- Email: [email protected]
A Validation of the Essen Stroke Risk Score in Outpatients with Ischemic Stroke
ARTICLE IN PRESS
A Validation of the Essen Stroke Risk Score in Outpatients with Ischemic Stroke Pan Chen, MD,* Yi Liu, MD,† Yilong Wang, MD,* Anxin Wang, MD,* Huaguang Zheng, MD,* Xingquan Zhao, MD,* Aoshuang Yan, MD,† and Yongjun Wang, MD*
Background: Little is known about the predictive accuracy of the Essen Stroke Risk Score (ESRS) for Chinese stroke outpatients. Our goal was to perform an external validation of the ESRS using a large multicenter cohort of outpatients with ischemic stroke (IS). Methods: We estimated the 1-year cumulative event rates for both recurrent stroke and combined vascular events for patients in different ESRS categories using data from a prospective cohort of 3316 outpatients with IS admitted to 18 hospitals in China. In addition, we evaluated the predictive accuracy of the ESRS for both recurrent stroke and combined vascular events using C statistic. Results: In the nonatrial fibrillation IS outpatients, the cumulative 1-year event rate was 2.47% (95% confidence interval [CI], 1.97%-3.06%) for recurrent stroke and 4.32% (95% CI, 3.65%-5.06%) for combined vascular events. The event rates were significantly higher in patients in higher ESRS categories. The ESRS had a predictive accuracy of .63 (.57-.69) for recurrent stroke and .63 (.58-.68) for combined vascular events. Conclusion: Among the Chinese outpatients with IS, the ESRS was able to stratify the risk of both recurrent stroke and combined vascular events equally well. A prediction model suitable for Chinese IS populations is needed. Key Words: Ischemic stroke—the Essen Stroke Risk Score—validation—outpatients. © 2016 Published by Elsevier Inc. on behalf of National Stroke Association.
From the *Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, Beijing, China; and †Department of Epidemiology and Health Statistics, School of Public Health, Capital Medical University, Beijing, China. Received December 9, 2015; revision received January 7, 2016; accepted February 1, 2016. P.C. and Y.L. contributed equally to this work. This study was funded by the National Health and Family Planning Commission of the People’s Republic of China (Grant No. 200902004), the Ministry of Science and Technology of the People’s Republic of China (Grant No. 2011BAI08B02), and Sanofi China. Address correspondence to Aoshuang Yan, MD, School of Public Health, Capital Medical University, No. 10 Xitoutiao, You’anmen Wai, Fengtai District, Beijing 100069, China. E-mail: [email protected].; Address correspondence to Yongjun Wang, MD, Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, No. 6 Tiantanxili, Dongcheng District, Beijing 100050, China. E-mail: [email protected]. 1052-3057/$ - see front matter © 2016 Published by Elsevier Inc. on behalf of National Stroke Association. http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2016.02.001
Introduction Recurrent stroke and subsequent cardiac events post stroke do the greatest harm to stroke victims. The recurrence rate is approximately 4%-12% in the first year after the onset of stroke.1-3 Therefore, we need tools to determine which ischemic stroke (IS) patients have a higher risk of recurring vascular events in clinical practice. With such tools, we can adopt a more positive and effective stratification approach to clinical care, making treatment decisions to achieve the best risk–benefit ratio. As a prediction model based on risk factors, the Essen Stroke Risk Score (ESRS) is an internationally recognized predictive tool for IS recurrence.4,5 The effectiveness of the ESRS was preliminarily established through the Reduction of Atherothrombosis for Continued Health (REACH) Registry and the European Stroke Prevention Study 2.6,7 Although the ESRS was validated for nonatrial fibrillation (AF) IS inpatients and outpatients in those
Journal of Stroke and Cerebrovascular Diseases, Vol. ■■, No. ■■ (■■), 2016: pp ■■–■■
1
ARTICLE IN PRESS P. CHEN ET AL.
2
studies, the performance of the ESRS in a large Chinese stroke population has not been examined. In 2011, the ESRS was validated for acute IS inpatients in China,8 using data from the China National Stroke Registry (CNSR).9 The results showed that the ESRS could accurately predict the occurrence of 1-year stroke recurrence and combined vascular events in acute IS patients, with area under the curve (AUC) values of .59 and .60. However, there is no large cohort study to verify the validity of the ESRS for outpatients in the nonacute phase of IS. Therefore, the goal of the present study was to perform an external validation of the ESRS in a large multicenter cohort of Chinese outpatients with IS so that the hierarchical management of secondary prevention strategies for IS patients in China could be improved and the recurrence of vascular events reduced without increasing the risk of bleeding. As a result, the prognosis for these patients could eventually be improved and medical costs could be reduced.
Materials and Methods Study Population Data from a substudy of Registry of Outpatients with Ischemic Stroke in Urban China (ROOTS) were analyzed in this report. The substudy was a multicenter prospective cohort study of consecutive outpatients with IS who were admitted to 18 hospitals in Beijing between July 1, 2010, and November 30, 2012.
Inclusion Criteria Patients who met the following criteria were included in the present study: 1) Age 18 years or older 2) Clinical diagnosis of IS, excluding cerebral hemorrhage, subarachnoid hemorrhage, and transient ischemic attack (TIA) 3) Occurrence of IS 1-6 months before enrollment IS is defined as stroke that consists of a sudden onset of focal or systematic clinical signs of neurological deficit that persists beyond 24 hours (World Health Organization definition). The diagnosis of IS was performed by neurologists from grade II or III urban hospitals. Computed tomography evaluation excluded other nonvascular causes of brain dysfunction (such as primary brain tumor, brain metastases, subdural hematoma, seizures, paralysis, or brain trauma). 4) Completion of 3-, 6-, and 12-month visits
Exclusion Criteria Exclusion criteria included the following: 1) Asymptomatic cerebral infarction
Figure 1. Flowchart of the subject enrollment. Abbreviation: AF, atrial fibrillation.
2) Noncerebrovascular disease (e.g., primary brain tumor, brain metastases, subdural hematoma, seizures, paralysis, or brain trauma) 3) Atrial fibrillation, which was defined as a previous diagnosis of atrial fibrillation with at least 1 confirmation via electrocardiogram (ECG) or current use of medication, or an ECG indication of atrial fibrillation confirmed and diagnosed in an outpatient clinic. 4) Lack of written informed consent 5) Subject involvement in other clinical trials These inclusion and exclusion criteria were evaluated by specialized neurologists and were evaluated for each subject using the subject’s medical records. The study was approved by the central Institutional Review Board at Beijing Tiantan Hospital, Capital Medical University, and all clinical investigations were conducted according to the principles expressed in the Declaration of Helsinki. Written informed consent was obtained from the patient or from the patient’s legally authorized representative.
Subject Enrollment A detailed patient recruitment flowchart is illustrated in Figure 1.
Baseline Data Collection and Data Management The subjects’ information was collected prospectively using unified paper-based registry forms. All of the neurologists or physicians who participated in the study went through professional program training. Each participating site collected the outpatients consecutively during the study period. Baseline information included the following: 1) Basic information: sex, age, height, weight, employment status before stroke onset, per capita
ARTICLE IN PRESS A VALIDATION OF ESRS IN OUTPATIENTS
income, medical insurance, modified Rankin Scale score, and blood pressure (mmHg); 2) Risk factors: history of stroke (on the basis of medical records or the patients’ described history of cerebral infarction, cerebral hemorrhage, or subarachnoid hemorrhage), history of TIA, hypertension (according to the medical records or the patients’ described history of high blood pressure or use of oral antihypertensive drugs), diabetes mellitus (according to the medical records or the patients’ described history of high blood glucose level or use of oral hypoglycemic drugs), dyslipidemia (history of metabolic lipid disorders, including various types of dyslipidemia), coronary heart disease (previous history of coronary heart disease, angina pectoris, or myocardial infarction), history of peripheral artery disease and current smoking status; and 3) Medications (e.g., antiplatelet drugs).
Follow-Up Data Collection and Data Management The follow-up data were collected by telephone by trained research personnel from the Tiantan stroke research follow-up and data management center. Followup information included the following: 1) stroke-associated death and all-cause death; 2) stroke recurrence (including cerebral infarction, cerebral hemorrhage and subarachnoid hemorrhage); and 3) other vascular events (including myocardial infarction, angina pectoris, TIA, pulmonary embolism, or peripheral artery disease).
3
of the case reports. The sampling rate was greater than or equal to 20%, and the reports were signed by both a representative of the CRO and the lead researcher of the subcenter. 2) Data input: a database was established through “back-to-back” double entry using EpiData3.1 (The EpiData Association, Odense, Denmark). Then, the data were exported to SAS 9.3 (SAS Institute, Inc., Cary, NC) for analysis.
Statistical Analysis Proportions were used for categorical variables, while the mean and standard deviation was used to report continuous variables. Chi-square tests were used to compare categorical variables. The Kruskal–Wallis test was used to compare continuous, non-normally distributed variables. In accordance with previous studies, to determine the boundary value, we divided the patients into low-risk (0-2 points) and high-risk (3-9 points) groups according to ESRS. The cumulative event rates and the corresponding 95% confidence intervals (CIs) of the outcomes were calculated based on a binomial distribution. We estimated the discriminatory power of the ESRS to predict the 1-year cumulative recurrence of stroke and combined vascular events using the AUC calculated by the C statistic. The 95% CI for the AUC was estimated using a nonparametric method. All of the tests were 2 tailed, and a P value less than .05 was considered statistically significant. The statistical analyses were conducted using SAS 9.3 (SAS Institute, Inc.).
Results Outcome Assessments The outcome assessments used included the following: 1) Stroke or TIA recurrence, rehospitalization, death. Death was classified as vascular death (including fatal stroke, fatal myocardial infarction, and other cardiovascular death) or death due to any other cause. Stroke recurrence included nonfatal stroke and fatal stroke recurrence (including IS and hemorrhagic stroke recurrence); and 2) Combined vascular events (including recurrent stroke, myocardial infarction, vascular death, angina pectoris, and TIA).
Research Quality Control Research quality controls included the following: 1) Data monitoring: a third-party contract research organization (CRO) was responsible for data monitoring. The CRO carried out on-site quality monitoring at every subcenter and monitored the integrity and authenticity of the case reports. The CRO reported on the integrity of all of the cases and checked the truthfulness of a sample
There were 3316 eligible subjects with IS in the present study, after excluding 230 (6.2%) patients with a history of atrial fibrillation and 151 (4.4%) patients without complete follow-up (Fig 1). The average age of the patients was 63.8 ± 12.0 years old, and 68.1% of the patients were male. Comparing the distribution of risk factors of the patients with and without complete follow-up, we found that the patients without complete follow-up were 64.8 ± 12.3 years old on average, and 59.6% of them were male. The patients with and without complete followup were comparable with respect to the prevalence of baseline risk factors (except the sex difference) and their clinical characteristics (Table 1). The cumulative 1-year occurrence rate was 2.47% (95% CI, 1.97%-3.06%) for recurrent stroke and 4.32% (95% CI, 3.65%-5.06%) for combined vascular events (Table 2). Both event rates increased significantly with increasing ESRS (Fig 2). We estimated the discriminatory power of the ESRS using the AUC values and found that the ESRS AUC values could predict the 3-, 6-, and 12-month cumulative occurrence of recurrent stroke and combined vascular events. The ESRS exhibited a similar predictive accura-
ARTICLE IN PRESS P. CHEN ET AL.
4
Table 1. Prevalence of baseline vascular risk factors and the clinical characteristics of patients with and without complete 12-month follow-up
Characteristics
ESRS
Sex, male Age (years) Younger than 65 65-75 Older than 75 Hypertension Diabetes mellitus Previous MI Other cardiovascular disease (except MI and AF) Peripheral arterial disease Smoker Previous IS or TIA ESRS mean (SD)
0 1 2 1 1 1 1 1 1 1 2.9 (1.4)
With follow-up n = 3316 (%)
Without follow-up n = 151 (%)
P value
2259 (68.12)
90 (59.60)
.0285
1740 (52.47) 929 (28.02) 647 (19.51) 2529 (76.26) 1136 (34.26) 128 (3.86) 438 (13.21) 91 (2.74) 1519 (45.81) 1498 (45.17) 2.9 (1.4)
73 (48.34) 46 (30.46) 32 (21.19) 106 (70.20) 54 (35.76) 3 (1.99) 16 (10.60) 4 (2.65) 72 (47.68) 56 (37.09) 3.1 (1.5)
.3018
.1129 .7094 .2410 .3619 .9517 .6659 .0568 .3119
Abbreviations: AF, atrial fibrillation; ESRS, Essen Stroke Risk Score; IS, ischemic stroke; MI, myocardial infarction; SD, standard deviation; TIA, transient ischemic attack.
cy of .63 for either recurrent stroke or combined vascular events at the 12-month visit (Table 3).
Discussion The present study provides the first external validation of the ESRS in a large multicenter cohort of Chinese outpatients with IS. The ROOTS cohort subgroups included 18 secondary and tertiary hospitals in Beijing; these hospitals were distributed in both urban areas and suburbs. To date, ROOTS is China’s largest prospective cohort study of outpatients with nonacute IS. Due to the study’s large size, the ROOTS cohort partially represents the healthcare status of Chinese outpatients with IS.
The ESRS was Universally Applicable The ESRS originated from a randomized, doubleblind, international clinical trial known as Clopidogrel versus Aspirin in Patients at Risk of Ischaemic Events (CAPRIE).4 Later, the ESRS was verified in 2 studies. One of these studies was the REACH study,6 which was an international, prospective, observational clinical registry.10 The REACH study included 44 countries (African countries and China were excluded), and the study population was globally representative. The results suggested that the ESRS could be applied to outpatients. The second study was a prospective cohort study in Germany (Systemic Risk Score Evaluation in Ischemic Stroke Patients [SCALA]
Table 2. One-year cumulative occurrence rates of recurrent stroke and combined vascular events stratified by ESRS category Recurrent stroke
Combined vascular events
ESRS
Total no.
No.
% (95% CI)
No.
% (95% CI)
0 1 2 3 4 5 6 7 8 9 Total
76 415 881 931 607 293 88 22 2 1 3316
0 7 11 26 18 11 5 3 1 0 82
.00 (.00-.00) 1.69 (.45-2.93) 1.25 (.52-1.98) 2.79 (1.74-3.86) 2.97 (1.62-4.31) 3.75 (1.58-5.93) 5.68 (.85-10.52) 13.64 (.00-27.98) 50.00 (.00-100) 0 (.00-.00) 2.47 (1.97-3.06)
2 9 21 46 28 22 11 3 1 0 143
2.63 (.00-6.23) 2.17 (.77-3.57) 2.38 (1.38-3.39) 4.94 (3.55-6.34) 4.61 (2.94-6.28) 7.51 (4.49-10.53) 12.50 (5.59-19.41) 13.64 (.00-27.98) 50.00 (.00-100) 0 (.00-.00) 4.32 (3.65-5.06)
Abbreviations: CI, confidence interval; ESRS, Essen Stroke Risk Score.
ARTICLE IN PRESS A VALIDATION OF ESRS IN OUTPATIENTS
Figure 2.
5
Rates of recurrent stroke and combined vascular events in patients in different ESRS categories. Abbreviation: ESRS, Essen Stroke Risk Score.
study), which included acute stroke patients from 85 stroke units.11,12 The results using the ROOTS study data indicate that the ESRS can be used to stratify Chinese patients with IS and may effectively predict stroke recurrence and combined vascular events. Above all, the ESRS was universally applicable for outpatients, acute stroke patients, and general inpatients. The ESRS was able to stratify the risk of recurrent stroke in Western countries and in China based on our study. This study did not include patients with atrial fibrillation because in the CAPRIE study, patients with atherosclerotic IS were included, whereas patients with cardiac IS caused by atrial fibrillation were excluded. In ROOTS registry, patients with AF were as little as 6.2%. The ratios of stroke patients with AF were not high in China and there were only 5.5% of hospitalized patients in the CNSR study.9 The detection rates of patients with new AF onset were relatively lower. In the CNSR the present study, the related data of outpatients were collected by ECG or AF history reported by the patients themselves; thus, there were possibilities of undervaluation of the proportion of patients with AF. Currently, China is conducting a research to increase the detection rates of IS patients with AF, which is of great
significance for the treatment of these patients for secondary prevention. In addition, several special models or scores for the recurrence of AF patients have been established, such as the CHA2DS2 and CHA2DS2-VASc scores, which have better predictive power for patients with AF.13
Recurrence of Stroke in Different Research Populations According to the data from Western countries, the risk of recurrence is higher (13%-14%) in the first year after the onset of acute stroke.3 A Chinese prospective registry study recruited 1913 consecutive cases of stroke14; of these, 1314 (68.7%) were IS, and the 1-year stroke recurrence rate was 10.5%. However, in the CAPRIE study, the annual stroke recurrence rate was 6.3% and that of combined vascular events was 8.0%. In the REACH study, the 1-year occurrence rates of recurrent stroke and combined vascular events were 5.8% and 7.9%, respectively. In contrast, in the ROOTS subgroup, the 1-year cumulative occurrence rates for recurrent stroke and combined vascular events were 2.47% and 4.32%, respectively. These rates are lower than those observed in previous studies. The patients in the ROOTS cohort subgroup came from
Table 3. Predictive accuracy of ESRS at the 3-, 6-, and 12-month visit Recurrent stroke
Combined vascular events
ESRS
AUC
95% CI
AUC
95% CI
3 months 6 months 12 months
.6303 .6156 .6283
.5308-.7298 .5350-.6952 .5683-.6883
.6079 .6256 .6295
.5310-.6848 .5661-.6851 .5836-.6754
Abbreviations: AUC, area under the curve; CI, confidence interval; ESRS indicates Essen Stroke Risk Score.
ARTICLE IN PRESS P. CHEN ET AL.
6
18 hospitals in Beijing (the most developed region), where the quality of health care is higher than in the rest of the country. In Beijing, patient compliance with treatment has also increased significantly over the last few years. Moreover, the ROOTS study included nonacute IS patients; patients with a history of stroke and TIA accounted for more than 40% of the total number of patients. These factors may all affect the occurrence rates of recurrent stroke and associated vascular events.
The ESRS could Identify High-Risk Groups In the CAPRIE and European Stroke Prevention Study 2 studies, the annual recurrence risk of stroke was more than 4% when the corresponding ESRS was 3 points or higher, and the patients in the clopidogrel group had a lower recurrence rate than the patients in the aspirin group. Therefore, patients with an ESRS of 3 points or higher were defined as high-risk patients. In the CAPRIE study, the 1-year recurrent stroke and combined vascular event rates increased significantly from lower to higher ESRS scores. Among those patients, 50% were high-risk patients with an ESRS of 3 points or higher. In the REACH study, the annual recurrence risk of stroke was more than 4% when the corresponding ESRS was higher than 3 points; those high-risk patients comprised 43.8% of the entire population. In the SCALA study, patients with an ESRS of 3 points or higher accounted for 58% of the entire population. In our study, the patients with a baseline ESRS of 3 points or higher accounted for 55.8% of the total. These results indicate that more than half of the studied population was at high risk for recurrent stroke, so we need to strengthen prevention of secondary stroke. The ESRS helps to identify high-risk patients; thus, it is a valuable tool to help evaluate patient risk stratification and to guide medication choices. Secondary prevention of stroke should pay attention to risk stratification and more positive interventions should be adopted for patients at high risk of recurrent stroke. At present, antiplatelet drugs are recommended for the use of noncardiac IS patients to reduce stroke recurrence. Aspirin alone, aspirin plus extendedrelease dipyridamole, and clopidogrel alone are all optional schemes. Several studies suggested that the high-risk patients assessed by ESRS scores administering a combination of aspirin and dipyridamole, or taking clopidogrel alone have better effects than patients taking aspirin alone.5,7
The Predictive Value of the ESRS The AUC values of our study (.63, .63) were similar to those previously obtained in the REACH (.59, .56), SCALA (.62), and CNSR (.59, .60) studies. Therefore, compared with previous research, the validity of the ESRS for the Chinese population was similar to the ESRS validity for populations in other Western countries. Generally, for a diagnostic study, AUC values between .5 and .7 are categorized as having low diagnostic value. A midlevel
diagnostic value falls between .7 and .9, and a high diagnostic value is above .9. Because the ESRS is a predictive model based on risk factors, the ideal AUC values would be above .7. However, similar to the results obtained using other comparable models and other study populations, the AUC value achieved remained below .7.
Study Limitations The study sites were selected from urban areas and may represent hospitals with better outpatient care; thus, the rates of recurrent stroke and combined vascular events may be underestimated. In addition, the ROOTS study recruited patients with IS that occurred 1-6 months before enrollment, and the recurrence of IS generally occurred within 3 months after onset. Therefore, the inclusion criteria may result in a lower stroke recurrence rate and a lower incidence of vascular events. The CAPRIE study had an average follow-up time of 1.91 years (1-3 years). The data used in our study included only 1 year of followup information, so we were unable to verify whether the ESRS could predict the risk of stroke recurrence and combined vascular events after 1 year. The ESRS does not take into account other risk factors of IS, such as lipid levels, blood pressure control levels, and treatment for secondary prevention. Therefore, the prediction accuracy of the ESRS was low; the missed diagnosis rate was 31%, and the misdiagnosis rate was 49%.15 A more accurate predictive tool for recurrent vascular events is urgently needed. Hopefully, future research will focus on the risk factors in Chinese populations and establish a more effective prediction model of stroke recurrence.
Conclusion Among Chinese outpatients with IS, the ESRS was able to stratify the risk of both recurrent stroke and combined vascular events. A prediction model suitable for IS in the Chinese population is urgently needed. Acknowledgments: We thank members, researchers, and collaborators of this study.
References 1. Davis SM, Donnan GA. Clinical practice. Secondary prevention after ischemic stroke or transient ischemic attack. N Engl J Med 2012;366:1914-1922. 2. Giles MF, Rothwell PM. Risk of stroke early after transient ischaemic attack: a systematic review and meta-analysis. Lancet Neurol 2007;6:1063-1072. 3. Feng W, Hendry RM, Adams RJ. Risk of recurrent stroke, myocardial infarction, or death in hospitalized stroke patients. Neurology 2010;74:588-593. 4. CAPRIE Steering Committee. A randomised, blinded, trial of clopidogrel versus aspirin in patients at risk of ischaemic events (CAPRIE). CAPRIE Steering Committee. Lancet 1996;348:1329-1339.
ARTICLE IN PRESS A VALIDATION OF ESRS IN OUTPATIENTS 5. Diener HC, Ringleb PA, Savi P. Clopidogrel for the secondary prevention of stroke. Expert Opin Pharmacother 2005;6:755-764. 6. Weimar C, Diener HC, Alberts MJ, et al. The Essen stroke risk score predicts recurrent cardiovascular events: a validation within the REduction of Atherothrombosis for Continued Health (REACH) registry. Stroke 2009;40:350354. 7. Diener HC. Modified-release dipyridamole combined with aspirin for secondary stroke prevention. Aging Health 2005;1:19-26. 8. Meng X, Wang Y, Zhao X, et al. Validation of the Essen Stroke Risk Score and the Stroke Prognosis Instrument II in Chinese patients. Stroke 2011;42:3619-3620. 9. Wang Y, Cui L, Ji X, et al. The China National Stroke Registry for patients with acute cerebrovascular events: design, rationale, and baseline patient characteristics. Int J Stroke 2011;6:355-361. 10. Ohman EM, Bhatt DL, Steg PG, et al. The REduction of Atherothrombosis for Continued Health (REACH) Registry: an international, prospective, observational
7
11.
12.
13.
14.
15.
investigation in subjects at risk for atherothrombotic events-study design. Am Heart J 2006;151:786.e1-e10. Weimar C, Goertler M, Rother J, et al. Predictive value of the Essen Stroke Risk Score and Ankle Brachial Index in acute ischaemic stroke patients from 85 German stroke units. J Neurol Neurosurg Psychiatry 2008;79:1339-1343. Weimar C, Goertler M, Rother J, et al. Systemic risk score evaluation in ischemic stroke patients (SCALA): a prospective cross sectional study in 85 German stroke units. J Neurol 2007;254:1562-1568. Giralt-Steinhauer E, Cuadrado-Godia E, Ois A, et al. CHA(2)DS(2)-VASc score and prognosis in ischemic strokes with atrial fibrillation. J Neurol 2012;259:745-751. Feng SJ, Liu M, Li WZ, et al. A prospective study of stroke recurrence and the risk factors. Nan Fang Yi Ke Da Xue Xue Bao 2009;29:983-985. Weimar C, Benemann J, Michalski D, et al. Prediction of recurrent stroke and vascular death in patients with transient ischemic attack or nondisabling stroke: a prospective comparison of validated prognostic scores. Stroke 2010;41:487-493.
A Validation of the Essen Stroke Risk Score in Outpatients with Ischemic Stroke Pan Chen, MD,* Yi Liu, MD,† Yilong Wang, MD,* Anxin Wang, MD,* Huaguang Zheng, MD,* Xingquan Zhao, MD,* Aoshuang Yan, MD,† and Yongjun Wang, MD*
Background: Little is known about the predictive accuracy of the Essen Stroke Risk Score (ESRS) for Chinese stroke outpatients. Our goal was to perform an external validation of the ESRS using a large multicenter cohort of outpatients with ischemic stroke (IS). Methods: We estimated the 1-year cumulative event rates for both recurrent stroke and combined vascular events for patients in different ESRS categories using data from a prospective cohort of 3316 outpatients with IS admitted to 18 hospitals in China. In addition, we evaluated the predictive accuracy of the ESRS for both recurrent stroke and combined vascular events using C statistic. Results: In the nonatrial fibrillation IS outpatients, the cumulative 1-year event rate was 2.47% (95% confidence interval [CI], 1.97%-3.06%) for recurrent stroke and 4.32% (95% CI, 3.65%-5.06%) for combined vascular events. The event rates were significantly higher in patients in higher ESRS categories. The ESRS had a predictive accuracy of .63 (.57-.69) for recurrent stroke and .63 (.58-.68) for combined vascular events. Conclusion: Among the Chinese outpatients with IS, the ESRS was able to stratify the risk of both recurrent stroke and combined vascular events equally well. A prediction model suitable for Chinese IS populations is needed. Key Words: Ischemic stroke—the Essen Stroke Risk Score—validation—outpatients. © 2016 Published by Elsevier Inc. on behalf of National Stroke Association.
From the *Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, Beijing, China; and †Department of Epidemiology and Health Statistics, School of Public Health, Capital Medical University, Beijing, China. Received December 9, 2015; revision received January 7, 2016; accepted February 1, 2016. P.C. and Y.L. contributed equally to this work. This study was funded by the National Health and Family Planning Commission of the People’s Republic of China (Grant No. 200902004), the Ministry of Science and Technology of the People’s Republic of China (Grant No. 2011BAI08B02), and Sanofi China. Address correspondence to Aoshuang Yan, MD, School of Public Health, Capital Medical University, No. 10 Xitoutiao, You’anmen Wai, Fengtai District, Beijing 100069, China. E-mail: [email protected].; Address correspondence to Yongjun Wang, MD, Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, No. 6 Tiantanxili, Dongcheng District, Beijing 100050, China. E-mail: [email protected]. 1052-3057/$ - see front matter © 2016 Published by Elsevier Inc. on behalf of National Stroke Association. http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2016.02.001
Introduction Recurrent stroke and subsequent cardiac events post stroke do the greatest harm to stroke victims. The recurrence rate is approximately 4%-12% in the first year after the onset of stroke.1-3 Therefore, we need tools to determine which ischemic stroke (IS) patients have a higher risk of recurring vascular events in clinical practice. With such tools, we can adopt a more positive and effective stratification approach to clinical care, making treatment decisions to achieve the best risk–benefit ratio. As a prediction model based on risk factors, the Essen Stroke Risk Score (ESRS) is an internationally recognized predictive tool for IS recurrence.4,5 The effectiveness of the ESRS was preliminarily established through the Reduction of Atherothrombosis for Continued Health (REACH) Registry and the European Stroke Prevention Study 2.6,7 Although the ESRS was validated for nonatrial fibrillation (AF) IS inpatients and outpatients in those
Journal of Stroke and Cerebrovascular Diseases, Vol. ■■, No. ■■ (■■), 2016: pp ■■–■■
1
ARTICLE IN PRESS P. CHEN ET AL.
2
studies, the performance of the ESRS in a large Chinese stroke population has not been examined. In 2011, the ESRS was validated for acute IS inpatients in China,8 using data from the China National Stroke Registry (CNSR).9 The results showed that the ESRS could accurately predict the occurrence of 1-year stroke recurrence and combined vascular events in acute IS patients, with area under the curve (AUC) values of .59 and .60. However, there is no large cohort study to verify the validity of the ESRS for outpatients in the nonacute phase of IS. Therefore, the goal of the present study was to perform an external validation of the ESRS in a large multicenter cohort of Chinese outpatients with IS so that the hierarchical management of secondary prevention strategies for IS patients in China could be improved and the recurrence of vascular events reduced without increasing the risk of bleeding. As a result, the prognosis for these patients could eventually be improved and medical costs could be reduced.
Materials and Methods Study Population Data from a substudy of Registry of Outpatients with Ischemic Stroke in Urban China (ROOTS) were analyzed in this report. The substudy was a multicenter prospective cohort study of consecutive outpatients with IS who were admitted to 18 hospitals in Beijing between July 1, 2010, and November 30, 2012.
Inclusion Criteria Patients who met the following criteria were included in the present study: 1) Age 18 years or older 2) Clinical diagnosis of IS, excluding cerebral hemorrhage, subarachnoid hemorrhage, and transient ischemic attack (TIA) 3) Occurrence of IS 1-6 months before enrollment IS is defined as stroke that consists of a sudden onset of focal or systematic clinical signs of neurological deficit that persists beyond 24 hours (World Health Organization definition). The diagnosis of IS was performed by neurologists from grade II or III urban hospitals. Computed tomography evaluation excluded other nonvascular causes of brain dysfunction (such as primary brain tumor, brain metastases, subdural hematoma, seizures, paralysis, or brain trauma). 4) Completion of 3-, 6-, and 12-month visits
Exclusion Criteria Exclusion criteria included the following: 1) Asymptomatic cerebral infarction
Figure 1. Flowchart of the subject enrollment. Abbreviation: AF, atrial fibrillation.
2) Noncerebrovascular disease (e.g., primary brain tumor, brain metastases, subdural hematoma, seizures, paralysis, or brain trauma) 3) Atrial fibrillation, which was defined as a previous diagnosis of atrial fibrillation with at least 1 confirmation via electrocardiogram (ECG) or current use of medication, or an ECG indication of atrial fibrillation confirmed and diagnosed in an outpatient clinic. 4) Lack of written informed consent 5) Subject involvement in other clinical trials These inclusion and exclusion criteria were evaluated by specialized neurologists and were evaluated for each subject using the subject’s medical records. The study was approved by the central Institutional Review Board at Beijing Tiantan Hospital, Capital Medical University, and all clinical investigations were conducted according to the principles expressed in the Declaration of Helsinki. Written informed consent was obtained from the patient or from the patient’s legally authorized representative.
Subject Enrollment A detailed patient recruitment flowchart is illustrated in Figure 1.
Baseline Data Collection and Data Management The subjects’ information was collected prospectively using unified paper-based registry forms. All of the neurologists or physicians who participated in the study went through professional program training. Each participating site collected the outpatients consecutively during the study period. Baseline information included the following: 1) Basic information: sex, age, height, weight, employment status before stroke onset, per capita
ARTICLE IN PRESS A VALIDATION OF ESRS IN OUTPATIENTS
income, medical insurance, modified Rankin Scale score, and blood pressure (mmHg); 2) Risk factors: history of stroke (on the basis of medical records or the patients’ described history of cerebral infarction, cerebral hemorrhage, or subarachnoid hemorrhage), history of TIA, hypertension (according to the medical records or the patients’ described history of high blood pressure or use of oral antihypertensive drugs), diabetes mellitus (according to the medical records or the patients’ described history of high blood glucose level or use of oral hypoglycemic drugs), dyslipidemia (history of metabolic lipid disorders, including various types of dyslipidemia), coronary heart disease (previous history of coronary heart disease, angina pectoris, or myocardial infarction), history of peripheral artery disease and current smoking status; and 3) Medications (e.g., antiplatelet drugs).
Follow-Up Data Collection and Data Management The follow-up data were collected by telephone by trained research personnel from the Tiantan stroke research follow-up and data management center. Followup information included the following: 1) stroke-associated death and all-cause death; 2) stroke recurrence (including cerebral infarction, cerebral hemorrhage and subarachnoid hemorrhage); and 3) other vascular events (including myocardial infarction, angina pectoris, TIA, pulmonary embolism, or peripheral artery disease).
3
of the case reports. The sampling rate was greater than or equal to 20%, and the reports were signed by both a representative of the CRO and the lead researcher of the subcenter. 2) Data input: a database was established through “back-to-back” double entry using EpiData3.1 (The EpiData Association, Odense, Denmark). Then, the data were exported to SAS 9.3 (SAS Institute, Inc., Cary, NC) for analysis.
Statistical Analysis Proportions were used for categorical variables, while the mean and standard deviation was used to report continuous variables. Chi-square tests were used to compare categorical variables. The Kruskal–Wallis test was used to compare continuous, non-normally distributed variables. In accordance with previous studies, to determine the boundary value, we divided the patients into low-risk (0-2 points) and high-risk (3-9 points) groups according to ESRS. The cumulative event rates and the corresponding 95% confidence intervals (CIs) of the outcomes were calculated based on a binomial distribution. We estimated the discriminatory power of the ESRS to predict the 1-year cumulative recurrence of stroke and combined vascular events using the AUC calculated by the C statistic. The 95% CI for the AUC was estimated using a nonparametric method. All of the tests were 2 tailed, and a P value less than .05 was considered statistically significant. The statistical analyses were conducted using SAS 9.3 (SAS Institute, Inc.).
Results Outcome Assessments The outcome assessments used included the following: 1) Stroke or TIA recurrence, rehospitalization, death. Death was classified as vascular death (including fatal stroke, fatal myocardial infarction, and other cardiovascular death) or death due to any other cause. Stroke recurrence included nonfatal stroke and fatal stroke recurrence (including IS and hemorrhagic stroke recurrence); and 2) Combined vascular events (including recurrent stroke, myocardial infarction, vascular death, angina pectoris, and TIA).
Research Quality Control Research quality controls included the following: 1) Data monitoring: a third-party contract research organization (CRO) was responsible for data monitoring. The CRO carried out on-site quality monitoring at every subcenter and monitored the integrity and authenticity of the case reports. The CRO reported on the integrity of all of the cases and checked the truthfulness of a sample
There were 3316 eligible subjects with IS in the present study, after excluding 230 (6.2%) patients with a history of atrial fibrillation and 151 (4.4%) patients without complete follow-up (Fig 1). The average age of the patients was 63.8 ± 12.0 years old, and 68.1% of the patients were male. Comparing the distribution of risk factors of the patients with and without complete follow-up, we found that the patients without complete follow-up were 64.8 ± 12.3 years old on average, and 59.6% of them were male. The patients with and without complete followup were comparable with respect to the prevalence of baseline risk factors (except the sex difference) and their clinical characteristics (Table 1). The cumulative 1-year occurrence rate was 2.47% (95% CI, 1.97%-3.06%) for recurrent stroke and 4.32% (95% CI, 3.65%-5.06%) for combined vascular events (Table 2). Both event rates increased significantly with increasing ESRS (Fig 2). We estimated the discriminatory power of the ESRS using the AUC values and found that the ESRS AUC values could predict the 3-, 6-, and 12-month cumulative occurrence of recurrent stroke and combined vascular events. The ESRS exhibited a similar predictive accura-
ARTICLE IN PRESS P. CHEN ET AL.
4
Table 1. Prevalence of baseline vascular risk factors and the clinical characteristics of patients with and without complete 12-month follow-up
Characteristics
ESRS
Sex, male Age (years) Younger than 65 65-75 Older than 75 Hypertension Diabetes mellitus Previous MI Other cardiovascular disease (except MI and AF) Peripheral arterial disease Smoker Previous IS or TIA ESRS mean (SD)
0 1 2 1 1 1 1 1 1 1 2.9 (1.4)
With follow-up n = 3316 (%)
Without follow-up n = 151 (%)
P value
2259 (68.12)
90 (59.60)
.0285
1740 (52.47) 929 (28.02) 647 (19.51) 2529 (76.26) 1136 (34.26) 128 (3.86) 438 (13.21) 91 (2.74) 1519 (45.81) 1498 (45.17) 2.9 (1.4)
73 (48.34) 46 (30.46) 32 (21.19) 106 (70.20) 54 (35.76) 3 (1.99) 16 (10.60) 4 (2.65) 72 (47.68) 56 (37.09) 3.1 (1.5)
.3018
.1129 .7094 .2410 .3619 .9517 .6659 .0568 .3119
Abbreviations: AF, atrial fibrillation; ESRS, Essen Stroke Risk Score; IS, ischemic stroke; MI, myocardial infarction; SD, standard deviation; TIA, transient ischemic attack.
cy of .63 for either recurrent stroke or combined vascular events at the 12-month visit (Table 3).
Discussion The present study provides the first external validation of the ESRS in a large multicenter cohort of Chinese outpatients with IS. The ROOTS cohort subgroups included 18 secondary and tertiary hospitals in Beijing; these hospitals were distributed in both urban areas and suburbs. To date, ROOTS is China’s largest prospective cohort study of outpatients with nonacute IS. Due to the study’s large size, the ROOTS cohort partially represents the healthcare status of Chinese outpatients with IS.
The ESRS was Universally Applicable The ESRS originated from a randomized, doubleblind, international clinical trial known as Clopidogrel versus Aspirin in Patients at Risk of Ischaemic Events (CAPRIE).4 Later, the ESRS was verified in 2 studies. One of these studies was the REACH study,6 which was an international, prospective, observational clinical registry.10 The REACH study included 44 countries (African countries and China were excluded), and the study population was globally representative. The results suggested that the ESRS could be applied to outpatients. The second study was a prospective cohort study in Germany (Systemic Risk Score Evaluation in Ischemic Stroke Patients [SCALA]
Table 2. One-year cumulative occurrence rates of recurrent stroke and combined vascular events stratified by ESRS category Recurrent stroke
Combined vascular events
ESRS
Total no.
No.
% (95% CI)
No.
% (95% CI)
0 1 2 3 4 5 6 7 8 9 Total
76 415 881 931 607 293 88 22 2 1 3316
0 7 11 26 18 11 5 3 1 0 82
.00 (.00-.00) 1.69 (.45-2.93) 1.25 (.52-1.98) 2.79 (1.74-3.86) 2.97 (1.62-4.31) 3.75 (1.58-5.93) 5.68 (.85-10.52) 13.64 (.00-27.98) 50.00 (.00-100) 0 (.00-.00) 2.47 (1.97-3.06)
2 9 21 46 28 22 11 3 1 0 143
2.63 (.00-6.23) 2.17 (.77-3.57) 2.38 (1.38-3.39) 4.94 (3.55-6.34) 4.61 (2.94-6.28) 7.51 (4.49-10.53) 12.50 (5.59-19.41) 13.64 (.00-27.98) 50.00 (.00-100) 0 (.00-.00) 4.32 (3.65-5.06)
Abbreviations: CI, confidence interval; ESRS, Essen Stroke Risk Score.
ARTICLE IN PRESS A VALIDATION OF ESRS IN OUTPATIENTS
Figure 2.
5
Rates of recurrent stroke and combined vascular events in patients in different ESRS categories. Abbreviation: ESRS, Essen Stroke Risk Score.
study), which included acute stroke patients from 85 stroke units.11,12 The results using the ROOTS study data indicate that the ESRS can be used to stratify Chinese patients with IS and may effectively predict stroke recurrence and combined vascular events. Above all, the ESRS was universally applicable for outpatients, acute stroke patients, and general inpatients. The ESRS was able to stratify the risk of recurrent stroke in Western countries and in China based on our study. This study did not include patients with atrial fibrillation because in the CAPRIE study, patients with atherosclerotic IS were included, whereas patients with cardiac IS caused by atrial fibrillation were excluded. In ROOTS registry, patients with AF were as little as 6.2%. The ratios of stroke patients with AF were not high in China and there were only 5.5% of hospitalized patients in the CNSR study.9 The detection rates of patients with new AF onset were relatively lower. In the CNSR the present study, the related data of outpatients were collected by ECG or AF history reported by the patients themselves; thus, there were possibilities of undervaluation of the proportion of patients with AF. Currently, China is conducting a research to increase the detection rates of IS patients with AF, which is of great
significance for the treatment of these patients for secondary prevention. In addition, several special models or scores for the recurrence of AF patients have been established, such as the CHA2DS2 and CHA2DS2-VASc scores, which have better predictive power for patients with AF.13
Recurrence of Stroke in Different Research Populations According to the data from Western countries, the risk of recurrence is higher (13%-14%) in the first year after the onset of acute stroke.3 A Chinese prospective registry study recruited 1913 consecutive cases of stroke14; of these, 1314 (68.7%) were IS, and the 1-year stroke recurrence rate was 10.5%. However, in the CAPRIE study, the annual stroke recurrence rate was 6.3% and that of combined vascular events was 8.0%. In the REACH study, the 1-year occurrence rates of recurrent stroke and combined vascular events were 5.8% and 7.9%, respectively. In contrast, in the ROOTS subgroup, the 1-year cumulative occurrence rates for recurrent stroke and combined vascular events were 2.47% and 4.32%, respectively. These rates are lower than those observed in previous studies. The patients in the ROOTS cohort subgroup came from
Table 3. Predictive accuracy of ESRS at the 3-, 6-, and 12-month visit Recurrent stroke
Combined vascular events
ESRS
AUC
95% CI
AUC
95% CI
3 months 6 months 12 months
.6303 .6156 .6283
.5308-.7298 .5350-.6952 .5683-.6883
.6079 .6256 .6295
.5310-.6848 .5661-.6851 .5836-.6754
Abbreviations: AUC, area under the curve; CI, confidence interval; ESRS indicates Essen Stroke Risk Score.
ARTICLE IN PRESS P. CHEN ET AL.
6
18 hospitals in Beijing (the most developed region), where the quality of health care is higher than in the rest of the country. In Beijing, patient compliance with treatment has also increased significantly over the last few years. Moreover, the ROOTS study included nonacute IS patients; patients with a history of stroke and TIA accounted for more than 40% of the total number of patients. These factors may all affect the occurrence rates of recurrent stroke and associated vascular events.
The ESRS could Identify High-Risk Groups In the CAPRIE and European Stroke Prevention Study 2 studies, the annual recurrence risk of stroke was more than 4% when the corresponding ESRS was 3 points or higher, and the patients in the clopidogrel group had a lower recurrence rate than the patients in the aspirin group. Therefore, patients with an ESRS of 3 points or higher were defined as high-risk patients. In the CAPRIE study, the 1-year recurrent stroke and combined vascular event rates increased significantly from lower to higher ESRS scores. Among those patients, 50% were high-risk patients with an ESRS of 3 points or higher. In the REACH study, the annual recurrence risk of stroke was more than 4% when the corresponding ESRS was higher than 3 points; those high-risk patients comprised 43.8% of the entire population. In the SCALA study, patients with an ESRS of 3 points or higher accounted for 58% of the entire population. In our study, the patients with a baseline ESRS of 3 points or higher accounted for 55.8% of the total. These results indicate that more than half of the studied population was at high risk for recurrent stroke, so we need to strengthen prevention of secondary stroke. The ESRS helps to identify high-risk patients; thus, it is a valuable tool to help evaluate patient risk stratification and to guide medication choices. Secondary prevention of stroke should pay attention to risk stratification and more positive interventions should be adopted for patients at high risk of recurrent stroke. At present, antiplatelet drugs are recommended for the use of noncardiac IS patients to reduce stroke recurrence. Aspirin alone, aspirin plus extendedrelease dipyridamole, and clopidogrel alone are all optional schemes. Several studies suggested that the high-risk patients assessed by ESRS scores administering a combination of aspirin and dipyridamole, or taking clopidogrel alone have better effects than patients taking aspirin alone.5,7
The Predictive Value of the ESRS The AUC values of our study (.63, .63) were similar to those previously obtained in the REACH (.59, .56), SCALA (.62), and CNSR (.59, .60) studies. Therefore, compared with previous research, the validity of the ESRS for the Chinese population was similar to the ESRS validity for populations in other Western countries. Generally, for a diagnostic study, AUC values between .5 and .7 are categorized as having low diagnostic value. A midlevel
diagnostic value falls between .7 and .9, and a high diagnostic value is above .9. Because the ESRS is a predictive model based on risk factors, the ideal AUC values would be above .7. However, similar to the results obtained using other comparable models and other study populations, the AUC value achieved remained below .7.
Study Limitations The study sites were selected from urban areas and may represent hospitals with better outpatient care; thus, the rates of recurrent stroke and combined vascular events may be underestimated. In addition, the ROOTS study recruited patients with IS that occurred 1-6 months before enrollment, and the recurrence of IS generally occurred within 3 months after onset. Therefore, the inclusion criteria may result in a lower stroke recurrence rate and a lower incidence of vascular events. The CAPRIE study had an average follow-up time of 1.91 years (1-3 years). The data used in our study included only 1 year of followup information, so we were unable to verify whether the ESRS could predict the risk of stroke recurrence and combined vascular events after 1 year. The ESRS does not take into account other risk factors of IS, such as lipid levels, blood pressure control levels, and treatment for secondary prevention. Therefore, the prediction accuracy of the ESRS was low; the missed diagnosis rate was 31%, and the misdiagnosis rate was 49%.15 A more accurate predictive tool for recurrent vascular events is urgently needed. Hopefully, future research will focus on the risk factors in Chinese populations and establish a more effective prediction model of stroke recurrence.
Conclusion Among Chinese outpatients with IS, the ESRS was able to stratify the risk of both recurrent stroke and combined vascular events. A prediction model suitable for IS in the Chinese population is urgently needed. Acknowledgments: We thank members, researchers, and collaborators of this study.
References 1. Davis SM, Donnan GA. Clinical practice. Secondary prevention after ischemic stroke or transient ischemic attack. N Engl J Med 2012;366:1914-1922. 2. Giles MF, Rothwell PM. Risk of stroke early after transient ischaemic attack: a systematic review and meta-analysis. Lancet Neurol 2007;6:1063-1072. 3. Feng W, Hendry RM, Adams RJ. Risk of recurrent stroke, myocardial infarction, or death in hospitalized stroke patients. Neurology 2010;74:588-593. 4. CAPRIE Steering Committee. A randomised, blinded, trial of clopidogrel versus aspirin in patients at risk of ischaemic events (CAPRIE). CAPRIE Steering Committee. Lancet 1996;348:1329-1339.
ARTICLE IN PRESS A VALIDATION OF ESRS IN OUTPATIENTS 5. Diener HC, Ringleb PA, Savi P. Clopidogrel for the secondary prevention of stroke. Expert Opin Pharmacother 2005;6:755-764. 6. Weimar C, Diener HC, Alberts MJ, et al. The Essen stroke risk score predicts recurrent cardiovascular events: a validation within the REduction of Atherothrombosis for Continued Health (REACH) registry. Stroke 2009;40:350354. 7. Diener HC. Modified-release dipyridamole combined with aspirin for secondary stroke prevention. Aging Health 2005;1:19-26. 8. Meng X, Wang Y, Zhao X, et al. Validation of the Essen Stroke Risk Score and the Stroke Prognosis Instrument II in Chinese patients. Stroke 2011;42:3619-3620. 9. Wang Y, Cui L, Ji X, et al. The China National Stroke Registry for patients with acute cerebrovascular events: design, rationale, and baseline patient characteristics. Int J Stroke 2011;6:355-361. 10. Ohman EM, Bhatt DL, Steg PG, et al. The REduction of Atherothrombosis for Continued Health (REACH) Registry: an international, prospective, observational
7
11.
12.
13.
14.
15.
investigation in subjects at risk for atherothrombotic events-study design. Am Heart J 2006;151:786.e1-e10. Weimar C, Goertler M, Rother J, et al. Predictive value of the Essen Stroke Risk Score and Ankle Brachial Index in acute ischaemic stroke patients from 85 German stroke units. J Neurol Neurosurg Psychiatry 2008;79:1339-1343. Weimar C, Goertler M, Rother J, et al. Systemic risk score evaluation in ischemic stroke patients (SCALA): a prospective cross sectional study in 85 German stroke units. J Neurol 2007;254:1562-1568. Giralt-Steinhauer E, Cuadrado-Godia E, Ois A, et al. CHA(2)DS(2)-VASc score and prognosis in ischemic strokes with atrial fibrillation. J Neurol 2012;259:745-751. Feng SJ, Liu M, Li WZ, et al. A prospective study of stroke recurrence and the risk factors. Nan Fang Yi Ke Da Xue Xue Bao 2009;29:983-985. Weimar C, Benemann J, Michalski D, et al. Prediction of recurrent stroke and vascular death in patients with transient ischemic attack or nondisabling stroke: a prospective comparison of validated prognostic scores. Stroke 2010;41:487-493.