Is positive airway pressure therapy overused?

By Nic Butkov, RPSGT

The potential for overscoring, overdiagnosing, and overtreating sleep-disordered breathing has grown exponentially within the past few years. The use of oversimplified scoring formulas and automated scoring functions has set the stage for inflated respiratory disturbance indices and possible misrepresentation of event etiology. The problem is compounded when positive airway pressure (PAP) titrations are performed in a haphazard manner or when inappropriate PAP modalities are used. The term “complex sleep apnea” has become popular in describing events that emerge during PAP titration or remain unresponsive to PAP therapy. However, complex sleep apnea is a term that has not been adequately defined and has been used indiscriminately.1 The following case review presents an example of a patient who was unnecessarily prescribed with several different PAP modalities, all of which proved to be ineffective.


A 62-year-old female patient was seen in our sleep center for consultation regarding her previous prescription for adaptive servo-ventilation (ASV). The ASV unit had been prescribed on the basis of several sleep studies conducted at other sleep centers. Her initial diagnosis was obstructive sleep apnea with an apnea/hypopnea index (AHI) of 29. She was placed on continuous positive airway pressure (CPAP) therapy and was titrated to 9 cm of water pressure (CWP). After experiencing difficulties tolerating CPAP, she was restudied, at which time her prescribed CPAP level was increased to 14 CWP. Reportedly, she developed central apneas on CPAP, which were described as “being consistent with complex sleep apnea.” This was treated with bi-level PAP, which proved to be ineffective, and subsequently she was prescribed an ASV unit. Upon arrival at our sleep center, the patient was still attempting to use bi-level PAP. She complained that the therapy did not help her symptoms and was seeking advice regarding the use of ASV.

The patient’s medical examination revealed a “talkative, well-developed, well-nourished 62-year-old female with appropriate mood and affect in no distress.” Her height was measured at 62 inches; her weight was 192 pounds with body mass index (BMI) of 35 and neck circumference of 34 inches. She had a moderately large tongue and crowded oropharynx, but no micrognathia, adenopathy, or thyromegaly. Her medical history included possible asthma, fatigue, fibromyalgia, irritable bowel syndrome, hypothyroidism, and gastroesophageal reflux disease. Her prescribed medications included Tenormin and Xanax.

The patient’s sleep history included complaints of numbness sensations in her arms at night and episodes of waking up with tachycardia. She stated that she did not sleep well and was using Xanax as needed for her insomnia. She also complained of daytime fatigue. Her Epworth Sleep Scale score was 7/24.


Figure 1. Example of an artifact that might be misinterpreted as a respiratory event.

Figure 2. Normal respiratory fluctuations during REM sleep.
Credit: Butkov N. Atlas of Clinical Polysomnography. 2nd ed.

Because of the patient’s history and the apparent confusion regarding her various PAP prescriptions, it was suggested that a repeat diagnostic sleep study be conducted without the use of any PAP device. She was instructed to discontinue the use of her bi-level PAP for a period of 2 weeks prior to the study to rule out any carryover effects from the therapy. A standard overnight polysomnogram was performed, which included the recordings of the electroencephalogram (EEG), electrooculogram (EOG), chin and limb electromyogram (EMG), electrocardiogram (ECG), nasal and oral airflow, chest and abdomen motion, snoring sounds, and pulse oximetry. The study was performed in an attended laboratory setting with direct behavioral observation by an experienced sleep technologist. The patient was allowed to sleep according to her normal sleep-wake schedule without constraints regarding wake-up time.

The resulting polysomnogram was visually scored according to standard criteria; however, all respiratory irregularities were carefully examined within the context of sleep-wake physiology before arbitrary scores of apnea or hypopnea were applied. Thus, any perceived “events” that occurred during periods of wakefulness, or any obvious recording artifacts in the respiratory channels or oximetry tracings due to body movement or sensor displacement, were not scored.

The immediate impression derived from a visual review of the raw data was that the patient had difficulty initiating and maintaining sleep. However, when she did sleep, her sleep was well-consolidated with very few respiratory disturbances. The technical report yielded a sleep efficiency index of 59%, based on a total sleep time of 288 minutes within a total recording time of 485.5 minutes. The patient had ample slow wave sleep for her age—nearly 25% of total sleep time; however, her total REM sleep time was reduced to only 5% of total sleep time. She slept in both supine and lateral decubitus positions, with all of her REM sleep occurring while supine. In terms of actual respiratory disturbance, the study demonstrated only eight scorable hypopneas and no obstructive or central apneas. Based on the patient’s total sleep time, this equated to an apnea/hypopnea index of less than 2. The scored hypopneas did result in some mild O2 desaturation, with a nadir of 85%; however, this was nonsustained, and the patient’s overall O2 saturation levels remained above 90% during 98% of her total sleep time. There was no evidence of hypoventilation, Cheyne-Stokes respiration, Biot’s respiration, or any other condition that would suggest the need for bi-level ventilatory support or ASV. Based on her relatively low AHI, this patient did not even qualify for CPAP.

It was concluded that this particular patient’s sleep disorder was not based on physiological abnormalities. Although she did present with anthropometric markers for obstructive sleep apnea, her overall AHI was sufficiently low and the results of her sleep study did not warrant any type of PAP therapy. The patient was advised that her daytime symptoms of fatigue were likely unrelated to sleep-disordered breathing. Her diagnosis was sleep maintenance insomnia, and our recommendation was to pursue insomnia counseling, with emphasis on sleep hygiene.


The results of the sleep study raise the question of why this patient was prescribed with CPAP, then bi-level PAP, and finally with adaptive servo-ventilation. Without access to the original raw data, one can only speculate on the reasons. It may be argued that differences in the AHI can occur due to night-to-night variability. However, given such a wide discrepancy, this is unlikely. A more probable explanation is that the original study was overscored. The fact that the patient did not respond to PAP therapy further suggests that the patient’s symptoms were not related to obstructive sleep apnea.

One can also speculate that the patient was overtitrated with CPAP, leading to an increased incidence of central apnea. According to previous reports, her CPAP was initially titrated to 9 CWP, but later increased to 14 CWP. The reason for such a wide range in prescribed CPAP levels remains unclear. It is not surprising that the patient developed central apneas; this phenomenon has been described2 and is likely to occur at higher PAP levels. However, according to her subsequent reports, the finding of central apneas on CPAP was instead interpreted as “complex sleep apnea,” which then led to further intervention attempts in the form of bi-level PAP and ASV.

In retrospect, a great deal of time and cost could have been saved by a more conservative and rational approach to the patient’s sleep disorder. An in-depth review of the patient’s sleep habits, medication use, and possibly a reevaluation of her initial sleep study may have yielded a different outcome and circumvented much frustration for the patient. However, the scenario described in this case is not unusual and appears to be part of a growing trend as sleep medicine continues to move toward oversimplified, rule-driven protocols that emphasize rapid treatment assignment in lieu of careful diagnosis.


During the formative years of sleep medicine, progressive sleep centers continually refined their approach to scoring and interpreting the polysomnogram. This especially pertained to the analysis of sleep-disordered breathing. Based on a growing bank of knowledge and experience, technologists and clinicians learned to differentiate between clinically significant events, normal physiological variants, and recording artifacts. Polysomnograms were visually scored by viewing the raw data in a large uncompressed format, which allowed the scorer to discern essential detail and subsequently interpret all events within the context of sleep-wake physiology. Although the scoring process has always been based on specific scoring conventions, it was well-tempered by sound clinical judgment and common sense.

In contrast, today’s scoring methods are increasingly based on rigid rules, which do not take into account normal physiological variants and recording artifacts. The respiratory scoring rules presented by the American Academy of Sleep Medicine (AASM) scoring manual3 are based almost exclusively on respiratory data and oximetry, without reference to other physiological measures. There is only brief mention of arousal in the context of the alternative hypopnea definition (B) and the description of respiratory effort related arousal (RERA); however, neither of these definitions is recognized by current Medicare criteria. The AASM scoring manual does not discuss normal respiratory variants and their possible resemblance to respiratory events, nor does it address the issue of artifacts—a common finding in almost every sleep study (Figure 1). Under these circumstances, inflated respiratory disturbance indices and misinterpretation of event etiology are highly likely, especially in patients who have difficulty initiating and maintaining sleep. Patient movement, sighing, yawning, and other behaviors can be misinterpreted as respiratory events if scored out of context. Other normal phenomena that might potentially elevate the AHI include sleep-onset central apneas, post-arousal central apneas, and respiratory fluctuations that normally occur during REM sleep (Figure 2).

Neither the AASM nor Medicare has provided any concrete statement regarding the use of automated respiratory event scoring—a practice that has never been validated, yet is increasingly used for the analysis of both in-lab and home sleep testing. Inflated indices are possible when automated scoring functions are employed, especially in the context of sleep testing modalities that do not record physiological measures of sleep.

The problem is further compounded by present-day PAP titration protocols. Common problems occur when titrations are performed in a robotic manner without attention to sleep-wake physiology, when patients are overtitrated, when patients are inappropriately fitted with full-face masks (which often necessitate higher pressures), or when patients are arbitrarily placed on alternative PAP modalities without determining event etiology or providing the rationale for doing so. At the present time, there are no standards of practice regarding the use of bi-level PAP for ventilatory support, or for the use of any of the newer PAP modalities. Current AASM and Medicare guidelines for bi-level PAP application simply state that patients who fail CPAP can be titrated with bi-level PAP. Not surprisingly, this has become common practice in many sleep centers. Lacking any substantive parameters in the use of advanced PAP modalities, many sleep centers rely increasingly on industry for guidance in the use of their products, gradually obscuring the fine line between scientific versus market-driven clinical practice.


Scoring rules and treatment protocols are a necessary foundation to foster consistency in patient care; however, our primary goal is to help patients become well, not simply to perform robotic functions. A little extra time spent evaluating each patient’s individual situation, instead of rapidly assigning treatment based on “numbers,” goes a long way toward achieving that goal.

Nic Butkov, RPSGT, is education coordinator at the Rogue Valley Sleep Center, CEO of Synapse Media Inc, and director of the School of Clinical Polysomnography in Medford, Ore.

  1. Malhotra A, Bertisch S, Wellman A. Complex sleep apnea: it isn’t really a disease. J Clin Sleep Med. 2008;4(5):406-8.
  2. Butkov N, Nino-Murcia G. CPAP therapy: the technician’s role. Journal of Polysomnographic Technology. Fall 1987;9:11.
  3. Iber C, Ancoli-Israel S, Chesson A, Quan SF. The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications. Westchester, Ill: American Academy of Sleep Medicine; 2007.