Sleep Medicine is Not Rocket Science

One of the very odd things about the field of sleep medicine is its alleged reliance on metrics deemed to be quantitative when in fact so many measurements in our evaluation of sleep are more aptly described as qualitative. Nowhere is this conundrum more evident than with the apnea-hypopnea index (AHI) or the respiratory disturbance index (RDI). Neither of these metrics designates much beyond the diagnosis of OSA or UARS with the exception of those cases that are truly in the severe category. Even then, what does it mean scientifically or clinically when a patient exhibits an AHI> 40 events/hour, constantly desaturates oxygen levels into the 80% or 70% range, and yet reports absolutely no symptoms, no sleepiness, no hypertension, no cardiac disease? Although this combination is not a frequent occurrence, it is certainly not rare among cohorts of elderly men. What, then, in this case does the quantitative metric define? The answer would appear to be nothing at all. 

For example, would we really subject an 80 year-old man to PAP therapy if he were to present with the asymptomatic AHI>40? If so, what would our reasoning be—maybe prevent strokes, heart attacks and high blood pressure? If that argument were accepted, would we offer the patient the option of starting blood thinners, beta-blockers, and calcium channel blockers instead of PAP therapy? As I described in a recent post, some of these patients really do suffer from some symptoms, but only after they use PAP are we likely to sort out the nature of the symptoms. Yet, as implausible as it may seem, most sleep doctors have encountered severe OSA patients without any symptoms whatsoever, which raises important questions about what we think we are measuring in these patients.

By comparison, consider a standard metric of diabetes, hemoglobin A1C, which by measuring the glycosylation of hemoglobin molecules we come very close not only to gauging how well blood sugar is controlled, but also we can predict the progression of diabetes-related morbidity that may ensue. If the A1C is 8 or above, expect trouble over the long and short-haul. Establish and maintain an AIC in the 6 to 7 range and expect far less diabetic-related complications.  The A1C is a proven quantitative metric, so much so many sleep researchers and clinicians for a good many years have been clamoring for the OSA equivalent of the A1C. Clearly AHI and RDI are not such a metric.

Take the single most problematic circumstances when using these breathing event indices, the so-called cases of “mild OSA.” Due to their poor adaptability to PAP, documented by some research, sleep doctors are dissuaded from recommending attempts at the device. Yet, rarely is there discussion on the potential harm to these patients for not using PAP. The common refrain to this cohort is “why bother?” Many sleep facilities do not even acknowledge mild OSA as a diagnosis. Instead, they write off the patient and discourage him or her from returning for the titration study. This problem is an inglorious example of a physician committing a cardinal sin of medical practice—treating the number and not the patient.

Twenty-one years ago, when I first learned about UARS, before the advent of the nasal cannula pressure transducer, I started noticing cases where objective findings on the breathing signal did not match subjective complaints. The thermistor tracing was not normal but then again criteria were not met for hypopneas. If, however, you looked closely at sleep architecture, sleep fragmentation was pervasive despite the lack of obvious, scorable breathing events. So, we put these patients on PAP, the airflow and sleep consolidation improved, and most importantly, these patients reported decreased daytime fatigue and sleepiness. Later, as the technology advanced, it was obvious these patients suffered UARS.

Nowadays, many UARS cases are not even scored, because apparently a sizeable proportion (percentage unknown) of sleep centers do not report the RDI, which means they are not scoring RERAS, which means they are ignoring the full extent of the sleep breathing condition in such patients. Instead, a case of this sort where the AHI is also low mild OSA is the diagnosis, despite the patient presenting with severe sleepiness, a history of drowsy driving, and cognitive impairment. Such patients may then be diagnosed with idiopathic hypersomnia and placed on stimulants. Contrary to the conventional wisdom, as described in several posts, most of these patients will tolerate an advanced PAP device such as ABPAP or ASV and achieve dramatic improvements in their outcome measures. In such instances, the true severity of the case should be scrutinized by the change in symptoms. If we eliminated the problem of drowsy driving, would we still be thinking about the case as one of mild OSA? Seems doubtful. As we all know, it is a very big deal when someone attains the capacity to stay awake behind the wheel and thus becomes a significantly safer driver. If PAP therapy yields this reversal of sleepiness, it is difficult to imagine anyone defining the original problem as mild. 

Making matters worse are insurance carriers and sometimes other non-sleep, medical specialists that demand an AHI or RDI as a statement of severity level without any interest whatsoever in the clinical picture of the whole patient. This incomplete perspective further compromises the patient’s ability to understand the full breadth of his or her sleep breathing disorder, because of the conflicting messages from non-sleep professionals who communicate about breathing event indexes as if sleep gospel. 

In the worst case scenario, I have worked with several patients who showed primarily sustained flow limitations, the class of events in which the airflow signal is flattened for long stretches of time without the typical arousals or awakenings after say 30 or 60 seconds. Instead, breathing is abnormal for 5 to 10 to 15 consecutive minutes before an arousal emerges. The upshot is the AHI or RDI is dramatically and by definition artificially reduced to numbers well below standard diagnostic cut-offs. In adults, it is quite difficult, sometimes impossible, to persuade an insurance company to accept the diagnosis and offer coverage. Yet, when we placed these patients on PAP therapy, they demonstrated normalization of the airflow curve and a marked reduction in their daytime symptoms, particularly large improvements in their sleepiness. These patients are required to purchase the PAP device out of pocket, and most will do so in light of the successful response to the device.   Although I fully accept an insurance carrier has to draw a line on the diagnostic parameters of an illness so as to maintain the financial viability of their business, it remains a serious quirk in sleep medicine that the inexactitude of the AHI or RDI metrics are not more frequently challenged.

The standard should be to develop the broader clinical case report on any patient to reflect a much more accurate picture of the patient’s sleep disorder. Measuring and reporting the AHI and RDI are a requirement of the business model we find ourselves engaged in with the patients, other healthcare providers and insurance carriers. However, I would strongly advise for the need to step back from these metrics from time to time and notice how poorly they reflect on most sleep-related impairment. By doing so, I believe it increases awareness of the broader ramifications of a sleep breathing disorder, which over time should lead to a better and more in-depth treatment regimen.   

Barry Krakow MD


Dr Krakow’s 27 years of sleep research have focused on the complex relationship between physiological and psychological sleep disorders. Dr Krakow currently operates private sleep medical center, Maimonides Sleep Arts & Sciences, Ltd., and serves as Classic SleepCare’s paid Medical Director.

Leave a comment

Comments have to be approved before showing up.