乌鸦传媒

CHICAGO -- Supplemental oxygen routinely given to acute myocardial infarction (MI) patients from the ambulance through to the recovery room might actually be hurting their hearts, a trial suggested.

Infarct size estimated by creatinine kinase levels was greater with oxygen than with none for patients with normal oxygen saturation (ratio of geometric mean peaks 1.26, P=0.01), Dion Stub, MBBS, PhD, of the Baker IDI Heart and Diabetes Institute in Melbourne, Australia, and colleagues found.

Infarct size measured by the more accurate cardiac MRI in a subset of patients after 6 months showed the same pattern (median 20.3 versus 13.1 g, P=0.04), the researchers reported here at the American Heart Association meeting.

Their AVOID trial wasn't powered for clinical endpoints but did show significantly more recurrent myocardial infarction (5.5% versus 0.9%, P<0.01) and a trend for more cases of significant arrhythmia (40.4% versus 31.4%, P=0.05) in the group that got oxygen.

At the very least, the study demonstrated that routine oxygen isn't necessary for heart attack patients who aren't hypoxic, said study discussant , co-director of Sarver Heart Center at the University of Arizona in Tucson.

There are implications across acute coronary syndromes too, commented , an emergency physician at the University of Pennsylvania in Philadelphia.

"If it's worse for STEMI, it's highly unlikely -- highly unlikely -- that oxygen would be good for anything less," he said. "If it's good for anybody [with chest pain], it's going to be the people with the real injuries. ... The simple thing would be don't give it to anyone with chest pain."

"The great thing about this trial is it's totally actionable," Abella told 乌鸦传媒 in an interview monitored by his institution's media relations. "This could be adopted tomorrow."

Kern wasn't ready to "break up with" this key component of MONA [morphine, oxygen, nitrates, aspirin] in emergency care of acute MI, but he predicted "this may be the beginning of her demise."

While there's limited evidence to support a benefit from oxygen therapy in the absence of hypoxemia, there is mounting evidence from basic science that it may reduce coronary blood flow, increase coronary vascular resistance, and fuel reperfusion injury, Stub noted.

The question hanging over the trial was how the results would apply to 2 to 4 L/minute oxygen flow more typical of U.S. practice, commented James L. Januzzi Jr., director of the cardiac ICU at Massachusetts General Hospital in Boston. Another trial would be needed to provide the answer, he said.

Still, Januzzi said his practice will change based on AVOID to stop giving any oxygen in acute MI unless the patient is hypoxic.

The AVOID trial included 638 patients with ST-elevation MI and an oxygen saturation of at least 94% randomized in the ambulance to high-flow oxygen (8 L/minute via face mask) or none.

Stub noted that the pragmatic trial followed informed consent procedures according to Australian frameworks and with approval of all participating ethics committees.

Ambulance-based randomized trials are notoriously difficult, commented , director of the cardiac intensive care unit at Duke University Medical Center.

It was possible for AVOID because the emergency system there is experienced in conducting research, Stub explained.

Randomization was maintained through primary percutaneous coronary intervention (PCI) in the cath lab.

Control group patients could receive oxygen if their saturation fell below 94%, and about 35% did end up getting it, largely during PCI or in the hours immediately following it.

Creatinine kinase-estimated infarct was also greater by area under the curve, and median peak compared with no oxygen (both P=0.04).

However, the other primary endpoint on infarct size, measured by troponin I trended in the same direction numerically but showed no significant associations, which Stub suggested could have been because one site had trouble with ascertainment.

Geometric mean infarct size and proportion of left ventricular mass by MRI showed nonsignificant trends, with ratios of means of 1.43 and 1.38 (P=0.06).

Notably, despite discussion of symptomatic benefit for patients with oxygen, there was no impact on pain scores or administration of painkillers in the trial, Stub pointed out.

One limitation was lack of actual blood oxygen saturation data instead of just pulse oximetry measurements, Kern noted, pointing to cardiac arrest literature suggesting that anything over 300 torr oxygen is detrimental.

Another limitation was that the analysis was by intent to treat, without too much information on crossovers, he said.

"These findings certainly need to be confirmed in larger randomized trials powered for hard clinical endpoints," Stub concluded. "But the AVOID study investigators really question the practice of giving oxygen to all patients, and certainly those that have normal oxygen levels."

Comment