. 2020 Oct 14;10(10):e040466. doi: 10.1136/bmjopen-2020-040466

Table 2.

Secondary end points for the trial and the statistical test to be used for each

Secondary safety/tolerability end points	Statistical test of choice
1. Discontinuation prior to 30 days	Fisher’s exact test
2. Proportion completing the run-in period and entering the randomisation phase	Fisher’s exact test
3. Physical examination signs of tissue or neurovascular injury resulting from RIC treatment at 30 days	Fisher’s exact test
4. Development of symptomatic upper extremity deep vein thrombosis at 30 days and 90 days	Fisher’s exact test
5. Peak and end-cycle pain levels reported by the participant using the Visual Analogue Scale during the 30-day treatment period	Repeated measures analysis with linear mixed models will be used to estimate the mean VAS per session, using all VAS data and including the subject as a random effects term to account for within-subject correlation. Peak and end VAS will be analysed in separate models. The proportion with intolerable pain, defined as estimated mean VAS >8, will be compared by Fisher’s exact test. Subjects with insufficient VAS data, defined as <3 recorded VAS peak or <3 recorded VAS end levels, will be excluded from these analyses
Secondary efficacy end points
1. Change in MRI WMH volume at 30 days and 90 days	Volumes at baseline and follow-up will be logarithmically transformed (natural log) to give a more normal distribution. Then differences between each group will be compared using a linear mixed model
2. Change in MRI diffusion tensor imaging peak skeletonised mean diffusivity⁴¹ at 30 days and 90 days	Linear mixed model, testing difference at 30 days and 90 days
3. Number of new MRI infarcts at 30 days and 90 days	Fisher’s exact test
4. Number of new MRI DWI-positive lesions at 30 days and 90 days	Fisher’s exact test
5. Change in MRI ASL grey matter cerebral blood flow at 30 days and 90 days	Linear mixed model, testing difference at 30 days and 90 days
6. Change in MoCA³⁶ score at 30 days and 90 days	Linear mixed model, testing difference at 30 days and 90 days
7. Change in Trail-Making A and B³⁷ at 30 days and 90 days	Volumes at baseline and follow-up will be logarithmically transformed (natural log) to give a more normal distribution. Linear mixed model, testing difference at 30 days and 90 days
8. Change in Controlled Oral Word Association^{38 39} score at 30 days and 90 days	Linear mixed model, testing difference at 30 days and 90 days
9. Change in CERAD 10-item word list learning⁴⁰ score at 30 days and 90 days	Linear mixed model, testing difference at 30 days and 90 days
10. Change in total score on MBI Tracking Tool, adapted from the MBI checklist,⁴⁹ at 30 days and 90 days	Linear mixed model, testing difference at 30 days and 90 days
11. Change in BADLS⁴⁸ at 30 days and 90 days	Linear mixed model, testing difference at 30 days and 90 days
12. Difference in candidate blood biomarkers at 30 days and 90 days	Linear mixed model, testing difference at 30 days and 90 days

ASL, arterial spin labelling; BADLS, Bristol Activities of Daily Living Scale; DWI, diffusion-weighted imaging; MBI checklist, mild behavioural impairment checklist; MoCA, Montreal Cognitive Assessment; RIC, remote ischaemic conditioning; VAS, Visual Analog Scale; WMH, white matter hyperintensity.