ARTICLE IN BRIEF
A trial investigating the use of normobaric oxygen therapy (NBO) for reducing damage from acute ischemic stroke was stopped by trial monitors after 17 out of 43 subjects in the NBO arm of the trial died compared with seven out of 42 in the room air cohort. The investigators offer more detail about the trial being halted, and suggest that the NBO concept deserves further research.
A phase 2B study investigating the use of normobaric oxygen therapy for reducing damage from acute ischemic stroke and potentially widening the window for thrombolytic treatment was stopped by trial monitors after a disproportionate number of treated patients died, even though none of the deaths were attributed to therapy.
Seventeen out of 43 subjects in the normobaric oxygen (NBO) arm of the trial died compared with seven out of 42 in the room air cohort, but most patients who died were removed from life support due to massive baseline infarcts and were above 80 years old. An additional 29 percent died from pre-existing conditions between days 21-106, including liver cancer and lung cancer, according to principal investigator Aneesh Singhal, MD, associate professor of neurology at Harvard Medical School and director of neurology quality and safety at Massachusetts General Hospital and Brigham and Women's Hospital.
The study was stopped after 85 patients out of a target of 240 had been enrolled.
“When the overall baseline health status of patients on admission was considered, there was no association between treatment and deaths. The NBO group simply had more patients with more severe strokes and other pre-existing illnesses,” he told Neurology Today. “The data and safety monitoring board (DSMB) was not aware of the severity of enrollees' admission status leading to withdrawal of care when they halted the study — they just saw the high rate and unexplained imbalance of deaths.”
Dr. Singhal presented the clinical trial results and brain imaging findings in two platform sessions at the AAN annual meeting in San Diego. He and his colleagues in the Specialized Programs of Translational Research in Acute Stroke (SPOTRIAS) network found no significant difference in primary safety outcomes between the two groups, as measured by changes in National Institutes of Health Stroke Scale (NIHSS) or lesion growth by MRI or CT scans. They also found no evidence of oxygen toxicity, and no difference in the rates of adverse events such as brain edema. Secondary analysis showed slight improvement in brain tissue recovery as measured by MRI voxel-based algorithms.
SPOTRIAS was funded by the National Institute of Neurological Disorders and Stroke.
In a 2005 pilot study in 16 highly selected patients, these investigators found transient improvement of clinical deficits and MRI abnormalities in stroke patients treated with NBO. Out of nine NBO-treated patients, relative DWI lesion volumes decreased in six patients during therapy at four hours, with decreases of more than 20 percent in three subjects. NIH stroke scores were similar at baseline between the treated patients and a group of six control subjects, but tended to improve at four hours and one week in NBO patients, with significant improvements at 24 hours.
In the current study, participants were randomly selected to receive either room air or NBO at flow rates of 30-45L/min administered via a facemask for eight hours in double-blind fashion. Neurological function scores and magnetic resonance imaging (MRI) or computed tomography (CT) scans were obtained before, during, and after therapy for 90 days. There was no upper limit for age, NIHSS, or stroke volume, and NIHSS and MRI scans were obtained at baseline, at four hours during therapy, 24 hours and 48 hours, and again at three months.
Patients were randomized several hours after stroke symptom onset, but all were outside the time window for tissue plasminogen activator (tPA) and most were outside the window for intra-arterial thrombolysis. For NBO to be effective the time window is probably closer to 45 minutes, based on animal studies, and not nine-plus hours, which was the time window for inclusion in the study, Dr. Singhal said.
“The proper way to test NBO would be in the form of a pre-hospital trial where paramedics administer NBO within 45-60 minutes, and then NBO is continued in the hospital for a few hours. Unfortunately, we were unable to go directly for a pre-hospital trial and had to undertake this first step.”
There was no significant difference in pre-specified outcomes for safety as determined by 0-24 hour change in NIHSS, 0-48 hour percent lesion growth on CT or MRI, or efficacy as gauged by 0-4 hour absolute change in NIHSS or 0-4 hour percent MRI lesion growth. Secondary outcomes were assessed as changes in NIHSS, percent lesion growth over three months, and tissue reperfusion. There was no significant difference in the rates of serious adverse events such as pneumonia, cardiac failure, brain edema, or brain hemorrhage.
“In this study of oxygen therapy we accepted as many patients as possible, rather than carefully selecting them which is common practice in pharmaceutical drug trials. Unfortunately, many of them had much more serious stroke damage,” Dr. Singhal said. “With the first two looks at the data after around 50 patients had been enrolled, everything went well, but when the monitoring board took a third look they halted the trial. They just saw the large number of deaths without full knowledge of their cause. We were surprised that they terminated the trial, rather than temporarily suspending it.”
He said there are no plans to continue the trial without funding from NIH.
“We learned a lot of lessons. I was surprised, but that is how it works. I still believe NBO can be effective but there is no money to be made from it because it is so inexpensive, so obtaining funding from private companies is unlikely and starting from scratch again would require a lot of resources. In hindsight, maybe we should have been more restrictive in selecting patients.”
Other recent studies of NBO in humans and animal models of ischemic stroke have shown promise, but mostly in transient cerebral ischemia and usually when started early (within 30 minutes). Harmful effects have been reported, notably vasoconstriction and the release of oxygen free radicals, but not in the setting of focal stroke. Death has not been a concern in numerous previous trials of oxygen therapy in diverse conditions like acute myocardial infarction, traumatic brain injury, and stroke. A 2006 Taiwanese trial of severe ischemic stroke patients showed reduced mortality and fewer adverse events in the group treated with inspired oxygen.
Animal studies have suggested that NBO might reduce serious sequelae and extend the treatment window for thrombolysis with tPA. Some small studies have also reported advantages in treating stroke patients in hyperbaric oxygen chambers, but these, typically used to decompress scuba divers, are considered too costly and cumbersome for treating stroke.
One 2010 study by Indian researchers found that 20 patients with anterior circulation ischemic stroke who received normobaric oxygen (10 L/min for 12 hours) showed a trend toward improvement in NIHSS, modified Rankin scale (mRS), and modified Barthel index at days 0, 1, 7, and at 3 month follow up. The benefits were seen until the 7th day.
Another study, published in Stroke in 2009, found that combined NBO and tPA in a rat model of ischemic stroke significantly reduced the mortality rate from 70 percent to 11 percent, and reduced brain edema and hemorrhage when compared to rats treated with tPA alone.
Louis R. Caplan, MD, professor of neurology at Harvard Medical School and senior neurologist at Beth Israel Deaconess Medical Center, said the trial was undone not by evidence but by circumstance.
“I think the idea is great, but Dr. Singhal just got terrible luck. There was an unfortunate balance of severe cases, and the DSMB stopped the trial without looking at the data,” he commented. “Oxygen is cheap, ubiquitous, plentiful, and widely used in the field, so it would be terrible kill the idea just because of bad luck. It's really a pity,” he told Neurology Today.
He said that he believes NBO may yet offer some benefit for stroke patients, but that it will take more research with a healthier group of stroke patients.
“I think this work should continue, but we will need more preliminary data in order to balance the two groups. This might be difficult because there is no money in oxygen treatment, so you can't get commercial funding. Oxygen might work, not in the long run but it might buy more time.”
TUNE IN: Why was a trial of normobaric oxygen in acute ischemic stroke halted early? Is the therapeutic concept feasible? In a video interview, Kevin N. Sheth, MD, chief of the Division of Neurocritical Care and Emergency Neurology and director of the Neurosciences Intensive Care Unit at Yale New Haven Hospital, offers key clinical takeaways — why the concept may be feasible and what it would take to make it so: http://bit.ly/XgS6E5. Dr. Sheth is a member of the Neurology Today editorial advisory board.