Oxygen therapy in ischemic stroke.

To the Editor: Previous studies of therapeutic oxygen in stroke have had conflicting results, and the latest study by Rusyniak et al1 seems to suggest that hyperbaric oxygen therapy (HBO) is harmful in ischemic stroke. The authors have already discussed how several inadequacies in trial design might explain their negative results. We wish to draw attention to additional features in their study that might offer important insights into the design of future studies of oxygen in stroke. The authors evaluated the safety, feasibility, and efficacy of using HBO in a sham-controlled pilot study of 33 patients with acute ( 24 hours) ischemic stroke. The HBO group (17 patients) received 100% oxygen at 2.5 atm absolute (ATA), and the sham group (16 patients) received 100% oxygen at 1.14 ATA, both for 60 minutes. Outcomes were measures at 24 hours and 90 days using several clinical stroke scales. Complications related to the use of pressurized chambers were comparable in both groups. There was no difference in stroke outcomes at 24 hours; however, at 90 days a significantly higher percentage of patients in the sham group had good outcome. The authors conclude that their HBO protocol appears feasible and safe (ie, without significant pressure-related complications) but state that HBO might be harmful in patients with acute ischemic stroke. As acknowledged by the authors, this study had a small number of patients, which limits data interpretation, and the negative results are probably attributable to factors like late timing of therapy and use of excessively high chamber pressures. It is well known that higher pressures, late timing, and longer duration of HBO produce harmful effects.2–5 In a recent animal study we found that HBO at 3 ATA for 60 minutes (similar to the present study) reduced infarct volumes and improved neurological outcomes when applied at 3 and 6 hours, but enhanced infarct volumes and aggravated neurological deficits when applied at 12 or 24 hours after the onset of reperfusion.6 If most patients in this pilot study were treated 12 hours after the onset of stroke (only 15% were treated within 6 hours), it is not surprising that a harmful effect of HBO was observed. Whether early HBO therapy (within 3 to 6 hours) affords benefit remains to be determined. It is extremely important to note that the sham group was actually treated with 100% oxygen. There was no control group treated with room air at elevated atmospheric pressures, as in previous studies.2,3 We have compared the results of different oxygen pressures on brain protection in a neonatal rat model and found that oxygen therapies at 1, 1.5, and 3 ATA produced similar brain protection (measured by brain weight) in pups (Figure; for detailed methods see Ref. 7). Since this study only compared oxygen therapy at different pressures, it is misleading to conclude that HBO itself might be harmful—the only fair conclusion is that (within the limits of this protocol), oxygen therapy at 2.5 ATA is harmful as compared with oxygen therapy at 1.14 ATA. We were impressed by the excellent results in the sham group: 80% to 91% of patients had good outcome at 90 days, depending on the stroke scale used. Spontaneous reperfusion is common in patients with low NIHSS, and since 15 of 33 patients had NIHSS 7 at onset, factors such as reperfusion might have contributed. However, these excellent outcomes might not be simply fortuitous—an alternate possibility is that low-pressure oxygen therapy is, in fact, highly beneficial. Several recent animal studies have shown that shortduration normobaric oxygen therapy can be highly neuroprotective and does not increase oxidative stress, if started early after stroke onset.8–10 Maximum benefit was observed in the cerebral cortex.8–10 We ask whether the authors have any information regarding tissue perfusion status, stroke subtypes, and immediate response to oxygen treatment. This information would be extremely useful in analyzing why the sham group had such excellent outcomes.