Remote ischaemic conditioning: individual patient data meta-analysis of acute stroke trials and prospective MRI evaluation in healthy adults

Background: Remote ischaemic conditioning (RIC) is a promising neuroprotective method, with preclinical studies showing improved neurological outcome following acute stroke. However, early phase randomised controlled trials (RCTs) have given rise to conflicting results. Furthermore, the optimal strategy to apply RIC in humans is unknown. Clinical studies have utilised a variety of doses and methods, ranging from a single ‘dose’ using one limb, to daily application for 300 days using two limbs, decisions that are not based on meaningful pre-clinical data. A reliable way to directly measure end-organ effects of RIC would allow optimal methods of application and ‘dose’ to be determined, thereby informing clinical trial design. Aims: To prospectively pool and analyse individual patient data (IPD) from RCTs on the effects of RIC on safety and outcomes in acute stroke. The aim was to also test a small group of healthy young adults to establish the feasibility and tolerability of administering RIC during a non-invasive magnetic resonance imaging (MRI) scan, and to collect preliminary data on using Phase-contrast (PC) and Arterial Spin Labelling (ASL) for detecting RIC-induced cerebral haemodynamic changes. Methods: For the individual patient data meta-analysis, we systematically searched electronic databases up to December 2020, including PubMed/MEDLINE, EMBASE, Cochrane, and clinical trial registries using pre-specified keywords. RCTs were included when RIC (intervention) and sham therapy (control) were administered within 24 hours of stroke. We extracted individual patient data obtained from the invited trial investigators. The primary functional outcome (mRS) was assessed by ordinal analysis at the end of the trials. The secondary outcomes included early and late neurological deterioration based on the National Institute of Health Stroke Severity (NIHSS) scores, adverse vascular events, death at 90 days, and changes in infarct volume, NIHSS scores (between baseline and end-of-trial) and serum biomarkers. Unadjusted and multivariable regression analysis adjusted for age, sex, baseline NIHSS and systolic blood pressure, and time-to-treatment were conducted. For the healthy human volunteer feasibility study, we recruited 6 young (18-40 years) healthy males to undergo a single ‘dose’ of 4 cycles of intermittent arm ischaemia - alternating 5 minutes inflation (200 mmHg) followed by 5 minutes deflation of an upper arm blood pressure cuff during an MRI scan. ASL & Phase contrast MRI of the brain were performed before and after RIC, as well as 24 hours later. Measurements of blood flow and perfusion were compared to baseline pre-RIC values. Results: Seven RCTs from four countries comprising 556 patients (281 RIC, 275 sham) were included: age 66.3years (SD 13.9), 61% male, NIHSS 7 (IQR 5-12) and 43% randomised ≤3 hours. Final NIHSS scores significantly improved following RIC therapy (OR=-0.85, 95%CI -1.55 to -0.16, p=0.01), but there was no statistical difference in the mRS scores (OR=0.97, 95%CI 0.71-1.31,p=0.83), overall major cardiovascular adverse events (OR=0.76, 95%CI 0.42-1.39,p=0.38), or mortality (OR=1.44, 95%CI 0.69-2.98,p=0.33). A possible treatment interaction (p=0.08) occurred with time-to-treatment (mRS: ≤3 hours, OR 0.71, 95%CI 0.44-1.16; >3 hours, OR 1.23, 95%CI 0.82-1.85). Study heterogeneity and risk of bias (except one study) were low. In the healthy human volunteer pilot study, RIC administration during the MRI scan appeared feasible, safe, and was well tolerated by the study participants. There were no significant changes observed in the blood flow velocity of the major intracranial vessels (bilateral internal carotid arteries and basilar artery). Although ASL analysis was planned in this cohort, due to extenuating circumstances caused by the COVID-19 pandemic, no formal analysis could be performed by the time of the thesis submission. Conclusions: RIC appeared safe and significantly improved the final NIHSS, but did not improve the functional outcome (mRS) in acute stroke. RIC therapy may be beneficial when administered ≤3 hours from stroke onset. The preliminary findings of the human volunteer pilot study showed RIC administration during MRI scanning appeared to be feasible, safe, and well tolerated. It also allowed quantitative haemodynamic measurements during RIC, though no significant change in blood flow velocity was demonstrated in this small study sample. Future work could exploit this method further, in a larger sample size, and potentially assess the treatment ‘dose’ and administration parameters in target populations