Supplementary Material for: Use of Antidepressants and Risk of Incident Stroke: A Systematic Review and Meta-Analysis

Background: Both depression and use of antidepressants have been reported to be risk factors for stroke, but results from the literature are still not conclusive regarding the risk attributable to antidepressants rather than to the underlying disease. Objective: To estimate the risk of incident stroke associated with use of antidepressants, a meta-analysis was performed. Methods: PubMed, Medline, Cochrane, ProQuest, Scopus, and bibliographies of articles were searched up to September 2018. The final meta-analysis included 31 observational studies. STROBE statement-checklist and GRADE guidelines were used for quality assessment. Results: The random-effects meta-analysis on the association between use of any antidepressant and risk of any stroke resulted in meta-risk ratio (RR) of 1.41 (95% CI 1.13–1.69, I2 = 93, 7%). The pooled estimate for selective serotonin reuptake inhibitors (SSRIs) resulted in a meta-RR of 1.41 (95% CI 1.13–1.69, I2 = 94, 5%) and for tricyclic antidepressants (TCAs) of 1.08 (95% CI 0.93–1.22, I2 = 0%). SSRI users displayed a higher risk of ischemic (1.57, 95% CI 1.06–2.09, I2 = 96.4%) than hemorrhagic stroke (1.34, 95% CI 1.15–1.53, I2 = 72.9%). Meta-RRs were lower for TCA, although with smaller heterogeneity (ischemic 1.22, 95% CI 0.97–1.46; I2 = 0%; hemorrhagic: 1.00, 95% CI 0.83–1.18, I2 = 0%). Restricting to studies on depressed individuals, both SSRI and TCA remained associated with an increased risk of any stroke type (meta-RR for SSRI: 1.27, 95% CI 1.11–1.43, I2 = 76.6%; meta-RR for TCA: 1.21 (95% CI 1.02–1.40, I2 = 47, 3%). Conclusions: Despite the high heterogeneity, these results demonstrate that even after adjusting for depression, use of antidepressants retains an independent increased risk of stroke.

背景：已有研究显示抑郁症与抗抑郁药物使用均为脑卒中的危险因素，但现有文献针对抗抑郁药物本身而非基础疾病所导致的卒中风险，尚未得出明确结论。目的：为评估抗抑郁药物使用与新发脑卒中的关联风险，本研究开展了一项荟萃分析（meta-analysis）。方法：检索截至2018年9月的PubMed、Medline、Cochrane、ProQuest、Scopus数据库及相关文献的参考文献列表，最终纳入31项观察性研究进行荟萃分析。采用STROBE声明核查清单与GRADE指南开展质量评价。结果：针对任意抗抑郁药物使用与任意类型脑卒中风险关联的随机效应荟萃分析显示，合并相对风险（risk ratio, RR）为1.41（95%置信区间confidence interval, CI：1.13~1.69，I²=93.7%）。针对选择性5-羟色胺再摄取抑制剂（selective serotonin reuptake inhibitors, SSRIs）的合并估计值显示，合并RR为1.41（95%CI：1.13~1.69，I²=94.5%）；针对三环类抗抑郁药（tricyclic antidepressants, TCAs）的合并RR为1.08（95%CI：0.93~1.22，I²=0%）。使用选择性5-羟色胺再摄取抑制剂的人群，其缺血性脑卒中风险（1.57，95%CI：1.06~2.09，I²=96.4%）高于出血性脑卒中风险（1.34，95%CI：1.15~1.53，I²=72.9%）。三环类抗抑郁药对应的合并RR相对更低，且异质性更小：缺血性脑卒中为1.22（95%CI：0.97~1.46，I²=0%）；出血性脑卒中为1.00（95%CI：0.83~1.18，I²=0%）。在仅针对抑郁症人群的亚组分析中，选择性5-羟色胺再摄取抑制剂与三环类抗抑郁药均与任意类型脑卒中的风险升高相关（选择性5-羟色胺再摄取抑制剂的合并RR：1.27，95%CI：1.11~1.43，I²=76.6%；三环类抗抑郁药的合并RR：1.21，95%CI：1.02~1.40，I²=47.3%）。结论：尽管存在较高的异质性，本研究结果表明，即便在校正抑郁症混杂因素后，抗抑郁药物的使用仍会独立增加脑卒中的发病风险。