HS‐CRP Modifies the Prognostic Value of Platelet Count for Clinical Outcomes After Ischemic Stroke

Abstract

Background

We examined whether the relationship between baseline platelet count and clinical outcomes is modulated by HS‐CRP (high‐sensitivity C‐reactive protein) in patients with ischemic stroke.

Methods and Results

A total of 3267 patients with ischemic stroke were included in the analysis. The primary outcome was a combination of death and major disability at 1 year after ischemic stroke. Secondary outcomes included major disability, death, vascular events, composite outcome of vascular events or death, and an ordered 7‐level categorical score of the modified Rankin Scale at 1 year. Multivariate logistic regression and Cox proportional hazards regression models were used to assess the association between the baseline platelet count and clinical outcomes stratified by HS‐CRP levels when appropriate. There was an interaction effect of platelet count and HS‐CRP on the adverse clinical outcomes after ischemic stroke (all P _interaction<0.05). The elevated platelet count was significantly associated with the primary outcome (odds ratio [OR], 3.14 [95% CI, 1.77–5.58]), major disability (OR, 2.07 [95% CI, 1.15–3.71]), death (hazard ratio [HR], 2.75 [95% CI, 1.31–5.79]), and composite outcome of vascular events or death (HR, 2.57 [95% CI, 1.38–4.87]) among patients with high HS‐CRP levels (all P _trend<0.05).

Conclusions

The HS‐CRP levels had a modifying effect on the association between platelet count and clinical outcomes in patients with ischemic stroke. Elevated platelet count was significantly associated with adverse clinical outcomes in patients with ischemic stroke with high HS‐CRP levels, but not in those with low HS‐CRP levels. These findings suggest that strategies for anti‐inflammatory and antiplatelet therapy should be developed according to the results of both platelet and HS‐CRP testing.

Nonstandard Abbreviations and Acronyms

CATIS

China Antihypertensive Trial in Acute Ischemic Stroke

mRS

modified Rankin Scale

Clinical Perspective.

What Is New?

• Our prospective study demonstrated that high HS‐CRP (high‐sensitivity C‐reactive protein) levels could significantly modify the prognostic value of the platelet count for clinical outcomes after ischemic stroke, and provide valuable insights into adding anti‐inflammatory therapy to the prevention of poor outcomes after ischemic stroke when administering antiplatelet therapy.

What Are the Clinical Implications?

• Our findings have important clinical implications not only for improving the risk stratification of ischemic stroke, but also for suggesting a combination of both antiplatelet and anti‐inflammatory therapies in patients with ischemic stroke, especially in patients with elevated platelet and HS‐CRP levels.

Stroke is still the second leading cause of mortality and the main cause of disability worldwide, with 12.2 million incident stroke cases in 2019, ¹ leading to a serious disease burden in low‐ and middle‐income countries. ² Ischemic stroke accounts for >70% of incident strokes. ³ , ⁴

Thrombosis and inflammatory reactions are important mechanisms in the occurrence and development of ischemic stroke. ⁵ , ⁶ , ⁷ , ⁸ , ⁹ Platelets play a critical role in the formation of thrombosis and embolism, and HS‐CRP (high‐sensitivity C‐reactive protein) is an important indicator of inflammation, which contributes to the development of cardiocerebrovascular atherosclerosis and the initiation of ischemic stroke development. ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ It has been reported that increased platelet counts are associated with cardiovascular events and mortality in the general population. ¹⁷ , ¹⁸ In a Mendelian randomization study, a higher genetically determined platelet count increased the risk of ischemic stroke. ¹⁹ However, the relationship between platelet count and ischemic stroke prognosis remains inconclusive, and 2 studies ²⁰ , ²¹ showed a significant association of increased platelet count with the development of functional outcomes in ischemic stroke, whereas 1 study showed no significant association. ²² Mounting evidence suggests that an elevated HS‐CRP level reflects the instability of atherosclerotic plaques ¹³ and is associated with the risk of cardiovascular diseases, stroke, and adverse clinical outcomes after stroke. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ²³ Elevated HS‐CRP levels have been reported to increase plasma von Willebrand factor levels, ²⁴ which, in turn, affect the activity of platelets. ¹¹ , ²⁵ However, the effect of HS‐CRP on the association between platelet count and clinical outcomes of ischemic stroke is unclear. Thus, the aim of the present study was to evaluate the prognostic value of platelet count in patients with ischemic stroke stratified by HS‐CRP levels.

METHODS

Study Participants

The data that support the findings of this study are available from the corresponding author on reasonable request. This study was conducted on the basis of the CATIS (China Antihypertensive Trial in Acute Ischemic Stroke), a multicenter, single‐blind, blinded end point randomized clinical trial that was conducted in 26 hospitals across China. More details on the rational design and major results of the CATIS have been reported in previous publications. ²⁶ Briefly, 4071 patients were recruited for the CATIS. The inclusion criteria for the CATIS were as follows: (1) age ≥22 years, (2) ischemic stroke confirmed by computed tomography or magnetic resonance imaging of the brain within 48 hours of symptom onset, and (3) elevated systolic blood pressure (BP) between 140 and <220 mm Hg. The exclusion criteria were as follows: (1) systolic BP ≥220 mm Hg or diastolic BP ≥120 mm Hg, (2) severe heart failure, (3) acute myocardial infarction or unstable angina, (4) atrial fibrillation, (5) aortic dissection, (6) cerebrovascular stenosis (≥70%), (7) resistant hypertension, (8) deep coma, and (9) treatment with intravenous thrombolytic therapy. Furthermore, 804 patients were excluded because of a lack of baseline HS‐CRP or platelet count recordings (N=644) or loss to follow‐up at 1 year (N=160). Finally, 3267 patients with acute ischemic stroke were included in the present analysis (Figure 1).

Figure 1. Study participant flowchart.

CATIS indicates China Antihypertensive Trial in Acute Ischemic Stroke; and HS‐CRP, high‐sensitivity C‐reactive protein.

This study was approved by the Institutional Review Boards of Soochow University in China and Tulane University in the United States, as well as by the ethical committees of the 26 participating hospitals. Written informed consent was obtained from all study participants or their immediate family members.

Data Collection and Measurements

Baseline data on demographic characteristics, medication history, and clinical features were collected at enrollment. The National Institutes of Health Stroke Scale was used to evaluate the stroke severity. ²⁷ According to the symptoms and imaging data of the patients, ischemic stroke was classified as large‐artery atherosclerosis, cardiac embolism, and small‐artery occlusion lacunae (lacunar). ²⁸ Three baseline BP measurements were obtained by trained nurses while the patient was in the supine position using a standard mercury sphygmomanometer. ²⁹ The mean of the 3 BP records was used in the analyses.

Blood samples were collected within 24 hours of hospital admission after at least 8 hours of fasting. Routine laboratory analyses (those of blood glucose, blood lipids, and platelet count) were performed for all enrolled patients at each participating hospital on admission. Body weight and height were measured using standard methods, and the body mass index was calculated as weight in kilograms divided by the square of height in meters (kg/m²). The HS‐CRP concentration was measured using commercially available immunoassays (R&D Systems). Dyslipidemia was defined as total cholesterol level of ≥5.2 mmol/L, triglycerides level of ≥1.7 mmol/L, low‐density lipoprotein cholesterol level of ≥3.4 mmol/L, or high‐density lipoprotein cholesterol level of <1.0 mmol/L, according to Chinese guidelines on the prevention and treatment of dyslipidemia in adults. ³⁰ The estimated glomerular filtration rate was calculated using the Chronic Kidney Disease Epidemiology Collaboration creatinine equation with an adjusted coefficient of 1.1 for the Chinese population. ³¹

Outcome Assessment

Participants were followed up at 1 year after stroke by trained neurologists unaware of random assignment. ²⁷ The primary outcome was a 1‐year poor functional outcome (modified Rankin Scale [mRS] score, 3–6). The secondary outcomes were major disability (mRS score of 3–5), death (mRS score of 6), vascular events (ie, recurrent nonfatal stroke, nonfatal myocardial infarction, hospitalized and treated angina, hospitalized and treated congestive heart failure, and hospitalized and treated peripheral arterial disease), and a composite outcome of vascular events or death. We further included an ordered 7‐level categorical score of the mRS at 1 year as an outcome of the neurologic functional status. ³² The death certificates were obtained for all deceased participants, and hospital data were abstracted for vascular events. The study outcome assessment committee, blinded to random assignment, reviewed and adjudicated subsequent outcomes based on the criteria established in the ALLHAT (Antihypertensive and Lipid‐Lowering Treatment to Prevent Heart Attack Trial). ³³

Statistical Analysis

First, the patients were divided into low HS‐CRP level (<4.8 mg/L) and high HS‐CRP level (≥4.8 mg/L) subgroups based on the cutoff value of the third quartile of baseline HS‐CRP levels. Second, the platelet count was divided into 4 groups according to the baseline platelet count quartiles within each HS‐CRP subgroup. Tests for linear trends in baseline characteristics across the platelet count quartiles in each HS‐CRP subgroup were performed using generalized linear regression analysis for continuous variables and the Cochran‐Armitage trend χ² test for categorical variables. We inferred robustness of the model by robust SE estimate. The baseline characteristics of the patients included in the present study and excluded from CATIS were compared using Student t test, Wilcoxon rank‐sum test, or the χ² test as appropriate (Table S1). Multivariate logistic regression and Cox proportional hazards regression models were used to assess the association between the baseline platelet count and clinical outcomes stratified by HS‐CRP levels when appropriate. Multivariate ordinal logistic regression was used to calculate the influence of platelet counts on the ordered 7‐level categorical score of the mRS at 1 year in each HS‐CRP level subgroup. Odds ratios (ORs) or hazard ratios (HRs) and 95% CIs were calculated for the upper quartiles of platelet counts compared with the lowest quartile. Model 1 was adjusted for age and sex. Model 2 was further adjusted for current cigarette smoking, current alcohol drinking, the National Institutes of Health Stroke Scale score at admission, time to randomization after admission, body mass index, randomized antihypertensive status, dyslipidemia, estimated glomerular filtration rate, history of hypertension, family history of stroke, history of diabetes, use of lipid‐lowering drugs, coronary heart disease, blood glucose, white blood cell count, systolic BP at baseline, and the ischemic stroke subtype based on model 1. Model 3 was further adjusted for antiplatelet therapy after admission based on model 2. The effect of the interaction between the HS‐CRP level and platelet count on study outcomes was also tested using the likelihood ratio test in models with multiplicative interaction terms. Nonparametric restricted cubic splines were used to explore the shape of the association between the platelet count and adverse clinical outcomes after ischemic stroke with 4 knots (at the 5th, 35th, 65th, and 95th percentiles of the subgroup‐specific distribution of platelet counts). ³⁴

To test the robustness of our findings, thrombocytopenia (platelet count <100×10⁹/L; N=79) and thrombocytosis (platelet count >450×10⁹/L; N=10) were further excluded from the sensitivity analyses conducted in multivariate‐adjusted logistic regression models or multivariate‐adjusted Cox proportional hazards models.

All P values were 2 tailed, and a significance level of 0.05 was used. Statistical analyses were performed using SAS statistical software version 9.4 (SAS Institute, Cary, NC).

RESULTS

Baseline Characteristics

The baseline characteristics of the patients included in this study were essentially balanced with those excluded from CATIS population (Table S1). A total of 3267 individuals (2093 men and 1174 women) with an average age of 62.4 years were included. Among the patients with high HS‐CRP levels, those with higher platelet counts tended to be women, had a lower rate of alcohol drinking and smoking, had a higher prevalence of diabetes, and had higher low‐density lipoprotein cholesterol, total cholesterol, and blood glucose levels and white blood cell counts. Among the patients with low HS‐CRP levels, those with higher platelet counts were more likely to be younger and women and have lower systolic BP levels, lower rates of alcohol drinking and diabetes, higher white blood cell counts and triglycerides, total cholesterol, low‐density lipoprotein cholesterol, and high‐density lipoprotein cholesterol levels, and a longer time from onset to randomization (Table 1).

Table 1.

Baseline Characteristics of Study Participants With Different Baseline Platelet Counts Stratified by HS‐CRP Levels

Characteristics*	HS‐CRP <4.8 mg/L				P trend value	HS‐CRP ≥4.8 mg/L				P trend value
	Quartile 1	Quartile 2	Quartile 3	Quartile 4		Quartile 1	Quartile 2	Quartile 3	Quartile 4
No. of subjects	604	614	630	602		197	215	172	233
Demographics
Age, y	64.0 (11.0)	61.5 (10.8)	61.0 (9.9)	59.4 (10.9)	<0.001	66.7 (10.7)	64.7 (10.7)	65.8 (10.7)	64.5 (11.0)	0.106
Men, N (%)	449 (74.3)	416 (67.8)	399 (63.3)	305 (50.7)	<0.001	143 (72.6)	153 (71.2)	102 (59.3)	126 (54.1)	<0.001
Current cigarette smoking, N (%)	242 (40.1)	223 (36.3)	247 (39.2)	205 (34.1)	0.088	78 (39.6)	91 (42.3)	56 (32.6)	75 (32.2)	0.031
Current alcohol drinking, N (%)	204 (33.8)	215 (35.0)	200 (31.8)	166 (27.6)	0.010	61 (31.0)	74 (34.4)	44 (25.6)	57 (24.5)	0.040
Time to randomization after admission, h	10 (4–24)	12 (5–24)	10 (5–24)	12 (5–24)	0.006	7 (4–24)	10 (5–24)	12 (5–24)	7 (3–20)	0.461
Medical history, N (%)
Hypertension	467 (77.3)	490 (79.8)	484 (76.8)	473 (78.6)	0.927	150 (76.1)	176 (81.9)	143 (83.1)	189 (81.1)	0.220
Hyperlipidemia	44 (7.3)	47 (7.7)	41 (6.5)	50 (8.3)	0.696	12 (6.1)	9 (4.2)	8 (4.7)	18 (7.7)	0.838
Diabetes	122 (20.2)	108 (26.0)	97 (15.4)	89 (14.8)	0.007	29 (14.7)	36 (16.7)	33 (19.2)	55 (23.6)	0.014
Clinical features
Baseline NIHSS score	4 (2–7)	4 (2–7)	4 (2–7)	4 (2–7)	0.686	6 (3–11)	6 (3–10)	7 (3–11)	6 (4–10)	0.980
BMI, kg/m2	24.8 (3.1)	24.9 (3.0)	24.9 (3.2)	24.9 (3.0)	0.527	24.9 (3.2)	24.9 (3.1)	25.0 (3.3)	25.3 (3.3)	0.161
Systolic BP, mm Hg	167.1 (17.6)	166.1 (16.8)	165.6 (15.6)	165.0 (17.0)	0.030	166.9 (17.6)	167.4 (17.1)	167.7 (18.0)	167.5 (16.8)	0.704
Diastolic BP, mm Hg	96.5 (11.2)	96.7 (11.3)	96.6 (9.8)	96.7 (11.5)	0.829	95.7 (10.9)	97.7 (11.3)	97.3 (11.7)	96.1 (11.5)	0.920
Triglycerides, mmol/L	1.4 (1.0–2.1)	1.5 (1.0–2.0)	1.5 (1.1–2.2)	1.6 (1.1–2.2)	0.029	1.3 (0.9–2.0)	1.4 (1.0–2.1)	1.5 (1.1–2.0)	1.4 (1.0–2.1)	0.379
TC, mmol/L	4.7 (4.0–5.4)	5.0 (4.3–5.6)	5.1 (4.3–5.8)	5.3 (4.5–5.9)	<0.001	4.7 (4.1–5.5)	4.8 (4.3–5.6)	5.1 (4.6–5.9)	5.1 (4.3–5.9)	0.001
LDL‐C, mmol/L	2.6 (2.1–3.3)	2.9 (2.3–3.5)	2.9 (2.3–3.6)	3.0 (2.4–3.6)	<0.001	2.7 (2.1–3.4)	2.8 (2.3–3.5)	3.0 (2.4–3.5)	3.0 (2.4–3.6)	<0.001
HDL‐C, mmol/L	1.2 (1.0–1.4)	1.2 (1.0–1.5)	1.2 (1.0–1.5)	1.3 (1.1–1.5)	0.006	1.3 (1.0–1.5)	1.2 (1.0–1.5)	1.3 (1.0–1.5)	1.2 (1.0–1.5)	0.589
Blood glucose, mmol/L	5.8 (5.1–7.5)	5.7 (5.0–6.8)	5.7 (5.0–7.0)	5.6 (5.0–6.8)	0.059	5.7 (5.1–7.1)	6.0 (5.2–7.3)	6.2 (5.4–8.1)	6.5 (5.5–8.2)	0.009
White blood cell, 109/L	5.9 (4.9–7.1)	6.3 (5.2–7.5)	6.9 (5.7–8.2)	7.4 (6.1–8.7)	<0.001	6.9 (5.5–8.9)	7.5 (6.1–9.2)	8.4 (6.4–8.4)	8.5 (6.8–10.0)	<0.001
Ischemic stroke subtype, N (%)
Thrombotic	470 (77.8)	471 (76.7)	495 (78.6)	462 (76.7)	0.866	134 (68.0)	169 (78.6)	130 (75.6)	168 (72.1)	0.598
Embolic	26 (4.3)	22 (3.6)	17 (2.7)	25 (4.2)	0.686	20 (10.2)	15 (7.0)	9 (12.2)	30 (12.9)	0.333
Lacunar	120 (19.9)	134 (21.8)	140 (22.2)	131 (21.8)	0.414	46 (23.4)	37 (17.2)	38 (22.1)	38 (16.3)	0.167
Randomized antihypertensive, N (%)	302 (50.0)	294 (47.9)	336 (53.3)	279 (46.4)	0.561	107 (54.3)	109 (50.7)	92 (53.5)	118 (50.6)	0.583

Quartile 1, platelet count <171×10⁹/L; quartile 2, 171×10⁹/L≤platelet count<209×10⁹/L; quartile 3, 209×10⁹/L≤platelet count<248×10⁹/L; and quartile 4, platelet count ≥248×10⁹/L. BMI indicates body mass index; BP, blood pressure; HDL‐C, high‐density lipoprotein cholesterol; HS‐CRP, high‐sensitivity C‐reactive protein; LDL‐C, low‐density lipoprotein cholesterol; NIHSS, National Institutes of Health Stroke Scale; and TC, total cholesterol.

^*Continuous variables are expressed as mean (SD) or median (interquartile range). Categorical variables are expressed as frequency (percentage).

Baseline Platelet Count and Clinical Outcomes

At 1 year after ischemic stroke, 419 (17.10%) patients in the low HS‐CRP group and 317 (38.80%) patients in the high HS‐CRP group met the primary end point of poor functional outcomes. In the high HS‐CRP group, the multivariable‐adjusted ORs (95% CIs) or HRs (95% CIs) for the primary outcome, major disability, death, composite outcome of vascular events or death, and the ordered 7‐level categorical score of the mRS at 1 year for patients in the highest quartile of platelet count compared with the lowest quartile of platelet count were 3.14 (1.77–5.58), 2.07 (1.15–3.71), 2.75 (1.31–5.79), 2.57 (1.38–4.87), and 2.01 (1.33–3.04), respectively. The multivariable‐adjusted ORs (95% CIs) or HRs (95% CIs) for the primary outcome, major disability, death, composite outcome of vascular events or death, and the ordered 7‐level categorical score of the mRS at 1 year for patients in the highest quartile of platelet count compared with the lowest quartile were 1.13 (0.75–1.71), 1.34 (0.87–2.08), 0.48 (0.19–1.24), 0.79 (0.45–1.37), and 1.17 (0.91–1.49; all P _trend>0.05) in those with low HS‐CRP levels, which showed an interaction effect of the platelet count and HS‐CRP on the primary outcome (OR, 1.41 [95% CI, 1.14–1.75]; P _interaction=0.002), death (HR, 1.84 [95% CI, 1.32–2.56]; P _interaction<0.001), vascular events (HR, 1.40 [95% CI, 1.09–1.80]; P _interaction=0.008), composite outcome of vascular events or death (HR, 1.38 [95% CI, 1.08–1.77]; P _interaction=0.011), and the ordered 7‐level categorical score of the mRS (OR, 1.20 [95% CI, 1.04–1.39]; P _interaction=0.016) at 1 year. HS‐CRP significantly modified the prognostic value of the platelet count for the clinical outcomes of ischemic stroke (Table 2). Each 100×10⁹/L increment in platelet counts was associated with a 1.97‐fold increase in risk for the primary outcome, 1.55‐fold increase in risk for major disability, 1.84‐fold increase in risk for death, 1.77‐fold increase in risk for vascular events, 1.85‐fold increase in risk for composite of vascular events or death, and 1.64‐fold increase in risk for a 1‐unit higher modified Rankin score in high HS‐CRP level group (Table 3). After further exclusion of thrombocytopenia and thrombocytosis, our findings were consistent with the original results, confirming the robustness of our results (Tables S2 and S3).

Table 2.

ORs or HRs and 95% CIs of Outcomes for Quartile of Baseline Platelet Counts Stratified by HS‐CRP Levels at Baseline

Characteristics	HS‐CRP <4.8 mg/L					HS‐CRP ≥4.8 mg/L					Risks of interaction	P interaction value
	Quartile 1	Quartile 2	Quartile 3	Quartile 4	P trend value	Quartile 1	Quartile 2	Quartile 3	Quartile 4	P trend value
Primary outcome at 1 y (mRS score 3–6, multiple‐adjusted OR)											1.41 (1.14–1.75)	0.002
Cases, N (%)	107 (17.7)	110 (17.9)	109 (17.3)	93 (15.5)		66 (33.5)	73 (34.0)	68 (39.5)	110 (47.2)
Model 1	1.00	1.13 (0.84–1.53)	1.12 (0.83–1.51)	1.02 (0.74–1.41)	0.893	1.00	1.14 (0.75–1.74)	1.38 (0.89–2.15)	2.08 (1.37–3.15)	<0.001
Model 2	1.00	1.14 (0.77–1.68)	1.13 (0.76–1.67)	1.13 (0.75–1.71)	0.592	1.00	1.46 (0.81–2.64)	1.88 (1.00–3.48)	3.19 (1.80–5.67)	<0.001
Model 3	1.00	1.15 (0.77–1.70)	1.13 (0.76–1.68)	1.13 (0.75–1.71)	0.586	1.00	1.43 (0.79–2.57)	1.85 (0.99–3.45)	3.14 (1.77–5.58)	<0.001
Secondary outcomes
Major disability at 1 y (mRS score 3–5, multiple‐adjusted OR)											1.16 (0.93–1.45)	0.187
Cases, N (%)	80 (13.3)	89 (14.5)	86 (13.7)	80 (13.3)		46 (23.4)	49 (22.8)	47 (27.3)	70 (30.0)
Model 1	1.00	1.21 (0.87–1.68)	1.15 (0.82–1.68)	1.16 (0.82–1.64)	0.480	1.00	1.02 (0.64–1.62)	1.27 (0.79–2.05)	1.51 (0.97–2.36)	0.038
Model 2	1.00	1.18 (0.77–1.80)	1.25 (0.82–1.91)	1.34 (0.86–2.07)	0.188	1.00	1.33 (0.73–2.44)	1.61 (0.85–3.02)	2.01 (1.15–3.68)	0.012
Model 3	1.00	1.19 (0.78–1.81)	1.26 (0.83–1.93)	1.34 (0.87–2.08)	0.184	1.00	1.33 (0.73–2.44)	1.60 (0.85–3.02)	2.07 (1.15–3.71)	0.011
Death within 1 y (multiple‐adjusted HR)											1.84 (1.32–2.56)	<0.001
Cases, N (%)	27 (4.5)	21 (3.4)	23 (3.7)	13 (2.2)		20 (10.2)	24 (11.2)	21 (12.2)	40 (17.2)
Model 1	1.00	0.90 (0.51–1.60)	1.01 (0.57–1.78)	0.57 (0.28–1.14)	0.199	1.00	1.19 (0.65–2.17)	1.09 (0.58–2.06)	1.87 (1.08–3.23)	0.027
Model 2	1.00	1.06 (0.53–2.13)	0.80 (0.37–1.73)	0.49 (0.19–1.25)	0.200	1.00	1.41 (0.76–3.92)	1.72 (0.76–3.92)	2.94 (1.40–6.17)	0.002
Model 3	1.00	1.07 (0.53–2.14)	0.81 (0.37–1.73)	0.48 (0.19–1.24)	0.128	1.00	1.34 (0.59–3.05)	1.69 (0.74–3.84)	2.75 (1.31–5.79)	0.004
Vascular events within 1 y (multiple‐adjusted HR)											1.40 (1.09–1.80)	0.008
Cases, N (%)	31 (5.1)	32 (5.2)	37 (5.9)	26 (4.3)		13 (6.6)	13 (6.1)	10 (5.8)	25 (10.7)
Model 1	1.00	1.00 (0.46–2.17)	0.95 (0.41–2.16)	1.85 (0.93–3.66)	0.810	1.00	1.01 (0.46–2.17)	0.95 (0.41–2.16)	1.85 (0.93–3.66)	0.066
Model 2	1.00	1.08 (0.61–1.92)	1.02 (0.57–1.80)	0.89 (0.48–1.65)	0.687	1.00	0.93 (0.37–2.33)	1.09 (0.43–2.73)	1.92 (0.87–4.26)	0.068
Model 3	1.00	1.08 (0.61–1.92)	1.02 (0.57–1.80)	0.89 (0.48–1.65)	0.684	1.00	0.87 (0.35–2.17)	1.07 (0.43–2.67)	1.77 (0.80–3.93)	0.100
Composite outcome of death or vascular events within 1 y (multiple‐adjusted HR)											1.38 (1.08–1.77)	0.011
Cases, N (%)	47 (7.8)	40 (6.5)	50 (7.9)	34 (5.7)		25 (12.7)	29 (13.5)	24 (14.0)	46 (19.7)
Model 1	1.00	0.91 (0.60–1.39)	1.13 (0.75–1.69)	0.80 (0.51–1.27)	0.614	1.00	1.20 (0.70–2.06)	1.18 (0.67–2.07)	1.78 (1.08–2.92)	0.024
Model 2	1.00	1.00 (0.60–1.64)	1.02 (0.63–1.68)	0.79 (0.45–1.37)	0.471	1.00	1.47 (0.73–2.96)	1.85 (0.90–3.73)	2.76 (1.45–5.22)	0.001
Model 3	1.00	1.00 (0.61–1.65)	1.03 (0.63–1.68)	0.79 (0.45–1.37)	0.473	1.00	1.38 (0.68–2.79)	1.80 (0.89–3.62)	2.57 (1.38–4.87)	0.002
mRS score at 1 y (multiple‐adjusted OR)*											1.20 (1.04–1.39)	0.016
Model 1	1.00	0.91 (0.74–1.12)	1.01 (0.83–1.24)	1.09 (0.88–1.35)	0.283	1.00	1.17 (0.83–1.65)	1.25 (0.87–1.80)	1.66 (1.18–2.33)	0.004
Model 2	1.00	0.87 (0.69–1.10)	1.03 (0.81–1.30)	1.17 (0.91–1.49)	0.098	1.00	1.32 (0.87–2.00)	1.55 (0.99–2.42)	2.03 (1.34–3.06)	<0.001
Model 3	1.00	0.88 (0.69–1.11)	1.03 (0.81–1.30)	1.17 (0.91–1.49)	0.103	1.00	1.29 (0.85–1.96)	1.56 (1.00–2.44)	2.01 (1.33–3.04)	<0.001

Model 1, adjusted for age and sex; model 2, further adjusted for current cigarette smoking, current alcohol drinking, admission National Institutes of Health Stroke Scale score, time to randomization after admission, body mass index, randomized antihypertensive, dyslipidemia, estimated glomerular filtration rate, history of hypertension, family history of stroke, history of diabetes, coronary heart disease, blood glucose, white blood cell count, systolic blood pressure at baseline, ischemic stroke subtype, and use of lipid‐lowering drugs based on model 1; model 3, further adjusted for antiplatelet therapy after admission based on model 2. For the mRS score, 0 indicates no symptoms; 1, no significant disability despite symptoms; 2, slight disability; 3, moderate disability; 4, moderately severe disability; 5, severe disability; and 6, death. Quartile 1, platelet count <171×10⁹/L; quartile 2, 171×10⁹/L≤platelet count<209×10⁹/L; quartile 3, 209×10⁹/L≤platelet count<248×10⁹/L; and quartile 4, platelet count ≥248×10⁹/L. HR indicates hazard ratio; HS‐CRP, high‐sensitivity C‐reactive protein; mRS, modified Rankin Scale; and OR, odds ratio.

^*Odds of a 1‐unit higher mRS score.

Table 3.

ORs or HRs and 95% CIs of Study Outcomes for Each 100×10⁹/L of Baseline Platelet Count Stratified by HS‐CRP Level

Outcome	HS‐CRP <4.8 mg/L		HS‐CRP ≥4.8 mg/L
	OR or HR (95% CI)	P value	OR or HR (95% CI)	P value
Death or major disability at 1 y (multiple‐adjusted OR)	1.06 (0.85–1.32)	0.613	1.97 (1.46–2.65)	<0.001
Major disability at 1 y (multiple‐adjusted OR)	1.19 (0.94–1.50)	0.140	1.55 (1.15–2.09)	0.004
Death within 1 y (multiple‐adjusted HR)	0.65 (0.42–1.01)	0.055	1.84 (1.32–2.56)	<0.001
Vascular events within 1 y (multiple‐adjusted HR)	0.86 (0.62–1.19)	0.349	1.77 (1.21–2.59)	0.003
Death or vascular events within 1 y (multiple‐adjusted HR)	0.84 (0.63–1.12)	0.238	1.85 (1.40–2.45)	<0.001
Modified Rankin Scale score at 1 y (multiple‐adjusted OR)	1.11 (0.97–1.27)	0.131	1.64 (1.32–2.03)	<0.001

In the multivariate models adjusted for age, sex, current cigarette smoking, current alcohol drinking, admission National Institutes of Health Stroke Scale score, time to randomization after admission, body mass index, randomized antihypertensive, dyslipidemia, estimated glomerular filtration rate, history of hypertension, family history of stroke, history of diabetes, coronary heart disease, blood glucose, white blood cell count, systolic blood pressure at baseline, ischemic stroke subtype, use of lipid‐lowering drugs, and antiplatelet therapy after admission. HR indicates hazard ratio; HS‐CRP, high‐sensitivity C‐reactive protein; and OR, odds ratio.

Multivariate‐adjusted spline regression models showed a linear association between the platelet count and primary outcome (P _linearity<0.001) and composite outcome of vascular events or death (P _linearity<0.001) in the high HS‐CRP group (Figure 2A and 2B). Moreover, there was a dose‐dependent relationship between the platelet count and 1‐year mRS score (P _trend<0.001; Figure 3 and Table 2).

Figure 2. Linear test of the association between baseline blood platelet count and 1‐year clinical outcomes.

Odds ratio or hazard ratio and 95% CI derived from restricted cubic spline regression, with knots placed at the 5th, 35th, 65th, and 95th percentiles of platelet count. Odds ratios or hazard ratios were adjusted for the same variables as model 3 in Table 2. A, Death or major disability. B, Death or vascular events. HS‐CRP indicates high‐sensitivity C‐reactive protein.

Figure 3. Baseline platelet count and 1‐year modified Rankin Scale (mRS) score.

Multivariable‐adjusted odds ratio of ordinal logistic regression analysis was 2.01 (95% CI, 1.33–3.04) for patients in the highest quartile (Q) of platelet count compared with the patients in the lowest quartile in the high HS‐CRP (high‐sensitivity C‐reactive protein) group (P _trend<0.001). Multivariable model adjusted for the same variables as model 3 in Table 2.

DISCUSSION

To our knowledge, this is the first large prospective study to assess the prognostic value of the baseline platelet count for clinical outcomes according to HS‐CRP stratification among patients with ischemic stroke. In the present study based on the CATIS, we found that an elevated platelet count was independently associated with an increased risk of primary outcome, death, and composite outcome of vascular events or death, and the ordered 7‐level categorical score of the mRS within 1 year after ischemic stroke among patients with high HS‐CRP levels, but not in those with low HS‐CRP levels. Thus, these results strongly support the theory that high HS‐CRP levels could significantly modify the prognostic value of the platelet count for clinical outcomes after ischemic stroke, and provide valuable insights into adding anti‐inflammatory therapy to the prevention of poor outcomes after ischemic stroke when administering antiplatelet therapy. ³⁵ In fact, anti‐inflammatory treatments, such as use of colchicine, have been reported in patients with chronic coronary disease, ³⁶ and other studies have already demonstrated the potential benefits of statin therapy in reducing the risk of vascular events by lowering HS‐CRP levels. ³⁷

Previous studies have reported the involvement of platelets in maintaining vascular homeostasis and mediating immune responses, inflammation, and atherosclerosis. ¹⁰ In an observational study, the platelet count showed a U‐shaped relationship with future cardiovascular events and mortality in the general population. ¹⁷ However, the relationship between the platelet count and stroke prognosis has been inconsistent. In a cohort of 16 842 participants with stroke, platelet counts within normal ranges were associated with long‐term stroke recurrence, mortality, and poor functional outcome. ²⁰ In contrast, Du et al ³⁸ reported that an elevated platelet count was not significantly associated with 30‐days prognosis after ischemic stroke. Recently, a study investigated the prognostic value of the platelet count for clinical outcomes in patients with ischemic stroke and transient ischemic attack. Patients in the top quintile of platelet count (249–450× 10⁹/L) had an increased risk of death (adjusted HR, 1.43 [95% CI, 1.19–1.73]) and poor functional outcome (adjusted OR, 1.49 [95% CI, 1.28–1.74]) at 1‐year follow‐up compared with those in the low platelet count quintile (186–212×10⁹/L). ²⁰ However, in a prospective study of 553 patients with first‐ever ischemic stroke, Ghodsi et al reported that increased mean platelet volume, not platelet count, was significantly associated with mortality within 3 months (OR, 3.88 [95% CI, 2.04–7.38]) and 1 year (OR, 3.32 [95% CI, 1.91–5.78]), as well as poor function at 3 months (OR, 3.25 [95% CI, 1.80–5.86]) and 1 year (OR, 4.35 [95% CI, 2.36–8.02]). ²² Among the above studies, some included transient ischemic attack ³⁸ or used small samples, ²² , ³⁸ which did not consider the effect of HS‐CRP on the relationship between the platelet count and clinical outcome of stroke. Both thrombosis and inflammatory reactions are important mechanisms for the occurrence and development of ischemic stroke, ¹³ , ³⁹ , ⁴⁰ which may affect the prognosis after stroke in the form of their interaction. Our study found a linear relationship between the baseline platelet count and poor outcomes of ischemic stroke only in patients with high HS‐CRP levels, but not in those with low HS‐CRP levels, and the association persisted in a sensitivity analysis in which patients with thrombocytopenia or thrombocytosis were excluded. Our results suggest that HS‐CRP modifies the prognostic value of the platelet count for the clinical outcome of ischemic stroke.

As an inflammatory factor, HS‐CRP is considered an important marker of atherosclerotic rupture ¹³ , ³⁹ , ⁴⁰ and is associated with incidence risk ¹⁶ , ²³ and poor prognosis risk of ischemic stroke. ⁴¹ , ⁴² Several mechanisms may underlie the observed interaction between the HS‐CRP level and platelet count in the prognosis of ischemic stroke. The platelet count and HS‐CRP level are important indicators of thrombosis and inflammation in the development of ischemic stroke. ¹² , ¹⁶ Moreover, several studies have found that elevated HS‐CRP levels not only promote an increase in the plasma von Willebrand factor concentration ²⁴ but also stimulate the expression of matrix metalloprotein‐2 and matrix metalloprotein‐9, ¹³ , ³⁹ , ⁴⁰ which are directly involved in the rupture of atherosclerotic plaques. Ruptured plaques can promote the adhesion of von Willebrand factor and reduce platelet velocity, thereby stimulating the activation of platelet function, enhancing platelet adhesion and aggregation, and forming thrombosis, ¹¹ , ²⁵ which may increase the risk of adverse clinical outcomes after ischemic stroke. Our findings support the above‐mentioned relationship between HS‐CRP and platelets and suggest that there is an effect of the interaction between HS‐CRP and platelet count on adverse clinical outcomes after ischemic stroke.

Antiplatelet therapy is an important component of the prevention of poor outcome after ischemic stroke ³⁵ ; therefore, we adjusted for antiplatelet therapy in a multivariable model as an important confounder. Thus, the effect of antiplatelet therapy on the credibility of the present results is minimal. At present, there are no randomized clinical trials on anti‐inflammatory drugs for ischemic stroke prognosis, but Ridker et al ⁴³ found that the use of statins reduced the incidence of stroke and all‐cause mortality in a healthy population with elevated HS‐CRP levels, but without hyperlipidemia. This provides a perspective on the use of statins to reduce inflammation in patients with stroke. Laboratory testing for the platelet count and HS‐CRP level is universal and inexpensive, and the combined detection of the 2 indexes can better predict the prognosis of ischemic stroke; our findings suggest that antiplatelet and anti‐inflammatory therapies should be taken simultaneously when the 2 indexes increase. Therefore, our findings have important clinical implications not only for improving the risk stratification of ischemic stroke, but also for suggesting a combination of both antiplatelet and anti‐inflammatory therapies in patients with ischemic stroke, especially in patients with elevated platelet and HS‐CRP levels.

Our study has several limitations. First, patients with BP ≥220/120 mm Hg or severe cardiovascular disease or those undergoing intravenous thrombolytic therapy were excluded because of the CATIS design; thus, this study might have some selection bias. However, the proportion of patients with BP ≥220/120 mm Hg or those treated with intravenous thrombolytic therapy is low in China, ²⁶ , ⁴⁴ and the baseline characteristics of participants in this study were similar to those from the China National Stroke Registry (CNSR), ⁴⁵ suggesting that the selection bias may be minimal. Second, we only collected data on serum HS‐CRP levels and platelet counts at admission, not their follow‐up data; thus, the association between the changes in these biomarkers and stroke prognosis cannot be estimated. Third, this was an observational study; thus, the possibility of residual confounding could not be fully eliminated, although we did adjust for almost all of the important confounders.

CONCLUSIONS

We found that the prognostic value of the platelet count was modified by HS‐CRP in patients with ischemic stroke. Elevated platelet counts were associated with adverse clinical outcomes only in patients with ischemic stroke with high HS‐CRP levels, suggesting that strategies for anti‐inflammatory and antiplatelet therapy should be developed according to the results of both platelet and HS‐CRP testing. Further prospective studies from other populations and randomized clinical trials are required to verify our findings.

Sources of Funding

This study was supported by the National Natural Science Foundation of China (grant 82020108028).

Disclosures

None.

Supporting information

Tables S1–S3