In prior work on this site and elsewhere we have put forth the necessity to use sleep lab retitration protocols as a mainstay of practice to solve the problem of CPAP failure. In this post, we will delve into the history on the development of this concept, which has culminated most recently in the acceptance of our first peer-reviewed, research paper on the REPAP approach (REPAP stands for repeat, rescue, retitrations to reverse CPAP failure). Our paper will appear online in the next few months in the journal Respiratory Care, after more than ten years of effort to bring to fruition.1
Our story begins in 2005 in the third year of operations at our new private sleep center, Maimonides Sleep Arts & Sciences. Initially, we were treating all patients with CPAP, auto-CPAP, and a few patients with CFLEX. With careful analysis of the airflow curve in the sleep lab, we eventually recognized a specific objective finding (described below) to explain the inferior results generated by fixed CPAP pressures. Going forward, we only used bilevel devices in the vast majority of patients. During this timeframe, we were already receiving an influx of patients seeking second opinions, and nearly all of them were using or had previously failed CPAP; these patients in particular guided us in narrowing our focus to explore the advantages of bilevel. The linchpin of all this work was the objective finding that CPAP patients could not tolerate fixed pressures when breathing out; bilevel alleviated the problem, providing considerable relief of both discomfort and distress. What surprised us most about our insights in 2005 and thereafter was the rarity of any discussion on the topic in the scientific literature. Specifically, we found few recommendations or strategies on how patients should be directed to proceed when they fail CPAP and want a second opinion.
In fact, we could only find a few articles on the potential value in conducting retitrations in some patients, but it was rarely spelled out that these patients were seeking second opinions. To this day, we are uncertain whether or not other papers or guidelines cover the concept of second opinions in the field of sleep medicine for failing CPAP patients. When you consider this omission from the scientific literature, we believe it should cause concern and possibly apprehension about the workings of the field of sleep medicine. After all, is it not a standard of care in virtually every aspect of medicine, surgery or psychiatry to educate and inform patients on the value of seeking second opinions for a host of medical, surgical and psychiatric disorders? I daresay virtually every person reading this post has either personally experienced the need for a second opinion or knows of a friend or family member who required one, if not more of these encounters to solve assorted health problems.
Why then are there no obvious tracking systems, research articles, or formal policy statements about second opinions in the field of sleep medicine? I believe no immediate or satisfactory answers to this question are forthcoming other than our field is relatively young compared to many specialties. Nevertheless, dispensing with diplomacy for a moment, I also wonder whether this lacuna is more representative of a large blind spot among sleep experts, a phenomenon likely to occur when a group of medical professionals form consensus opinions and policies about specific diseases or health issues, which then become set in stone. From my vantage point as something of an outsider who left the academic-oriented university environment and who has offered a number of criticisms of the American Academy of Sleep Medicine (an organization largely comprised of professionals with continuing ties to academia), I remain disconcerted by the lack of attention to second opinions in the field of sleep medicine.
In 2007, with this framework in mind, we began to collect cases from our database of patients who had come to us specifically seeking a second opinion, nearly all of them were using CPAP. As our caseload grew, we re-examined the scientific literature as well as policies and procedures from the American Academy of Sleep Medicine (AASM). To our surprise, there was almost nothing to read on the topic of second opinions except a few researchers describing how they had successfully switched patients from CPAP to auto-CPAP or bilevel. From the AASM the only salient point in their published procedures was the suggestion to consider bilevel in patients with complaints of expiratory pressure intolerance, usually due to high CPAP pressures defined as > 15 cm H20.
In 2008, we started writing our first draft of a manuscript to submit for peer-review publication. Our main intent at the time was to publish data on how switching patients from CPAP to BPAP not only improved comfort but also improved use and adherence rates. At this same point, we were also moving on from bilevel to auto-adjusting bilevel (ABPAP) devices and soon thereafter the even more advanced auto-bilevel device known as adaptive servo-ventilation (ASV). Because of this transition in the types of modes applied at our center, we paused temporarily in our manuscript preparation, because we were seeing better results with ABPAP and ASV compared to fixed BPAP.
Another issue that appeared, one that took a few years to iron out, involved how to define CPAP failure. By 2010, we had already seen more than 1000 second opinion patients, but with greater scrutiny we realized many CPAP failure cases actually represented various stages of noncompliance, which encouraged us to developed an internal “CPAP failure timeline.” For example, obviously someone who completely rejected CPAP before ever trying it or perhaps trying it once in the sleep lab during a titration is different than a patient who filled a prescription for a device, took it home and attempted to use it for a few months or longer. Still more differentiation could be found among patients using CPAP irregularly for a long time with fair results versus those using regularly for a long time with poor results. Then, some patients reported great early results for several months and then a steady deterioration in sleep caused by other sleep problems or simply a decline in the response to PAP.
You can imagine many other permutations, but there was a larger problem embedded within our efforts to define CPAP failure, which we soon realized would make things more difficult in trying to publish a research paper. The problem, which we noted around 2011, was most other sleep experts, clinicians or researchers were not investigating CPAP failure in this much detail. Because so many sleep professionals by this time were entranced and obsessed in dealing with Medicare’s definition of CPAP adherence, the compliance concept of use > 4 hrs per night for 70% of nights permeated medical judgment and decision-making in the treatment of OSA/UARS patients.
Consider the most obvious example: a patient reports sleeping with CPAP six hours per night for 6 nights per week, which by the way is the standard we strive for in our patients. However, the patient reports persisting symptoms of insomnia and nocturia at night and sleepiness and fatigue during the daytime. Did you know that among a fair number of sleep professionals, the first approach to this patient might assume the co-occurrence of some other health problem, sleep-related or otherwise? To be sure, nowadays many sleep specialists would want to review a data download of the patient’s CPAP use, but viewing these data and interpreting the findings often yield cognitive dissonance because of the following: many data downloads do not properly record the discrete breathing events (flow limitations) of UARS; elevated leak values are frequently deemed in the “normal” range; many sleep staff ignore the presence of central apneas if only 1 or 2 centrals per hour are recorded; and finally, even today, the majority of sleep physicians do not even consider the potential value of a retitration study as an important component of the work-up to determine the cause of the persistent daytime and nighttime symptoms.
Thus, from our vantage point of trying to “sell” this research to sleep professionals who might be unwilling to define CPAP failure with greater precision, we knew our efforts would be wasted unless we could unequivocally convince reviewers of the absolute certainty our patients had failed CPAP.
To be clear, we think someone using CPAP regularly for many hours per night yet reporting poor outcomes is experiencing CPAP failure. This scenario does not mean we are not attentive to other possible causes of the problems, for example, two of the most common co-occurring issues observed in these types of poor responders would be depression or leg jerks. In various ways, we look for both subjective and objective evidence for these conditions in a large proportion of second opinion, failing patients, because our sleep center specializes in the treatment of mental health patients with sleep disorders. Nevertheless, while these problems emerge as a key factor in CPAP failure in a small proportion of cases, actual CPAP failure caused by none other than “CPAP Failure” is in fact the leading cause of…you guessed it…CPAP failure!
Sorry, just trying to emphasize a point that remains largely ignored in our field, namely CPAP does not really work as well as many sleep professionals profess, and perceptions that mislead sleep docs into thinking CPAP works well are often based on their lack of experience in trying out other modes of PAP therapy in exactly the kinds of patients who report obvious residual symptoms.
By 2013, we treated more than 2000 second opinion patients and spent the next several months sorting them into different categories of CPAP failure developed using our timeline of failure. During this exploration, we realized a majority of patients were currently attempting CPAP but with either poor results or constant struggles to adapt or both. Because these patients might be using the device—at any level of use—we sensed it might be too difficult to explain their failure problems to sleep experts reviewing our research manuscript. To solve this possible confound, we selected only patients with the most extreme or obvious forms of CPAP failure, which included patients who originally underwent a titration at another sleep center, during which CPAP was attempted, but then:
In our minds, there was no confusion about such patients designated as CPAP failure, yet we continued to polish our manuscript during the next three years in ways that informed us of new barriers we would face in trying to publish a paper on the REPAP protocol. Over that time, we predicted that two large obstacles would hamper our efforts, and these barriers related to how other experts and clinicians in the field of sleep medicine would view our aggressive attempts to titrate RERAs as well as our use of the sleep lab to manually titrate the auto-adjusting technology of ABPAP and ASV. Despite our small attempts to head these problems off at the pass, we would learn later once we completed our first attempt at a submission that reviewers could not grasp or simply rejected both ideas about treating RERAs and manually titrating ABPAP and ASV.
Although our foresight was vindicated in the long-run, in the near term we decided to directly attack these two obstacle with projects that would ultimately strengthen the REPAP manuscript, one a very in-depth study on the neglect of RERAs manifesting in the Journal of Clinical Sleep Medicine (JCSM) and the other a brief commentary explaining several factors that prevented past researchers from demonstrating the value of auto-adjusting PAP modes.
In the first effort, we examined seven years-worth (2006 to 2012) of original research papers in JCSM (219 that provided objective data on sleep breathing disorders) to determine whether or not other sleep researchers were holding up the AASM standard to score RERAs on the diagnostic or titration studies. (2) In the study, we showed how few published papers were adhering to these AASM recommendations as only 36 of 219 paper reported RERAs, and only 25 of 219 included the RERAs in the respiratory disturbance index (RDI). Moreover, another sign of high quality research is whether or not a paper acknowledges important omissions in a study design or in results to provide insight to improve future investigations; but, in the 179 papers that did not reports RERAs or RDI data, 157 made no mention of their neglect in the limitation sections.2
These proportions (< 20% use of RERAs or RDI; >85% failing to acknowledge the omission) were striking and indicated very weak endorsement of AASM recommendations described between 1999 and 2005. That RERAs were ignored by leading researchers in the field of sleep medicine, despite all the evidence compiled by the AASM indicating the essential need to treat RERAs, demonstrated the vast majority of sleep researchers publishing in JCSM were working at sub-standard levels of care in the diagnosis and treatment of OSA/UARS. The publication of our paper led to a commentary on the topic by Dr. Nancy Collop who in nutshell called out Medicare for not addressing this issue accurately and stated that regardless of what name is given to these more subtle breathing events, they need to be scored and titrated.3 Last, in a follow-up commentary in response to Dr. Collop and another researcher’s letter to the editor,4 we reiterated how failing to attend to RERAs in a titration sleep study would most likely lead to residual daytime symptoms in most sleep-disordered breathing patients.5
The second effort was also a commentary, and it specifically addressed the work of Powell et al in 2012, “A Pilot Study Assessing Adherence to Auto-Bilevel following a Poor Initial Encounter with CPAP,”6 which claimed ABPAP was no better than CPAP in attempting to improve adherence in patients doing poorly with CPAP. This paper represented a perfect opportunity for us to highlight the flaws in this type of research, which employed only the auto-mode of therapy instead of manually titrating the device in the sleep lab. In the same volume of JCSM that published their paper, the editor and colleagues, Quan et al, authored a commentary about the Powell et al’s paper, concluding that routine use of an ABPAP device was not indicated,7 yet Quan et al completely ignore the two primary flaws in the study design: 1) the lack of manual titration of ABPAP; and 2) the failure to address residual RERAs. The most telling quote from Quan et al indicated a strong aversion to technology solving CPAP adherence problems: “Finally, such negative device trials [auto-adjusting devices, etc.] suggest that any improvement in PAP adherence rates will more likely occur by addressing social and psychological obstacles to usage rather than newer technical innovations in PAP delivery or interface design.”7
In response, our commentary published a few months later, “Driving on ‘auto’: Hands-on is more effective than Hands-free,” we explain the flawed premises in both Powell et al and Quan et al works.8 We highlighted the inappropriate use of auto-mode without sleep technologist intervention as poor quality because it does not meet standards of AASM titration guidelines. Further, we explained how this approach fails to titrate out RERAs and prevent expiratory pressure intolerance. Last, we explained only through a manual titration overriding the auto mode can such approaches achieve successful outcomes.
With our background papers published in 2012, we returned to the drawing board to prepare our best effort at a new manuscript to describe the REPAP process, and the next three years proved highly instructive as we went through various rejections and appeals in two leading sleep medicine journals before landing a home for the manuscript in a respiratory journal in 2016.1 That story will follow in the next post.