Associations Between Estimated Glomerular
Filtration Rate and Stroke Outcomes in Diabetic Versus
Nondiabetic Patients
Yang Luo, MD, PhD*; Xianwei Wang, MD*; Kunihiro Matsushita, MD, PhD;
Chunxue Wang, MD, PhD; Xingquan Zhao, MD, PhD; Bo Hu, PhD; Liping Liu, MD, PhD;
Hao Li, PhD; Gaifen Liu, PhD; Qian Jia, MD, PhD; Yilong Wang, MD, PhD; Yongjun Wang, MD;
on behalf of the CNSR Investigators†
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Background and Purpose—Low estimated glomerular filtration rate (eGFR) is known to be associated with clinical adverse
outcomes. However, whether diabetes mellitus influences the association between eGFR and prognosis of stroke is still
not elucidated.
Methods—Prospective cohort of 17 280 (nondiabetic 12 498 and diabetic 4782) Chinese patients with acute stroke from
the China National Stroke Registry and from abnormal glucose regulation in patients with acute stroke across China
(ACROSS) between 2007 and 2009 were followed-up for 1 year for all-cause mortality, stroke recurrence, and stroke
disability related to baseline eGFR in the presence and absence of diabetes mellitus.
Results—Among nondiabetic patients, as compared with eGFR of 90 to 119 mL/min per 1.73 m2, the adjusted odds ratio
of lower eGFR of < 45 mL/min per 1.73 m2 was 2.79 (95% confidence interval, 2.09–3.73) for all-cause mortality, 2.28
(1.74–2.98) for stroke recurrence, and 1.53 (1.16–2.01) for stroke disability; higher eGFR of ≥120 mL/min per 1.73 m2
was just significantly associated with higher risk of all-cause mortality (odds ratio, 1.38; 95% confidence interval, 1.02–
1.86) but not with other outcomes. In diabetic patients, the adjusted odds ratios of all-cause mortality, stroke recurrence,
and stroke disability in lower eGFR were 2.16 (1.51–3.08), 1.43 (1.02–2.00), and 1.38 (0.98–1.95), respectively; higher
eGFR was significantly associated with higher risks of all stroke outcomes.
Conclusions—Decreased eGFR (<45 mL/min per 1.73 m2) is a strong predictor of all-cause mortality, stroke recurrence,
and stroke disability in diabetic and nondiabetic patients with acute stroke. Increased eGFR (≥120 mL/min per 1.73 m2) is
associated with all of stroke outcomes in diabetic patients and linked to all-cause mortality in nondiabetic patients. (Stroke.
2014;45:2887-2893.)
Key Words: assessment, outcomes ◼ diabetes mellitus ◼ stroke
S
troke has become the leading cause of death and adult
disability in China.1 Identifying and controlling strokerelated risk factors would be extremely important in secondary prevention of stroke.2 Kidney dysfunction, characterized
by low estimated glomerular filtration rate (eGFR), is not
only associated with all-cause and cardiovascular mortality
in general population but also linked with long-term clinical
outcomes in patients with stroke.3–5 However, whether the link
between reduced eGFR and outcomes after acute stroke could
be modified by the presence or absence of some particular diseases is still uncertain.
The prevalence of diabetes mellitus in China has been
increasing significantly in the past 2 decades and is now
reaching epidemic proportions.6,7 Because diabetes mellitus
has been regarded as a potent risk factor for both stroke and
kidney dysfunction,8,9 we hereby hypothesized that the association of kidney dysfunction and adverse outcomes after acute
stroke could be modified by the existence of diabetes mellitus.
We tested this hypothesis by assessing the pattern and magnitude of associations of eGFR with outcomes after acute stroke
in diabetic versus nondiabetic patients with stroke recruited
from the China National Stroke Registry.
Received March 5, 2014; final revision received July 16, 2014; accepted July 18, 2014.
From the Department of Nephrology (Y.L.) and Department of Neurology (X.W., C.W., X.Z., L.L., H.L., G.L., Q.J., Yilong Wang, Yongjun Wang),
Beijing Tiantan Hospital, Capital Medical University, Beijing, China; Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health,
Baltimore, MD (K.M.); and Department of Quantitative Health Sciences, Cleveland Clinic, Cleveland, OH (B.H.).
†A list of all CNSR Investigators is given in the Appendix.
*Drs Luo and X. Wang contributed equally.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.114.
005380/-/DC1.
Correspondence to Yilong Wang, MD, PhD, or Yongjun Wang, MD, Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, No
6 Tiantanxili, Dongcheng District, Beijing 100050, China. E-mail [email protected] or [email protected]
© 2014 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org
DOI: 10.1161/STROKEAHA.114.005380
2887
2888 Stroke October 2014
Methods
Study Participants
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
The study population was from the China National Stroke Registry
(CNSR) between 2007 and 2008 and from abnormal glucose regulation in patients with acute stroke across China (ACROSS) between
2008 and 2009. ACROSS was a derivative study of CNSR, which
meant that ACROSS framework was based on the pre-existing CNSR
study. Briefly, CNSR is a national hospital-based, prospective cohort
study designed to evaluate the quality of care for hospitalized patients
with stroke and measure the clinical and functional outcomes at 12
months after disease onset. The rationale and design of CNSR have
been described previously.10 ACROSS is a prospective study aimed
at investigating the prevalence and distribution of abnormal glucose
regulation and assessing the impact of abnormal glucose regulation
on the 1-year outcome among hospitalized patients with ischemic or
hemorrhagic stroke.11 The reason for combination of these studies was
that the same basic characteristic variables were collected and that the
same outcome measures were assessed in the 2 cohorts. Of 25 666
patients from CNSR (22 216) and ACROSS (3450), 2508 (9.8%) patients were excluded because of undetermined diagnoses, incomplete
information at baseline, hospitals transfers, and missing information
of diabetes mellitus, leaving 23 158 patients with stroke for study sample (Figure I in the online-only Data Supplement). Of 23 158 patients
with or without diabetes mellitus, 5878 were excluded because of lack
of serum creatinine on admission, missing 1-year follow-up information, and baseline age or sex (Table 1; Figure I in the online-only Data
Supplement). Written informed consent was signed by patients or their
legally authorized representatives. The study was approved by the central institutional review board at Beijing Tiantan Hospital.
Data Collection
Trained research coordinators of CNSR and ACROSS at each institute collected baseline information, including patient demographics,
vascular risk factors, stroke severity (National Institutes of Health
Stroke Scale [NIHSS]), medication use, imaging studies, diagnosis,
and complications. Vascular risk factors included history of stroke,
hypertension, dyslipidemia, diabetes mellitus, atrial fibrillation, coronary heart disease, current or previous smoking, and moderate or
Table 1. Demographic and Clinical Characteristics Between
Patients Included and Patients Excluded
Characteristics
Age (mean±SD), y
Patients
Included (17 280)
63.5±12.8
Patients
Excluded (5878)
64.0±12.8
P Value
0.02
6664 (38.6)
2260 (38.5)
0.97
History of stroke, n (%)
4790 (27.7)
1770 (30.1)
<0.001
Hypertension, n (%)
10 890 (63.0)
3710 (63.1)
0.90
Dyslipidemia, n (%)
1659 (9.6)
619 (10.5)
0.04
933 (5.4)
332 (5.6)
0.47
Coronary heart
disease, n (%)
2083 (12.1)
685 (11.7)
0.41
Ever smoking, n (%)
6754 (39.2)
2257 (38.6)
0.43
Heavy alcohol, n (%)
1828 (10.6)
578 (9.8)
0.11
BMI at admission,
median (IQR), kg/m2
24.2 (22.0–26.3) 24.15 (22.0–26.2)
Vascular risk factors
Baseline NIHSS,
median (IQR)
Pneumonia and urethral
infection, n (%)
Kidney Function Measurement
eGFR was calculated using the Chronic Kidney Disease Epidemiology
Collaboration creatinine equation with adjusted coefficient of 1.1 for
the Asian population.12,13 Serum creatinine level was measured on
admission using the Jaffe method and was reduced by 5% for the
Chronic Kidney Disease Epidemiology Collaboration equation.14
Outcome Assessment
Follow-up was done by telephone interview. Data collection was performed by trained research personnel who were blinded to patients’
baseline clinical status. They completed training modules on NIHSS
and modified Rankin Scale and were certified. Patients were asked
the standardized follow-up questions at 12 months after stroke onset.
Outcome data collected included all-cause mortality, stroke recurrence, and stroke disability. Death was confirmed either by the death
certificates from local citizen registries or from the treating hospitals. If such records were unavailable, death was confirmed if death
was reported on 2 consecutive follow-up contacts with ≥2 different proxies. Recurrent stroke included ischemic stroke, intracranial
hemorrhage, and subarachnoid hemorrhage. Recurrent stroke end
points comprised fatal stroke and nonfatal stroke. We also evaluated
the composite of death or recurrent stroke in the prognostic analysis.
Stroke disability was defined as modified Rankin Scale of 2 to 6.
Statistical Analysis
Women, n (%)
Atrial fibrillation, n (%)
heavy alcohol consumption (≥2 standardized alcohol drinks per day).
Diabetes mellitus was defined as fasting glucose level ≥7.0 mmol/L
(126 mg/dL), the nonfasting glucose concentration ≥11.1 mmol/L
(200 mg/dL) with classic symptoms of hyperglycemia or hyperglycemic crisis, any use of glucose-lowering drugs, or any self-reported
history of diabetes mellitus. Hypertension was defined as systolic
blood pressure ≥140 mm Hg, diastolic blood pressure ≥90 mm Hg,
any use of antihypertensive drug, or self-reported history of hypertension. Dyslipidemia was defined as serum triglyceride ≥150 mg/
dL, low-density lipoprotein cholesterol ≥130 mg/dL, high-density
lipoprotein cholesterol ≤40 mg/dL, any use of lipid-lowering drugs,
or any self-reported history of dyslipidemia. Atrial fibrillation was
defined as history of atrial fibrillation confirmed by ≥1 ECG or presence of the arrhythmia during hospitalization. Height and weight
were measured with participants wearing a scrub suit and no shoes,
and then body mass index was calculated by dividing weight in kilograms by the square of height in meters.
0.32
5 (2–10)
4 (1–9)
<0.001
2707 (15.7)
774 (13.2)
<0.001
BMI indicates body mass index; IQR, interquartile range; and NIHSS, National
Institutes of Health Stroke Scale.
eGFR values were categorized into 5 categories: <45, 45 to 59, 60 to
89, 90 to 119, and ≥120 mL/min per 1.73 m2 according to the new
classification for chronic kidney disease assessment and management; patients with eGFR of <45 mL/min per 1.73 m2 were pooled
together because of having relatively poor outcomes and patients with
high eGFR of 120 mL/min per 1.73 m2 were separated from patients
with eGFR of ≥90 mL/min per 1.73 m2 because of muscle loss related
to malnutrition or kidney hyperfiltration in diabetics.15 eGFR of 90 to
119 mL/min per 1.73 m2 was served as the reference group in analysis. We compared baseline and clinical characteristics of patients
according to eGFR levels in both diabetes mellitus and no diabetes
mellitus groups. Continuous variables are expressed as mean with
SD or medians with interquartile ranges, as appropriate. Categorical
data are presented as proportions. Medians were imputed for missing
continuous variables (body mass index and NIHSS) and modes were
used to replace categorical variables (smoking status). Distributions
of eGFR in patients with or without diabetes mellitus were estimated
using kernel method with the Sheather–Jones plug-in approach for
selecting the bandwidth (bandwidth multiplier=1).
We evaluated the association between eGFR and clinical outcomes
among all patients stratified by diabetes mellitus using a logistic regression model with study sites as random effects adjusting for covariates of age, sex, history of stroke, hypertension, dyslipidemia,
atrial fibrillation, coronary heart disease, smoking, alcohol, body
mass index at admission, baseline NIHSS, pneumonia, and urethral
infection. To examine effect modification by diabetes mellitus, we
Luo et al eGFR and Stroke Outcomes 2889
tested the statistical significance of eGFR category×diabetes mellitus
in multivariable-adjusted logistic model by a postestimation Wald test
to get an omnibus P value for interaction between eGFR categories
and diabetes mellitus. We also assessed the association of eGFR with
stroke outcomes among patients with ischemic stroke or hemorrhagic
stroke stratified by diabetes mellitus in multivariable analysis.
Finally, we further evaluated the pattern and magnitude of associations between eGFR and risks of adverse stroke outcomes using a logistic regression model with restricted cubic splines for eGFR among
diabetic or nondiabetic patients adjusting for covariates. eGFR of 90
mL/min per 1.73 m2 in patients with or without diabetes mellitus was
treated as the individual reference and the 6 knots for spline were
placed at 30, 45, 60, 90, 120, and 150 mL/min per 1.73 m2.
All analyses were conducted with SAS Version 9.2 software (SAS
Institute). Two-tailed P values <0.05 were considered to be statistically significant.
Results
Study Participants and Characteristics
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
A total of 17 280 (nondiabetic 12 498 and diabetic 4782)
patients were included in the final analysis. Overall, regardless of diabetes mellitus status at baseline, participants with
lower eGFR were more likely to be women, nonsmoker,
nondrinker, and to have lower body mass index at baseline,
higher NIHSS on admission, higher prevalence of hypertension, atrial fibrillation, coronary heart disease, history of
stroke, and infection, whereas participants with higher eGFR
had reverse characteristics of those with lower eGFR, when
compared with normal range of eGFR 90 to 119 mL/min per
1.73 m2 (Table 2). The distribution of eGFR values based on
the Kernel density estimation in both diabetic and nondiabetic
participants is observed in Figure 1.
Association of eGFR With Outcomes
Table 3 shows the 1-year incidence of stroke outcomes including all-cause death, stroke recurrence, composite of death
or stroke recurrence, and stroke disability, across eGFR categories. Risks of all end points increase with reduced eGFR
and, moreover, these rates in diabetic or nondiabetic patients
are similar within groups of eGFR <60 mL/min per 1.73 m2.
However, nondiabetic participants with higher eGFR tend to
have slightly lower proportions of end points, whereas diabetic
patients with higher eGFR have relatively higher incidence, as
compared with those with normal range of eGFR. Incidences
in participants stratified by different subtypes of stroke are
reported in Table I in the online-only Data Supplement.
Odds ratios (ORs) of stroke outcomes based on categories
of eGFR and stratified by diabetes mellitus status are reported
in Table 4. Among nondiabetic patients, as compared with
eGFR of 90 to 119 mL/min per 1.73 m2, the adjusted ORs of
eGFR of <45 mL/min per 1.73 m2 were 2.79 (95% confidence
interval, 2.09–3.73) for all-cause mortality, 2.28 (1.74–2.98)
Table 2. Baseline Characteristics of Nondiabetic and Diabetic Patients According to eGFR Categories
No Diabetes
Mellitus (n=12 498)
Diabetes
Mellitus (n=4872)
eGFR Categories, mL/min per 1.73 m2
45–59
60–89
90–119
(n=849) (n=4281) (n=5852)
≥120
(n=1096)
Age (mean±SD), y
62.9±13.3 63.0±13.3 72.0±14.3 73.4±10.4 67.8±11.4
60.4±11.4
65.0±11.3 71.0±11.4 71.8±10.1 68.6±9.7
62.6±10.2
50.4±10.0
Women, n (%)
4634 (37.1) 200 (47.6) 392 (46.2) 1613 (37.7) 2040 (34.9)
389 (35.5)
2030 (42.5) 149 (55.6) 199 (52.2) 706 (44.4) 857 (38.8)
119 (35.7)
History of
stroke, n (%)
3299 (26.4) 142 (33.8) 304 (35.8) 1308 (30.6) 1371 (23.4)
174 (15.9)
1491 (31.2) 106 (39.6) 132 (34.6) 550 (34.6) 635 (28.7)
68 (20.4)
Ischemic stroke
8926 (71.4) 294 (70.0) 646 (76.1) 3321 (77.6) 4067 (69.5)
598 (54.6)
3958 (82.8) 230 (85.8) 314 (82.4) 1383 (86.9) 1792 (81.1)
239 (71.8)
ICH and SAH
3572 (28.6) 126 (30.0) 203 (23.9) 960 (22.4) 1785 (30.5)
498 (45.4)
Hypertension, n (%)
7478 (59.8) 309 (73.6) 583 (68.7) 2779 (64.9) 3349 (57.2)
458 (41.8)
Characteristics
Overall
<45
(n=420)
Overall
<45
(n=268)
45–59
60–89
90–119
(n=381) (n=1591) (n=2209)
≥120
(n=333)
Subtypes of stroke
824 (17.2)
38 (14.2)
67 (17.6) 208 (13.1) 417 (18.9)
94 (28.2)
3412 (71.4) 218 (81.3) 301 (79.0) 1186 (74.5) 1516 (68.6)
191 (57.4)
Dyslipidemia, n (%)
961 (7.7)
29 (6.9)
66 (7.8)
370 (8.6)
431 (7.4)
65 (5.9)
698 (14.6)
50 (18.7)
49 (12.9) 232 (14.6) 333 (15.1)
Atrial fibrillation,
n (%)
700 (5.6)
49 (11.7)
98 (11.5) 304 (7.1)
219 (3.7)
30 (2.7)
233 (4.9)
23 (8.6)
30 (7.9)
82 (19.5) 152 (17.9) 597 (13.9) 474 (8.1)
30 (2.7)
748 (15.6)
59 (22.0)
91 (23.9) 298 (18.7) 288 (13.0)
12 (4.5)
14 (3.7)
CHD, n (%)
1335 (10.7)
CHF, n (%)
167 (1.3)
4 (0.4)
82 (1.7)
Ever smoking,* n (%) 5028 (40.4) 141 (33.6) 263 (31.1) 1646 (38.6) 2488 (42.7)
490 (44.8)
1726 (36.1)
Heavy alcohol, n (%) 1388 (11.1)
191 (17.4)
440 (9.2)
BMI,* median
(IQR), kg/m2
NIHSS,* median (IQR)
Infection,† n (%)
15 (3.6)
15 (3.6)
26 (3.1)
42 (4.9)
63 (1.5)
381 (8.9)
59 (1.0)
759 (13.0)
83 (5.2)
27 (1.7)
91 (4.1)
27 (1.2)
76 (28.4) 101 (26.6) 518 (32.6) 884 (40.1)
12 (4.5)
14 (3.7)
108 (6.8)
239 (10.8)
34 (10.2)
6 (1.8)
12 (3.6)
2 (0.6)
147 (44.3)
67 (20.1)
24.0
23.4
23.9
24.0
24.1
23.7
24.7
24.2
24.2
24.7
24.7
25.0
(21.9–26.2) (20.8–25.8) (21.5–26.1) (21.8–26.1) (22.0–26.0) (21.7–26.1) (22.6–27.0) (21.5–26.8) (22.5–26.5) (22.6–27.1) (22.7–26.7) (22.8–27.2)
4.0 (2–10)
4.0 (1–11)
5 (2–10)
6 (2–13)
1873 (15.0) 143 (34.0) 220 (25.9) 647 (15.1) 743 (12.7)
5 (2–10) 6.5 (2–15) 6.0 (2–13) 4.0 (2–9)
120 (10.9)
834 (17.4)
88 (32.8)
5 (2–9)
5 (2–10)
89 (23.4) 291 (18.3) 322 (14.6)
6 (2–11)
5 (2–9)
44 (13.2)
eGFR, median (IQR),
93.4
34.9
54.3
77.3
103.3
126.2
92.2
34.4
53.4
77.2
103.5
125.6
mL/min per 1.73 m2 (75.5–107.4) (24.6–40.6) (50.1–57.1) (69.8–83.8) (96.8–110.3) (122.6–132.6) (72.7–106.5) (23.6–40.4) (50.0–56.3) (69.8–83.6) (96.6–110.4) (122.5–130.8)
BMI indicates body mass index; CHD, coronary heart disease; CHF, congestive heart failure; eGFR, estimated glomerular filtration rate; ICH, intracranial hemorrhage;
IQR, interquartile range; NIHSS, National Institutes of Health Stroke Scale; and SAH, subarachnoid hemorrhage.
*Smoking is missing in 45 patients, BMI in 1876 patients, and NIHSS in 97 patients.
†Infection included pneumonia and urinary tract infection.
2890 Stroke October 2014
Figure 1. Distributions of estimated glomerular filtration rate (eGFR) calculated using Chinese modification of the Chronic Kidney Disease Epidemiology
Collaboration equation based on kernel density
estimation. DM indicates diabetes mellitus.
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
for stroke recurrence, and 1.53 (1.16–2.01) for stroke disability; higher eGFR of ≥120 mL/min per 1.73 m2 was also
significantly associated with higher risk of all-cause mortality (adjusted OR, 1.38; 95% confidence interval, 1.02–1.86)
but not with stroke recurrence or stroke disability. In diabetic
patients, the adjusted ORs of all-cause mortality, stroke recurrence, and stroke disability in lower eGFR of <45 mL/min
per 1.73 m2 were 2.16 (1.51–3.08), 1.43 (1.02–2.00), and 1.38
(0.98–1.95); in higher eGFR of ≥120 mL/min per 1.73 m2,
ORs of all stroke outcomes were increased significantly compared with eGFR of 90 to 119 mL/min per 1.73 m2: 2.55 (1.75–
3.74), 1.57 (1.13–2.18) and 1.89 (1.41–2.53), respectively. We
also identified significantly statistical interactions between
diabetes mellitus and categories of eGFR for all stroke end
points (P=0.03, 0.01, 0.001, and 0.03 for death, recurrence,
composite end point, and stroke disability, respectively). ORs
for stroke outcomes in patients stratified by different subtypes
of stroke are presented in Table II (ischemic stroke) and Table
III (hemorrhagic stroke) in the online-only Data Supplement.
Other predictors for each of 4 stroke outcomes among entire
population were also reported with OR in Table IV in the
online-only Data Supplement.
Finally, we used the logistic regression model with restricted
cubic spline to evaluate the pattern and magnitude of association between eGFR and stroke outcomes adjusted for covariates. We found a reverse J-shaped association between eGFR
and all-cause death in nondiabetic and diabetic patients, but the
risk gradient in increased level of eGFR is steeper in patients
with diabetes mellitus than those without diabetes mellitus
(Figure 2A and 2B). J shape was not observed for association
between eGFR and stroke recurrence in nondiabetics, and yet
still remained in diabetics (Figure 2C and 2D). With regard to
combined end point or stroke disability, the results are similar
to those of all-cause death or stroke recurrence (Figure II in
the online-only Data Supplement).
Discussion
In this national cohort of patients with acute stroke, decreased
eGFR was associated with all-cause mortality, stroke recurrence, and stroke disability in both diabetic and nondiabetic
Table 3. One-Year Incidence of Clinical Outcomes in Nondiabetic and Diabetic Patients According to eGFR Categories
No Diabetes Mellitus (n=12 498)
Diabetes Mellitus (n=4872)
eGFR Categories, mL/min per 1.73 m
2
<45
45–59
60–89
90–119
≥120
<45
45–59
60–89
90–119
≥120
Characteristics Overall (n=420) (n=849) (n=4281) (n=5852) (n=1096) P Value Overall (n=268) (n=381) (n=1591) (n=2209) (n=333) P Value
All-cause
mortality
1425
(11.4)
153
(36.4)
193
(22.7)
480
(11.2)
515
(8.8)
84
(7.7)
<0.001
704
(14.7)
88
(32.8)
82
(21.5)
229
(14.4)
255
(11.5)
50
(15.0)
<0.001
Stroke
recurrence
1952
(17.2)
149
(40.3)
212
(28.0)
697
(17.9)
775
(14.5)
119
(12.0)
<0.001
1003
(23.1)
88
(36.7)
93
(27.0)
344
(23.9)
413
(20.5)
65
(21.9)
<0.001
Combined
end point
2307
(18.5)
178
(42.4)
259
(30.5)
808
(18.9)
915
(15.6)
147
(13.4)
<0.001
1243
(26.0)
110
(41.0)
114
(29.9)
416
(26.1)
515
(23.3)
88
(26.4)
<0.001
Stroke
disability
5152
(41.5)
268
(64.0)
492
(58.4)
1811
(42.5)
22 365
(38.5)
345
(31.7)
<0.001
2402
(50.7)
175
(66.0)
236
(62.8)
827
(52.4)
1009
(46.1)
155
(46.8)
<0.001
Values represent no. of participants (%). Combined end point including all-cause mortality and stroke recurrence. eGFR indicates estimated glomerular filtration rate.
Luo et al eGFR and Stroke Outcomes 2891
Table 4. Odds Ratios for the Association Between eGFR Levels and Clinical Outcomes in Nondiabetic and Diabetic Patients
With Stroke
eGFR Categories, mL/min per 1.73 m2
<45
45–59
60–89
90–119
≥120
All-cause mortality
P Value for
Interaction
0.03
No diabetes mellitus
Unadjusted
5.94 (4.77–7.39)
3.05 (2.54–3.67)
1.31 (1.15–1.49)
1.00 (reference)
0.86 (0.68–1.09)
...
Adjusted
2.79 (2.09–3.73)
1.44 (1.14–1.83)
0.93 (0.79–1.09)
1.00 (reference)
1.38 (1.02–1.86)
...
Diabetes mellitus
Unadjusted
3.75 (2.81–4.99)
2.10 (1.59–2.77)
1.29 (1.06–1.56)
1.00 (reference)
1.35 (0.98–1.88)
...
Adjusted
2.16 (1.51–3.08)
1.18 (0.85–1.65)
0.92 (0.74–1.16)
1.00 (reference)
2.55 (1.75–3.74)
...
Stroke recurrence
0.01
No diabetes mellitus
Unadjusted
3.98 (3.19–4.97)
2.30 (1.93–2.74)
1.29 (1.15–1.44)
1.00 (reference)
0.81 (0.66–0.99)
...
Adjusted
2.28 (1.74–2.98)
1.39 (1.12–1.71)
1.06 (0.93–1.20)
1.00 (reference)
1.10 (0.86–1.39)
...
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Diabetes mellitus
Unadjusted
2.24 (1.69–2.98)
1.44 (1.11–1.86)
1.22 (1.03–1.43)
1.00 (reference)
1.09 (0.81–1.46)
...
Adjusted
1.43 (1.02–2.00)
1.04 (0.77–1.39)
1.00 (0.83–1.19)
1.00 (reference)
1.57 (1.13–2.18)
...
Combined end point
0.001
No diabetes mellitus
Unadjusted
3.97 (3.23–4.88)
2.37 (2.01–2.79)
1.26 (1.13–1.39)
1.00 (reference)
0.84 (0.69–1.01)
...
Adjusted
2.21 (1.72–2.84)
1.40 (1.15–1.70)
1.01 (0.90–1.14)
1.00 (reference)
1.14 (0.91–1.42)
...
Diabetes mellitus
Unadjusted
2.29 (1.76–2.98)
1.40 (1.10–1.79)
1.16 (1.00–1.35)
1.00 (reference)
1.18 (0.91–1.54)
...
Adjusted
1.44 (1.06–1.96)
0.97 (0.74–1.27)
0.91 (0.77–1.08)
1.00 (reference)
1.87 (1.40–2.52)
...
Stroke disability
0.03
No diabetes mellitus
Unadjusted
2.83 (2.30–3.48)
2.24 (1.93–2.59)
1.18 (1.09–1.28)
1.00 (reference)
0.74 (0.65–0.85)
...
Adjusted
1.53 (1.16–2.01)
1.12 (0.92–1.36)
0.89 (0.80–1.00)
1.00 (reference)
1.07 (0.89–1.29)
...
Unadjusted
2.28 (1.74–2.98)
1.97 (1.58–2.47)
1.29 (1.13–1.47)
1.00 (reference)
1.03 (0.82–1.30)
...
Adjusted
1.38 (0.98–1.95)
1.27 (0.96–1.67)
0.99 (0.84–1.16)
1.00 (reference)
1.89 (1.41–2.53)
...
Diabetes mellitus
Values represent odds ratio (95% confidential interval). Multivariable analysis adjusted for age, sex, history of stroke, hypertension, dyslipidemia, atrial fibrillation,
coronary heart disease, smoking, alcohol, body mass index at admission, baseline National Institutes of Health Stroke Scale, pneumonia, and urethral infection. eGFR
indicates estimated glomerular filtration rate.
patients, whereas increased eGFR was associated with all
adverse stroke outcomes in diabetic patients and only linked
to all-cause mortality in nondiabetic patients; the magnitudes
of association of stroke outcomes with high eGFR level were
more pronounced in diabetic group than that in nondiabetic
group. Taken together, these findings highlight the importance
of reduced eGFR as a predictor of stroke outcomes in the presence or absence of diabetes mellitus. Besides, increased eGFR
might have a potential role in predicting stroke outcomes. To
our knowledge, this is the first study in the national level to
compare the influence of kidney dysfunction on outcomes after
acute stroke in diabetic versus nondiabetic patients from China.
A growing body of evidence suggests that eGFR is an
important predictor of all-cause and cardiovascular mortality,
kidney failure, and stroke in the general population.16–18 The
pattern of the link between eGFR and all-cause mortality was
described as J-shaped in the Cardiovascular Health Study.19 In
contrast, Mostofsky20 reported U-shaped relationship between
eGFR and all-cause mortality in 1175 patients with acute ischemic stroke, they thought that both decreased and increased
eGFR were associated with a higher risk of mortality compared with the normal range of eGFR. As the results of previous studies were inconsistent, we further explored the pattern
and magnitude of this association. Our study demonstrated
that the patterns of these associations between decreased
eGFR and all outcomes are much the same in the presence
or absence of diabetes mellitus, despite those curves had different characteristics between nondiabetics and diabetics in
higher level of eGFR (Table 4; Figure 2). Our study not only
confirms the results of previous meta-analysis that patients
with or without diabetes mellitus have similar risks of mortality linked with decreased eGFR21 but also extends the association of reduced eGFR with other clinical end points like stroke
recurrence or stroke disability. Thus, our results suggest that
decreased eGFR would be able to be applied into clinical risk
stratification for the prognosis of patients with acute stroke.
2892 Stroke October 2014
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Figure 2. Adjusted odds ratios for stroke recurrence and all-cause mortality according to estimated glomerular filtration rate (eGFR) in
individuals with and without diabetes mellitus (DM); (A) all-cause mortality in patients without DM; (B) all-cause mortality in patients with
DM; (C) stroke recurrence in patients without DM; (D) stroke recurrence in patients with DM. The solid line indicates adjusted odds ratio
and the dashed lines the 95% confidence interval bands. Reference is eGFR of 90 mL/min per 1.73 m2 in patients with or without DM.
Data were fitted using a logistic model of restricted cubic spline with 6 knots (30, 45, 60, 90, 120, and 150 mL/min per 1.73 m2) for eGFR,
adjusting for confounding factors.
Our study also showed that increased eGFR was linked
with higher risk of all-cause mortality regardless of diabetes mellitus status, yet the risk gradient is steeper in patients
with diabetes mellitus than those without diabetes mellitus
(Table 4; Figure 2). In addition, it is only in diabetic patients
that association between increased eGFR and other 2 end
points including stroke recurrence and stroke disability was
found to be statistical significant, but not in nondiabetics. The
relationship between increased eGFR and all-cause mortality in our study was in line with that of a recent cohort in
general populations.22 Although we have seen the association
between increased eGFR and stroke outcomes, these results
should be interpreted cautiously with the reason that increased
eGFR may not always correspond to real high level of eGFR
in some situations, such as lower muscle mass, inflammation,
differences in body composition, or undiagnosed malignancy
accounting for the excess mortality.23,24 It is likewise worth
mentioning that, except for those confounding factors, kidney
hyperperfusion is an important pathophysiological feature in
diabetic patient, which is usually manifested as higher levels
of eGFR.25 This might be one of the important reasons of the
association of increased eGFR and stroke outcomes existing
in diabetic patients. However, we cannot exclude the possible
influence by residual confounders that we mentioned above.
Therefore, future work should try to validate our analyses in
representative samples with less confounding factors and even
include applying new biomarkers of kidney function like cystatin C in future study.
Strengths of our study include a large sample size of patients
with stroke from China. It also contains outcomes of interest
such as stroke recurrence and stroke disability in our study to
reflect the prognosis of stroke. In addition, we used a new eGFR
calculating equation to improve risk prediction.14 Nonetheless,
some limitations should be noticed while interpreting our
Luo et al eGFR and Stroke Outcomes 2893
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
results. First, albuminuria has proved to be an important independent risk factor for the adverse clinical outcomes in many
diseases and was not included in our study, thus we were unable
to evaluate eGFR and albuminuria simultaneously along with
other factors in our patients. Second, as our cohort only comprised Chinese adult patients with stroke, these results may not
be generalizable to other races and ethnicities. Third, patients
with missing baseline serum creatinine or lost follow-up within
1 year were not included in the study, so a selection bias might
exist, such as discrepancy in previous stroke and infection.
In conclusion, decreased eGFR is a strong predictor of allcause mortality, stroke recurrence, and stroke disability in
diabetic and nondiabetic patients with acute stroke. Although
increased eGFR is associated with all of stroke outcomes in
diabetic patients and linked to all-cause mortality in nondiabetic patients, the results should be explained carefully because
of the confounding like lower muscle mass and inflammation.
Our study suggests that kidney dysfunction should be paid
more attention in secondary prevention of stroke, especially in
patients with diabetes mellitus.
Appendix
A full list of the CNSR investigators is available at http://onlinelibrary.wiley.com/doi/10.1111/j.1747-4949.2011.00584.x/
suppinfo.
Acknowledgments
We thank all participating hospitals, colleagues, nurses, imaging and
laboratory technicians.
Sources of Funding
The CNSR is funded by the Ministry of Science and Technology and
the Ministry of Health of the People’s Republic of China. The Grant
Nos. are 2008ZX09312-008, 2009CB521905, 2011BAI08B02, and
D101107049310005.
Disclosures
None.
References
1. Yang G, Wang Y, Zeng Y, Gao GF, Liang X, Zhou M, et al. Rapid health
transition in China, 1990-2010: findings from the Global Burden of
Disease Study 2010. Lancet. 2013;381:1987–2015.
2. Liu M, Wu B, Wang WZ, Lee LM, Zhang SH, Kong LZ. Stroke in China:
epidemiology, prevention, and management strategies. Lancet Neurol.
2007;6:456–464.
3. Putaala J, Haapaniemi E, Gordin D, Liebkind R, Groop PH, Kaste M, et al.
Factors associated with impaired kidney function and its impact on longterm outcome in young ischemic stroke. Stroke. 2011;42:2459–2464.
4. Tsagalis G, Akrivos T, Alevizaki M, Manios E, Stamatellopoulos K,
Laggouranis A, et al. Renal dysfunction in acute stroke: an independent
predictor of long-term all combined vascular events and overall mortality. Nephrol Dial Transplant. 2009;24:194–200.
5. Oh SW, Baek SH, Kim YC, Goo HS, Heo NJ, Na KY, et al. Mild decrease
in estimated glomerular filtration rate and proteinuria are associated with
all-cause and cardiovascular mortality in the general population. Nephrol
Dial Transplant. 2012;27:2284–2290.
6. Yang W, Lu J, Weng J, Jia W, Ji L, Xiao J, et al; China National Diabetes
and Metabolic Disorders Study Group. Prevalence of diabetes among
men and women in China. N Engl J Med. 2010;362:1090–1101.
7.Xu Y, Wang L, He J, Bi Y, Li M, Wang T, et al; 2010 China
Noncommunicable Disease Surveillance Group. Prevalence and control
of diabetes in Chinese adults. JAMA. 2013;310:948–959.
8. Kissela B, Air E. Diabetes: impact on stroke risk and poststroke recovery.
Semin Neurol. 2006;26:100–107.
9. De Cosmo S, Rossi MC, Pellegrini F, Lucisano G, Bacci S, Gentile S, et
al; AMD-Annals Study Group. Kidney dysfunction and related cardiovascular risk factors among patients with type 2 diabetes. Nephrol Dial
Transplant. 2014;29:657–662.
10. Wang Y, Cui L, Ji X, Dong Q, Zeng J, Wang Y, et al; China National
Stroke Registry Investigators. The China National Stroke Registry for
patients with acute cerebrovascular events: design, rationale, and baseline patient characteristics. Int J Stroke. 2011;6:355–361.
11. Jia Q, Zheng H, Zhao X, Wang C, Liu G, Wang Y, et al; Investigators
for the Survey on Abnormal Glucose Regulation in Patients With Acute
Stroke Across China (ACROSS-China). Abnormal glucose regulation in
patients with acute stroke across China: prevalence and baseline patient
characteristics. Stroke. 2012;43:650–657.
12. Teo BW, Xu H, Wang D, Li J, Sinha AK, Shuter B, et al. GFR estimating equations in a multiethnic Asian population. Am J Kidney Dis.
2011;58:56–63.
13. Wang X, Luo Y, Wang Y, Wang C, Zhao X, Wang D, et al; China National
Stroke Registry Investigators. Comparison of associations of outcomes
after stroke with estimated GFR using Chinese modifications of the
MDRD study and CKD-EPI creatinine equations: results from the China
National Stroke Registry. Am J Kidney Dis. 2014;63:59–67.
14.Skali H, Uno H, Levey AS, Inker LA, Pfeffer MA, Solomon SD.
Prognostic assessment of estimated glomerular filtration rate by the new
Chronic Kidney Disease Epidemiology Collaboration equation in comparison with the Modification of Diet in Renal Disease Study equation.
Am Heart J. 2011;162:548–554.
15. Levey AS, de Jong PE, Coresh J, El Nahas M, Astor BC, Matsushita K, et
al. The definition, classification, and prognosis of chronic kidney disease:
a KDIGO Controversies Conference report. Kidney Int. 2011;80:17–28.
16. Matsushita K, van der Velde M, Astor BC, Woodward M, Levey AS, de
Jong PE, et al. Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality in general population
cohorts: a collaborative meta-analysis. Lancet. 2010;375:2073–2081.
17. Astor BC, Matsushita K, Gansevoort RT, van der Velde M, Woodward M,
Levey AS, et al; Chronic Kidney Disease Prognosis Consortium. Lower
estimated glomerular filtration rate and higher albuminuria are associated
with mortality and end-stage renal disease. A collaborative meta-analysis
of kidney disease population cohorts. Kidney Int. 2011;79:1331–1340.
18. van der Velde M, Matsushita K, Coresh J, Astor BC, Woodward M,
Levey A, et al; Chronic Kidney Disease Prognosis Consortium. Lower
estimated glomerular filtration rate and higher albuminuria are associated
with all-cause and cardiovascular mortality. A collaborative meta-analysis of high-risk population cohorts. Kidney Int. 2011;79:1341–1352.
19. Shlipak MG, Sarnak MJ, Katz R, Fried LF, Seliger SL, Newman AB,
et al. Cystatin C and the risk of death and cardiovascular events among
elderly persons. N Engl J Med. 2005;352:2049–2060.
20.Mostofsky E, Wellenius GA, Noheria A, Levitan EB, Burger MR,
Schlaug G, et al. Renal function predicts survival in patients with acute
ischemic stroke. Cerebrovasc Dis. 2009;28:88–94.
21. Fox CS, Matsushita K, Woodward M, Bilo HJ, Chalmers J, Heerspink HJ, et
al; Chronic Kidney Disease Prognosis Consortium. Associations of kidney
disease measures with mortality and end-stage renal disease in individuals
with and without diabetes: a meta-analysis. Lancet. 2012;380:1662–1673.
22. Tonelli M, Klarenbach SW, Lloyd AM, James MT, Bello AK, Manns
BJ, et al. Higher estimated glomerular filtration rates may be associated
with increased risk of adverse outcomes, especially with concomitant
proteinuria. Kidney Int. 2011;80:1306–1314.
23. Barraclough K, Harris M, Montessori V, Levin A. An unusual case of
acute kidney injury due to vancomycin lessons learnt from reliance on
eGFR. Nephrol Dial Transplant. 2007;22:2391–2394.
24. Inker LA, Levey AS. Pro: estimating GFR using the chronic kidney disease epidemiology collaboration (CKD-EPI) 2009 creatinine equation:
the time for change is now. Nephrol Dial Transplant. 2013;28:1390–1396.
25. Premaratne E, Macisaac RJ, Tsalamandris C, Panagiotopoulos S, Smith
T, Jerums G. Renal hyperfiltration in type 2 diabetes: effect of age-related
decline in glomerular filtration rate. Diabetologia. 2005;48:2486–2493.
Associations Between Estimated Glomerular Filtration Rate and Stroke Outcomes in
Diabetic Versus Nondiabetic Patients
Yang Luo, Xianwei Wang, Kunihiro Matsushita, Chunxue Wang, Xingquan Zhao, Bo Hu,
Liping Liu, Hao Li, Gaifen Liu, Qian Jia, Yilong Wang and Yongjun Wang
on behalf of the CNSR Investigators
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Stroke. 2014;45:2887-2893; originally published online August 12, 2014;
doi: 10.1161/STROKEAHA.114.005380
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2014 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/45/10/2887
Data Supplement (unedited) at:
http://stroke.ahajournals.org/content/suppl/2014/08/12/STROKEAHA.114.005380.DC1
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click
Request Permissions in the middle column of the Web page under Services. Further information about this
process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Stroke is online at:
http://stroke.ahajournals.org//subscriptions/
SUPPLEMENTAL MATERIAL
Table I. One-year incidence of clinical outcomes in diabetic and non-diabetic patients
according to eGFR categories
Table II. Odds ratios for the association between eGFR levels and clinical outcomes
in diabetic and non-diabetic patients with Ischemic stroke
Table III. Odds ratios for the association between eGFR levels and clinical outcomes
in diabetic and non-diabetic patients with ICH or SAH
Table IV. Independent predictors for 1-year all-cause mortality, stroke recurrence,
combined endpoint and stroke disability among the entire population
Figure I. Flow chart showing the patient selection;
Figure II. Adjusted odds ratios for the combined endpoint and stroke disability
according to eGFR in individuals with and without diabetes
Table I. One-year incidence of clinical outcomes in diabetic and non-diabetic patients according to eGFR categories
No Diabetes (12498)
Diabetes (4872)
eGFR Categories, ml/min/1.73m2
Characteristics
<45
45-59
60-89
90-119
≥120
<45
45-59
60-89
90-119
≥120
All-cause mortality
109(37.1)
129(20.0)
326(9.8)
269(6.6)
28(4.7)
75(32.6%)
62(19.7%)
175(12.7%)
186(10.4%)
32(13.4%)
Stroke recurrence
109(41.8)
143(25.0)
535(17.6)
498(13.4)
55(10.3)
77(37.0%)
72(25.5%)
287(23.0%)
335(20.5%)
49(22.8%)
Combined endpoint
129(43.9)
177(27.4)
615(18.5)
585(14.4)
71(11.9)
95(41.3)
90(28.7)
348(25.2)
415(23.2)
64(26.8)
Stroke disability
184(62.8)
354(55.1)
1333(40.3)
1416(35.1)
174(29.4)
148(65.2)
191(61.6)
699(51.0)
789(44.4)
108(45.4)
All-cause mortality
44/(34.9)
64(31.5)
154(16.0)
246(13.8)
56(11.2)
13(34.2%)
20(29.9%)
54(26.0%)
69(16.5%)
18(19.1%)
Stroke recurrence
40(36.7)
69(37.5)
162(18.9)
277(16.9)
64(14.0)
11(34.4)
21(33.9)
57(29.8)
78(20.6)
16(19.5)
Patients with Ischemic Stroke
Patients with ICH or SAH
Combined endpoint
49(38.9)
82(40.4)
193(20.1)
330(18.5)
76(15.3)
15(39.5)
24(35.8)
68(32.7)
100(24.0)
24(25.5)
Stroke disability
84(66.7)
138(68.7)
478(50.3)
820(46.3)
171(34.5)
27(71.1)
45(68.2)
128(61.5)
220(53.1)
47(50.5)
eGFR indicates estimated Glomerular Filtration Rate; ICH, intracranial Hemorrhage; SAH, Subarachnoid Hemorrhage; Combined endpoint including stroke or
mortality.
Table II. Odds ratios for the association between eGFR levels and clinical outcomes in diabetic and non-diabetic patients with Ischemic stroke
OR (95% CI)
P value for
eGFR Categories, ml/min/1.73m2
interaction
<45
45-59
60-89
90-119
≥120
0.03
All-cause mortality
No Diabetes
Unadjusted
8.32(6.37-10.86)
3.52(2.80-4.43)
1.54(1.30-1.82)
1.00 (reference)
0.69(0.47-1.03)
Adjusted
2.89(2.00-4.19)
1.39(1.03-1.87)
0.98(0.80-1.21)
1.00 (reference)
1.59(0.99-2.54)
Unadjusted
4.18(3.05‐5.72) 2.12(1.55‐2.92) 1.25(1.00‐1.56) 1.00 (reference)
1.33(0.89-2.00)
Adjusted
2.39(1.61‐3.54)
1.24(0.85‐1.80)
0.87(0.67‐1.12)
1.00 (reference)
2.57(1.63-4.07)
Diabetes
Stroke recurrence
0.01
No Diabetes
Unadjusted
4.64(3.56-6.03)
2.15(1.74-2.66)
1.38(1.21-1.58)
1.00 (reference)
0.75(0.56-1.00)
Adjusted
2.25(1.62-3.12)
1.19(0.93-1.54)
1.10(0.94-1.28)
1.00 (reference)
1.14(0.82-1.59)
Unadjusted
2.28(1.68‐3.10) 1.33(0.99‐1.78) 1.16(0.97‐1.38) 1.00 (reference)
1.14(0.81‐1.61) Adjusted
1.50(1.05‐2.15)
0.98(0.71‐1.36)
0.94(0.77‐1.15)
1.00 (reference)
1.62(1.12‐2.35)
Diabetes
0.002
Combined endpoint
No Diabetes
Unadjusted
4.65(3.64‐5.95) 2.25(1.85‐2.73) 1.35(1.20-1.53)
1.00 (reference)
0.80(0.62-1.04)
Adjusted
2.24(1.64‐3.05)
1.24(0.98‐1.57)
1.05(0.91-1.21)
1.00 (reference)
1.20(0.89-1.62)
2.33(1.76‐3.10) 1.33(1.02‐1.74) 1.12(0.95‐1.31) 1.00 (reference)
1.21(0.89‐1.65) Diabetes
Unadjusted
Adjusted
1.47(1.05‐2.05)
0.94(0.70‐1.27)
0.87(0.72‐1.04)
1.00 (reference)
1.90(1.35‐2.66)
0.37
Stroke disability
No Diabetes
Unadjusted
3.12(2.44-3.99)
2.27(1.92-2.69)
1.25(1.13-1.37)
1.00 (reference)
0.77(0.64-0.93)
Adjusted
1.39(0.99-1.95)
1.07(0.86-1.35)
0.93(0.82-1.05)
1.00 (reference)
1.31(1.03-1.67)
Unadjusted
2.35(1.76-3.13)
2.01(1.57-2.57)
1.30(1.13-1.50)
1.00 (reference)
1.04(0.79-1.37)
Adjusted
1.32(0.91-1.92)
1.25(0.92-1.68)
0.97(0.81-1.15)
1.00 (reference)
1.94(1.39-2.71)
Diabetes
eGFR indicates estimated Glomerular Filtration Rate; OR, Odds Ratio; CI, Confidential Interval; *Adjusted for age, sex, history of stroke, hypertension, dyslipidemia, atrial fibrillation, coronary heart disease, smoking, alcohol, body mass index at admission, baseline national institutes of health stroke scale, pneumonia and urethral infection. Table III. Odds ratios for the association between eGFR levels and clinical outcomes in diabetic and non-diabetic patients with ICH or SAH
OR (95% CI)
P value for
eGFR Categories, ml/min/1.73m2
interaction
<45
45-59
60-89
90-119
≥120
0.12
All-cause mortality
No Diabetes
Unadjusted
3.36(2.27‐4.96) 2.88(2.08-3.99)
1.20(0.96-1.49)
1.00 (reference)
0.79(0.58-1.08)
Adjusted
2.35(1.46‐3.79)
1.67(1.12-2.49)
0.90(0.69-1.16)
1.00 (reference)
1.09(0.74-1.60)
Unadjusted
2.62(1.28‐5.38) 2.15(1.20‐3.85) 1.77(1.18‐2.65) 1.00 (reference)
1.19(0.67-2.12)
Adjusted
1.36(0.53‐3.45)
1.13(0.53‐2.42)
1.36(0.85‐2.16)
1.00 (reference)
2.48(1.25-4.91)
Diabetes
Stroke recurrence
0.13
No Diabetes
Unadjusted
2.85(1.89-4.29)
2.95(2.13-4.08)
1.15(0.92-1.42)
1.00 (reference)
0.80(0.59-1.07)
Adjusted
2.29(1.42-3.70)
2.10(1.42-3.10)
0.93(0.72-1.19)
1.00 (reference)
1.00(0.70-1.42)
Unadjusted
2.02(0.94‐4.37) 1.98(1.10‐3.54) 1.64(1.10‐2.44) 1.00 (reference)
0.94(0.51‐1.71) Adjusted
1.11(0.41‐3.04)
1.36(0.66‐2.81)
1.40(0.89‐2.22)
1.00 (reference)
1.64(0.80‐3.37)
Diabetes
0.13
Combined endpoint
Unadjusted
2.81(1.92‐4.09) 2.99(2.20‐4.05) 1.11(0.91‐1.35) 1.00 (reference)
0.79(0.61‐1.04) Adjusted
2.09(1.34‐3.24) 2.03(1.41‐2.91) 0.91(0.72‐1.14) 1.00 (reference)
1.04(0.75‐1.44) Unadjusted
2.07(1.04‐4.12) 1.77(1.02‐3.06) 1.54(1.07‐2.22) 1.00 (reference)
1.09(0.65‐1.82) Adjusted
1.30(0.56‐3.05)
1.15(0.58‐2.30)
1.27(0.83‐1.93)
1.00 (reference)
2.03(1.10‐3.76)
No Diabetes
Diabetes
0.29
Stroke disability
No Diabetes
Unadjusted
2.32(1.58-3.40)
2.54(1.86-3.47)
1.17(1.00-1.37)
1.00 (reference)
0.61(0.50-0.75)
Adjusted
1.86(1.14-3.01)
1.37(0.91-2.05)
0.83(0.68-1.01)
1.00 (reference)
0.80(0.61-1.06)
Unadjusted
2.16(1.05-4.48)
1.89(1.09-3.28)
1.41(1.00-1.98)
1.00 (reference)
0.90(0.57-1.41)
Adjusted
1.66(0.66-4.17)
1.42(0.67-3.01)
1.06(0.69-1.62)
1.00 (reference)
1.79(0.98-3.25)
Diabetes
eGFR indicates estimated Glomerular Filtration Rate; OR, Odds Ratio; CI, Confidential Interval; ICH, intracranial Hemorrhage; SAH, Subarachnoid Hemorrhage; *Adjusted for age, sex, history of stroke, hypertension, dyslipidemia, atrial fibrillation, coronary heart disease, smoking, alcohol, body mass index at admission, baseline national institutes of health stroke scale, pneumonia and urethral infection. Table IV. Independent predictors for 1-year all-cause mortality, stroke recurrence, combined
endpoint and stroke disability among the entire population
Outcomes / predictors
Odds Ratio (95% CI)
All-cause mortality
Age
1.04 (1.04-1.05)
Diabetes Mellitus
1.35 (1.20-1.52)
Dyslipidemia
0.69 (0.56-0.85)
Atrial fibrillation
1.53 (1.26-1.85)
Baseline NIHSS
1.09 (1.08-1.09)
Infection
2.09 (1.84-2.36)
eGFR < 45 ml/min/1.73m2
2.55 (2.04-3.19)
eGFR 45-59 ml/min/1.73m2
1.36 (1.13-1.65)
eGFR 60-89 ml/min/1.73m2
0.93 (0.82-1.06)
eGFR ≥120 ml/min/1.73m2
1.74 (1.37-2.19)
Stroke recurrence
Age
1.02 (1.02-1.03)
Diabetes Mellitus
1.44 (1.31-1.59)
Coronary heart disease
1.26 (1.11-1.44)
Atrial fibrillation
1.73 (1.45-2.06)
Baseline NIHSS
1.05 (1.05-1.06)
Infection
1.70 (1.51-1.90)
eGFR < 45 ml/min/1.73m2
1.92 (1.56-2.36)
eGFR 45-59 ml/min/1.73m2
1.27 (1.07-1.50)
eGFR 60-89 ml/min/1.73m2
1.04 (0.94-1.16)
eGFR ≥120 ml/min/1.73m2
1.22 (1.01-1.48)
Combined endpoint
Age
1.02 (1.02-1.03)
Diabetes Mellitus
1.54 (1.40-1.68)
Coronary heart disease
1.23 (1.09-1.39)
Atrial fibrillation
1.74 (1.47-2.04)
Baseline NIHSS
1.05 (1.05-1.06)
Infection
1.69 (1.52-1.88)
eGFR < 45 ml/min/1.73m2
1.89 (1.55-2.29)
eGFR 45-59 ml/min/1.73m2
1.25 (1.06-1.46)
eGFR 60-89 ml/min/1.73m2
0.98 (0.89-1.08)
eGFR ≥120 ml/min/1.73m2
1.34 (1.12-1.59)
Stroke disability
Age
1.04 (1.03-1.04)
Female
1.19 (1.10-1.28)
Diabetes mellitus
1.48 (1.36-1.61)
Congestive heart failure
1.41 (1.02-1.96)
Baseline NIHSS
1.18 (1.17-1.19)
Infection
2.20 (1.95-2.47)
eGFR < 45 ml/min/1.73m2
1.48 (1.20-1.84)
eGFR 45-59 ml/min/1.73m2
1.18 (1.01-1.38)
eGFR 60-89 ml/min/1.73m2
0.93 (0.85-1.01)
eGFR ≥120 ml/min/1.73m2
1.25 (1.07-1.46)
CI denotes Confidence Interval.
Figure I. Flow chart showing the patient selection; CNSR, China National Stroke Registry;
ACROSS, abnormal glucose regulation in patients with acute stroke across China; DM, Diabetes
Mellitus;
Figure II. Adjusted odds ratios for the combined endpoint and stroke disability according to eGFR in
individuals with and without diabetes
A, combined endpoint in patients without DM; B, combined endpoint in patients with DM; C, stroke
disability in patients without DM; D, stroke disability in patients with DM. The solid line indicates
adjusted odds ratio and the dashed lines the 95% confidence interval bands. eGFR of 90
ml/min/1.73m2 was served as reference value in patients with or without DM. Data were fitted using a
logistic model of restricted cubic spline with six knots (30, 45, 60, 90, 120, 150 ml/min/1.73m2)
adjusting for confounding factors. DM, Diabetes Mellitus; eGFR, estimated Glomerular Filtration Rate;