Exercise After Bone Marrow Transplants Shows Real Gains, But Only in Some Patients

Surviving a blood and marrow transplant has never been the finish line it once seemed. Adults who receive a haematopoietic stem cell transplant, a procedure that replaces a patient’s diseased blood-forming system with healthy stem cells, routinely carry a heavy load of long-term problems: about two-thirds report at least one chronic condition affecting their hormones, heart, lungs, muscles or bones. Fatigue can linger for years, cardiovascular risk factors accumulate, and psychological distress is common. A new systematic review published in eJHaem has now pulled together the best randomised trial evidence on whether structured exercise rehabilitation can meaningfully change that trajectory, and its findings are more nuanced than the optimistic headlines that often accompany cancer fitness research.

The review, conducted by researchers at the University of Winchester and published in March 2026, followed the PRISMA 2020 reporting standard and searched seven bibliographic sources for randomised controlled trials published between January 2019 and February 2026. To qualify, trials had to compare structured exercise-based rehabilitation against usual care or a control condition in adult transplant recipients, with outcomes measured at least three months after transplant or after the intervention ended. The screening process whittled 1,212 records down to seven trials comprising 529 participants across seven countries, spanning inpatient, outpatient, home-based and telehealth delivery. Two trials enrolled exclusively allogeneic recipients, who receive donor stem cells and typically remain on immunosuppressive drugs; two enrolled exclusively autologous recipients, who receive their own cells; and three enrolled mixed cohorts. Because the trials differed so much in modality, timing, population and outcome instruments, the authors abandoned meta-analysis and synthesised the findings narratively, rating certainty with the GRADE framework.

The clearest signal concerned walking ability. In the largest trial, an Australian telehealth-delivered programme of exercise and mindfulness reported in the Journal of Clinical Oncology, patients who exercised improved their six-minute walk distance by 51.4 metres more than controls at three months, an effect that persisted at 59.3 metres at twelve months. That is a substantial mechanical difference: a published threshold for clinically meaningful deterioration in the six-minute walk, borrowed from lung cancer research, sits between 22 and 42 metres, so the observed gains meet or exceed it, though the review authors are careful to note that no such threshold has been formally established in transplant survivors. A United States prehabilitation trial reported a non-significant between-group difference of 62.8 metres at day 100 post-transplant, and a Spanish telehealth trial found that the exercise group recovered to its pre-admission walking distance while controls remained 41 metres below baseline.

Lower-body strength told an even more consistent story. Across four trials that measured sit-to-stand or chair-stand performance, exercise groups gained roughly 2.5 to 3.8 extra repetitions in thirty seconds compared with controls, and no trial reported findings contradicting the direction of benefit. In the German whole-body vibration trial, which had patients training on a vibrating platform for about twenty minutes five times a week during their transplant admission, chair-rising power improved significantly in the intervention group. Grip strength, by contrast, showed no consistent between-group advantage anywhere, suggesting that the benefits of exercise after transplant concentrate in the large muscle groups of the legs rather than the hands.

The picture darkens for the outcomes patients often care about most. Effects on cardiorespiratory fitness, measured directly as peak oxygen uptake, were inconsistent: the German trial found a within-group gain in the exercisers, while the large Dutch EXIST trial, which tested an eighteen-week supervised programme in autologous recipients, found no between-group difference immediately after the programme or twelve months later, and the intervention was not cost-effective. Fatigue results were similarly mixed, with benefits reported in trials that began during or close to the transplant period but null findings in late-survivorship cohorts. Global quality of life was largely unchanged across trials, although physical functioning and daily-activity domains did improve in several. The GRADE ratings reflect this patchwork: low certainty for strength and walking benefits, very low certainty for everything else.

Perhaps the most provocative pattern in the review is the split between transplant types. Autologous recipients, whose immune systems typically engraft without long-term immunosuppression and whose main late risk is relapse, showed essentially no durable benefit in the only fully powered trial, EXIST. Allogeneic recipients, by contrast, face chronic graft-versus-host disease, steroid-related bone loss, late infections and years of deconditioning, and it is precisely in these patients that exercise appeared to deliver the most. The Spanish trial’s allogeneic subgroup showed a 77-metre between-group walking difference, and the allogeneic peri-transplant trials preserved strength, fatigue levels and even body cell mass while controls declined. The biology makes sense: patients with more deconditioning have more room to recover. The practical implication is that exercise referrals and future trials should be designed differently for the two populations rather than pooling them, as most mixed-cohort studies have done.

Two trials dipped a toe into immunometabolism, the least-studied domain and arguably the most tantalising. The Australian trial measured 27 serum cytokines and found, in an exploratory analysis of 69 of its 139 participants, a significant between-group reduction in soluble intercellular adhesion molecule-1, a marker of vascular inflammation and endothelial activation, at three months. No other cytokine differed, and samples were collected only at baseline and three months, so the finding is a hypothesis rather than a conclusion. The German trial reported that body cell mass, a proxy for metabolically active lean tissue, fell significantly in controls over the six months after transplant while exercisers regained their baseline values. Strikingly, no trial measured lymphocyte subset reconstitution, T-cell receptor diversity, vaccination responses, lipid profiles or glycaemic control, leaving the question of whether exercise can actually reshape immune recovery after transplant entirely open.

Safety, at least, is settled. No serious exercise-related adverse events were reported in any of the seven trials. Whole-body vibration was performed at platelet counts as low as 10,000 per microlitre without bleeding complications, the American prehabilitation trial reported zero exercise-related adverse events with 92 percent adherence, and the only serious event in any trial, a fall during a baseline walking test that caused a subarachnoid haemorrhage, occurred in a participant who had not been randomised and was unrelated to exercise.

Methodological honesty tempers all of these findings. Using the Cochrane Risk of Bias 2.0 tool, the reviewers judged six of the seven trials at high risk overall, mostly because of large and often health-related losses to follow-up: the Canadian feasibility trial retained roughly a third of its participants by one year, and the German trial analysed 44 of 71 randomised patients. Because exercise cannot be blinded, every trial was open-label, and none used blinded outcome assessment. Trials at higher risk of bias tend to report larger effects, which means the conservative null results from the largest study, EXIST, may in some respects represent the more realistic estimate of what exercise achieves in unselected survivors.

The review’s authors conclude that structured exercise appears safe and may improve objective physical function after transplant, with the greatest benefit in allogeneic and more impaired recipients, but that cardiorespiratory fitness, fatigue and quality-of-life effects remain uncertain and immunometabolic evidence is preliminary. Their prescriptions for the field are specific: future trials should report outcomes separately by transplant type, adopt harmonised core outcome sets, pre-specify immunological endpoints such as lymphocyte subsets and inflammatory panels alongside metabolic measures like fasting lipids and HbA1c, and follow patients well beyond twelve months to test effects on cardiometabolic risk, immune reconstitution and survival. For now, the practical message for the growing population of transplant survivors is measured but real: if you are still rebuilding strength and walking distance months after your transplant, structured exercise, particularly through telehealth models that reach patients at home, is a reasonable and safe bet, and the sooner after treatment it starts, the more deconditioning there may be to reverse.

Subject of Research: Exercise-based rehabilitation outcomes and immunometabolic parameters in adult haematopoietic stem cell transplant recipients

Article Title: Long‐Term Outcomes of Exercise‐Based Rehabilitation and Immunometabolic Parameters in Adult Haematopoietic Stem Cell Transplant Recipients: An Updated Systematic Review of Randomised Controlled Trials

Article References: Bayquen, D. D., & Narasimhan, P. (2026). Long‐Term Outcomes of Exercise‐Based Rehabilitation and Immunometabolic Parameters in Adult Haematopoietic Stem Cell Transplant Recipients: An Updated Systematic Review of Randomised Controlled Trials. eJHaem, 7(5), Article e70420. https://doi.org/10.1002/jha2.70420

Image Credits: AI Generated

DOI: 10.1002/jha2.70420

Keywords: haematopoietic stem cell transplantation, exercise rehabilitation, systematic review, randomised controlled trials, allogeneic transplant, autologous transplant, physical function, six-minute walk distance, fatigue, quality of life, immune reconstitution, sICAM-1