We discussed Gupta et al.'s trial of continuous positive airway pressure (CPAP) in patients who have had a stroke and with obstructive sleep apnea (OSA)1 at our Twitter-based journal club (@ respandsleepjc, #rsjc) on May 17, 2018. Lead author Dr. Anupama Gupta participated online.
Our participants highlighted that the study was underpowered. Seventy patients were randomized, with an a priori power analysis requiring 160 subjects to achieve 80% power. Dr. Gupta acknowledged the power limitation in our discussion.2 While the results align with our expectations, an underpowered study is unable to provide definitive answers to the research question. We understand government budget limitations prevented the use of sham-CPAP for a double-blinded protocol.3
An important and avoidable problem, however, was failure to perform an intention-to-treat (ITT) analysis. Following randomization, 4 patients crossed from the CPAP to the non-CPAP group and were analyzed in the non-CPAP group. The author's response to our concern was that a per-protocol analysis was preferred to ensure uniformity of the treatment arms and to avoid confounders.4 This explanation seems invalid. An ITT analysis is invariably preferred in superiority trials because failing to analyze participants in the groups to which they were randomized throws away randomization and all its benefits, including control of unmeasured confounders, and because not knowing whether patients will adhere to therapy imitates clinical decision making.
An additional concern was lack of clarity regarding how the authors identified the primary outcome of new vascular events. Dr. Gupta explained the primary outcome was evaluated by a single investigator assessing patients based on self-reported adverse events.5 Most clinical trials of stroke would assess this outcome with regular brain imaging as an objective measure of recurrent strokes.
There is a general understanding that untreated OSA leads to systemic hypoxia and global cardiovascular stress. There are established cardiovascular benefits to OSA treatment in patients who have had a stroke, before even considering stroke recurrence rates. Our participants unanimously concluded they would treat OSA in patients who have had a stroke even if this was a validated negative trial.
All authors have seen and approved the manuscript. The authors report no conflicts of interest.
Budd L, Stanbrook MB, Anand A. Twitter discussions from a respirology journal club: role of positive airway pressure therapy for obstructive sleep apnea in patients with stroke. J Clin Sleep Med. 2018;14(10):1817.
Gupta A, Shukla G, Afsar M, et al. Role of positive airway pressure therapy for obstructive sleep apnea in patients with stroke: a randomized controlled trial. J Clin Sleep Med. 2018;14(4):511–521. [PubMed Central][PubMed]
@ANUPAMA59329210. #rsjc agree and discussed in limitation section of paper. https://twitter.com/ANUPAMA59329210/status/997675646187864064. Posted May 18, 2018.
@ANUPAMA59329210. #rsjc This trial was funded by Department of Science and technology, Government of India. Initially sham PAP was proposed in study design but not sanctioned, because of budget limitation. https://twitter.com/ANUPAMA59329210/status/997264107681165312. Posted May 17, 2018.
@ANUPAMA59329210. #rsjc Thank you! For your valuable inputs. Ours was Superiority Trial and per-protocol analysis is preferred method to ensure uniformity of treatment arm and all noise removed. https://twitter.com/ANUPAMA59329210/status/997675411596283904. Posted May 18, 2018.
@ANUPAMA59329210. #rsjc No it was not chart review. Every patient who was enrolled in study has to report our staff as soon any adverse event occurs. Each outcome was assessed by Dr. Garima Shukla, who was completely blinded by patient PAP therapy status. https://twitter.com/ANUPAMA59329210/status/997268899430199297. Posted May 17, 2018.