Outcomes of Argatroban Treatment in Patients With Atherothrombotic Stroke
Observational Nationwide Study in Japan
Background and Purpose—Argatroban, a selective thrombin inhibitor, is recommended for the use in patients with atherothrombotic stroke by the Japanese Guidelines for the Management of Patients with Acute Ischemic Stroke. We performed a nationwide Japanese study to investigate whether argatroban improved early stroke outcomes in patients with acute atherothrombotic stroke.
Methods—This retrospective observational study, using the Diagnosis Procedure Combination database in Japan, included patients who were hospitalized from July 1, 2010, to March 31, 2012, with a diagnosis of atherothrombotic stroke within 1 day of stroke onset. Patients were divided into 2 groups: those receiving argatroban on admission (argatroban group), and those who did not receive argatroban during hospitalization (control group). To balance the baseline characteristics and concomitant treatments during hospitalization between the 2 groups, one-to-one propensity-score matching analyses were performed. The main outcomes were the modified Rankin Scale score at discharge and the occurrence of hemorrhagic complications during hospitalization. An ordinal logistic regression analysis evaluated the association between argatroban use and modified Rankin Scale at discharge.
Results—After propensity-score matching, 2289 pairs of patients were analyzed. There were no significant differences in modified Rankin Scale at discharge between the argatroban and the control groups (adjusted odds ratio, 1.01; 95% confidence interval, 0.88–1.16). The occurrence of hemorrhagic complications did not differ significantly between the argatroban and the control groups (3.5% versus 3.8%; P=0.58).
Conclusions—The present study suggested that argatroban was safe, but had no added benefit in early outcomes after acute atherothrombotic stroke.
In Western countries, there are no specific recommendations for anticoagulant drug use in patients with acute ischemic stroke.1,2 In the Japanese guidelines, it is stated that there is no scientific evidence for the use of heparin, heparinoid drugs, or low molecular weight heparin in patients with acute ischemic stroke. In contrast, argatroban, a selective thrombin inhibitor, is recommended in Japan for the use in patients with acute atherothrombotic stroke except those with lacunar infarction.2
To the best of our knowledge, there are only 3 randomized controlled trials on argatroban use in patients with acute ischemic stroke: 1 from North America3 and 2 from Japan.4,5 These studies suggested that argatroban was effective and safe for the use as early anticoagulant therapy in acute ischemic stroke. However, the sample sizes of these studies were small. A previous meta-analysis of 23 748 patients from 24 studies, including the 3 studies on argatroban, concluded that early anticoagulant therapies were not beneficial in patients with acute ischemic stroke.6 However, patients from the 3 studies on argatroban accounted for only 1.3% (303 patients) of all patients included in the meta-analysis.3–5 Thus, it is still uncertain whether argatroban is beneficial in patients with acute ischemic stroke.
A previous in vivo study suggested that argatroban was a neuroprotective agent and an anticoagulant.7 In fact, argatroban directly inhibits thrombin, which damages neurovascular units during acute ischemia.8 Thus, there is a possibility that argatroban could improve outcomes in patients with acute ischemic stroke.
The aim of the present study was to investigate whether argatroban was effective and safe in patients with acute ischemic stroke, using a database of Japanese inpatients.
Study Design and Setting
This was a retrospective observational cohort study using data from a nationwide administrative claims and discharge abstract database in Japan, the Diagnosis Procedure Combination database. The Diagnosis Procedure Combination database has been described in detail previously.9 In brief, >1000 hospitals, including all 82 academic hospitals in Japan, contribute to the Diagnosis Procedure Combination database. The annual number of cases added to the database is ≈7 000 000 and encompasses ≈50% of all hospital admissions in Japan.
The Diagnosis Procedure Combination database contains the following information for each patient: patient demographics, a unique hospital identifier, diagnoses, outcomes, drugs used, procedures performed, healthcare costs, and several disease-specific data. Diagnoses include the main diagnosis on admission, preexisting comorbidities, and postadmission complications, which are recorded separately with International Classification of Diseases-10th revision codes and text data in Japanese. Drugs and procedures are recorded along with the date of use or implementation. For patients with stroke, the date of occurrence and the modified Rankin Scale (mRS) and Japan Coma Scale scores10 are also documented. These scores are evaluated on admission and at discharge. An attending physician is responsible for clinical data entry for each patient. Drugs and procedures are recorded based on receipt data of medical care.
The present study was approved by the ethical committee of The University of Tokyo. Informed consent was waived because of the anonymous nature of the data.
Use of Argatroban for Acute Ischemic Stroke in Japan
Argatroban is recommended for use in patients with atherothrombotic stroke within 48 hours after onset, except in those with lacunar infarction. The Ministry of Health Labour, and Welfare of Japan approved argatroban in patients with stroke according to the following schedule: argatroban infused continuously at a dose of 60 mg/d for the first 2 days, followed by twice daily infusion, at a dose of 10 mg over 3 hours, for 5 days.4,5,11
Selection of Participants
The inclusion criteria were as follows: hospital admission with a main diagnosis of atherothrombotic stroke identified with the International Classification of Diseases-10th revision code of I633 between July 1, 2010, and March 31, 2012; admission within 1 day of stroke onset; and the receipt of oral antiplatelet agents during hospitalization. When patients were admitted to the same hospitals with ischemic stroke twice or more during the study period, only data from their first admission were used for subsequent analyses.
Patients who met at least one of the following criteria were excluded from the present study because of the possibility that their stroke subtype was not atherothrombotic: age <40 years; pregnancy; preexisting comorbidities of malignancy, or coagulopathy; and pre-existing atrial fibrillation or the receipt of oral anticoagulants including warfarin and dabigatran during hospitalization. With reference to a previous study conducted in North America,3 we also excluded patients with liver failure, those who received intravenous antihypertensive therapy or heparin on admission, and those who received thrombolysis with recombinant tissue-type plasminogen activator (r-tPA) or who underwent endovascular therapy during hospitalization. Patients were excluded if they were referred from another hospital. Finally, patients were excluded if they initially received argatroban at ≥1 day after hospital admission.
The included patients were then divided into 2 groups: those who received argatroban on admission (argatroban group), and those who did not receive argatroban during hospitalization (control group).
Data Collection and Processing
Data on demographics, preexisting comorbidities, the Japan Coma Scale, and mRS scores on admission, concomitant therapies, and several hospital characteristics were extracted for each patient. From the International Classification of Diseases-10th revision code, pre-existing comorbidities included hypertension, dyslipidemia, diabetes mellitus, heart failure, chronic obstructive pulmonary diseases, chronic renal failure, and previous stroke. Chronic renal failure was also identified by the receipt of hemodialysis within 4 days of hospital admission. Previous stroke was also identified by a history of previous admission to the same hospital with ischemic stroke. Coding for diseases is shown in Table I in the online-only Data Supplement. The following concomitant therapies were also identified: administration of ozagrel, edaravone, and hyperosmolar drugs on admission; use of an oral antiplatelet drug within 3 days of hospital admission; antacid use for ulcer prophylaxis within 3 days of hospital admission; and statin use within 7 days of hospital admission. Use of magnetic resonance imaging on admission was also identified. Characteristics of the healthcare facilities, including hospital type, availability of thrombolysis with r-tPA, and hospital volume were identified. Hospital volume was defined as the total number of patients with acute ischemic stroke admitted to each hospital during the study period. Stroke care unit admission was also determined. We identified whether patients were admitted on a weekend or a weekday, and whether admission occurred on the day of stroke onset or the day following onset.
Main outcomes were 7-day mortality, the occurrence of hemorrhagic complications during hospitalization, and mRS score at discharge. The occurrence of hemorrhagic complications was identified as follows: postadmission complications of hemorrhagic stroke or gastrointestinal bleeding; having a craniotomy for hematoma removal or gastrointestinal endoscopy; receiving carbazochrome or tranexamic acid, which is frequently used in patients with hemorrhage; and having a transfusion. Other outcomes included mechanical ventilation during hospitalization, length of hospital stay, and readmission with stroke within 90 days.
One-to-one propensity-score matching analysis was performed in the present study. This can minimize confounding in an observational study, which investigates the effects of treatment on outcomes. To estimate a patient’s propensity score for argatroban use, a binary multivariable logistic regression model was used with adjustment for possible confounders. The C-statistic was calculated to evaluate the goodness-of-fit of the model. Thereafter, a patient in the argatroban group was matched with a patient in the control group in accordance with the nearest neighbor rule within a caliper width of 0.2 of the SD of the propensity score.12
Standardized differences were used to evaluate the balance between the baseline variables. A standardized difference >10% was regarded as imbalanced. In the matched populations, continuous variables were compared between the argatroban and control groups with the Mann–Whitney U test, whereas categorical variables were compared with Fisher exact test or the χ2 test. In the matched population, an ordinal logistic regression analysis was also performed for mRS scores at discharge. This regression model was fitted with a generalized estimating equation to adjust for patient clustering within hospitals.
All tests of significance were 2 tailed, and P<0.05 was considered significant. All statistical analyses were performed using SPSS Statistics version 20.0 (IBM Corp, Armonk, NY).
The process of patient selection is illustrated in Figure 1. Finally, 20 980 patients were analyzed as prematched patients, 9552 patients in the argatroban group, and 11 428 in the control group.
In the prematched population, the following variables were imbalanced between the argatroban and the control groups: the presence of dyslipidemia; ambulance service use; magnetic resonance imaging on admission; edaravone and ozagrel use on admission; oral antiplatelet and histamine-2 receptor blocker use within 3 days after admission; statin use within 7 days after admission; hospital characteristics; and stroke care unit admission (Table 1). After one-to-one propensity-score matching, 2289 pairs of patients were included (Figure 1). In the matched population, the baseline characteristics were successfully balanced between the argatroban and the control groups (Table 1). The C-statistic for goodness-of-fit was 0.896.
In the matched populations, there were no significant differences between the argatroban and the control groups in 7-day mortality, the occurrence of hemorrhagic complications, the use of mechanical ventilation, length of hospital stay, and readmission for stroke recurrence within 90 days (Table 2). Details of the hemorrhagic complications during hospitalization are described in Table 3.
The mRS scores at discharge did not differ significantly between the argatroban and the control group in the χ2 test. Similarly, ordinal logistic regression analysis showed that argatroban use had no significant association with mRS at discharge (adjusted odds ratio, 1.01; 95% confidence interval, 0.88–1.16; Figure 2).
The present study, based on a nationwide database in Japan, analyzed 2289 pairs of patients with acute atherothrombotic stroke to compare outcomes between argatroban and control groups. The propensity-score matching successfully balanced the baseline characteristics and treatments between the 2 groups. The study did not show any added benefit of argatroban on outcomes in these patients.
One of the strengths in the present study was using propensity-score matching to balance patient backgrounds between the argatroban and the control groups to compensate for the retrospective nature of the study. In the prematched population, several tendencies were found in patients treated with argatroban. First, they were more likely to use the ambulance service, indicating that argatroban was more likely to be used in more severely ill patients. Argatroban patients were also more likely to receive histamine-2 receptor blockers, which are recommended for use in older patients and patients with severe stroke in the Japanese guidelines for treatment of acute ischemic stroke.2 This tendency would also indicate that argatroban was used for patients with severe stroke in Japan.
Second, oral antiplatelet drugs and edaravone were more likely to be used in the argatroban group than in the control group. These therapies are recommended for patients with acute ischemic stroke in the Japanese guidelines.2 Thus, patients may be more likely to receive argatroban when treated by physicians who comply with the guidelines.
Third, the majority of patients in the control group received ozagrel, whereas only 3% of patients in the argatroban group received ozagrel. In fact, both argatroban and ozagrel are recommended for use in patients with atherothrombotic stroke in the Japanese guidelines.2 Because the argatroban group seemed to include patients with more severe stroke than the control group, ozagrel may tend to be used in patients with milder stroke.
Finally, patients tended to receive argatroban when they were admitted to academic, higher-volume, or well-equipped hospitals. In fact, more patients in the argatroban group than in the control group were admitted to hospitals where r-tPA was available and were more likely to receive magnetic resonance imaging on admission. These tendencies could lead to confounding between argatroban use and outcomes. However, in the present study, these possible confounders were successfully adjusted between the argatroban and the control groups by using propensity-score matching analyses.
A previous randomized controlled trial with a small sample size demonstrated that argatroban use in acute ischemic stroke was safe,3 and this was consistent with the present study, where mortality and the occurrence of hemorrhagic complications did not differ significantly between the argatroban and the control groups. However, the previous study was too small to establish the efficacy of argatroban on outcomes. The present study analyzed 2289 pairs of patients and suggested that argatroban was not associated with more favorable outcomes at least according to its recommended use in Japan.2
In the present study, we evaluated the effect of argatroban monotherapy on stroke outcomes. On the contrary, the Argatroban tPA Stroke Study (ARTSS-1) evaluated the safety of argatroban use in combination with r-tPA in patients with ischemic stroke.13 The effectiveness or safety of the combination therapy could not be evaluated in the present study. A recently conducted randomized controlled trial on combination therapy (ARTSS-2) should provide further information on the efficacy and safety of combination therapy.
Several limitations should be acknowledged. First, because the present study used an administrative claims database, there may be a risk of miscoding. Second, several important factors could not be evaluated. There were no data on National Institutes of Health Stroke Scale scores and findings of diagnostic imaging tests. Even though the argatroban and control groups were balanced in many measured variables, unmeasured confounders may still exist. Third, there is a possibility that the study population included some patients with ischemic stroke with a subtype of stroke other than atherothrombotic. However, all patients included in the present study had a main diagnosis of atherothrombotic stroke, which was identified with the International Classification of Diseases-10th revision code of I633. Furthermore, we excluded patients who were treated with oral anticoagulant therapy or received heparin because they might have cardioembolic stroke. Nevertheless, we could not deny that some of the patients in the study population had a type of ischemic stroke other than atherothrombotic. Fourth, the differences in hospital length of stay may affect mRS at discharge between the argatroban and the control groups. However, the difference in the median length of stay was 2 days and was statistically insignificant. Fifth, long-term stroke outcomes could not be evaluated in the present study. We could only evaluate outcomes during hospitalization, and readmission within 90 days for stroke recurrence. Finally, about readmission for stroke recurrence, we could only identify patients who were readmitted to the same hospital at which they received their initial treatment.
In conclusion, the present study suggested that argatroban was safe for use in patients with acute atherothrombotic stroke. However, argatroban may not have an added benefit on outcomes in these patients.
Sources of Funding
This work was supported by Grants for Research on Policy Planning and Evaluation from the Ministry of Health, Labour, and Welfare, Japan (grant numbers: H27-Policy-Designated-009 and H27-Policy-Strategy-011). Drs Yasunaga, Horiguchi, and Fushimi received grants for Research on Policy Planning and Evaluation from the Ministry of Health, Labour, and Welfare, Japan.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.115.011250/-/DC1.
- Received August 26, 2015.
- Revision received November 3, 2015.
- Accepted November 16, 2015.
- © 2015 American Heart Association, Inc.
- Jauch EC,
- Saver JL,
- Adams HP Jr,
- Bruno A,
- Connors JJ,
- Demaerschalk BM,
- et al
- LaMonte MP,
- Nash ML,
- Wang DZ,
- Woolfenden AR,
- Schultz J,
- Hursting MJ,
- et al
- Lyden P,
- Pereira B,
- Chen B,
- Zhao L,
- Lamb J,
- Lei IF,
- et al
- Chen B,
- Friedman B,
- Whitney MA,
- Winkle JA,
- Lei IF,
- Olson ES,
- et al
- Shigematsu K,
- Nakano H,
- Watanabe Y.
- Fukuuchi Y,
- Tohgi H,
- Shinohara Y,
- Kobayashi S,
- Fujishima M,
- Tazaki Y,
- et al
- Barreto AD,
- Alexandrov AV,
- Lyden P,
- Lee J,
- Martin-Schild S,
- Shen L,
- et al