长期护理机构居民中有计划地停用降压药:随机证据的系统评价。
完整摘要 · 中文译文
背景:支持老年人降压治疗的证据主要来自与长期护理机构居民不同的人群,在这些居民中,衰弱、痴呆、多重疾病、多重用药以及相互竞争的临床优先事项可能改变预防获益与治疗负担之间的平衡。 目的:旨在综合关于在机构性长期护理中有计划地撤减降压药的随机证据,并确定临床上和方法学上可比的效果是否支持一个共同的疗效估计。 方法:对长期护理中有计划撤减降压药的随机研究或整群随机研究进行了系统评价。检索了MEDLINE/PubMed、Embase、CENTRAL、CINAHL Complete、Web of Science核心合集、试验注册库和PROSPERO,检索时间自建库起至2026年8月20日更新。仅对至少有两个临床上和方法学上兼容、可解释共同效果的估计值的结局进行荟萃分析;否则采用结构化的研究层面综合。采用修订版Cochrane偏倚风险工具(RoB 2)评估各结局的特异性偏倚风险,并采用推荐分级的评估、制定与评价(GRADE)工具评估证据确定性。 结果:共纳入三项随机长期护理研究,涉及1548名居民。其人群、撤减策略、监测、重启标准及随访存在实质性差异,未有任何结局被认为在临床上和方法学上具有足够的可比性以进行荟萃分析。RETREAT-FRAIL重新 报告了在全因死亡率为1.02(95%置信区间0.86-1.21)的调整后风险比(HR),中位潜在随访时间为38.4个月,而DANTON报告至32周时HR为1.65(0.95-2.85);Gulla/COSMOS没有经整群调整的比较死亡率估计。在DANTON第16周时,减处方使收缩压升高11.0 mmHg(5.6-16.4),舒张压升高4.6 mmHg(1.1-8.1),并减少降压药物数量1.2种(1.0-1.4)。关于住院、跌倒、心血管事件、认知、功能和生活质量的证据仍然稀少、具有结局特异性或存在偏倚风险。 结论:在长期护理中,关于降压药减处方的随机证据仍然稀少且临床异质性大。在个别试验中,有监督的减处方减少了降压药暴露并升高了血压,但其幅度和临床后果因策略和随访而异。当前证据未能确立对死亡率、住院或其他对住民重要的结局的共同效应,使得试验特异性估计和减处方方案比单一合并治疗效应更具信息价值。 PROSPERO注册号:CRD420261370802。
完整摘要 · 英文原文
BACKGROUND: Evidence supporting antihypertensive treatment in older adults is derived largely from populations that differ from long-term care residents, in whom frailty, dementia, multimorbidity, polypharmacy, and competing clinical priorities may alter the balance between preventive benefit and treatment burden. OBJECTIVES: The aim was to synthesize randomized evidence on planned antihypertensive deprescribing in institutional long-term care and determine whether clinically and methodologically comparable effects support a common treatment-effect estimate. METHODS: Randomized or cluster-randomized studies of planned antihypertensive deprescribing in long-term care were systematically reviewed. MEDLINE/PubMed, Embase, CENTRAL, CINAHL Complete, Web of Science Core Collection, trial registries, and PROSPERO were searched from inception and updated to 20 August 2026. Meta-analysis was reserved for outcomes with at least two clinically and methodologically compatible estimates of a common, interpretable effect; otherwise, structured study-level synthesis was used. Result-specific risk of bias was assessed with the revised Cochrane risk-of-bias tool (RoB 2), and certainty with the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) tool. RESULTS: Three randomized long-term care studies involving 1548 residents were included. Their populations, deprescribing strategies, monitoring, restart criteria, and follow-up differed substantially, and no outcome was considered sufficiently clinically and methodologically comparable for meta-analysis. RETREAT-FRAIL reported an adjusted hazard ratio (HR) for all-cause mortality of 1.02 (95% confidence interval 0.86-1.21) over a median potential follow-up of 38.4 months, whereas DANTON reported an HR of 1.65 (0.95-2.85) to 32 weeks; Gulla/COSMOS had no cluster-adjusted comparative mortality estimate. At 16 weeks in DANTON, deprescribing increased systolic blood pressure by 11.0 mmHg (5.6-16.4), increased diastolic blood pressure by 4.6 mmHg (1.1-8.1), and reduced the number of antihypertensive drugs by 1.2 (1.0-1.4). Evidence for hospitalization, falls, cardiovascular events, cognition, function, and quality of life remained sparse, outcome-specific, or at risk of bias. CONCLUSIONS: Randomized evidence for antihypertensive deprescribing in long-term care remains sparse and clinically heterogeneous. Within individual trials, supervised deprescribing reduced antihypertensive exposure and raised blood pressure, but the magnitude and clinical consequences varied by strategy and follow-up. Current evidence does not establish a common effect on mortality, hospitalization, or other resident-important outcomes, making trial-specific estimates and deprescribing protocols more informative than a single pooled treatment effect. PROSPERO REGISTRATION: CRD420261370802.