and the recruiting criteria have been published in Stroke.6 A total of 3450 patients with new onset ischemic stroke, intracerebral
Neurological Research
2014
VOL .
36
NO .
5
Wu et al.
hemorrhage, or subarachnoid hemorrhage were sequentially recruited in ACROSS. The present study is a sub-analysis (only ischemic stroke patients included) of the ACROSS. Out of 2639 patients with ischemic stroke, 2186 patients (82.8%) had HbA1c values collected on admission. Among these patients, 1948 patients had follow-up at 3 months, 1810 patients were available for follow-up at 12 months. Among patients with 1-year mRS scores, 1779 (81.3%) patients were analyzed for death analysis, and 13 patients from this group were not included in the analysis for 1-year poor functional outcome due to lack of accurate mRS scores (Fig. 1).
Baseline clinical data Acute stroke was diagnosed according to World Health Organization (WHO) criteria7 and the neuroimaging information was obtained from brain CT or MRI. Stroke subtypes were determined based on the Trial of ORG 10172 in Acute Stroke Treatment (TOAST).8 Baseline HbA1c value as well as other blood biochemical variables were obtained following an 8–12 fasting state within the first 24 hours after admission. All the variables were analyzed in the certified central laboratory. HbA1c was measured using ‘high performance liquid chromatographic analysis (HPLC)’ with a Bio-Rad Variant II analyzer (Bio-Rad Laboratories, Hercules, CA, USA) with a reference value of 4.1–6.5%, which is aligned with the Diabetes Control and Complications Trial and National Glycohemoglobin Standardization Program (NGSP) standards.9 The intra-assay coefficient of variation (CV) was 2.5%, and the inter-assay CV was ,4.0%, both within NGSP acceptable limits.10 HbA1c was classified into four groups according to quartiles (,5.5% [,37 mmol/mol], 5.5 to ,6.1% [37 to 43 mmol/mol], 6.1 to ,7.2% [43 to ,55 mmol/mol], 7.2% [ 55 mmol/mol]). Data on age, gender, education status, alcohol intake, tobacco use, body mass index (BMI), systolic blood pressure (SBP), diastolic blood pressure (DBP), National Institutes of Health Stroke scale (NIHSS),11 and past medical history were collected at admission. Diabetes diagnosis was based on a self-reported history of diabetes confirmed by the medical records or using an oral hypoglycemic agent for at least 6 months.
Ethics The Ethics Committees of Beijing Tiantan Hospital at all participating centers approved the procedures. Written informed consent was obtained from all patients or from the designated family member when the patient was unable to complete it.
Clinical assessment during hospitalization and follow-up Blood biochemical samples were obtained within 24 hours after admission. The values of the homeostasis
HbA1c and all-cause mortality after ischemic stroke
model assessment of insulin resistance (HOMA2-IR)12 were calculated by HOMA2-IR model available from www.ocdem.ox.ac.uk. Telephone interviews were conducted for 3- and 12-month follow-up. Death was defined as all-cause mortality and confirmed through telephone interviewing by two different representatives. All interviewers were centrally trained on the usage of the interview protocol. Poor functional outcome was defined as mRS (2–6); Mortality was indicated by mRS 5 6.
Statistical analysis The clinical characteristics among HbA1c quartiles were compared using one-way ANOVA or Kruskal– Wallis H test for continuous variables. Categorical variables were analyzed by chi-square test. Univariate analysis was conducted followed by multivariate logistic regression analysis for poor outcomes. Risk factors were first compared among the four HbA1c quartiles, and then the significant ones were selected based on the initial assessment and further for comparison of the HbA1c level. The interaction effects analysis between HbA1c and the related risk factors were also assessed in the multivariate analysis. The association between the risk factors and the poor outcome were assessed by univariate logistic regression. Multivariate models were established for evaluating the association between HbA1c and outcomes. The association between HbA1c and 1-year all-cause mortality was further analyzed stratified by insulin use and history of DM, since insulin use is associated with death risk13 and both DM and insulin use are closely related to post-stroke glycemic control. The variables include traditional confounding factors, newly identified risk factors including HOMA2-IR, white blood cell count, insulin therapy,14,15 complications during hospitalization as well as factors that may alter HbA1c levels or outcomes including red blood cell count, hemoglobin, and medications (model 1 for 1-year allcause mortality, Model 2 and 3 for 3-month and 1-year poor functional outcome). Statistical analyses were conducted using the SPSS 17.0 software (SPSS Inc., Chicago, IL, USA).All P values are two-sided and considered significant when P , 0.05.
Results Demographics of the patients Table 1 shows clinical characteristics of patients according to HbA1c quartiles. The patients in the highest HbA1c quartile was much older, had more moderate alcohol intake, much higher BMI but not much severer ischemic strokes at baseline, and received more insulin therapy. Patients with the other three higher quartiles had much higher level of FG, triglyceride, SBP, DBP at baseline, and received more
Neurological Research
2014
VOL .
36
NO .
5
445
Wu et al.
HbA1c and all-cause mortality after ischemic stroke
Figure 1 Flow chart of patient selection mRS, modified Rankin scale.
446
Neurological Research
2014
VOL .
36
NO .
5
Wu et al.
therapy of anti-hypertension and lipid-lowering agents. The significant variables in Table 1 were further categorized into different levels for comparison of HbA1c level (Table 2). Female had a higher level of HbA1c than male. Most cardiovascular risk factors were significant associated with a higher level of HbA1c except moderate to severe drinking and current smoking in our data. Table 3 shows the association of risk factors with poor clinical outcome. Red cell counts and hemoglobin level were protective factors for all poor outcomes. While age of over 45, increasing white cell counts (per 16109/l), previous DM, previous hypertension, stroke severity at admission (NIHSS 9), and insulin use in hospital were all significant with poor clinical outcome.
Association between HbA1c and 1-year all-cause mortality Multivariate analysis indicated that the increased risk of all-cause mortality among higher HbA1c quartiles had a significant trend (P 5 0.009), particularly in HbA1c quartile 7.2% [ 55 mmol/mol]. In Model 1 (Fig. 3),
HbA1c and all-cause mortality after ischemic stroke
the increased risk in HbA1c 7.2% [ 55 mmol/mol] is two-fold higher than the lowest quartile (OR 5 2.45, P 5 0.006). When further stratified by history of DM and insulin use in hospital with the same variables as those adjusted in Model 1 (Table 4), in patients without previous DM and without insulin use in hospital, HbA1c levels were still significantly associated with 1-year all-cause mortality (P for trend 0.020), and particularly in HbA1c quartile 7.2% [ 55 mmol/mol] (P 5 0.004). However, among those with previous DM, whether insulin use in hospital or not, and those with insulin use but without previous DM, the all-cause mortality risk of patients in higher HbA1c levels was not significantly higher than that in lower HbA1c levels (all P . 0.05). All the interaction effect between HbA1c and risk factors significant in the univariate analysis (Table 1 and 2) including age levels, gender, tobacco use, drinking, BMI, stroke severity, past medical history of hypertension, dyslipidemia, coronary heart disease and DM, triglyceride, HOMA-IR, and insulin use in hospital were assessed in the above multivariate
Table 1 Clinical characteristics according to HbA1c quartiles HbA1c quartiles Baseline characteristics Age* (years) Male (%)* Tobacco use (%)* Never Quit Current Alcohol intake (%)* Never Moderate Severe BMI* NIHSS*, median(Q1–Q3) Past medical history Hypertension (%)* Hyperlipidemia (%)* Diabetes (%)* Coronary heart disease (%)* HbA1c (%)* Fasting glucose (mmol/l) * Triglyceride (mmol/l) * HOMA2-IR* Ischemic stroke subtypes (%) Cardioembolism Non-Cardioembolism Large artery atherothrombosis Small artery occlusion Other Medication in hospital Insulin therapy (%)* Anti-thrombotic agent (%) Anti-hypertension agent (%) Lipid-lowering agent (%)
,5.5% (n 5 540)
5.5 to ,6.1% (n 5 507)
6.1 to ,7.2% (n 5 579)
7.2% (n 5 560)
60 ¡ 13 367 (68.0)
63 ¡ 12 331 (65.3)
65 ¡ 12 339 (58.5)
63 ¡ 11 332 (59.3)
316 (58.5) 44 (8.1) 180 (33.3)
261 (51.5) 54 (10.7) 192 (37.9)
352 (60.8) 59 (10.2) 168 (29.0)
332 (59.3) 45 (8.0) 183 (32.7)
365 (67.6) 88 (16.3) 87 (16.1) 24.3 ¡ 3.2 5 (2–9)
327 (64.5) 87 (17.2) 93 (18.3) 24.6 ¡ 3.8 4 (4–8)
410 (70.8) 95 (16.4) 74 (12.8) 25.0 ¡ 3.6 4 (2–8)
318 (58.9) 38 (7.0) 34 (6.3) 48 (8.9) 4.8 ¡ 0.5 5.3 ¡ 1.3 1.56 ¡ 0.97 1.11 ¡ 0.71
296 (58.4) 66 (13.0) 38 (7.5) 55 (10.8) 5.8 ¡ 0.2 5.3 ¡ 1.1 1.72 ¡ 1.00 1.29 ¡ 0.80
355 (61.3) 81 (14.0) 134 (23.1) 99 (17.1) 6.5 ¡ 0.3 6.1 ¡ 1.8 1.77 ¡ 1.00 1.27 ¡ 0.79
377 (67.3) 104 (18.6) 79 (14.1) 25.3 ¡ 4.0 4 (2–7) 372 (66.4) 86 (24.4) 334 (59.6) 84 (15.0) 9.1 ¡ 1.7 9.6 ¡ 3.5 2.14 ¡ 1.61 1.39 ¡ 0.83
21 519 341 143 35
(7.6) (96.1) (63.1) (26.5) (6.5)
36 471 313 124 34
(7.1) (92.9) (61.7) (24.5) (6.7)
40 539 341 161 37
(6.9) (93.1) (58.9) (27.8) (6.4)
23 537 362 147 28
(4.1) (95.9) (64.6) (26.3) (5.0)
22 505 208 358
(4.1) (93.5) (38.5) (66.3)
16 468 227 367
(3.2) (92.3) (44.8) (72.4)
38 550 286 436
(6.6) (95.0) (49.4) (75.3)
235 527 262 426
(42.0) (94.1) (46.8) (76.1)
* Indicates P , 0.05 when comparing between four groups. HbA1c: glycated hemoglobin; mRS: modified Rankin scale; NIHSS: the National Institutes of Health Stroke scale; HOMA2-IR: the correctly solved computer model for homeostasis model assessment of insulin resistance; OTHER: other causes and stroke of undetermined etiology.
Neurological Research
2014
VOL .
36
NO .
5
447
Wu et al.
HbA1c and all-cause mortality after ischemic stroke
Table 2 Comparison of HbA1c level according to clinical characteristics Clinical characteristics
HbA1c (%)
Age (years)
Gender Smoking Drinking Body mass index (kg/m2) Stroke severity (NIHSS) Past medical histories Hypertension Dyslipidemia Diabetes mellitus Coronary heart disease Triglyceride (mmol/l) HOMA-IR Insulin therapy in hospital
,45 45–65 .65 Male Female Current No Never Moderate Severe 25 ,25 9 ,9
6.0 6.7 6.6 6.5 6.7 6.6 6.6 6.6 6.7 6.4 6.8 6.5 6.5 6.6
¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡
1.6 2.0 1.7 1.8 1.9 1.8 1.9 1.8 2.0 1.8 1.9 1.8 1.9 1.8
Yes No Yes No Yes No Yes No 1.7 ,1.7 .1 1 Yes No
6.7 6.4 7.0 6.5 8.1 6.1 6.8 6.6 6.9 6.4 6.8 6.2 8.7 6.3
¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡
1.9 1.8 1.9 1.8 2.0 1.5 1.8 1.8 2.0 1.7 2.0 1.5 2.3 1.5
P ,0.001
0.005 0.617 0.06
,0.001 0.011
0.004 ,0.001 ,0.001 0.015 ,0.001 ,0.001 ,0.001
analysis, and were not significant (all P values were .0.05).
Association between HbA1c and 3-month and 1-year poor functional outcome Figure 2 shows mRS ratios according to HbA1c quartiles. Higher HbA1c quartiles occurred more mRS from 2–6. The 3-month and 1-year poor functional outcome multivariate analysis (Model 2 and 3 in Fig. 3) indicated that compared to the lowest level of HbA1c ,5.5% [,37 mmol/mol], HbA1c in other three higher quartiles had no significant trend in increasing risk of poor functional outcome (Model 2 for 3-month and Model 3 for 1-year, P for trend were 0.165 and 0.718). All the interaction effect between HbA1c and the related risk factors were all not significant (all P . 0.05).
Other significant risk factor in multivariate analysis Age of over 65 was a risk factor of 1-year all-cause mortality and 3-month and 1-year poor functional outcome (OR with 95% CI, 1.12 (1.04–3.28), 1.06 (1.01– 4.13), and 1.13 (1.05–3.42), respectively). Previous hypertension and SBP were associated with 1-year allcause mortality (OR with 95% CI, 1.31 (1.02–3.58) for previous hypertension and 1.16 (1.08–4.22).
Discussion Given the annual morbidity and mortality rates caused by stroke have exceeded cardiovascular disease to become the leading cause of death and disability16 and
448
Neurological Research
2014
VOL .
36
NO .
5
the increased prevalence of abnormal glycometabolism in China,17 it is critical to have thorough understanding on how PSGS may influence prognosis of stroke. HbA1c is an accurate and stable parameter18 that reflects long-term glycemic control.5,19 Pre-stroke glycemic status (measured as HbA1c) was first proposed by the Japan Diabetes Society.4 Our current study indicated that even patients with a good PSGS (baseline HbA1c of 5.5% [ 37 mmol/mol]) was significantly associated with an increased risk of 1year all-cause mortality after acute first-ever ischemic stroke onset. In the highest HbA1c quartile ( 7.2% [ 55 mmol/mol]) the risk of 1-year all-cause mortality is dramatically increased. Such an association is prominent in acute ischemic stroke without a history of DM and without post-stroke insulin use.
HbA1c and 1-year all-cause mortality Our finding indicated HbA1c was significantly associated with increased risk for the all-cause mortality. Such an association started from HbA1c 5 5.5% [37 mmol/mol] and was sharply enlarged when HbA1c achieved 7.2% [55 mmol/mol] or higher. This is consistent with a previous study suggesting high ‘normal’ HbA1c is associated with all-cause mortality in all cardiovascular diseases.20 However, the association of HbA1c and death in the present study was somewhat different with that from other studies. It was indicated in the present study that an increased risk of all-cause mortality in the HbA1c level 7.2% [ 55 mmol/mol] which is somewhat inconsistent with a study from Action in Diabetes and Vascular disease: PreterAx and DiamicroN-MR Controlled Evaluation (ADVANCE). In their study, HbA1c over 7.0% [53 mmol/mol] was regarded as a significant risk for death from macrovascular diseases21. No association was observed in Kamouchi’s study between HbA1c and mortality during hospitalization5 while it is not the case when long-term outcomes are assessed in our studies. Monami M’s study with a mean 5.7 ¡ 3.5 year follow-up found HbA1c of 6.5–7.4% [48–57 mmol/mol] had the lowest death risk. They also stated that glycemic target should be individualized according to the characteristics of each type 2 diabetics because insulin treatment significantly increased the death risk associated with HbA1c of ,6.5% [48 mmol/ mol] in the study.13 Several reasons may account for the discrepancy. Our study focused on ischemic stroke (separate from cardiovascular diseases) while the ADVANCE trial included macro-and micro-vascular diseases; Shortterm poor outcomes were studied in Kamouchi’s study while much longer ones are assessed in ours; Not like Monami M’s study, ours enrolled both diabetics and non-diabetics and had a shorter followup period.
Wu et al.
HbA1c and all-cause mortality after ischemic stroke
Table 3 Association of clinical variables with poor clinical outcome 1-year all-cause 3-month poor 1-year poor mortality OR (95% CI) outcome OR (95% CI) outcome OR (95% CI)
Clinical characteristics Age (years)
Gender Smoking Drinking
Past medical histories Hypertension Dyslipidemia Diabetes mellitus Coronary heart disease Body mass index (kg/m2) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) NIHSS Red cell counts, per 161012/l White cell counts, per 16109/l Hemoglobin, per 1 g/l High density lipoprotein (mmol/l) Low density lipoprotein (mmol/l) Triglyceride (mmol/l) HOMA-IR In-hospital therapy Antihypertensive Lipid-lowering Antithrombotic Insulin therapy
.65 45–65 ,45 Female Male Current No Moderate to severe Mild Never
4.65 2.80 Ref 1.08 Ref 0.82 Ref 1.78
(1.86–11.60) (1.12–7.04)
1.51 (0.57–4.05) Ref
0.54 (0.41–0.70) Ref
0.77 (0.59–1.00) Ref
Yes No Yes No Yes No Yes No 25 ,25 140 ,140 90 ,90 9 ,9
1.55 ,1.55 3.37 ,3.37 1.7 ,1.7 .1 1
1.32 Ref 0.63 Ref 1.48 Ref 1.32 Ref 0.65 Ref 1.00 Ref 1.00 Ref 2.05 Ref 0.76 1.10 0.99 1.02 Ref 0.87 Ref 0.99 Ref 0.95 Ref
1.22 Ref 0.92 Ref 1.26 Ref 1.27 Ref 0.98 Ref 1.26 Ref 1.04 Ref 8.60 Ref 0.67 1.10 0.99 1.12 Ref 0.94 Ref 0.85 Ref 1.05 Ref
1.17 Ref 0.69 Ref 1.54 Ref 1.02 Ref 0.82 Ref 1.16 Ref 0.98 Ref 5.16 Ref 0.65 1.09 0.99 0.87 Ref 0.87 Ref 0.83 Ref 1.07 Ref
No Yes No Yes No Yes No Yes
1.00 Ref 1.50 Ref 1.86 Ref 0.53 Ref
(0.81–1.47) (0.60–1.10) (0.70–4.49)
(1.01–1.81) (0.37–1.06) (1.08–2.02) (0.88–1.99) (0.48–0.88) (0.73–1.39) (0.74–1.36) (1.50–2.81) (0.59–0.99) (1.05–1.16) (0.98–1.00) (0.64–1.63) (0.63–1.19) (0.73–1.33) (0.69–1.29)
(0.75–1.35) (1.09–2.05) (1.08–3.22) (0.37–0.75)
3.30 1.92 Ref 1.80 Ref 0.63 Ref 0.75
1.06 Ref 1.29 Ref 1.04 Ref 0.67 Ref
(2.15–5.05) (1.25–2.95) (1.48–2.17) (0.52–0.76) (0.58–0.97)
(1.00–1.48) (0.69–1.23) (1.02–1.55) (0.97–1.66) (0.81–1.19) (0.92–1.54) (0.86–1.25) (6.72–11.02) (0.57–0.79) (1.05–1.14) (0.98–1.00) (0.83–1.51) (0.77–1.15) (0.70–1.03) (0.88–1.27)
(0.88–1.28) (1.05–1.60) (0.68–1.59) (0.52–0.87)
4.36 2.40 Ref 1.64 Ref 0.62 Ref 0.72
0.88 Ref 1.20 Ref 1.06 Ref 0.56 Ref
(2.82–6.77) (1.55–3.71) (1.35–1.99) (0.51–0.76) (0.55–0.94)
(1.02–1.42) (0.52–1.03) (1.25–1.91) (0.78–1.35) (0.68–0.99) (0.94–1.43) (0.81–1.20) (4.62–6.62) (0.55–0.77) (1.04–1.13) (0.98–0.99) (0.64–1.18) (0.71–1.06) (0.68–1.00) (0.88–1.31)
(0.73–1.07) (0.96–1.48) (0.69–1.64) (0.43–0.74)
OR indicates odds ratio, CI indicates confidence interval. NIHSS indicates the National Institutes of Health Stroke scale; HOMA-IR indicates the correctly solved computer model for homeostasis model assessment of insulin resistance.
In the further stratified analysis, the association between baseline HbA1c (implying PSGS) and 1-year all-cause mortality was only observed in patients without a history of DM and without post-stroke insulin use in hospital. On the contrary, in those with a history of DM or accepting insulin therapy in hospital, higher HbA1c levels were not detrimental to 1-year all-cause mortality. Those without a history of DM but accepting insulin use in hospital were in a large possibility of newly-onset DM.22 Thus, we conclude that baseline HbA1c, indicating PSGS, is very likely associated with risk of mortality only in ischemic stroke patients without DM. The more prominent association between glycemia and poor outcome in non-diabetes than diabetes was partly in
line with Zsuga, et al.23 The difference was that poststroke hyperglycemia was also significant in DM group. The index was hyperglycemia post stroke in their research but PSGS in our study and the different following-up period might lead to the discordance. In DM, post stroke glycemic control was not only necessarily needed but also significant in the whole course of treatment and affecting later prognosis, however, it is not necessarily applied in each nondiabetic acute ischemic stroke patient, thus the effect of glycemia on mortality might be largely attributed to PSGS or acute hyperglycemia response22 but not the post-stroke glycemic control level in those without DM. As fasting plasma glucose, partly reflecting
Neurological Research
2014
VOL .
36
NO .
5
449
Wu et al.
HbA1c and all-cause mortality after ischemic stroke
Figure 2 Description for the 3-month and 1-year mRS ratios in HbA1c quartiles mRS, modified Rankin scale.
acute hyperglycemia response, was included in the multivariate analysis and it showed not significant, we believed that PSGS is significantly detrimental to 1-year all-cause mortality in non-diabetic ischemic stroke patients independently of acute hyperglycemia response. We also learned from our study that PSGS was not significant with 1-year all-cause mortality in acute ischemic stroke patients with DM. Concerning the consequence of post-stroke glycemic control in the treatment of DM and the following prognosis, we postulated that post-stroke glycemic control level has much more effect on the poor outcome than PSGS in acute ischemic stroke with DM. However, we still cannot well compare the effect of PSGS and poststroke glycemic control on 1-year all-cause mortality, since the longitude change of post-stroke glycemic level was not recorded in our study. Thus, further investigation is needed for comparison between poststroke glycemic control and PSGS in the association with post-stroke outcome in acute stroke patients.
HbA1c and poor functional outcome within 1 year following acute ischemic stroke onset A previous study from Kamouchi’s group has shown that increased HbA1c is associated with worsening of outcomes during hospitalization after acute ischemic
Circles in black stand for OR; diamond in black stand for 95% CI.; q2: HbA1c quartile 5.5 to ,6.1%;; q3: HbA1c quartile 6.1 to ,7.2%;; q4: HbA1c quartile 7.2%;; OR: odds ratio;; Model 1 is on 1-year mortality adjusted for age, gender, education status, tobacco use, alcohol intake, medical histories of hypertension and coronary heart disease, body mass index (BMI), systolic (SBP) and diastolic blood pressure (DBP), fasting glucose, high and low density lipoprotein, white blood cell counts, red blood cell counts, hemoglobin level, NIHSS, TOAST, and OCSP stroke subtypes, HOMA2-IR, triglyceride, the usages of insulin, anti-hypertension agent, lipid-lowering agent, and anti-thrombotic agent during hospitalization, urinary infection and pneumonia during hospitalization.; Model 2 is on 3-month poor functional outcomes analysis adjusted for age, gender, education status, alcohol intake, tobacco use, medical histories of hypertension and coronary heart disease, BMI, SBP and DBP, NIHSS, fasting glucose, high and low density lipoprotein, HOMA2-IR, triglyceride, white blood cell counts, red blood cell counts, hemoglobin level, OCSP, and TOAST subtypes, the usages of anti-hypertension agent, lipid-lowering agent, and anti-thrombotic agent during hospitalization, urinary infection, and pneumonia during hospitalization.; Model 3 is on 1-year poor functional outcomes analysis adjusted for variables in model 2 plus the usages of 3-month anti-hypertension agent, lipid-lowering agent, and antithrombotic agent, and 3-month stroke recurrence. Figure 3 HbA1c quartiles versus risk of all-cause mortality and poor functional outcome compared to HbA1c ,5.5%
stroke.5 These results indicated that poor functional outcome was higher in patients with poor PSGS when HbA1c is over 8.4% [68 mmol/mol]. That was not in line with our findings. Our finding did not suggest higher levels of HbA1c had a higher risk of poor functional outcome within 1 year after ischemic stroke. The longer follow up period in our study might partly explain the paradox.
Table 4 Association between HbA1c levels and 1-year all-cause mortality stratified by diabetes mellitus (DM) and insulin use* Diabetes mellitus HbA1c levels Q1 Q2 Q3 Q4 P for trend
(OR with 95% CI)
Insulin use
P Without insulin use
Ref 3.00 (0.25–16.32) 0.52 (0.04–6.58) 1.28 (0.15–10.88)
Ref 0.388 0.46 (0.04–5.41) 0.612 1.17 (0.23–5.81) 0.823 2.45 (0.54–11.15) 0.117
Without diabetes mellitus
P
0.534 0.851 0.245 0.377
OR indicates odds ratio, CI indicates confidence interval. * Adjusted for the same variables in Model 1. Q1: ,5.5%; Q2: 5.5 to ,6.1%; Q3: 6.1 to ,7.2%; Q4: 7.2%.
450
Neurological Research
2014
VOL .
36
NO .
5
Insulin use Ref 0.88 (0.06–12.97 2.62 (0.30–15.00) 0.97 (0.18–5.23)
(OR with 95% CI)
P
0.923 0.383 0.970 0.678
Without insulin use Ref 1.09 (0.60–1.97) 1.40 (0.79–2.45) 2.54 (1.34–4.82)
P
0.780 0.248 0.004 0.020
Wu et al.
HbA1c in ischemic stroke needs to be highlighted in clinician The present study showed that even a non-diabetic HbA1c value (,6.5%) has a trend for an increased risk of 1-year all-cause mortality in acute ischemic stroke, especially in patients without diabetes. HbA1c is not only a tool for assessing glycemic control but also meaningful to those without diabetes in acute ischemic stroke. Since its significance for secondary stroke prevention24 and poor clinical outcome, baseline HbA1c is recommended as a routine test in all acute ischemic stroke patients regardless of diabetes diagnosis. For patients whose HbA1c are in the high ‘normal’ level ( 6.5%), clinicians should be aware that the 1-year mortality risk might be increased 7% or so in this population. Pre-stroke glycemic status indicated by HbA1c at admission after acute ischemic stroke onset should be a vital notion to clinicians as the HbA1c level of over 7.2% predicts a two-fold or more increase in the risk for 1-year mortality. Since the association between PSGS and all-cause mortality differs in glycometabolism diagnosis and post stroke glycemic control status (use insulin or not), further investigation is needed for comparison between PSGS and post-stroke glycemic control in the association with poor clinical outcomes in acute stroke patients stratified by diagnosis of glycometabolism, including DM, prediabetes, and normoglycemia.
Other variables significant for poor clinical outcome Age was a traditional risk factor of poor outcome which was also observed in our study. Hypertension25 or SBP26 at admission were also related to higher risk of stroke and its poor clinical outcome which was also confirmed by others.
Limitations and strengths There were some limitations in our study. In total, 452 patients (17.2%) with ischemic stroke were excluded due to lack of HbA1c values, although this is slightly lower than that of a similar study from Japan (22.2%),5 a selective bias could not be avoided; Non-responder and non-mRS ratios were not as high as other clinical studies concerning HbA1c and CVD,27 which may potentially affect OR. Certain biochemical and physiological data, including HbA1c, SBP, and LDL, could not be collected before acute stroke or during the following-up period due to limitations of the current study’s design. Despite these limitations, we provided detailed and quantified evaluation for long-term prognosis of patients with ischemic stroke. Still many physiologic and biochemical parameters and traditional and newly established risk factors were accounted for multivariate analysis, such as blood pressure,28 insulin therapy, HOMA2IR, and complications during hospitalization.
HbA1c and all-cause mortality after ischemic stroke
Conclusions Elevated baseline HbA1c (from 5.5% [37 mmol/mol]) presenting pre-stroke glycemia status has a significant trend in increasing the risk of 1-year all-cause mortality. HbA1c (from 5.5% [37 mmol/mol]) is an independent risk factor for 1-year all-cause mortality after acute first-ever ischemic stroke onset. Such an association might be altered by glycometabolism status.
Disclaimer Statements Contributors SW drafted the manuscript; CW and KH helped to draft the manuscript; QJ helped to collect the data; GL, XW, AW and CW performed the statistical analysis; XZ participated in the design of the study and explanation of the data; LL, YW and YW participated in the design of the study. All authors read and approved the final manuscript. Funding This study was funded jointly by the Beijing Science and Technology Committee (grant no. 7102050, www.bjkw.gov.cn) and the National Science Foundation (grant no. 81071115, www.nsfc.gov.cn) Conflicts of interest None. Ethics approval The Ethics Committees of Beijing Tiantan Hospital at all participating centers approved the procedures. Written informed consent was obtained from all patients or from the designated family member when the patient was unable to complete it.
Acknowledgements We thank all participating hospitals, colleagues, patients, and caregivers involved in the study. This study was funded jointly by the Beijing Science and Technology Committee (grant no. 7102050, www.bjkw.gov.cn) and the National Science Foundation (grant no. 81071115, www.nsfc.gov.cn).
References 1 Jia Q, Zhao X, Wang C, Wang Y, Yan Y, Li H, et al. Diabetes and poor outcomes within 6 months after acute ischemic stroke: the China National Stroke Registry. Stroke. 2011;42:2758–62. 2 Fuentes B, Castillo J, San Jose´ B, Leira R, Serena J, Vivancos J, et al. The prognostic value of capillary glucose levels in acute stroke: the GLycemia in Acute Stroke (GLIAS) study. Stroke. 2009;40:562–8. 3 Chmielewska B, Hasiec T, Belniak-Legiec´ E, Stelmasiak Z. Usefulness of current and retrospective markers of glycemia for early prognostication of ischemic stroke in diabetics and nondiabetics. Ann Univ Mariae Curie Sklodowska Med. 1999;54:261–7. 4 Japan Diabetes Society. Treatment objectives and control indicators. In: Japan Diabetes Society, editor. Treatment guide for diabetes 2007. Tokyo: Bunkodo; 2007. p. 1–84. 5 Kamouchi M, Matsuki T, Hata J, Kuwashiro T, Ago T, Sambongi Y, et al. Prestroke glycemic control is associated with the functional outcome in acute ischemic stroke: the Fukuoka Stroke Registry. Stroke. 2011;42:2788–94. 6 Jia Q, Zheng H, Zhao X, Wang C, Liu G, Wang Y, et al. Abnormal glucose regulation in patients with acute stroke across China: prevalence and baseline patient characteristics. Stroke. 2012;43:650–7. 7 WHO Task Force on Stroke and other Cerebrovascular Disorders: Stroke-1989. Recommendations on stroke prevention, diagnosis, and therapy. Stroke. 1989;20:1407–31.
Neurological Research
2014
VOL .
36
NO .
5
451
Wu et al.
HbA1c and all-cause mortality after ischemic stroke
8 Adams HP Jr, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, et al. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke. 1993;24:35–41. 9 National Glycohemoglobin Standardization Program (NGSP). List of NGSP certified methods. Available from: http:// www.ngsp.org/certified.asp (accessed 2012 May 12). 10 Schwartz KL, Monsur JC, Bartoces MG, West PA, Neale AV. Correlation of same-visit HbA1c test with laboratory-based measurements: a MetroNet study. BMC Fam Pract. 2005;6:28. 11 Brott T, Adams HP Jr, Olinger CP, Marler JR, Barsan WG, Biller J, et al. Measurements of acute cerebral infarction: a clinical exami Renee Holleran, Robert Eberle, Vicki Hertzberg, Marvin nation scale. Stroke. 1989;20:864–70. 12 Levy JC, Matthews DR, Hermans MP. Correct homeostasis model assessment (HOMA) evaluation uses the computer program. Diabetes Care. 1998;21:2191–2. 13 Monami M, Vitale V, Lamanna C, Bartoli N, Martelli D, Zannoni S, et al. HbA1c levels and all-cause mortality in type 2 diabetic patients: epidemiological evidence of the need for personalized therapeutic targets. Nutr Metab Cardiovasc Dis. 2013;23:300–6. 14 Conway BN, May ME, Blot WJ. Mortality among low-income African Americans and whites with diabetes. Diabetes Care. 2012;35:2293–9. 15 Hayashi T, Kawashima S, Nomura H, Itoh H, Watanabe H, Ohrui T, et al. Age, gender, insulin and blood glucose control status alter the risk of ischemic heart disease and stroke among elderly diabetic patients. Cardiovasc Diabetol. 2011;10:86. 16 Liu L, Wang D, Wong KS, Wang Y. Stroke and stroke care in china: huge burden, significant workload, and a national priority. Stroke. 2011;42:3651–4. 17 Yang W, Lu J, Weng J, Jia W, Ji L, Xiao J, et al. Prevalence of diabetes among men and women in China. N Engl J Med. 2010;362:1090–101. 18 Derr R, Garrett E, Stacy GA, Saudek CD. Is HbA (1c) affected by glycemic instability? Diabetes Care. 2003;26:2728–33.
452
Neurological Research
2014
VOL .
36
NO .
5
19 Selvin E, Coresh J, Golden SH, Boland LL, Brancati FL, Steffes MW, et al. Glycemic control, atherosclerosis, and risk factors for cardiovascular disease in individuals with diabetes: the atherosclerosis risk in communities study. Diabetes Care. 2005;28:1965–73. 20 Smith SA. Higher ‘normal’ glycated hemoglobin levels were associated with increased risk for diabetes, CVD, stroke, and mortality in adults. Ann Intern Med. 2010;153:JC1–13. 21 Zoungas S, Chalmers J, Ninomiya T, Li Q, Cooper ME, Colagiuri S, et al. Association of HbA1c levels with vascular complications and death in patients with type 2 diabetes: evidence of glycaemic thresholds. Diabetologia. 2012;55: 636–43. 22 Jia Q, Zheng H, Liu L, Zhao X, Wang C, Jing J, et al. Persistence and predictors of abnormal glucose metabolisms in patients after acute stroke. Neurol Res. 2010;32:359–65. 23 Zsuga J, Gesztelyi R, Kemeny-Beke A, Fekete K, Mihalka L, Adrienn SM, et al. Different effect of hyperglycemia on stroke outcome in non-diabetic and diabetic patients – a cohort study. Neurol Res. 2012;34:72–9. 24 Wu S, Shi Y, Wang C, Jia Q, Zhang N, Zhao X, et al. Glycated hemoglobin independently predicts stroke recurrence within one year after acute first-ever non-cardioembolic strokes onset in A Chinese cohort study. PLoS One. 2013;8:e80690. 25 Yong H, Foody J, Linong J, Dong Z, Wang Y, Ma L, et al. A systematic literature review of risk factors for stroke in China. Cardiol Rev. 2013;21:77–93. 26 Rodrı´guez-Ya´n˜ez M, Castellanos M, Blanco M, Garcı´a MM, Nombela F, Serena J, et al. New-onset hypertension and inflammatory response/poor outcome in acute ischemic stroke. Neurology. 2006;67:1973–8. 27 Adams RJ, Appleton SL, Hill CL, Wilson DH, Taylor AW, Chittleborough CR, et al. Independent association of HbA1c and Incident cardiovascular disease in people without diabetes. Obesity (Silver Spring). 2009;17:559–63. 28 Jia Z, Wu S, Zhou Y, Wang W, Liu X, Wang L, et al. Metabolic syndrome and its components as predictors of stroke in middle-aged and elderly Chinese people. Neurol Res. 2011;33:453–9.
