Seasonal Variation in Neurological Severity and Clinical Outcomes in Ischemic Stroke Patients　― A 9-Year Study of 5,238 Patients ―

Abstract

Background: Because the effects of extreme weather conditions on stroke severity and outcomes are unclear, we evaluated seasonal variations in stroke severity and clinical outcomes.

Methods and Results: Between 2012 and 2020 we enrolled 5,238 patients with acute ischemic stroke, who were divided into 4 seasons according to stroke onset: spring, summer, autumn and winter. We analyzed the effect of season on the severity and outcomes of all subjects. Multivariable analysis showed that the winter group had 1.234-fold increased risk of moderate-to-severe neurological deficits than the summer group (95% confidence interval (CI): 1.034–1.472, P=0.020). Compared with the summer group, the winter and the spring groups experienced 1.243- and 1.251-fold the risk of suffering from worse outcomes among all patients at 6-month follow-up (95% CI 1.008–1.534, P=0.042, 95% CI 1.013–1.544, P=0.037). The 1-year follow-up revealed similar results. Further comparison of each season in the 2012–2015 and 2016–2020 periods found that the proportion of poor outcomes in the latter autumn group was lower than that in the former time period, with significant differences in both 6-month and 1-year follow-up.

Conclusions: The onset season was related to the severity and clinical outcomes of ischemic stroke. Patients with winter onset had more severe neurological deficits and worse outcomes than those with summer onset.

Worldwide, stroke is a leading cause of mortality and disability,¹ and China has a high prevalence and heavy burden of stroke, with ischemic stroke as the most common type, accounting for up to 70%.² In addition to the well-known common risk factors, including hypertension, diabetes mellitus, overweight, atrial fibrillation (AF) and smoking,³ the relationship between meteorological factors and ischemic stroke has attracted the attention of researchers in recent years.⁴ Several epidemiological and experimental studies investigated the strong influence of seasonality on stroke incidence, and most found higher stroke incidence during winter and spring and lower incidence during summer and autumn.^5,6 In addition, weather changes may lead to severe stroke or worsen the outcome of stroke, which is a public health problem of concern. However, the results of clinical studies differ, and some are even contradictory. A recent study in the USA demonstrated that worse outcomes of ischemic stroke were more frequent in cold seasons including winter and spring.⁷ Another study from Japan found no significant correlation between the 1-year clinical outcomes of ischemic stroke and seasonal changes.⁸ In a recent study in Saudi Arabia, Alghamdi et al found that weather or seasonal variations had no effect on stroke incidence or outcomes.⁹ However, previous studies have only covered short periods of 2–5 years, so in the present study we studied the relationship of season and patterns of neurological severity and outcomes of acute ischemic stroke in a hospital-based study with 9-year cohort data. The findings of this study can provide evidence for the formulation of effective strategies to deal with the adverse consequences of ischemic stroke.

Methods

Study Population

The study population was patients admitted to the Stroke Unit of the Second Hospital of Tianjin Medical University between January 2012 and December 2020. Patients aged ≥23 years old and admitted within 7 days of symptom onset were included. All the patients were permanent residents of Tianjin. Diagnosis of acute ischemic stroke was based on clinical symptoms as well as computed tomography (CT) scans and/or magnetic resonance imaging (MRI), according to the criteria of the 10th version of the International Classification of Diseases (I63.0). Patients who met any of the following criteria were excluded: (I) transient ischemic attack, and other non-stroke diseases (n=242); (II) serious systemic disease including malignant tumor (n=107), kidney disease (n=92), liver disease (n=56), mental health disorder (n=19), and cognitive impairment (n=25); (III) incomplete data (n=112); (IV) lost to follow-up (n=87). Finally, 5,238 acute ischemic stroke patients were included in the analysis (Figure 1). Recorded data were: age, sex, systolic blood pressure (SBP) and diastolic blood pressure (DBP) on admission, the date of the stroke event; neurological symptoms at the time of the event, past history (hypertension, coronary artery disease, diabetes mellitus, AF and stroke), smoking and drinking history, laboratory results and CT scan or MRI findings. Routine blood examination and coagulation function were performed within 24 h at admission after stroke onset; however, fasting blood tests, including blood fat, fasting blood glucose, glycated hemoglobin, and homocysteine (HCY), were performed on the day following hospitalization. The subtypes of ischemic stroke considered were: large artery atherosclerosis (LAA), cardiac embolism (CE), small vessel disease (SVD), stroke of other determined etiology (SOE), and stroke of undetermined etiology (SUE). These definitions were based on the subtypes considered at the Trial of ORG 10172 in Acute Stroke Treatment (TOAST).¹⁰ The presence of complicating pneumonia was also recorded.

Figure 1.

Flowchart of the study.

Stroke Severity and Outcomes

The severity of neurological deficits at admission was determined according to the National Institutes of Health Stroke Scale (NIHSS). The clinical outcomes, including death of any cause, disability and independent ambulatory status at discharge, 6 months and 1 year, were determined using the modified Rankin scale (mRS). An NIHSS score >7 was considered reflective of a moderate-to-severe neurological deficit, and a score of 0–7 was defined as mild.^11,12 A poor outcome was defined as a mRS score of 3–6, and a score of 0–2 was considered as a good outcome.

Seasonal and Meteorological Data

Tianjin is located in the eastern part of the North China Plain, between 116°43’–118°4’ east longitude and 38°34’–40°15’ north latitude, locating it in a typical warm-temperate, semi-humid monsoon climate zone in the north temperate zone with 4 distinct seasons. Patients were divided into 4 groups according to the season of stroke attack: spring (March–May); summer (June–August); autumn (September–November); and winter (December–February). The Tianjin Meteorological Bureau provided the meteorological observations during the period of the study, including the daily mean levels of air temperature, atmospheric pressure, relative humidity, wind speed etc.

Statistical Analysis

Differences in demographic characteristics, risk factors, the degree of neurological impairment, complicating pneumonia, blood pressure on admission and laboratory results were analyzed among patients by season. Continuous variables are described as the mean±standard deviation or median (interquartile range). Categorical variables are presented as counts (percentage). The one-way analysis of variance (ANOVA) was used for continuous variables with normal distribution and the Kruskal-Wallis rank test for continuous variables without normal distribution. The χ²-test was used for categorical variables. Post-hoc tests were performed by least-significant-difference test for differences between groups; further, Bonferroni correction was used to adjust test levels for pairwise comparisons. A binary logistic regression was performed to assess the effect of season on the severity and outcomes of all patients, and the results are described as odds ratios (OR) and 95% confidence intervals (CI). All tests were two-sided. The level of significance was set at 5%. All calculations were performed using SPSS 25.0 statistical software (IBM Corp., Armonk, NY, USA).

Results

Distribution of Baseline Clinical Characteristics According to Season

Of the 5,238 acute ischemic stroke patients enrolled in the study, 3,251 (62.1%) were male and 1,987 (37.9%) were female. Among all subjects, 1,289 (24.6%) ischemic strokes occurred in spring, 1,314 (25.1%) in summer, 1,251 (23.9%) in autumn and 1,384 (26.4%) in winter. The demographic characteristics of the study subjects are shown in Supplementary Table 1. There were differences in the frequency distributions across seasons for age, sex, SBP on admission, AF, smoking, drinking and complicating pneumonia. Among the 5,238 ischemic stroke patients, SBP and the percentage of complicating pneumonia were highest during winter, as compared with other seasons. The results of laboratory tests on admission in patients with acute ischemic stroke are presented in Table 1. Significant seasonal differences can be seen for hemoglobin (P=0.043), prothrombin time (PT; P<0.001), blood glucose (P=0.003) and HCY (P<0.001). There were no seasonal differences for white blood cells (WBC), platelets, fibrinogen, thrombin time (TT), blood fat level, albumin, creatinine or high-sensitivity C-reactive protein (hs-CRP), as shown in Table 1.

Table 1. Baseline Clinical Characteristics and Laboratory Results of Ischemic Stroke Patients by Seasons

Parameters	Spring (n=1,289)	Summer (n=1,314)	Autumn (n=1,251)	Winter (n=1,384)	P value
Age (years)	69.82±11.9	69.39±11.9	70.66±11.8§	70.46±12.1§	0.026*
Male (%)	821 (63.7)	840 (63.9)	775 (62.0)	815 (58.9)§	0.025*
SBP (mmHg)	153.89±24.3†	153.95±24.0†	154.09±23.9†	156.45±24.6	0.014*
DBP (mmHg)	86.97±14.3	86.97±13.5	86.70±13.6	87.63±14.1	0.353
Medical history
Hypertension (%)	966 (74.9)	961 (73.1)	918 (73.4)	1,049 (75.8)	0.336
Coronary artery disease (%)	480 (37.2)	492 (37.4)	453 (36.2)	537 (38.8)	0.589
Atrial fibrillation (%)	193 (15.0)	162 (12.3)	197 (15.7)	229 (16.5)§	0.015*
Diabetes (%)	439 (34.1)	434 (33.0)	385 (30.8)	459 (33.2)	0.339
Stroke history (%)	501 (38.9)	471 (35.8)	479 (38.3)	540 (39.0)	0.301
Smoking (%)	482 (37.4)	546 (41.6)	506 (40.4)	504 (36.4)§	0.019*
Drinking (%)	286 (22.2)	324 (24.7)	317 (25.3)	291 (21.0)	0.027*
Laboratory parameters
WBC (×109/L)	7.96±3.1	7.72±3.1	7.73±2.8	7.87±3.1	0.162
Hemoglobin (g/L)	136.54±21.4	137.57±20.1	135.7±20.9	135.4±20.8	0.043*
Platelet (×109/L)	220.86±74.5	215.69±69.3	217.88±64.3	221.98±73.6	0.100
PT (s)	11.29±1.0	11.37±1.3	11.10±1.0*,‡	11.40±1.3§,‡	<0.001**
Fibrinogen (g/L)	3.26±0.9	3.30±0.9	3.30±0.9	3.24±0.9	0.359
TT (s)	18.9±3.3	18.7±1.9	18.9±4.0	18.6±2.7	0.246
TCHO (mmol/L)	4.95±1.2	4.93±1.2	4.89±1.2	4.86±1.2	0.313
LDL-C (mmol/L)	3.14±0.9	3.08±0.9	3.12±0.9	3.15±1.0	0.238
Albumin (g/L)	39.83±5.1	40.07±5.0	39.70±4.6	39.71±5.0	0.213
Creatinine (μmol/L)	70.00 (58.85–85.60)	70.40 (58.60–84.45)	70.50 (59.10–86.80)	69.20 (56.25–86.3)	0.571
Glucose (mmol/L)	5.85 (4.99–7.70)	5.81 (4.92–7.65)†	5.81 (4.96–7.53)†	6.03 (5.13–7.85)	0.003**
HCY (μmol/L)	14.25 (11.57–19.51)	15.11 (11.80–21.33)†	14.56 (11.06–21.01)†	13.85 (10.78–19.76)	<0.001**
hs-CRP (mg/L)	2.99 (1.39–7.91)	2.87 (1.24–7.62)	3.75 (1.41–9.91)	3.49 (1.54–11.82)	0.059
Complicating pneumonia (%)	341 (26.5)†	311 (23.7)†	324 (25.9)†	449 (32.4)	<0.001**

Data are presented as mean±standard deviation, counts (percentages) or median with the 25th and 75th percentiles. *P<0.05; **P<0.01. ^†Compared with winter, ^‡compared with spring, ^§compared with summer. DBP, diastolic blood pressure; HCY, homocysteine; hs-CRP, high-sensitivity C-reactive protein; LDL-C, low density lipoprotein cholesterol; PT, prothrombin time; SBP, systolic blood pressure; TCHO, total cholesterol; TT, thrombin time; WBC, white blood cells.

Climate variables including temperature, air pressure, relative humidity and wind speed, showed statistically significant differences in the different seasons (Supplementary Table 2).

Seasonal Variation in Neurological Severity and Stroke Outcomes

The NIHSS and mRS scores by seasons are given in Table 2. There were significant discrepancies in the NIHSS at admission and moderate-to-severe neurological deficits among the 4 groups (P<0.001 and P=0.009, respectively). The distributions of admission NIHSS in the 12 months and 4 seasons are shown in Supplementary Figure 1A,B and Supplementary Figure 2. Short- and long-term clinical outcomes also differed across the 4 groups. Clinical outcomes at discharge distributed in the 12 months and a trend of worse outcomes in the 4 seasons are shown in Supplementary Figure 1C,D and Supplementary Figure 2. A 6-month follow-up showed that the proportion of patients with poor outcomes in winter was significantly higher than in summer or autumn. The clinical outcomes at 1 year were similar to those at the 6-month follow-up. Moreover, the proportion of adverse outcomes in spring was also higher than in summer at 1-year follow-up.

Table 2. Seasonal Variations in Neurological Severity and Stroke Outcomes

Parameters	Spring (n=1,289)	Summer (n=1,314)	Autumn (n=1,251)	Winter (n=1,384)	P value
Stroke subtypes					0.001**
LAA	393 (30.5)†	438 (33.3)	398 (31.8)	496 (35.8)
CE	200 (15.3)	158 (12.0)	199 (15.9)§	235 (17.0)§
SVD	660 (51.2)†	677 (51.5)†	618 (49.4)	620 (44.8)
SOE or SUE	36 (2.8)	41 (3.1)	36 (2.9)	33 (2.4)
NIHSS	4 (2–7)	3 (1–6)†	4 (2–7)	5 (2–8)	<0.001**
Severity level					0.009**
Mild	933 (72.4)	989 (75.3)	894 (71.5)	962 (69.5)§
Moderate-to-severe	356 (27.6)	325 (24.7)	357 (28.5)	422 (30.5)§
Clinical outcomes at discharge
mRS	2 (0–4)	1 (0–4)†	1 (0–4)†	2 (1–4)	<0.001**
Poor outcome	549 (42.6)	511 (38.9)	515 (41.2)	629 (45.4)§	0.006**
Death	71 (5.5)	67 (5.1)	59 (4.7)	94 (6.8)	0.101
Clinical outcomes at 6 months
mRS	2 (1–4)	2 (0–4)†	2 (1–4)†	2 (1–4)	<0.001**
Poor outcome	566 (43.9)	515 (39.2)†	524 (41.9)†	652 (47.1)	0.001**
Death	103 (8.0)	102 (7.8)	89 (7.1)	128 (9.2)	0.229
Clinical outcomes at 1 year
mRS	2 (1–4)§	2 (0–4)†	2 (1–4)†	2 (1–4)	<0.001**
Poor outcome	576 (44.7)§	515 (39.2)†	521 (41.6)†	661 (47.8)	<0.001**
Death	118 (9.2)	115 (8.8)	110 (8.8)	151 (10.9)	0.174

Data are presented as counts (percentages) or median with the 25th and 75th percentiles. *P<0.05; **P<0.01. ^§Compared with summer, ^†compared with winter. CE, cardiac embolism; LAA, large artery atherosclerosis; mRS, modified Rankin scale; NIHSS, National Institutes of Health Stroke Scale; SOE, stroke of other determined etiology; SUE, stroke of undermined etiology; SVD, small vessel disease.

Age, sex and variables with P<0.05 after univariate analysis were included in the multivariate analysis model. The dependent variables were 2 categories of neurological severity and stroke outcomes. After adjusting for age, sex, AF and other covariates, multivariate logistic regression demonstrated that the winter group had 1.234-fold risk of moderate-to-severe neurological deficits than the summer group (95% CI: 1.034–1.472, P=0.020). There were no significant group differences in poor outcomes at discharge after additionally adjusting for severity levels of stroke and TOAST classification. However, after comparison with the summer group, the winter and spring groups respectively showed 1.243- and 1.251-fold risk of suffering from worse outcomes in the overall ischemic stroke group at 6-month follow-up (95% CI 1.008–1.534, P=0.042, 95% CI 1.013–1.544, P=0.037). The 1-year follow-up had similar results (Table 3).

Table 3. Logistic Regression Analyses for NIHSS Categories, Poor Outcomes and Seasons

Dependent variable	Model 1		Model 2
	Crude OR	P value	Adjusted OR	P value
Moderate-to-severe neurological deficita
Summer (Ref.)	1.000		1.000
Spring	1.161 (0.975, 1.383)	0.094	1.111 (0.927, 1.333)	0.255
Autumn	1.215 (1.020, 1.448)	0.029*	1.137 (0.948, 1.365)	0.167
Winter	1.335 (1.127, 1.582)	0.001**	1.234 (1.034, 1.472)	0.020**
Poor outcome (mRS 3–6) at dischargeb
Summer (Ref.)	1.000		1.000
Spring	1.166 (0.997, 1.363)	0.055	1.171 (0.947, 1.447)	0.146
Autumn	1.100 (0.939, 1.288)	0.239	0.970 (0.778, 1.210)	0.787
Winter	1.309 (1.123, 1.526)	0.001**	1.138 (0.921, 1.406)	0.232
Poor outcome (mRS 3–6) at 6 monthsb
Summer (Ref.)	1.000		1.000
Spring	1.207 (1.032, 1.411)	0.018**	1.251 (1.013, 1.544)	0.037**
Autumn	1.111 (0.949, 1.301)	0.190	0.979 (0.786, 1.221)	0.852
Winter	1.365 (1.172, 1.591)	<0.001**	1.243 (1.008, 1.534)	0.042**
Poor outcome (mRS 3–6) at 1 yearb
Summer (Ref.)	1.000		1.000
Spring	1.237 (1.059, 1.446)	0.007**	1.302 (1.057, 1.603)	0.013**
Autumn	1.093 (0.934, 1.280)	0.268	0.948 (0.762, 1.180)	0.635
Winter	1.392 (1.195, 1.622)	<0.001**	1.286 (1.045, 1.583)	0.018**

*P<0.05; **P<0.01. Model 1 was unadjusted; ^aModel 2 was adjusted for age and sex, atrial fibrillation, smoking, drinking; ^bModel 2 was adjusted for age and sex, atrial fibrillation, smoking, drinking, severity level of stroke, TOAST classification, systolic blood pressure, glucose, hemoglobin. CI, confidence interval; OR, odds ratio.

Seasonal Variation in Neurological Severity and Stroke Outcomes by Stroke Subtype

Of the 5,238 patients, 1,725 (32.9%) were classified as the LAA subtype, 792 (15.1%) were classified as CE, and 2,575 (49.2%) as SVD. No significant difference in baseline NIHSS score and moderate-to-severe neurological deficits was shown among the 4 groups across each stroke subtype (LAA, CE, and SVD). At 1-year follow-up, among the 3 stroke subtypes, only LAA showed strong differences in the mRS scores and poor outcomes in the 4 seasons (P=0.032 and 0.035, respectively). After multivariable adjustment, poor outcomes were more common in winter than in summer (95% CI 1.216–2.470, P=0.002), as shown in Table 4 and Supplementary Table 3.

Table 4. Seasonal Variation in Neurological Severity and Clinical Outcomes by Stroke Subtype

LAA	Spring (n=393)	Summer (n=438)	Autumn (n=398)	Winter (n=496)	P value
NIHSS	7 (4–12)	7 (3–12)	8 (3–11)	8 (4–12)	0.255
Severity level					0.825
Mild	200 (50.9)	229 (52.3)	197 (49.5)	246 (49.6)
Moderate-to-severe	193 (49.1)	209 (47.7)	201 (50.5)	250 (50.4)
Clinical outcomes at discharge
mRS	4 (1–5)	3 (1–5)	4 (1–5)	3 (1–5)	0.106
Poor outcome	255 (64.9)	266 (60.7)	245 (61.6)	329 (66.3)	0.247
Death	32 (8.1)	39 (8.9)	30 (7.5)	52 (10.5)	0.435
Clinical outcomes at 6 months
mRS	4 (2–5)	3 (1–5)	4 (1–5)	4 (2–5)	0.150
Poor outcome	255 (64.9)	262 (59.8)	249 (62.6)	337 (67.9)	0.067
Death	42 (10.7)	56 (12.8)	42 (10.6)	64 (12.9)	0.563
Clinical outcomes at 1 year
mRS	4 (2–5)	3 (1–5)†	4 (1–5)	4 (2–5)	0.032**
Poor outcome	256 (65.1)	261 (59.6)†	247 (62.1)	339 (68.3)	0.035**
Death	44 (11.2)	61 (13.9)	52 (13.1)	78 (15.7)	0.265
CE	Spring (n=200)	Summer (n=158)	Autumn (n=199)	Winter (n=235)	P value
NIHSS	8 (4–14)	7.5 (3–15)	9 (3–15)	8 (3–15)	0.833
Severity level					0.816
Mild	92 (46.0)	79 (50.0)	90 (45.2)	112 (47.7)
Moderate-to-severe	108 (54.0)	79 (50.0)	109 (54.8)	123 (52.3)
Clinical outcomes at discharge
mRS	4 (1–5)	3.5 (1–5)	4 (1–5)	4 (1–5)	0.891
Poor outcome	125 (62.5)	92 (58.2)	125 (62.8)	147 (62.6)	0.792
Death	30 (15.0)	24 (15.2)	23 (11.6)	35 (14.9)	0.690
Clinical outcomes at 6 months
mRS	4 (2–5)	4 (1–5)	4 (1–5)	4 (1–5)	0.803
Poor outcome	132 (66.0)	93 (58.9)	126 (63.3)	151 (64.3)	0.560
Death	43 (21.5)	33 (20.9)	35 (17.6)	52 (22.1)	0.671
Clinical outcomes at 1 year
mRS	4 (2–5)	4 (1–5)	4 (1–5)	4 (1–5)	0.873
Poor outcome	133 (66.5)	95 (60.1)	126 (63.3)	153 (65.1)	0.629
Death	49 (24.5)	37 (23.4)	41 (20.6)	56 (23.8)	0.802
SVD	Spring (n=660)	Summer (n=677)	Autumn (n=618)	Winter (n=620)	P value
NIHSS	3 (1–4)	3 (1–4)	3 (1–4)	3 (1–4)	0.720
Severity level					0.280
Mild	616 (93.3)	647 (95.6)	578 (93.8)	581 (93.7)
Moderate-to-severe	44 (6.7)	30 (4.4)	40 (6.5)	39 (6.3)
Clinical outcomes at discharge
mRS	1 (0–2)	1 (0–2)	1 (0–2)	1 (0–2)	0.640
Poor outcome	152 (23.0)	142 (21.0)	130 (21.0)	135 (21.8)	0.787
Death	36 (0.9)	3 (0.4)	3 (0.5)	6 (1.0)	0.542
Clinical outcomes at 6 months
mRS	1 (0–2)	1 (0–2)	1 (0–2)	1 (0–2)	0.454
Poor outcome	161 (24.4)	148 (21.9)	134 (21.7)	146 (23.5)	0.590
Death	12 (1.8)	12 (1.8)	8 (1.3)	9 (1.5)	0.853
Clinical outcomes at 1 year
mRS	1 (0–2)	1 (0–2)	1 (0–2)	1 (0–2)	0.374
Poor outcome	169 (25.6)	147 (21.7)	133 (21.5)	151 (24.4)	0.222
Death	19 (2.9)	16 (2.4)	11 (1.8)	14 (2.3)	0.634

Data are presented as count (percentage) or median with the 25th and 75th percentiles. *P<0.05; **P<0.01. ^§Compared with summer, ^†compared with winter. Abbreviations as in Table 2.

Comparison of Neurological Severity and Stroke Outcomes in 2 Time Periods

There were significant seasonal differences in neurological severity and clinical outcomes in the 2012–2015 and 2016–2020 datasets, which was consistent with the results for the overall ischemic stroke group (Table 5). Further comparison of each season in the 2 time periods found that the proportion of poor outcomes in autumn of the later period was lower than that in the former time period, with significant differences in 6-month and 1-year follow-up. There was no significant difference in the other groups, but the proportion of poor outcomes in the later spring period was slightly increased (Figure 2).

Table 5. Seasonal Variations in Neurological Severity and Clinical Outcomes in 2012–2015 and 2016–2020

Parameters	2012–2015					2016–2020
	Spring (n=568)	Summer (n=524)	Autumn (n=505)	Winter (n=633)	P value	Spring (n=721)	Summer (n=790)	Autumn (n=746)	Winter (n=751)	P value
Severity level					0.036**					0.004**
Mild	418 (73.6)	396 (75.6)†	368 (72.9)	432 (68.2)		516 (71.6)	600 (75.9)†	540 (72.4)	508 (67.6)
Moderate-to-severe	150 (26.4)	128 (24.4)†	137 (27.1)	201 (31.8)		205 (28.4)	190 (24.1)†	206 (27.6)	243 (32.4)
Clinical outcomes at discharge
Poor outcome	232 (40.8)	206 (39.3)†	215 (42.6)	302 (47.7)	0.021**	317 (44.0)	305 (38.6)†	283 (37.9)†	344 (45.8)	0.003**
Death	39 (6.9)	33 (6.3)	22 (4.4)	49 (7.7)	0.131	32 (4.4)	34 (4.3)	37 (5.0)	45 (6.0)	0.416
Clinical outcomes at 6 months
Poor outcome	238 (41.9)	209 (39.9)†	235 (46.5)	298 (47.1)	0.039**	328 (45.5)§	3.6 (38.7)†	289 (38.7)†	354 (47.1)	<0.001**
Death	50 (8.8)	49 (9.4)	36 (7.1)	63 (10.0)	0.397	53 (7.4)	53 (6.7)	53 (7.1)	65 (8.7)	0.503
Clinical outcomes at 1 year
Poor outcome	243 (42.8)	206 (39.3)†	229 (45.3)	304 (48.0)	0.023**	331 (45.9)§	308 (39.0)†	292 (39.1)†	360 (47.9)	<0.001**
Death	58 (10.2)	54 (10.3)	44 (8.7)	75 (11.8)	0.392	60 (8.3)	61 (7.7)	66 (8.8)	76 (10.1)	0.393

Data are presented as count (percentage). *P<0.05; **P<0.01. ^§Compared with summer, ^†compared with winter.

Figure 2.

(A–D) Comparison of neurological severity and stroke outcomes in 2 time periods. *P<0.05 for trend.

Discussion

We found that ischemic stroke admissions were more frequent in winter than in other seasons. Our study identified that stroke severity and clinical outcomes showed a certain seasonal variation, which still existed after adjusting for age, sex, AF and other related variables. Poor outcomes at 1 year were more common in winter than in summer in the overall group of ischemic stroke patients, as well as in patients with the LAA subtype. Further comparison of each season in the 2012–2015 and 2016–2020 time periods found that the proportion of poor outcomes at 6 months and 1 year in autumn was significantly lower in the 2016–2020 dataset than in the 2012–2015 dataset.

The seasonal variations in stroke severity and long-term outcomes were generally consistent with previous studies,^13,14 including a recent cross-sectional study using guideline stroke data that revealed an association between season and stroke outcomes: cold seasons were associated with worse outcomes for ischemic stroke, and ischemic stroke hospitalizations were more frequent in winter.⁷ Similar results were obtained in our previous study.¹⁵ One study that assessed the seasonal variations in neurological severity and long-term clinical outcomes at 1 year based on a 5-year observational study found that ischemic stroke had more severe neurological deficits in winter and spring, but showed no significant difference in the distribution of poor outcomes among the seasons.⁸ The result of that study differs from our study in terms of outcomes, considering the different regions and sample sizes.

The mechanism underlying the seasonality of stroke severity and long-term clinical outcomes is worth exploring. Multiple meteorological factors are key to understanding the mechanism behind seasonality. The descriptive statistics of meteorological factors by season in this study support this. The main climatic characteristics in Tianjin City are high temperature and humidity in summer and a cold, dry winter. We found significant statistical differences in the distribution of meteorological factors in the 4 seasons. The daily mean/minimum/maximum temperatures were lowest in winter and highest in summer, but the daily mean relative humidity was highest in summer and lowest in winter. Thus, low temperature is considered a top cause of the high proportion of moderate-to-severe and poor outcomes in winter and spring, because exposure to cold weather may cause peripheral blood vessel contraction, increasing the tension in the vascular wall, and elevating blood pressure.¹⁶ Secondly, low temperature can also lead to changes in blood biochemistry. Some studies have shown that the concentration of total cholesterol, fibrinogen, platelets and coagulation factors in the blood increases in the cold months.¹⁷ We also found that the PT of participants with ischemic stroke onset in winter was higher than in other seasons. These factors can lead to aggravation of atherosclerosis and thrombosis, as well as insufficient blood supply to collateral branch vessels and secondary poor outcomes, which also partly explains why seasonal variations in long-term functional outcomes were more significant in the LAA subtype.

Besides these plausible biological mechanisms, it is also worth noting that the average age of the enrolled patients with ischemic stroke was 70.08 years. Elderly patients frequently have severe cerebrovascular atherosclerosis, increased vascular elasticity and fragility, large blood pressure fluctuations, and insufficient cerebral collateral circulation reserves.¹⁸ Furthermore, elderly patients with other vascular diseases, such as hypertension and diabetes, may have vasospasm or even occlusion after exposure to cold temperatures, resulting in aggravation of ischemic stroke.^19,20 Another interesting finding from our study was that ischemic stroke patients with onset in cold seasons had a higher proportion of complicating pneumonia than those with onset in summer and autumn. The most plausible explanation for this is that extremely cold weather is conducive to upper respiratory tract infections, pneumonia, and other inflammatory cascade reactions, resulting in pulmonary cardiovascular complications and poor prognosis in patients with ischemic stroke.^21,22

In addition, our study revealed that the proportions of poor outcomes at 6 months and 1 year in the autumn periods of 2016–2020 were lower than in the same time in 2012–2015, which may be due to the extension of home heating in Tianjin since 2016. The annual heating period was advanced to November 1, lasting for > 4 months, thus reducing the effect of the first wave of cold weather on elderly patients at the beginning of November in Tianjin City. Another possible reason is increased public health awareness and public health behavioral changes resulting from positive media coverage and government support. However, the concrete mechanism still needs further exploration.

Study Strengths

These include a time span of up to 9 years, a hospital-based study with detailed clinical information on stroke characteristics, laboratory examinations, and specific daily meteorological data provided by Tianjin Meteorological Bureau. Our results provide evidence for an association between seasonal changes and stroke severity and outcomes. In recent years, the Tianjin Municipal Government has used relevant research to deal with the adverse consequences of stroke by providing prioritized coping strategies including extending the period of heating. The Tianjin Meteorological Bureau and our research center jointly developed the “Stroke Weather Warning Systems” in July 2021, and issued different levels of early warning many times to put forward ideas and health protection suggestions for high-risk stroke groups.

Study Limitations

First, it was a single-center retrospective cohort study with the inevitable loss of some data for several variables, and a certain degree of unavoidable selection bias. Second, “season” is a general expression variable based on various meteorological factors, such as temperature, barometric pressure, relative humidity, wind speed, snowfall and air pollutants including PM2.5, NO₂ and other indicators, that were not adjusted in multivariate regression analysis. Third, in addition to potential biological factors, there may be some other factors that may affect the results, such as the time for patients to get to the emergency department in winter compared with summer, and snowy weather in the typical winter in Tianjin City may prevent individuals from going out of doors or to emergency treatment. Fourth, because the time range included in this study was large and treatment measures are constantly updated, treatment strategies (including prior reperfusion therapy, post-stroke statin and antiplatelet treatment), which may have affected the outcome of stroke to a certain degree, were not analyzed.

Conclusions

There were significant seasonal variations in the severity and clinical outcomes of acute ischemic stroke. Patients with onset in winter tended to show more severe symptoms and suffer worse outcomes at 6 months and at 1 year. Poor clinical outcomes in LAA subtype at 1 year were generally similar to the overall ischemic stroke group. Our study ascertained a seasonal variation in the severity and clinical outcomes of ischemic stroke, and further research on the triggering mechanism is warranted. Understanding the biological mechanisms underlying this correlation may be beneficial for resource allocation and useful for prevention and adaptive therapy strategies.

Acknowledgments

This study was funded by Research and Popularization of Appropriate Intervention Technology for the Stroke High Risk Group in China (grant no. 2020R0005).

Disclosures

The authors declare that there are no conflicts of interest.

IRB Information

The study was approved by the Institutional Review Board of the Second Hospital of Tianjin Medical University (KY2020K183).

Supplementary Files

Please find supplementary file(s);

https://doi.org/10.1253/circj.CJ-22-0801

References

• 1.   GBD 2016 Neurology Collaborators. Global, regional, and national burden of neurological disorders, 1990–2016: A systematic analysis for the Global Burden of Disease Study 2016. Neurol 2019; 18: 459–480, doi:10.1016/s1474-4422(18)30499-x.
• 2.   Wang W, Jiang B, Sun H, Ru X, Sun D, Wang L, et al. Prevalence, incidence, and mortality of stroke in China: Results from a nationwide population-based survey of 480 687 adults. Circulation 2017; 135: 759–771, doi:10.1161/circulationaha.116.025250.
• 3.   Poorthuis MH, Algra AM, Algra A, Kappelle LJ, Klijn CJ. Female- and male-specific risk factors for stroke: A systematic review and meta-analysis. JAMA Neurol 2017; 74: 75–81, doi:10.1001/jamaneurol.2016.3482.
• 4.   Kuzmenko NV, Galagudza MM. Dependence of seasonal dynamics of hemorrhagic and ischemic strokes on the climate of a region: A meta-analysis. Int J Stroke 2022; 17: 226–235, doi:10.1177/17474930211006296.
• 5.   Giroud M, Beuriat P, Vion P, D’Athis PH, Dusserre L, Dumas R. Stroke in a French prospective population study. Neuroepidemiology 1989; 8: 97–104, doi:10.1159/000110171.
• 6.   Turin TC, Kita Y, Murakami Y, Rumana N, Sugihara H, Morita Y, et al. Higher stroke incidence in the spring season regardless of conventional risk factors: Takashima Stroke Registry, Japan, 1988–2001. Stroke 2008; 39: 745–752, doi:10.1161/strokeaha.107.495929.
• 7.   Chu SY, Cox M, Fonarow GC, Smith EE, Schwamm L, Bhatt DL, et al. Temperature and precipitation associate with ischemic stroke outcomes in the United States. J Am Heart Assoc 2018; 7: e010020, doi:10.1161/jaha.118.010020.
• 8.   Toyoda K, Koga M, Yamagami H, Yokota C, Sato S, Inoue M, et al. Seasonal variations in neurological severity and outcomes of ischemic stroke: 5-year single-center observational study. Circ J 2018; 82: 1443–1450, doi:10.1253/circj.CJ-17-1310.
• 9.   Alghamdi SAM, Aldriweesh MA, Al Bdah BA, Alhasson MA, Alsaif SA, Alluhidan WA, et al. Stroke seasonality and weather association in a Middle East country: A single tertiary center experience. Front Neurol 2021; 12: 707420, doi:10.3389/fneur.2021.707420.
• 10.   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, doi:10.1161/01.str.24.1.35.
• 11.   DeGraba TJ, Hallenbeck JM, Pettigrew KD, Dutka AJ, Kelly BJ. Progression in acute stroke: Value of the initial NIH stroke scale score on patient stratification in future trials. Stroke 1999; 30: 1208–1212, doi:10.1161/01.str.30.6.1208.
• 12.   Cuadrado-Godia E, Jimena S, Ois A, Rodríguez-Campello A, Giralt-Steinhauer E, Soriano-Tarraga C, et al. Factors associated with early outcome in patients with large-vessel carotid strokes. J Neurol Neurosurg Psychiatry 2013; 84: 305–309, doi:10.1136/jnnp-2012-303638.
• 13.   Lorking N, Wood AD, Tiamkao S, Clark AB, Kongbunkiat K, Bettencourt-Silva JH, et al. Seasonality of stroke: Winter admissions and mortality excess: A Thailand National Stroke population database study. Clin Neurol Neurosurg 2020; 199: 106261, doi:10.1016/j.clineuro.2020.106261.
• 14.   Liu Y, Gong P, Wang M, Zhou J. Seasonal variation of admission severity and outcomes in ischemic stroke: A consecutive hospital-based stroke registry. Chronobiol Int 2018; 35: 295–302, doi:10.1080/07420528.2017.1369430.
• 15.   Qi X, Wang Z, Xia X, Xue J, Gu Y, Han S, et al. Potential impacts of meteorological variables on acute ischemic stroke onset. Risk Manag Healthc Policy 2020; 13: 615–621, doi:10.2147/RMHP.S253559.
• 16.   Kobayashi S, Sakakura K, Jinnouchi H, Taniguchi Y, Tsukui T, Watanabe Y, et al. Influence of daily temperature on the occurrence of ST-elevation myocardial infarction. J Cardiol 2022, doi:10.1016/j.jjcc.2022.12.005.
• 17.   Ueda Y, Kosugi S, Abe H, Ozaki T, Mishima T, Date M, et al. Transient increase in blood thrombogenicity may be a critical mechanism for the occurrence of acute myocardial infarction. J Cardiol 2021; 77: 224–230, doi:10.1016/j.jjcc.2020.08.007.
• 18.   Alpérovitch A, Lacombe JM, Hanon O, Dartigues JF, Ritchie K, Ducimetière P, et al. RelationshiP between blood pressure and outdoor temperature in a large sample of elderly individuals: The Three-City study. Arch Intern Med 2009; 169: 75–80, doi:10.1001/archinternmed.2008.512.
• 19.   Stewart S, Keates AK, Redfern A, McMurray JJV. Seasonal variations in cardiovascular disease. Nat Rev Cardiol 2017; 14: 654–664, doi:10.1038/nrcardio.2017.76.
• 20.   Takahashi S, Ishizuka K, Hoshino T, Mizuno T, Nishimura A, Toi S, et al. Long-term outcome in patients with acute ischemic stroke and heart failure. Circ J 2023; 87: 401–408, doi:10.1253/circj.CJ-21-0937.
• 21.   Liu Y, Chen Y, Kong D, Liu X, Fu J, Zhang Y, et al. Short-term effects of cold spells on hospitalisations for acute exacerbation of chronic obstructive pulmonary disease: A time-series study in Beijing, China. BMJ Open 2021; 11: e039745, doi:10.1136/bmjopen-2020-039745.
• 22.   Turin TC, Kita Y, Rumana N, Murakami Y, Ichikawa M, Sugihara H, et al. Stroke case fatality shows seasonal variation regardless of risk factor status in a Japanese population: 15-year results from the Takashima Stroke Registry. Neuroepidemiology 2009; 32: 53–60, doi:101159/000170907.