Thirty-second polysomnography epochs showing normal breathing (a), upper airway resistance syndrome (UARS) (b), hypopnea (c), obstructive sleep apnea (OSA) (d), central sleep apnea (CSA) (e), and expiratory pressure intolerance (EPI) (f)
Common medical comorbidity associated with either PTSD or OSA
Defined theories on clinically relevant relationships between OSA as well as UARS and PTSD appear to have been introduced into the scientific literature in 1998. In a study involving 156 female sexual assault survivors with nightmares and PTSD , 52% reported the combination of symptoms of snoring and daytime sleepiness, meeting screening criteria to test for sleep apnea based on the standards of the American Sleep Disorders Association (subsequently American Academy of Sleep Medicine) . At the time, we suggested that the concept of posttraumatic insomnia appeared to be a marker for much more complex and intrinsic sleep disorders in this cohort of PTSD patients. Using a more in-depth analysis of symptom reports, we discovered a strong potential for sleep breathing and/or sleep movement disorders among these sexual assault survivors complaining of nightmares and insomnia. When we compared those patients with these likely sleep physical disorders to those without such symptoms, the presumptive sleep disorders groups correlated significantly with PTSD symptom severity. We speculated that treatment of physiological sleep disorders might be associated with improvement in PTSD severity . And, in that same year, Youakim, Doghramji, and Schutte reported a single case of OSA in a PTSD patient, whose PTSD symptoms remitted when treated with positive airway pressure therapy (PAP-T ) . Earlier works had reported objective data on sleep breathing events in PTSD patients without suggesting a broader clinical relevance to posttraumatic stress [25, 26]. And some works had reported relationships between partner-assessed snoring or objective sleep apnea and posttraumatic anxiety dreams [27, 28].
From 2000 to 2002, our research team published six articles on the specific topic of sleep-disordered breathing and PTSD, developing a number of hypotheses as well as presenting key objective data [22, 29–33]. The main thrusts of these works showed that among female sexual assault survivors with nightmares and posttraumatic stress symptoms or diagnoses :
Sleep breathing symptoms were surprisingly common, often present in greater than 50% of the sample [22, 30, 31, 33].
Presumptive sleep breathing disorders presented more likely an insomnia disorder rather than classic sleep apnea [22, 30, 31].
Sleep-disordered breathing symptoms and disorders were associated with worse psychiatric distress [30, 31, 33].
Treatment of sleep breathing disorders in PTSD patients was associated with decreases in nightmares, insomnia, and posttraumatic stress .
The most important work from our initial effort to describe this phenomenon included a study objectively testing 44 consecutive crime victims with posttraumatic stress who were seeking treatment for nightmares and insomnia; 40 of the 44 patients were diagnosed by polysomnography (PSG ) with OSA or UARS . From these results, we coined the term “complex insomnia ” to describe patients who present with insomnia as their primary complaint while also suffering comorbid and usually covert sleep-disordered breathing . Subsequent to this work, we completed a review article to describe our hypotheses about the surprisingly high rates of sleep breathing problems among trauma survivors, which we were observing both in research and clinical practice . In particular, we hypothesized about the possibility of a bidirectional pathway (Fig. 21.2) in which posttraumatic stress creates sleep fragmentation, which is known to adversely impact the human airway and increase vulnerability to the subsequent development of OSA or UARS  and how sleep breathing problems, in turn, might worsen PTSD symptoms through further sleep fragmentation .
In the ensuing decade since the publication of these works, there has been a slow but steady increase in the interest in this unusual relationship. This chapter will review and highlight extant literature to provide current information about assessment and prevalence, potential clinical relevance, treatment options, and recommendations for future research on the relationship between PTSD and sleep-disordered breathing.
Two entrenched paradigms have diluted research on the relationships between sleep breathing disorders and PTSD, and both reflect problems of conventional wisdom inhibiting the adoption of new perspectives and with technology to assess trauma survivors with sleep complaints . Foremost among these barriers is the predominant view that PTSD is strictly a psychological or psychiatric disorder [36–38]. Despite efforts several years ago to introduce concepts about sleep and PTSD into the scientific literature [39–42], a clinical sleep medicine perspective on PTSD has only emerged during the past decade [43–55]. But these efforts have not pervaded the psychiatric literature, where most PTSD research excludes any primary interest in sleep, and only rare mention is made of sleep breathing problems. In fact, if one completes a Medline search with the single term, “PTSD,” the results generate more than 20,000 citations, but if the terms “PTSD and sleep” are used, there are fewer than 1000 entries; and of these thousands, the vast majority do not assess sleep as a primary variable of interest.
More specific to our discussion is the second entrenched paradigm that adversely influences research on sleep and PTSD in which diagnostic PSG is conducted rarely even when sleep is a variable of interest. When PSG was performed, it was unusual that breathing was formally assessed. Finally, even when breathing was assessed, not until 2001 was a study conducted using the now current standard technology in the field, known as nasal cannula pressure transducer . This gap in technology produces a substandard level of assessment—based on the nosology of the American Academy of Sleep Medicine [56, 57]—which is best exemplified in Breslau et al.’s widely cited study on objective sleep findings in patients with chronic PTSD . This study measured breathing solely with thermistor technology, a qualitative tool that can only measure one (apnea) of the three common types of obstructive breathings. Yet, 5 years prior to the Breslau study, the American Academy of Sleep Medicine gave thermistors a “D” grade for its capacity to measure hypopneas accurately, and it was incapable of measuring RERAs, both of which require pressure transducer technology [56, 57]. Another unusual element of the Breslau et al. study was the use of the threshold AHI (apnea-hypopnea index) of ten or greater, which is double the standard set in the field of sleep medicine of five events or greater . The authors reported only 7–10% received such a diagnosis (Breslau 2004). Yet, based on our research 3 years earlier, which used the standard sensors for respiratory assessment, we found that 91% of PTSD patients were diagnosed with either OSA or UARS. Thus, the absence of appropriate technology in Breslau’s study would appear to have vastly underestimated both hypopneas and RERAs, leading to a clear-cut underestimation in the number of patients diagnosed with sleep breathing disorders in their PTSD sample.
This barrier to assessment persists. As recently as 2007  and 2010 , two research studies using polysomnography in PTSD cohorts did not adopt the appropriate technology to measure hypopneas and RERAs, and both reported the absence of breathing disorders in their samples, an unlikely possibility in light of other research using the appropriate technology.
Prevalence and Clinical Relevance
Prevalence rates for the comorbidity of SDB and PTSD are difficult to discern as the literature reflects highly mixed results. However, there is a general agreement that individuals with PTSD have a disproportionately higher rate of SDB as compared to the general population . In a study of elderly war veterans (n = 59) with and without PTSD, the PTSD group reported poorer sleep quality, but PSG showed minimal differences. However, in the recruitment for this study, many in the PTSD group were excluded due to positive screens for sleep apnea . In a sample of 156 female sexual assault victims with PTSD, Krakow and colleagues (2000) hypothesized that 52% suffered from some form of SDB , but this study was limited by relying on self-report only.
Several uncontrolled studies utilizing objective measures of SDB also support increased prevalence of SDB in PTSD patients. In a sample of 105 Vietnam-era veterans with PTSD, 69% had an apnea-hypopnea index >10 despite the fact this study used only thermistor technology . Dagan and colleagues found that 13 out of 24 combat veterans diagnosed with PTSD had sleep apnea . In a study of 44 crime victims with PTSD or clinically significant PTSD symptoms, 55% were found to have OSA and 45% were noted as having UARS .
To date, two controlled studies have used objective measures of SDB in patients with comorbid PTSD. In one study, nearly half of combat veterans with PTSD compared with 13% of a nonclinical sample were found to have apneas and hypopneas, although index scores were <10 . In a similar study, ten healthy controls were matched against ten combat veterans diagnosed with PTSD. Although no apneas or hypopneas were observed, more frequent arousals were present in the first half of the night in association with decreased slow-wave sleep in the PTSD group, which may indicate a potential for RERAs in this protocol that used only thermistor technology .
And from a different perspective in a review of the Veterans Health Administration healthcare database, a higher rate of PTSD was found in a group with sleep apnea compared to a group without sleep apnea (11.85% vs. 4.74%). However, similar findings were noted for mood and anxiety disorders , psychosis, and dementia indicating a potential general association between psychiatric disorders and sleep apnea . In a similar type of study, Raper et al. found a PTSD rate of 28% among combat veterans with sleep apnea .
The comorbidity of PTSD and SDB is more than just an academic issue. There are various clinical implications. Past studies have shown that both SDB and PTSD lead to a variety of physical health consequences (Table 21.1). For example, the medical literature provides clear and overwhelming evidence that SDB is linked to hypertension, dyslipidemia, and insulin resistance [8, 19]. OSA in particular is believed to put the individual at greater risk for diabetes and cardiovascular disease . Similarly, individuals with PTSD are more likely to suffer from diabetes, noncirrhotic liver disease, angina pectoris, gastritis, arthritis, and other various physical ailments [7, 68].
Less clear in the literature is the combined impact of comorbid PTSD and SDB on physical health and functioning. Although there are virtually no data to support the idea that physical health problems are exacerbated in individuals with comorbid PTSD and SDB and vice versa, one can reasonably assume, based on studies of depression and chronic pain  and stress and cardiovascular disease , that a bidirectional process may occur leading to increased symptom severity, decreased global functioning, and less than robust treatment effects.
One study indirectly addressed this issue in 187 sexual assault survivors with posttraumatic stress symptoms (majority with PTSD) who were divided into two groups: those with likely sleep apnea (n = 168) and those unsuspected of having a sleep breathing disorder (n = 19) . Of the 168 suspected SDB cases, all 21 participants from this group who had completed objective tests were confirmed to be suffering an SDB diagnosis. Thus, as all their demographic, sleep, and psychiatric histories were no different than the remaining 147 who did not complete objective testing, an assumption was made that likely all 168 from this group of suspected cases suffered from sleep breathing disorders. When using the SF-36 health scales  to compare the SDB group with the 19 women for whom SDB was unsuspected, all eight physical and mental health scales showed significantly worse outcomes (lower scores) for the SDB group with medium to large effects (Cohen’s d):
Physical functioning: 67.53 (27.51) vs. 83.42 (16.92); d = 0.59
Role physical: 40.06 (41.09) vs. 65.79 (40.15); d = 0.62
Bodily pain: 47.47 (26.09) vs. 63.37 (19.27); d = 0.61
General health perception: 47.69 (25.03) vs. 69.95 (18.58); d = 0.88
Energy/vitality: 26.63 (19.02) vs. 50.53 (21.53); d = 1.16
Social functioning: (%) 48.87 (24.84) 67.76 (26.46); d = 0.74
Role emotional: 21.28 (30.74) vs. 52.63 (39.00); d = 0.95
Mental health: 57.86 (12.89) vs. 64.63 (9.45); d = 0.52
Whether separate processes are involved (e.g., high levels of cortisol in PTSD; sympathetic nervous system activation in SDB) leading to increased physical health problems or some unknown syngergistic effect, future studies will need to be undertaken to address these questions. Certainly, the interesting comorbidity associated with either OSA or PTSD warrants further research (Table 21.1).
Treatment Adherence and Outcomes
Treatmen t of OSA/UARS among PTSD patients follows standard practice parameters in so far as the use of evidence-based medical technology or procedures , but there are scant data on samples of trauma survivors undergoing treatment for sleep apnea.
In a case-controlled prospective design, El-Solh et al. (2010) showed at 30-day follow-up, PAP adherence was significantly lower in a PTSD group compared to a control group (41% versus 70%). Greater sleepiness at intake appeared to promote adherence, whereas more frequent reports of nightmares diminished adherence. Mask comfort and claustrophobia were also reported as barriers to care as well as too little or too much air pressure . In a study by Means and colleagues, black veterans with mental health diagnoses (of which nearly 60% included those with PTSD) showed significantly worse adherence to CPAP than white veterans with similar mental health diagnoses . In a randomized controlled study by Lajos et al. (2004), there were no differences in CPAP adherence between the PTSD and non-PTSD group, but anxiety was the strongest predictor of difficulty with CPAP adaptation . In another study, adherence rates measured by number of nights was not different between PTSD patients and controls, but the actual number of hours of PAP usage was markedly lower in the PTSD group .
Adherence rates and outcomes data are also limited in scope and duration, but preliminary studies point to potential problems among trauma survivors in learning to adapt to the device and possibly receiving suboptimal responses to standard PAP therapy. Like all PAP therapy patients, there are always a series of adjustments involving mask fit and comfort, adjusting to pressurized airflow, and global acceptance of learning to sleep with a cumbersome and foreign device attached to one’s face all night long for years on end. See Bollig (2010) for an outstanding review on the factors that contribute most to the barriers and enhance the compliance with PAP therapy in the general population .
In PTSD patients, mask-triggered claustrophobia is a potential barrier to treatment, but in recent years, the technological expansion in the types of masks along with specific comfort enhancements has created opportunities for virtually every patient with rare exception to find a usable, well-fitting, and comfortable mask. There is little research conducted about masks, perhaps due to such rapid changes in this essential component of PAP therapy, but the general consensus or research findings are as follows:
Full face masks alleviate mouth breathing.
Chinstraps may alleviate mouth breathing in patients using nasal masks; and in some cases chinstraps help patients using full face masks.
Nasal pillow masks (inserted into nostrils) may yield the most comfort by the limited contact with the facial surface.
Various cushions and pads now marketed for use between the face and mask may further alleviate irritation or sores due to masks tightened down onto the skin, but these same tools may increase air leak.
In our clinical experience, we have found that the most distinctive adaptation problem for PTSD patients in particular and those with other forms of anxiety in general is the problem of expiratory pressure intolerance (Fig. 21.1) [78, 79], wherein the patient tends to either report direct discomfort when breathing out against fixed pressure devices (CPAP)  or engages in attention amplification when first learning to use CPAP and thereby develops an adverse psychosomatic response to the sensation of pressurized airflow (almost exclusively on exhalation) [78–80]. This sensation usually can only be eliminated by either a change in pressure delivery mode (e.g., dual pressure devices that utilize lower pressure on expiration)  or specific distraction coaching to learn to divert attention away from the sensation itself [78, 79].
To address many of these issues in our insomnia or anxiety patients, we developed the PAP-NAP procedure , a daytime desensitization or “test-drive” of the PAP device, which provides a less stressful experience to initiate usage to patients with greater complexity . Our original paper on the PAP-NAP was conducted on a sample of chronic insomnia patients with elevated psychiatric distress and sleep-disordered breathing, who showed a much greater percentage of adherence to the PAP device compared to a historical control group of insomnia patients with SDB who did not undergo a PAP-NAP. The PAP-NAP is conducted without most of the typical sensors of polysomnography, but instead it focuses on having the patient try out various styles of masks and most importantly try out various modes of pressure delivery, which in most situations leads to a more comfortable or palatable introduction to this unique therapeutic modality. By working one on one with a sleep technologist for upward of 3–4 h, the vast majority of patients build self-efficacy for having worn the mask for an hour or two and experienced pressurized airflow for a similar time interval .
Medical and psychiatric comorbidity is another area of concern about PAP therapy, because there are so many potential ways in which another health condition can adversely interfere with the use of the device . In fact, this comorbidity argues against the unrealistic position taken by Medicare policy-makers who imagine, quite fancifully we would add, that there exists something called “the typical sleep apnea patient.” Such patients are extremely rare when the issue of comorbidity is appreciated. Consider just a few anecdotal examples from a potential list of hundreds of co-occurring conditions :