“Hospital readmission and wound separation were also more common in the 80% group,” she said. “None of the numbers reached statistical significance because the study was not powered for that, but the data [were trending] in the wrong direction.” Because the P value exceeded the P value for futility, suggesting the observed differences were unlikely to reach statistical significance with continued recruitment, the trial was stopped.
The study was limited by a number of considerations. The median duration of oxygen therapy was less than planned, there was significant overlap in the median level of venous oxygen in the 80% group and the 30% group, and it's possible a therapeutic level was not achieved, and the dorsum of the foot was used as a proxy measure of oxygen delivery. Finally, the outcome may have been 'muddled by antibiotic overuse in the study population, Dr. Gardella said.