Five Days of Acute Intermittent Hypoxia Produced Measurable Gains in Multiple Sclerosis
A short course of acute intermittent hypoxia improved somatic motor function in ambulatory adults with multiple sclerosis while remaining well tolerated across a clinically mixed disability range. In this randomized, double-masked, placebo-controlled crossover preprint, five daily sessions of intermittent low-oxygen exposure were followed by stronger ankle plantarflexion and longer walking distance, without serious adverse events related to the intervention.
The result matters because walking impairment is one of the most common and disabling features of MS. Many rehabilitation strategies can improve performance in the moment, but durable neuroplastic gains are harder to elicit in a disease marked by demyelination and variable conduction failure. Acute intermittent hypoxia has attracted attention in neurorehabilitation because repeated brief hypoxic episodes can trigger serotonin-dependent spinal plasticity, a mechanism that may amplify motor output when paired with training. This preprint tested a simpler question first: whether repeated daily dosing is safe in people with MS, and whether the intervention has stand-alone effects before it is layered onto gait training.
Safety, Tolerability, and Physiologic Recovery After Each Session
Twenty-three ambulatory adults with MS completed the crossover study. Disability ranged from Expanded Disability Status Scale scores of 3.0 to 6.5, and the cohort included both relapsing-remitting and secondary-progressive disease. Participants received AIH and sham exposure in counterbalanced order, separated by at least four weeks, which helped reduce sequence effects in a small mechanistic trial.
The hypoxic dose was tightly controlled. Each session consisted of 15 brief episodes of approximately 60 seconds at an inspired oxygen fraction of 0.09, with oxygen saturation targeted to 84% to 88%. Mean nadir oxygen saturation during AIH was 84.7%, and cardiovascular variables returned to baseline within minutes after each session ended. No serious adverse events related to the intervention were reported, and follow-up MRI showed no new MS activity in either arm.
Daily exposure to controlled hypoxia did not trigger visible clinical destabilization in this cohort.
That safety profile is important for a population in which fatigue, autonomic symptoms, and concern about disease activity often narrow the practical room for experimentation. In a demyelinating disorder, even modest physiologic stress can raise concern about relapse, exacerbation, or poor tolerance. Here, repeated daily sessions were completed without evidence of such a signal in the prespecified safety assessments.
Plantarflexion Strength Rose More Than Endurance Metrics Changed
The clearest functional effect appeared in ankle plantarflexion torque. Compared with sham, AIH produced a significant increase in plantarflexion strength, with a beta coefficient of 8.74 Nm and a p value below 0.001. Mean plantarflexion torque increased 17.5% at post-intervention and 29.2% at follow-up, suggesting that the effect was not restricted to an immediate post-session bump.
That pattern fits the expected biology. Plantarflexors contribute substantially to late-stance propulsion during gait, and even modest improvements in this muscle group can alter walking economy and push-off mechanics. If hypoxia-induced spinal plasticity enhances descending drive or motor-unit recruitment, the plantarflexors are a plausible site where the change would first appear because they carry a major load in locomotor force generation.
Walking endurance moved in the same direction. Six-minute walk distance was significantly greater after AIH than after sham, with a beta of 10.14 m and a p value of 0.039. The gain was modest, but it points toward a real functional transfer beyond isolated dynamometry. In a rehabilitation context, even small shifts in endurance can matter if they translate into longer household or community ambulation.
What Did Not Change, and Why That May Matter
Not every mobility measure improved. Dorsiflexion torque did not differ significantly between conditions, nor did Timed 25-Foot Walk performance or MS Walking Scale-12 scores. Those null findings are not surprising in a short intervention that was not paired with task-specific practice. Dorsiflexors are critical for foot clearance, but the physiologic pathway engaged by AIH may not influence every muscle group equally, and walking speed over a short corridor can be driven by factors that a five-day protocol is too brief to alter.
The absence of change in patient-reported walking impact also deserves careful reading. Self-report tools often require larger or more global functional shifts before patients notice them, especially over a brief timeframe. A 10-meter increase in six-minute walk distance and a torque gain at the ankle can be biologically meaningful without immediately crossing the threshold of subjective daily-life recognition.
Greater heart rate elevation during hypoxic episodes was associated with larger strength gains at follow-up, with an r value of 0.58 and a p value of 0.009. That relationship raises a mechanistic clue rather than a clinical directive. Sympathetic responsiveness, ventilatory effort, or individual hypoxic reactivity may be indexing who is most likely to engage the plasticity pathway, although this remains an associative signal rather than a tested predictor.
Why This Preprint Fits the Next Phase of MS Rehabilitation Research
This study is the first multi-day AIH trial in people with MS, and it moves the field beyond proof-of-concept safety into preliminary efficacy. Prior work in spinal cord injury suggested that pairing AIH with walking training can amplify locomotor gains. In MS, the immediate question is not whether hypoxia alone can replace rehabilitation. The more relevant question is whether AIH can prime the nervous system so task-specific training produces larger and more durable improvements than training alone.
The preprint supports that strategy, but only as an early signal. The sample size was small, the intervention window was short, and the trial was not designed to establish long-term functional benefit or define which MS subgroups respond best. Disability was clinically heterogeneous, which strengthens feasibility, but also makes it harder to isolate disease-stage effects. The outcome pattern suggests that lower-limb force generation may be more responsive than short-burst speed or patient-reported walking impact in this dosing framework.
For rehabilitation clinicians, the immediate implication is cautious optimism around feasibility rather than adoption. For researchers, the more important takeaway is that repeated AIH can be delivered safely in ambulatory MS without provoking obvious acute safety concerns in this cohort, and it can produce measurable gains in plantarflexion strength and walking endurance. The next trial will need to test whether those gains become larger, more durable, and more functionally relevant when AIH is combined with gait training, and whether the heart-rate response observed here can help identify responders before treatment is scaled more widely.