Emergency Medicine News:
Letter to the Editor
Ialways look forward to reading EMN, and find it very informative.
Just after the study discussed in the article, “Titrated Oxygen Better than High Flow for Patients with COPD” (EMN 2011;33:1; http://bit.ly/Titrated) appeared, I was asked by the training coordinator for one of our regional EMS offices whether we should change our practice in Alaska. I found that this study has some major methodologic flaws. The glaring problem is that the high-flow group and the titrated group were not well matched. The methodology randomized providers rather than patients, which is unusual, to say the least, and may have introduced bias into the study.
While the conclusions may be valid, the study has some major methodologic problems, and does not present crucial data needed to evaluate the efficacy of “titrated oxygen” compared with high-flow oxygen.
The first problem is that all of their conclusions are based on 28 deaths: 21 in the high-flow arm and seven in the titrated arm. If one more patient had died in the titrated arm, it would have changed the results. In the group with confirmed COPD, there were 11 deaths in the high-flow and two deaths in the titrated arm. It is not difficult to see that an additional death in the titrated arm would have markedly changed the statistics. Remember that the high-flow arm was more than 25 percent larger than the titrated arm for all patients and more than 20 percent larger for the confirmed COPD patients, so the raw numbers of deaths are misleading. If the high-flow arm had been the same size as the titrated arm, the corrected numbers would be 16 and seven. Still different but not quite as lopsided.
The biggest problem with the study is randomization. They randomized paramedics to giving one treatment or the other rather than randomizing patients to receiving one treatment or the other. That the patient groups were not perfectly matched is evident by the fact that they were not the same size. There were 226 patients in the control group and only 179 in the treatment group. In a study in which patients are randomized, there should be no significant differences in patient characteristics between the two groups (and the groups should generally be the same size).
In all patients, the mean pretreatment SpO2 was 86% in the high-flow group and 88% in the titrated group. In the confirmed COPD patients, the mean pretreatment SpO2 was 84% in the high-flow and 87% in the titrated group. While these differences sound small, we don't know what the actual distribution was in each group. This leaves open the possibility of biased groups. If the high-flow group had just a few sicker patients, the increased death rate could have nothing to do with the differences in treatment. It may also be important that the range from 84% to 88% is the beginning of the steep portion of the oxyhemoglobin dissociation curve. There is a large physiological difference between an SpO2 of 84% and an SpO2 of 88%, while the difference between, say, 94% and 98% is less important.
I don't see a lot of harm in changing our way of doing things, but I'm not sure that this study is compelling evidence to do so. I hate to disagree with the legendary Ron Walls, but I am not yet convinced that we should change our practice.
Ken Zafren, MD
Dr. Zafren is the EMS Medical Director for Alaska.