Clinical characteristics
Figure 1 shows the patient enrollment process for the study. We included 301 males (51.3%) and 285 females (48.7%) with IS in this analysis. There were 190 cases in the AF group, of them 80% had known AF before admission, including 87 males and 103 females. There were 396 cases in the non-AF group, 214 males and 182 females. Gender distribution did not reach statistical significance in two groups. AF group was older (median, 80 vs 68 yo; P < 0.0001) and more likely to have anterior circulation stroke. NT-proBNP, international standardization ratio (INR), and LAD were significantly higher in the AF group. IS patients with AF were more likely to have a history of coronary artery disease (CAD), congestive heart failure (CHF), or chronic kidney disease (CKD) but less likely to have hyperlipidemia and diabetes mellitus (DM). Smoking and drinking were less prevalent in the AF group. Compared with non-AF group, AF group is more likely to be on anticoagulant (Warfarin, Dabigatran, or Rivaroxaban) before hospitalization. The AF group had a greater in-hospital mortality (24.2% vs 4.8%, P < 0.0001), and higher hospitalization costs.
Severity of IS
The severity of IS was divided into five levels, according to the NIHSS. In the AF group, more patients were in the levels of 11-15, 16-22, and ≥ 23 and less in the level of 0-6. Overall, compared with non-AF group, AF group had greater stoke severity at admission (NIHSS median, 11 vs 3; P < 0.0001) and discharge (median, 7 vs 2; P < 0.0001, Table 1). A linear relationship was present between AF and stoke severity (Mantel-Haenszel c2 test, c2=100.400, P < 0.0001; Pearson correlation, R = 0.414, P < 0.001). Multivariate logistic regression revealed that the following factors were associated with severe IS (NIHSS scores, ≥ 16; Table 2): advanced age (odds ratio [OR], 1.033 per year), history of stroke (OR, 1.679), and AF (OR, 4.926).
Table 1 Clinical Characteristics of IS Patients
Characteristics
|
AF group(n=190)
|
non-AF group(n=396)
|
P Value
|
Age, year, median (IQR)
|
80(74,85)
|
68(60,78)
|
<0.0001
|
Gender, female number, (%)
|
103(54.2)
|
182(46)
|
0.061
|
Oneset time (hour)
|
20.07±30.83
|
51.55±83.31
|
<0.0001
|
Smoking (%)
|
15(7.9)
|
86(21.7)
|
<0.0001
|
Alcohol consumption (%)
|
6(3.2)
|
43(10.9)
|
0.002
|
Stroke localizations
|
|
|
0.008
|
Anterior circulation stroke (%)
|
156(82.1)
|
285(72.0)
|
|
Posterior circulation stroke (%)
|
34(17.9)
|
111(28.0)
|
|
Admission NIHSS score
|
11(3,19)
|
3(1,7)
|
<0.0001
|
|
0-6, n (%)
|
70(36.8)
|
292(73.7)
|
<0.0001
|
|
7-10, n (%)
|
23(12.1)
|
47(11.9)
|
0.934
|
|
11-15, n (%)
|
29(15.3)
|
28(7.1)
|
0.002
|
|
16-22, n (%)
|
42(22.1)
|
21(5.3)
|
<0.0001
|
|
≥23, n (%)
|
26(13.7)
|
8(2.0)
|
<0.0001
|
NIHSS score at discharge
|
7(2,14)
|
2(1,5)
|
<0.0001
|
Laboratory findings
|
|
|
|
LDL-C (moml/L)
|
2.57±0.9
|
2.86±0.9
|
0.001
|
Hcy (mmol/L)
|
15.8±6.9
|
14.9±8.8
|
0.216
|
NT-proBNP (ng/ml)
|
1570.0(980.2,3476.2)
|
177.2(59.6,1071.0)
|
<0.0001
|
HbA1c (%)
|
5.9(5.6,6.8)
|
6.1(5.6,7.7)
|
0.014
|
INR
|
1.1±0.3
|
1.0±0.1
|
<0.0001
|
LAD (mm)
|
41.6±6.7
|
36.7±5.2
|
<0.0001
|
LVEF (%)
|
62(59,65)
|
63(60,66)
|
0.001
|
Comorbid conditions
|
|
|
|
History of stroke (%)
|
59(31.1)
|
114(28.8)
|
0.574
|
Coronary artery disease (%)
|
35(18.4)
|
155(9.3)
|
0.002
|
Congestive heart failure (%)
|
37(9.6)
|
19(4.8)
|
<0.0001
|
Hypertension (%)
|
139(73.2)
|
298(82.1)
|
0.066
|
Diabetes mellitus (%)
|
46(24.2)
|
153(38.6)
|
0.001
|
Hyperlipidaemia (%)
|
61(32.1)
|
202(51.0)
|
<0.0001
|
Vascular disease (%)
|
1(0.5)
|
7(1.8)
|
0.405
|
Chronic kidney disease (%)
|
20(10.5)
|
12(3.0)
|
<0.0001
|
Preadmission medications (%)
|
35(18.4)
|
64(16.2)
|
0.494
|
|
Antiplatelets
|
27(14.2)
|
60(15.2)
|
0.764
|
|
Oral anticoagulant
|
8(4.2)
|
0(0)
|
<0.0001
|
In-hospital interventions
|
|
|
|
|
Oral anticoagulant
|
27(14.2)
|
3(0.8)
|
<0.0001
|
|
Thrombolysis with r-tPA
|
38(20.0)
|
40(10.1)
|
0.001
|
|
Interventional therapy
|
23(12.1)
|
18(4.5)
|
0.001
|
In-hospital death (%)
|
46(24.2)
|
19(4.8)
|
<0.0001
|
Hospital stays, days
|
11(8-17)
|
10(8-12)
|
0.001
|
Hospitalization costs, CNY
|
19182.64(11368.45,40808.35)
|
13329.54(10850.33,18629.34)
|
<0.0001
|
Values are expressed as number of subjects (%) or mean ± SD, of median (interquartile range).
IS, ischemic stroke; AF, atrial fibrillation; IQR, interquartile range; NIHSS, National Institutes of Health Stroke Scale; LDL-C, low density lipoprotein cholesterol; Hcy, homocycteine; NT-proBNP, N-terminal pro-B-type natriuretic peptide; HbA1c, Glycosylated Hemoglobin; INR, international standardization ratio; LAD, left atrial diameter; LVEF, left ventricular ejection fraction; CNY, China Yuan.
Table 2 Multivariate logistic regression analysis models for severe ischemic stroke (IS)
Variable
|
OR
|
95%CI
|
P Value
|
Age
|
1.034
|
1.009 to 1.060
|
0.008
|
Female
|
1.090
|
0.672 to 1.768
|
0.728
|
Oneset time
|
0.994
|
0.987 to 1.001
|
0.088
|
Smoking
|
0.924
|
0.419 to 2.039
|
0.846
|
Coronary artery disease
|
1.093
|
0.528 to 2.263
|
0.810
|
Congestive heart failure
|
1.526
|
0.719 to 3.235
|
0.271
|
History of stroke
|
1.757
|
1.062 to 2.906
|
0.028
|
Atrial fibrillation
|
4.266
|
2.497 to 7.286
|
<0.0001
|
OR, odds ratio; CI, confidence interval.
In-hospital mortality
Totally, 65 patients (11.1% mortality) died during hospitalization. The mortality rate was 24.2% in the AF group and 4.8% in the non-AF group (P < 0.0001, Table 1). Comparing patients with and without AF in the same age groups 60-69 (Figure 2, P < 0.0001), 70-79 (P = 0.012), and over 80 (P = 0.005), the mortality rate remained greater for the AF patients. Among 46 death cases in the AF group, 3 cases died of sudden malignant arrhythmia and the rest died of respiratory failure. Among the 19 death cases in the non-AF group, all of them died of massive cerebral infarction involving the respiratory center. Univariate analyses were performed for variables that might affect mortality, the variables with P < 0.1 were included in the multivariate logistic regression model. Advanced age, a higher NIHSS score at admission and AF were independent risk factors for in-hospital death, as shown in Table 3.
Table 3 Multivariate logistic regression analysis models for In-hospital mortality
Variable
|
OR
|
95%CI
|
P Value
|
Age
|
1.037
|
1.005 to 1.071
|
0.023
|
Female
|
0.572
|
0.313 to 1.045
|
0.069
|
Smoking
|
0.835
|
0.226 to 3.084
|
0.787
|
Coronary artery disease
|
2.127
|
0.947 to 4.774
|
0.067
|
Congestive heart failure
|
0.924
|
0.376 to 2.268
|
0.863
|
NIHSS score
|
1.104
|
1.066 to 1.143
|
<0.0001
|
Atrial fibrillation
|
2.150
|
1.088 to 4.248
|
<0.0001
|
History of stroke
|
0.904
|
0.477 to 1.713
|
0.757
|
NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio; CI, confidence interval.
Brain infarct volume after PSM
Of 586 patients, 170 patients did not complete the brain MRI and were excluded from the analysis. Among the rest, 107 patients in the AF group and 309 in non-AF group received brain MRI scan. After propensity score matching, 101 patients in each group were included in analysis. Baseline variables including age, gender, smoking, and drinking habits, comorbid conditions were similar between the two groups (Table 4, P > 0.05). The AF group had greater infarction volume [Table 5, 25.8(7.9–83.8) vs 1.4 (0.7–9.5) ml; P < 0.0001] than the non-AF group. Patients with AF were more prone to hemorrhagic transformation (37.6% vs 6.9%, P < 0.0001). Patients with AF also had greater NIHSS on admission (median, 8 vs 3; P < 0.0001) and at discharge (median, 6 vs 2; P < 0.0001). The stroke localizations (83.2% vs 79.2%, P = 0.471) were not statistically significant between the two groups. Additionally, in the PSM groups, the mortality rate was 10.9% in the AF group and 4.0% in the non-AF group (P = 0.06).
Table 4 Baseline patient characteristics after propensity score matching
|
AF group (n=101)
|
Controls (n=101)
|
P Value
|
Female
|
54 (53.5)
|
49 (48.5)
|
0.482
|
Age, years
|
78(71.5,84)
|
79(68,84)
|
0.835
|
Smoking
|
12(11.9)
|
8(7.9)
|
0.346
|
Alcohol
|
5 (5)
|
5 (5)
|
1.000
|
History of stroke
|
29 (28.7)
|
28 (27.7)
|
0.876
|
Coronary artery disease
|
10 (9.9)
|
12 (11.9)
|
0.651
|
Congestive heart failure
|
8 (7.9)
|
7(6.9)
|
0.788
|
Hypertension
|
76(75.2)
|
79 (78.2)
|
0.617
|
Diabetes mellitus
|
26 (25.7)
|
24 (23.8)
|
0.744
|
Chronic kidney disease
|
5(5.0)
|
5(5.9)
|
0.757
|
Values are expressed as number of subjects (%) or median (interquartile range).
AF, atrial fibrillation.
Table 5 Stroke severity and acute outcome for the propensity-matched groups
|
AF group (n=101)
|
Controls (n=101)
|
P Value
|
Stroke localizations: ACS
|
84(83.2)
|
80(79.2)
|
0.471
|
Haemorrhagic stroke
|
38(37.6)
|
7(6.9)
|
<0.0001
|
DWI volume, ml
|
25.8(7.9,83.8)
|
1.4(0.7,9.5)
|
<0.0001
|
Admission NIHSS
|
8(3,17)
|
3(1,5)
|
<0.0001
|
NIHSS at discharge
|
6(2,13)
|
2(1,4)
|
<0.0001
|
In-hospital death
|
11(10.9)
|
4(4.0)
|
0.06
|
Values are expressed as number of subjects (%) or median (interquartile range).
AF, atrial fibrillation; ACS, anterior circulation stroke; DWI, diffusion-weighted imaging; NIHSS, National Institutes of Health Stroke Scale.
The frequency of OACs use
In this study, only 7 of the 178 of IS patients with CHA2DS2-VASc ≥ 2 in AF group received anticoagulant therapy before the stroke events. Specifically, 5 were on warfarin, 1 was on dabigatran, and 1 was on rivaroxaban. Moreover, 26 patients took antiplatelet agents (21 were on aspirin, and 5 were on clopidogrel). The anticoagulation rate was only 3.9%, the INR was less than 2.0 in all 5 patients taking warfarin.