2020 Volume 2 Issue 4 Pages 107-112
BACKGROUND
Cerebral vasospasm (CVS) occurs in 17% to 40% of patients with aneurysmal subarachnoid hemorrhage (aSAH). Various measures have been implemented to prevent CVS, including fasudil hydrochloride. However, data are lacking on the preventive effect of fasudil hydrochloride on CVS. The present study was performed to examine the association between postprocedural early administration of fasudil hydrochloride and outcomes in patients undergoing treatments for aSAH using a national inpatient database in Japan.
METHODS
Patients aged ≥18 years who were admitted because of aSAH from July 2010 to March 2014 and underwent clipping surgery or intravascular coil embolization within 72 hours from admission were eligible. We defined preventive fasudil hydrochloride as that administered within 1 day after the intervention. The outcomes were 30-day in-hospital mortality and the modified Rankin scale (mRS) score at discharge. We used instrumental variable methods to analyze the differences in the risk of these outcomes between patients with and without preventive fasudil hydrochloride.
RESULTS
Of 23,843 eligible patients, 77.1% received preventive fasudil hydrochloride. The overall 30-day in-hospital mortality rate was 2.2%, and the proportion of a good neurological outcome (mRS score of ≤2) was 65.0%. Instrumental variable analyses showed no significant difference in 30-day in-hospital mortality (risk difference, −1.8%; 95% confidence interval, −3.8% to 2.0%), but demonstrated a significant difference in an mRS score of ≤2 at discharge (risk difference, 7.8%; 95% confidence interval, 3.4% to 12.3%).
CONCLUSION
Early administration of fasudil hydrochloride after treatment for aSAH could be associated with a better neurological outcome at discharge.
Aneurysmal subarachnoid hemorrhage (aSAH) is a life-threatening disease with a high mortality rate [1]. A previous study showed that about half of survivors of aSAH experience long-term neuropsychological impairment resulting in a deteriorated quality of life [2]. Population-based studies have shown that the incidence rates vary from 4 to 23 per 100,000 population [3]. Because aSAH develops even in young people and its incidence is higher in Japan than in other countries [3, 4], high-quality neurointensive care is essential for a good prognosis. Treatment of aSAH comprises three major parts: peri-ictal medical management, surgical intervention including endovascular clipping, and adequate postoperative management [5].
Cerebral vasospasm (CVS) is a lethal complication of aSAH. It occurs by about 3 weeks after the onset of aSAH and affects 70% of patients [5]. CVS reduces blood flow to the brain parenchyma, which causes irreversible cerebral ischemia and leads to neurologic complications in 17% to 40% of patients with CVS [6]. Therefore, prevention of CVS is important for improving patients’ outcomes.
Nimodipine, a calcium channel blocker, is the first-line treatment to prevent CVS [7]; however, it has not been approved in Japan. Instead, fasudil hydrochloride is frequently used for CVS. Fasudil hydrochloride, a selective Rho kinase inhibitor, was approved for prevention and treatment of CVS after aSAH in Japan in 1995. The Japanese Guidelines for the Management of Stroke recommend early administration of fasudil hydrochloride to prevent CVS after aSAH and have designated this a grade B recommendation based on the results of several clinical trials and one meta-analysis [8–10]. However, these clinical trials were based on limited numbers of samples. Because of limited data, treatments are not unified and variations in the use of this drug exist among facilities. Thus, a nationwide study to verify the effect of fasudil hydrochloride is required.
We therefore conducted this retrospective observational study using a Japanese national inpatient database to evaluate the preventive effect of fasudil hydrochloride on outcomes related to CVS after treatment for aSAH.
We used the Diagnostic Procedure Combination inpatient database in Japan. Details of the database have been described elsewhere [11]. Briefly, the database includes administrative claim and discharge data and covers approximately 92% of all tertiary-care emergency hospitals in Japan. The database includes data on patients’ age and sex; diagnoses, comorbidities at admission, and complications after admission, recorded in accordance with the International Classification of Diseases, 10th revision (ICD-10) codes and text data in Japanese; dates of using each drug; dates of performing surgical and nonsurgical procedures; the modified Rankin scale (mRS) score at admission and discharge; and the Japan Coma Scale (JCS) score at admission.
The JCS is a scale for impaired consciousness that is widely used in Japan. A score of 0 indicates alert consciousness, single-digit scores of 1 to 3 indicate mildly impaired consciousness with spontaneous eye-opening, double-digit scores of 10 to 30 indicate moderately impaired consciousness with eye-opening when patients receive stimuli, and triple-digit scores of 100 to 300 indicate coma. JCS is well-correlated with the Glasgow Coma Scale [12].
This study was based on a secondary analysis of de-identified data. The requirement for informed consent was waived because of the anonymous nature of the data. Study approval was obtained from the institutional review board of The University of Tokyo.
PATIENT SELECTIONFor this study, we identified patients aged ≥18 years who were admitted to the hospital with a diagnosis of aSAH (ICD-10 codes I60.0, I60.1, I60.2, I60.3, I60.4, and I60.5) and underwent surgical clipping or endovascular coiling within 72 hours from admission from July 2010 to March 2014.
We defined preventive fasudil hydrochloride as that administered within 1 day after surgical clipping or endovascular coiling. We assumed that starting administration of fasudil hydrochloride following a 1-day lapse after surgical clipping or endovascular coiling was not intended for prevention but for therapeutic use. Fasudil hydrochloride administration is approved at a dose of 30 mg two to three times a day for 14 days [9].
We excluded patients with missing mRS scores and those who died within 3 days after treatment to avoid immortal time bias. We also excluded patients with coma at admission because surgical clipping or endovascular coiling is not recommended for such patients in the Japanese Guidelines for the Management of Stroke, and the “preventive” effect of fasudil hydrochloride for patients with coma on arrival cannot be measured.
We extracted data on intravenous administration of antihypertensive drugs (diltiazem and nicardipine) on the day of admission as an indicator of high blood pressure. We also extracted the following data for each patient: age, sex, ambulance use, admission to a stroke care unit (SCU) or intensive care unit (ICU), catecholamine use, and mechanical ventilation before the intervention.
The primary outcome was 30-day in-hospital mortality, and the secondary outcome was the mRS score at discharge. We defined functional independence as an mRS score of ≤2.
STATISTICAL ANALYSISCategorical data were compared using the chi-squared test, and continuous data were compared using the Mann–Whitney U test. We performed a multivariable binary logistic regression analysis for 30-day in-hospital mortality and the mRS scores at discharge with adjustment for patient backgrounds (age, JCS score, mRS score at admission, use of blood pressure control agents, intubation, SCU or ICU admission, clipping or coiling, ambulance use, admission to an academic hospital, and duration from admission to intervention) while also adjusting for within hospital clustering using a generalized estimating equation. Multivariable regression analyses cannot remove hidden biases caused by unmeasured confounders; therefore, we performed instrumental variable analyses as confirmatory analyses. Instrumental variable analyses assume that instrumental variables are strongly associated with the treatment assignment, are not associated with any measured or unmeasured variables in patient characteristics, and do not affect outcomes except through treatment [13]. In the present study, the hospital’s preference regarding treatment with fasudil hydrochloride was used as an instrumental variable. If a hospital showed strong consistency in how fasudil was used for the prevention of CVS, the decision to administer a drug was assumed to have been made independent of individual patients’ characteristics. In this situation, the hospital’s preference acted as an instrumental variable. The hospital’s preference for fasudil hydrochloride was defined as the number of patients who received fasudil hydrochloride divided by the number of all patients with aSAH at each hospital. We divided the hospitals into those with a hospital preference at the ≥75th and <75th percentile of all hospitals and calculated the risk differences and their 95% confidence intervals (CIs). We estimated the risk differences and their 95% CIs in 30-day in-hospital mortality and an mRS score of ≤2 at discharge between the groups with and without early administration of fasudil hydrochloride using the two-stage least-square method with adjustment for patient backgrounds (same as those in the logistic regression analyses). For these analyses, we used the ivreg2 procedure of Stata/SE 13.0 (StataCorp, College Station, TX, USA). A partial F test was performed to confirm that the hospital’s preference for fasudil hydrochloride was not a weak instrument [14]. An F-statistic of >10 indicates that the instrument is not weak.
The threshold for significance was p < 0.05. IBM SPSS version 22 (IBM, Armonk, NY, USA) and Stata/SE 13.0 were used for all statistical analyses.
Of 33,973 patients aged ≥18 years with aSAH who underwent clipping or coiling, we excluded 10,130 patients (1,386 with missing data, 493 who died within 3 days after clipping or coiling, and 8,251 with coma at admission). The remaining 23,843 patients were eligible for this study. These patients were divided into the preventive fasudil hydrochloride group (n = 18,390) and the control group (n = 5,453).
The baseline characteristics of each group are shown in Table 1. Patients with preventive fasudil hydrochloride were more likely to use an ambulance, have hypertension, need intubation, undergo clipping, and be admitted to an ICU or SCU.
| Variables | Fasudil group (n = 18,390) | Control group (n = 5,453) | p value |
|---|---|---|---|
| Sex, male | 5,728 (31.1) | 1,709 (31.3) | 0.79 |
| Age, years | |||
| <65 | 10,221 (55.6) | 2,613 (53.4) | 0.02 |
| 65–74 | 4,263 (23.2) | 1,316 (24.1) | |
| ≥75 | 3,906 (21.2) | 1,224 (22.4) | |
| JCS score at admission | <0.01 | ||
| 0 | 5,433 (29.5) | 1,839 (33.7) | |
| 1–3 | 7,281(39.6) | 1,982 (36.3) | |
| 10–30 | 5,676 (30.9) | 5,453 (29.9) | |
| mRS score at admission | |||
| 0 | 6,825 (37.1) | 1,673 (30.7) | <0.01 |
| 1 | 2,306 (12.5) | 770 (14.1) | |
| 2 | 1,559 (8.5) | 557 (10.2) | |
| 3 | 1,492 (8.1) | 463 (8.5) | |
| 4 | 2,721 (14.8) | 797 (14.6) | |
| 5 | 3,487 (19.0) | 1,193 (21.9) | |
| Urgent use of antihypertensive drugs | 15,696 (85.4) | 2,766 (50.7) | <0.01 |
| Mechanical ventilation before intervention | 755 (4.1) | 114 (2.1) | <0.01 |
| ICU or SCU admission | 5,047 (27.4) | 827 (15.2) | <0.01 |
| Use of catecholamine | 74 (0.4) | 21 (0.4) | 0.86 |
| Type of surgery, clipping | 13,707 (74.5) | 3,694 (67.7) | <0.01 |
| Ambulance use | 13,904 (75.6) | 3,693 (67.7) | <0.01 |
| Admission to academic hospital | 3,705 (20.1) | 970 (17.8) | <0.01 |
| Duration from admission to intervention, days | 1 (1–2) | 1 (1–2) | <0.01 |
Data are expressed as n (%) or median (interquartile range). Percentages may not equal 100 because of rounding.
Abbreviations: JCS, Japan Coma Scale; mRS, modified Rankin scale; ICU, intensive care unit; SCU, stroke care unit.
The crude outcomes in each group are shown in Table 2. In the multivariable regression analyses, the use of fasudil hydrochloride was significantly associated with lower 30-day in-hospital mortality and a higher mRS score at discharge than those in the control group (Table 3).
| Variables | Fasudil group (n = 18,390) | Control group (n = 5,453) |
|---|---|---|
| 30-day in-hospital mortality | 534 (2.9) | 338 (6.2) |
| mRS score of ≤2 at discharge | 12,216 (66.4) | 3,282 (60.2) |
Data are presented as n (%).
Abbreviation: mRS, modified Rankin scale.
| Variables | Adjusted odds ratio | 95% CI | p value |
|---|---|---|---|
| 30-day in-hospital mortality | |||
| Fasudil group (vs. control group) | 0.38 | 0.32–0.45 | <0.01 |
| mRS score of ≤2 at discharge | |||
| Fasudil group (vs. control group) | 1.45 | 1.33–1.58 | <0.01 |
Abbreviations: CI, confidence interval; mRS, modified Rankin scale.
Adjusted for sex, age, Japan Coma Scale score at admission, mRS score at admission, urgent use of antihypertensive drugs, mechanical ventilation before intervention, admission to intensive care unit or stroke care unit, use of catecholamine, type of surgery (clipping or coiling), ambulance use, admission to an academic hospital, and duration from admission to intervention.
In the instrumental variable model, the F-statistic was 2,702 (p < 0.001). Preventive fasudil hydrochloride was not significantly associated with a reduction in 30-day in-hospital mortality, but it was significantly associated with a higher proportion of an mRS score of ≤2 at discharge (Table 4).
| Variables | Risk difference | 95% CI |
|---|---|---|
| 30-day in-hospital mortality | ||
| Fasudil group (vs. control group) | −1.8% | −3.8% to 2.0% |
| mRS score of ≤2 at discharge | ||
| Fasudil group (vs. control group) | 7.8% | 3.4% to 12.3% |
Abbreviations: CI, confidence interval; mRS, modified Rankin scale.
Adjusted for sex, age, Japan Coma Scale score at admission, mRS score at admission, urgent use of antihypertensive drugs, mechanical ventilation before intervention, admission to intensive care unit or stroke care unit, use of catecholamine, type of surgery (clipping or coiling), ambulance use, admission to an academic hospital, and duration from admission to intervention.
In this nationwide retrospective multicenter study of 23,843 patients with aSAH, we found a significant association between preventive fasudil hydrochloride use and a better mRS score at discharge in both the instrumental variable analysis and multivariable regression analysis, while there was a significant association between preventive fasudil hydrochloride use and 30-day in-hospital mortality in the multivariable regression analysis alone.
In agreement with our findings, the association of fasudil hydrochloride with good outcomes has been reported in other study settings and populations. For example, one small randomized controlled trial demonstrated that the effect of fasudil hydrochloride for CVS was similar to that of nimodipine [15]. However, the study included only 72 patients. In a prior meta-analysis, fasudil hydrochloride reduced the occurrence of CVS and cerebral infarction in patients with aSAH; however, the total number of patients included in this meta-analysis was only 772 [8]. The current study corroborates these earlier studies and extends them by demonstrating the association between preventive fasudil hydrochloride use and a better mRS score at discharge using nationwide retrospective multicenter data and instrumental variable analyses to control for measured and unmeasured confounding factors.
Our results showed that the difference in 30-day in-hospital mortality was not significant between the fasudil hydrochloride and control groups in the instrumental variable analysis but that the difference was significant in the multivariable regression analysis. A discrepancy was found between the results of the instrumental variable analysis and multivariable regression analysis. The multivariable regression analysis failed to adjust for many unmeasured confounding factors, including the severity of aSAH. In addition, factors related to mortality are more likely to involve the severity of bleeding itself and early intensive care rather than postoperative management. This is why the instrumental variable analysis did not show a significant difference in mortality. Conversely, the mRS score at discharge in the fasudil hydrochloride group was significantly better than that in the control group. Delayed cerebral ischemia due to CVS may have influenced the patients’ functional outcomes because it may have deteriorated the patients’ physical activity and led to cerebral edema that worsened the functional outcome. Fasudil hydrochloride has been regarded as a potential vasodilator for the treatment of vasospasm [16, 17]. This drug has wide-ranging effects via its inhibition of several protein kinases [9]. It directly halts smooth muscle contraction and prevents free radical formation. Thus, it could have effects of vasodilation and neuroprotection [18–20]. These effects may minimize brain damage, resulting in a better neurological outcome.
Several limitations of our study should be acknowledged. First, no data were available after discharge, and we only assessed short-term outcomes. However, the literature has indicated that CVS occurs by about 3 weeks from the onset of aSAH, and this period is sufficient to verify the preventive effect. Second, detailed clinical data were not available, including computed tomography findings and vital signs. The database does not include important scales such as the Hunt–Hess scale or Fisher score. Instead, we used the JCS score at admission for our analyses. In fact, previous studies have suggested that the JCS score at admission is a reliable predictor of stroke outcomes [21]. Third, one of the assumptions for instrumental variable analysis is that the instrumental variable is indirectly associated with the outcome only through the treatment; it does not have a direct effect on the outcome. However, it is impossible to prove non-existence of instrumental variable outcome confounders. For example, the quality of care at hospitals might be an instrumental variable outcome confounder. Finally, although the study population was large, the study was conducted retrospectively in an observational manner without randomization. To overcome bias of unmeasured confounding factors, we performed instrumental variable analyses.
Preventive administration of fasudil hydrochloride was not significantly associated with a reduction in 30-day in-hospital mortality; however, it was significantly associated with a better neurological outcome at discharge in patients who underwent clipping or coiling for aSAH.
None.
This work was supported by Grants for Research on Policy Planning and Evaluation from the Ministry of Health, Labour and Welfare, Japan (Grant Numbers: H30-Policy-Designated-004 and H29-Policy-General-004).
