Adding blood flow restriction to low-load resistance training for sarcopenia and cognitive function in oldest-old adults: an exploratory randomized controlled trial

Sarcopenia and mild cognitive impairment (MCI) frequently coexist in oldest-old adults, who may have limited tolerance for conventional resistance training (RT). Low-load blood flow restriction resistance training (BFR-RT) may offer an alternative, but evidence in this population remains limited. This exploratory, single-blind, three-arm randomized controlled trial included 36 oldest-old adults with coexisting sarcopenia and MCI, randomized equally to BFR-RT, matched low-load RT, or a no-exercise control (CT). Exercise was performed three times weekly for 8 weeks at approximately 30% of estimated one-repetition maximum; BFR-RT used 50% of estimated arterial occlusion pressure. Sarcopenia-related and cognitive outcomes were assessed at baseline, week 4, and week 8, and serum biomarkers at baseline and week 8. Intention-to-treat linear mixed-effects models were used, with Benjamini–Hochberg false discovery rate (FDR) correction within outcome domains. Constrained longitudinal data analysis and participant-level stratified cluster bootstrap analyses were conducted as sensitivity analyses. All 36 participants were included in the intention-to-treat analysis. Mean attendance was 96.9% in RT and 95.1% in BFR-RT, with no serious intervention-related adverse events. After FDR correction, significant group × time interactions remained for appendicular muscle mass index (P = 0.036), handgrip strength (P = 0.012), 4-m walk time (P = 0.044), CALT total score (P = 0.003), and CALT delayed recall (P = 0.002). Sensitivity analyses supported GS, 4 MW, CALT, and delayed recall, whereas the AMI finding was less robust. Compared with CT, BFR-RT showed favorable changes in GS, 4 MW, verbal learning, and delayed recall. Evidence for additional benefits of BFR over matched low-load RT was limited, with the clearest between-intervention signal for GS. No serum biomarker interaction remained significant after FDR correction. Supervised low-load BFR-RT may be a promising exercise strategy for oldest-old adults with coexisting sarcopenia and MCI. Favorable physical and memory-related changes were observed versus no exercise, but evidence for an additional benefit of BFR beyond matched low-load RT was limited. Biomarker findings were exploratory and did not support BFR-specific biological mechanisms. The broader parent project was registered with the Chinese Clinical Trial Registry (ChiCTR2600119054) on 14 February 2026 and is listed as retrospectively registered because other components of the parent project had already commenced. Recruitment for the present BFR-RT trial began after the registration date. Supervised low-load BFR-RT was well attended and was not associated with serious intervention-related adverse events in oldest-old adults with coexisting sarcopenia and mild cognitive impairment. After FDR correction and sensitivity analyses, BFR-RT showed favorable changes versus no exercise in handgrip strength, 4-m walk performance, verbal learning, and delayed recall, whereas the AMI finding was less robust. Evidence for additional benefits of BFR beyond matched low-load RT was limited for most outcomes, with the clearest signal observed for handgrip strength; no serum biomarker interaction remained significant after FDR correction.