Select Page

The popular trade magazine Sleep Review: The Journal for Sleep Specialists recentlypublished a story on the symposium last year in Bologna, Italy at the European Sleep Research Society meeting where Christian Guilleminault, Erla Björnsdóttir, Leon Lack and I each presented on the topic of comorbid insomnia and sleep-disordered breathing (aka “complex insomnia” or COMISA). The story was written by Alexander Sweetman, a graduate student working with Dr. Lack, and covers pragmatic issues regarding not only the connection between the two disorders, but also raises important questions on the ordering of treatment for these conditions.

Two key themes throughout Sweetman’s piece involved the extent to which PAP therapy might favorably influence different types of insomnia, and as a corollary, at what point in treatment should CBT-I (cognitive-behavioral therapy for insomnia) be offered to patients with co-morbid OSA/UARS. Some of the interesting points put forward in the presentations as well as during the Q & A following each speaker strongly indicated that CBT-I could almost always be offered early on in the treatment of a complex insomnia patient, the effects of which might actually improve the insomniac’s subsequent motivation to attempt PAP therapy. The research on this point was evident in showing that patients with severe OSA will experience clinically meaningful treatment gains for insomnia severity by applying CBT-I strategies prior to attempting treatment for sleep-disordered breathing. On the other side of the equation, the question was also raised as to how much more insomnia treatment (e.g. CBT-I, hypnotics) would be required for residual insomnia symptoms after successful application of PAP therapy.

Reading Sweetman’s article reminded me of several cases in the past few years where the timing of CBT-I implementation arose in complex insomnia patients and how circumstances and patient preferences often dictated the ordering of steps. To highlight the complexity of this issue, I will use a composite patient case description (i.e. merging details of several patients) to paint a comprehensive picture.

The patient is a middle-aged perimenopausal woman, who initially presented to us three years ago with complaints of severe insomnia, which manifested in classic ways with prolonged time in bed not sleeping, racing thoughts at bedtime, anxieties and fears about her sleep, hypnotic dependency, and excessive time monitoring behavior, the latter of which resulting in intense frustrations when lying awake in the middle of the night. She had been struggling with insomnia for nearly twenty years and once visited a sleep center that offered no CBT-I resources; instead, she was told to follow sleep hygiene instructions and was informed a sleep study would serve no purpose. For much of these 20 years, the patient was largely convinced her insomnia was a secondary effect of her anxiety and depressive disorders for which she used antidepressants and anxiolytic medications on a regular basis. Yet, neither these medications nor her benzodiazepine hypnotics solved the chronic bouts of sleeplessness often leading to more than 2 hours of wake time after initially going to sleep. In addition, the patient saw multiple therapists during this lengthy period, but while mental health gains were achieved, insomnia persisted and mild suicidal ideation was reported sporadically.

With some desperation the patient presented to our center, because her psychiatrist recently became aware of our work with severe insomnia cases and how we often found underlying pathophysiological problems that sometimes fueled the unwanted sleeplessness. The patient presented using two different benzodiazepines (Lorazepam, Clonazepam), both at low dosages at bedtime and reported being unable to fall asleep without the meds. She reported no symptoms of restless legs syndrome at bedtime or leg jerks during the night. Mild snoring was the only overt breathing symptom reported, but she occasionally to frequently complained of dry mouth in the morning and multiple episodes of nocturia each night.

With this background, we knew her risks were high for OSA/UARS, a finding we described in two papers published in 2010, both of which dealt with samples of insomniacs failing hypnotic meds at presentation to our center. The first paper appeared in the Journal of Nervous and Mental Disease and the second appeared in Primary Care Companion to the Journal of Clinical Psychiatry. Both articles demonstrated between 80 to 90% of these types of patients suffered underlying OSA/UARS, a condition that surprised nearly every patient in the samples studied. Moreover, one study specifically examined how insomnia was evaluated prior to referral to the sleep center. Again, the evidence was overwhelming that primary care physicians were not addressing insomnia problems with anything more than pharmacotherapy and rarely were discussions held regarding how to locate a CBT-I therapist or the necessity for polysomnography to assess for physical causes for insomnia.

We discussed our approach to insomnia with this patient and pointed out two pathways: the value in starting with CBT-I to overcome psychophysiological conditioning or PAP therapy to address the underlying nonrestorative component of her potential sleep-disordered breathing. Not only was the choice a slam dunk to rapidly target PAP, but the patient insisted she was not willing to stop her anxiolytic/hypnotics no matter what treatment she was provided. In her mind, PAP therapy sounded very expedient if in fact it was proven she suffered from a sleep breathing problem.

At this point the patient articulated a recurrent theme we had been noticing among COMISA patients dating back to the late 1990s. The perspective involved the strong belief that despite all the talk about psychological aspects of insomnia, many patients “felt” or “sensed” something physical was wrong with their bodies. In other words, in contrast to the obvious waking stress-related factors of anxiety or racing thoughts that plagued her along with nearly every other insomniac we encountered, a different insight was in play, though not readily accessible, indicating a belief in some other factor(s) as the destroyer of sleep quality. Now, this perspective is not necessarily a surprise in that virtually all these patients present with nonrestorative sleep or what we refer to as “bad and broken” sleep. And, therefore, it is also not a surprise when these patients express a strong feeling of resonance when they hear a sleep-breathing disorder might underlie their insomnia and an even stronger expression of relief once formally diagnosed.

As an aside, we ask a specific question on our online intake as to whether or not a person believes his or her sleep is broken, and among these types of patients greater than 90% respond affirmatively. It is true the wording of the question could simply indicate a patient’s perception of “sleep is broken up” during the night as an indicator of insomnia, but in our discussions we find the term also extends in some individuals to this concept of something feeling wrong or unhealthy with sleep itself.

The patient completed her overnight PSG and was diagnosed with a combination of moderate OSA and UARS events with only occasional oxygen desaturations during the minimal REM sleep she generated, probably due to her benzodiazepines or perhaps REM suppressant effects of her antidepressant. The study also showed a great deal of sleep fragmentation involving frequent bouts of wake time lasting 5 to 10 min or more, during which the patient showed some limb movements, perhaps suggestive of RLS since she was not sleeping at the time. The patient then returned for a retitration study, after which she was initiated on PAP therapy and was a regular user of the device for 3 years, during which she reported some benefits in both her insomnia and quality of sleep, but she was persuaded her results were not optimal.

The lack of optimal results led to subsequent follow-up clinic appointments and two additional retitration studies. At each encounter, the patient was encouraged to consider whether her continued use of benzodiazepines was interfering with the depth of her sleep and possibly decreasing the amount of time she spent in REM sleep. She was informed again that CBT-I might be a useful strategy if she elected to decrease or eliminate these hypnotic drugs. Instead, she focused on the retitrations, and with each tweak of her pressure settings coupled with aggressive attempts to manage persistent mask leak and mouth breathing issues further gains for insomnia and sleep quality were achieved, yet the patient continued to believe the response should have been better.

Two additional key observations manifested in this patient’s case. First, in her follow-up retitrations she continued to show scattered leg movements, but again only while awake, and still the patient continued to deny RLS or PLMD; her husband also reported no awareness of unusual leg activity during the night. Second, as noted, at no time during this three year interval was the patient motivated to consider CBT-I or the possibility of tapering off her bedtime benzodiazepines.

A change of circumstances serendipitously led to a change in the patient’s perspective about these bedtime medications. A change of jobs and a new psychotherapist apparently were critical factors in the patient developing a less rigid attitude about her meds, and in discussing things with her primary care physician (the current prescribing doctor), they developed a plan for tapering, which she completed during a two month period, all without her ever attempting CBT-I.

Due to the change in circumstances and the patient’s report of mild deterioration of sleep quality since successfully tapering off the benzodiazepines, the patient was eager return for a retitration to determine whether or not she could in fact obtain a deeper quality of sleep in the form of longer REM periods and more total REM sleep. Instead, the next study in fact revealed what we had been missing all along. The patient showed marked problems with periodic limb movements, now for the first time showing up frequently while asleep and clearly with disruption to sleep quality in the form of both arousals and frank awakenings.

While more REM sleep emerged due to the absence of the benzos, the findings also helped us understand the greater complexity of her case. All this time, including before she visited any sleep center and during her visits with us, everyone had missed the diagnosis of RLS/PLMD. Instead, somewhere at the outset, a prescribing doctor diagnosed anxiety and insomnia and placed the patient on benzodiazepines. Because they were marginally effective for insomnia and leg jerks, the patient maintained use for 3 years. Once the medications were removed from the treatment regimen, the leg jerks were unmasked, and the patient experienced more REM sleep.

Clinically, the patient reported sleeping better on this study, perhaps due to greater REM sleep, but she was not persuaded things were suddenly optimal. When she learned about the leg jerks, she was very interested in proceeding with treatment to clarify whether or not this movement problem was indeed the missing link.

In similar cases, a patient starts a leg jerk med and finally achieves the near optimal to fully optimal response he or she had so long had hoped for. One clinical take home message here is that if we had been able to persuade the patient to go directly into CBT-I as her first treatment, there’s a possibility she would have tapered off the medications sooner, which then would have unmasked the leg jerk problem much sooner. We anticipate that future research studies will be heavily focused on these complex insomnia cases, because it is so apparent these individuals have an assortment of pathways to choose from in starting their care. Hopefully, research will guides on predicting the paths to match each type of complex patient.



  1. Start with SR review of ESRS to segue into the complexities of starting with CBT I or PAP and describe the hegemony case as extending over 2 to 3 years.