The best magazine
Methods
Institutional Review Board Approval and Patient Consent
This study was approved by the Seoul National University Bundang Hospital Istitutional Review Board (SNUBH IRB) with waiver of informed consent because of its minimal risk and retrospective nature.
Study Participants
Based on a prospective stroke registry database, a consecutive series of patients 1) who were admitted to Seoul National University Bundang Hospital for acute ischemic stroke within 6 h of symptom onset between March 2004 and September 2011, 2) who underwent recanalization therapy and 3) who had symptomatic stenosis (>50 %) or occlusion of a major cerebral artery at initial angiographic evaluation were identified. According to the institutional stroke image protocols, patients who were potentially eligible for recanalization therapy underwent computed tomographic angiography (CTA) or magnetic resonance angiography (MRA) at presentation. In this study, the cerebral arteries of interest were the internal carotid artery (ICA), middle cerebral artery (MCA), anterior cerebral artery (ACA), posterior cerebral artery (PCA), basilar artery (BA), and vertebral artery (VA). Patients without a 3-month modified Rankin Scale (mRS) score were excluded.
Data Collection and Outcome Measures
Review of electronic medical records and the registry database provided clinical information on demographic factors, baseline stroke severity as measured by the National Institute of Heath Stroke Scale (NIHSS), vascular risk factors, and acute stroke management. Recanalization modalities were classified into the following types: IV thrombolysis (IV-only), intra-arterial treatment (IA-only), and combination of IV and IA treatments. IA treatment included IA use of chemical thrombolytic agents, clot maceration by multiple passages of microcatheter/microwire, and mechanical thrombectomy using devices such as the Penumbra system and the Solitaire. Stroke subtype was assigned by a vascular neurologist according to the Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria and validated at a weekly stroke registry meeting by consensus.
Premorbid statin use, defined as receiving statin therapy within 1 week before stroke onset, was ascertained by patient or proxy interview. Timing of statin administration after recanalization therapy, specific statin type, and initial dose were ascertained by review of the electronic medical records. Commencement of statin therapy was categorized according to the following scheme: statin administration on the first (D1), second (D2), and third day or later (D ≥ 3) of hospitalization, and no statin therapy (NU). Since statin was used in several forms and doses, we substituted a specific dose of a specific form of statin with an equivalent dose of atorvastatin (10 mg or less, 20 mg, 40 mg and 80 mg). The atorvastatin equivalent doses were then categorized into no use, low (less than atorvastatin equivalent dose of 40 mg) and high dose (40 mg or more).
After IRB approval, we prospectively collected NIHSS scores at baseline and on the second and seventh day of hospitalization or at discharge, as well as the 3-month mRS scores, as part of an institutional quality-of-care monitoring program for hospitalized stroke patients. The primary efficacy outcome of the study was good functional outcome at 3 months (mRS score of 0–1). A favorable outcome, defined as a 3-month mRS score of 0–2, was a secondary outcome.
Other secondary outcomes included neurologic improvement, neurologic deterioration, and ischemic stroke recurrence during hospitalization. Neurologic improvement was defined as a reduction in total NIHSS score of ≥4 points from baseline to discharge or a NIHSS score of 0–1 at discharge. Neurologic deterioration was defined as an increase in a total NIHSS score of 4 or more points from baseline. Ischemic stroke recurrence was defined as a significant change in neurologic symptoms and signs accompanied by corresponding new discrete lesions on diffusion-weighted magnetic resonance images. Symptomatic hemorrhagic transformation was included as a safety outcome and was defined as a local or remote parenchymal hemorrhage type 2 on a post-treatment brain image combined with an increase of 4 points or more in the NIHSS score from baseline or with the occurrence of death.
Statistical Analysis
Study variables are expressed as mean ± SD, median (interquartile range, IQR), or number of patients (percentage) according to variable characteristics. Statin therapy was characterized according to the commencement of statin therapy, type, and dose (Additional file 1: Figure S1, S2 and Table S1 http://www.biomedcentral.com/1471-2377/15/122/additional). Baseline characteristics were compared according to the statin starting time using the Pearson χ test, ANOVA, and Kruskal-Wallis test when appropriate (Table 1).
Dose-response relationships between the statin starting time and the primary and secondary outcomes were evaluated using the Mantel-Haenszel test for trend (Table 2). Regarding multivariable analysis, adjusted odds ratios (ORs) of the statin starting time (D1, D2 and D ≥ 3) compared to NU were estimated for various outcomes, and dose-response relationships were characterized by likelihood ratio tests for trend (Fig. 1). In cases where the event number of D2 or D ≥ 3 was less than 5, the statin starting time was re-categorized into three groups: D1, D ≥ 2 and NU.
(Enlarge Image)
Figure 1.
Adjusted odds ratios of the statin starting time with respect to various clinical outcomes. Statin starting time was defined as starting statin therapy at D1, D2, D ≥ 3, or no use, or as starting statin therapy at D1, D ≥ 2, and no use when event number of D2 or D ≥ 3 was less than 5. The adjusted odds ratios (circle) and 95 % confidence intervals (solid line) were estimated using multiple logistic regression models with adjustments for premorbid statin use, stroke history, atrial fibrillation, calendar year, stroke subtype, baseline NIHSS score, and recanalization modality. *Ps were calculated by log likelihood test for trend. mRS is the abbreviation for modified Rankin Score
Variables for adjustment were selected based on their p values (<0.2) and biological plausibility. To mitigate unmeasured confounding by recent advances in stroke care, calendar year was included in multivariable models (Additional file 1: Table S2 http://www.biomedcentral.com/1471-2377/15/122/additional). Because the number of patients with symptomatic hemorrhagic transformation was not enough to adjust all potential confounders at once, multivariable analyses were carried out using various sets of confounders as post-hoc analyses (Additional file 1: Table S3 http://www.biomedcentral.com/1471-2377/15/122/additional).
To investigate the heterogeneity of the early statin effect according to stroke subtype and recanalization modality, we performed subgroup analyses (cardioembolic vs. non-cardioembolic stroke and IV-only vs. IA-only vs. combined treatment) (Table 3 and Table 4). The statistical significance of the interaction between the statin starting time and stroke subtype or recanalization modality was examined in multivariable models. Finally, the associations of the statin dose with 3-month mRS 0–1 and symptomatic hemorrhagic transformation were analyzed (Additional file 1: Table S4 and S5 http://www.biomedcentral.com/1471-2377/15/122/additional). All statistical analyses were performed with SPSS version 18.0 (SPSS Inc., Chicago, IL). A two-sided p value of 0.05 was generally considered a minimum level of statistical significance.
Source: ...