Considering prehospital stroke trials: did RIGHT-2 get it right?
Considering prehospital stroke trials: did RIGHT-2 get it right?
复制标题
考虑院前卒中试验：RIGHT-2 做对了吗？
DOI：
10.1016/s0140-6736(19)30276-4
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Durkalski-Mauldin,ValerieL
中科院分区：
文献类型：
--
作者：
Johnston,KarenC;Durkalski-Mauldin,ValerieL
Comment964 www. thelancet. com Vol 393 March 9, 2019 p= 0· 083) or in cohort 2 (all patients—ie, intention to treat [ITT]; 1· 04 [0· 84–1· 29]; p= 0· 69). No difference was found between the two groups in deaths at day 4 in either cohort (cohort 1: acOR 1· 17 [0· 57–2· 39]; p= 0· 68; cohort 1: 1· 19 [0· 60–2· 35]; p= 0· 63). Additional analyses suggested potential harm, with worse outcomes in patients with intracerebral haemorrhage (p= 0· 057), the earliest enrolled (< 1 h) patients (p= 0· 014), and the most severe strokes (National Institutes of Health Stroke Scale> 12; 0· 044). We believe the clinical community now has data from numerous populations suggesting that blood pressure lowering in the acute and hyperacute setting is not beneficial to patients with stroke. Although the intervention did not improve clinical outcome, prehospital treatment with GTN used 184 ambulance stations in eight ambulances services in England and Wales as the trial sites, and 516 paramedics as the front-line recruiting investigators. This trial6 is the first in the UK to our knowledge to show the ability of the emergency medical service teams to successfully enrol in a hyperacute stroke clinical trial in the field. The FAST-MAG trial7 in the USA has previously shown such a model, with the enrolment of 1700 patients in a phase 3 trial assessing magnesium (4 g intravenous magnesium sulphate) in acute stroke in the Los Angeles area. Such trials are changing the paradigm of acute stroke research by showing the feasibility of very early treatment initiation within a median of 45 min for FAST-MAG7 and 73 min for RIGHT-2. 6The limitations of a prehospital enrolment model are also shown in RIGHT-2. This trial had an original sample size of 850 patients (425 per group), which provided 90% power to detect an ordinal shift in the mRS. This sample size took into account a 3% loss to follow-up and a 20% stroke mimic rate. During the trial, the non-stroke diagnosis rate exceeded 30% in patients who were randomly assigned, therefore reducing the power to assess the efficacy of the intervention in the target population. Thus, the sample size was increased to 1050 to maintain statistical power at the predefined clinically relevant difference. Additionally, a decision was made by the trial steering committee to specify a hierarchical analysis with the first analysis of patients with stroke or transient ischaemic attack (cohort 1) and the second analysis of all patients (the ITT population; cohort 2), which presumably was the primary analysis population initially. The
