In the last 30 years one of the most prolific fields of research in clinical medicine–and particularly in neurology–has been sleep disorders. Sleep medicine has long established its own specialty with board certification and its own academy. Just in Austin, TX we have at least four sleep laboratories. Most are directed by board certified sleep neurologists who do this work full-time.
When I was a neurology resident at the University of Minnesota in Minneapolis in 1969, one of our neurology professors had very severe sleep apnea with all its clinical features, and he underwent a tracheostomy for his symptoms. He encouraged one of my fellow residents, Dr. Mark Mahowald, to specialize in this field and helped start one of the first clinical sleep laboratories in the US. Dr. Mahowald has been an expert in the field and has published many important clinical papers.
In no way do I (or most of our readers) profess to be sleep experts, but we hear all too often about disordered sleep in our patients from their families and caregivers. We are also aware of the cognitive impairments that can occur in the normal population and especially the elderly, as well as in those already diagnosed with dementia, such as Alzheimer's disease. I have chosen to discuss obstructive sleep apnea only in this article because it has the greatest capacity to produce hypoxia and sleep fragmentation, both of which cause brain dysfunction and major problems in quality of life of our patients, families and caregivers.
What the Data Show
Obstructive sleep apnea (OSA) is a sleep-related breathing disorder in which the upper airway collaps- es repeatedly during sleep, resulting in intermittent hypoxia and sleep fragmentation due to recurrent nocturnal arousals necessary to relieve the upper air- way obstruction. If we use this definition of obstruc- tive sleep apnea and include a defining threshold of at least five episodes of apnea or hypopnea per hour of sleep (AHI >5), and excessive daytime sleepiness, OSA appears to occur in four percent of men and two percent of women.1
Much of what we know about the cognitive consequences of this disorder in the normal population has come from neuropsychologists who have studied this subject in great detail. Beebe, et al.,2 reviewed all the cognitive studies in sleep apnea, both controlled and uncontrolled, up to 2002. They found that general intelligence and verbal/speech abilities were typically unaffected by OSA. The primary areas of brain func- tion consistently affected by OSA were vigilance and executive function. OSA markedly impairs the ability to sustain attention for extended periods, which has a major adverse affect on driving, occupational functioning, and interaction with people. Executive function, which is defined as the ability to develop, sustain and organize goal-directed and flexible approach to problem solving, is moderately to severely impaired. Most case-controlled studies have also shown impaired fine motor coordination and drawing, but very simple motor speed and visual perception appear to be normal. The authors found inconclusive results in regard to memory function, especially visual and verbal short and long-term memory. This important review discussed the clinical implications of these cognitive difficulties. Their conclusion was that if sleep apnea is suspected in a patient, it is important to inquire about vigilance, executive function, and fine motor coordination. Working memory may be impaired because of problems with vigilance and executive dysfunction. They also stated that the standard mini-mental examination and other simple office cognitive tests are likely to miss these cognitive abnormalities, and thorough neuropsychological testing may be necessary. Patients who come to the neurologist's office with cognitive complaints, are then referred for neuropsychological testing for one reason or another, and found to have abnormalities in executive function, working memory, and vigilance, should be screened for sleep apnea.
Verstnaeten, et al., in 20043 studied executive function in sleep apnea specifically. In a well-controlled study, their main findings were slow information processing speed and diminished working memory. They conclude, as others have suggested, that cognitive slowing, working memory impairment (immediate short-term memory), and vigilance decline are not due to damage to the frontal lobe, but likely resembles cognitive decline in people with normal sleep deprivation. This conclusion was supported by the correction of these deficits by continuous positive air- way pressure treatment (CPAP) and different cognitive impairment patterns that have been described in patients with chronic hypoxia from chronic obstruc- tive pulmonary disease (COPD.)
Naegele, et al., in 20054 studied the memory processes specifically in obstructive sleep apnea patients versus controls. There were 95 patients in each group. The study evaluated three separate memory systems: 1.) verbal episodic memory (recollection of specific experiences) 2.) procedural memory (motor and cognitive learning skills) and 3.) working memory (immediate access memory.) The study found that verbal episodic memory was impaired due to a problem with retrieval. The ability to store the information, consolidate, and even encode the memory appeared to be normal. This suggested a possible pre-frontal and subcortical dysfunction. Procedural memory was affected not by the cognitive slowness, but by impairment in fine motor skills. The working memory was also found to be abnormal, especially the inability to maintain both auditory and visual information while simultaneously processing them in one task.
The team also found that other features of working memory, including maintenance and processing of information that is stored in short-term memory and coordinating mental steps and problem solving, were all normal. They concluded in their study that these memory abnormalities were overall mild despite moderate to severe sleep apnea in most of their cases.
Now that we have some idea of the cognitive impairment in OSA, how effective is treatment? The standard recognized treatment for OSA is continuous positive airway pressure (CPAP) delivered by a machine and a facemask. Many studies follow CPAP being used continuously for one week, two weeks, and six months after the diagnosis. Cognitive testing was given both pre- and post-use of a CPAP system in each study.
Bardwell, et al., in 20015 randomized 36 OSA patients in equalnumbers to regular CPAP equipment and placebo CPAP equipment. The placebo CPAP equipment had a mask that had multiple holes so that the air could escape while the mask was on. They studied the cognitive testing results of patients prior to the use of CPAP and followed by a week of CPAP use. On neuropsychological testing, they carefully studied attention, word fluency, vigilance, and mental set switching. At the end of the week these patients were restudied, and the CPAP treatment group scored significantly better than placebo in all areas of testing.
Lim, et al., in 2007 6 studied 46 OSA patients: 15 with regular CPAP, 15 with CPAP and oxygen, and 16 with placebo CPAP. Subjects were on treatment for two weeks. Neuropsychological testing was done before the treatment with careful study of attention, working memory, executive function, and alertness. At the end of two weeks they found that all areas of cognitive function had improved, even in the placebo group, but more improvement occurred in the category of vigilance, and only in the regular CPAP treatment group.
Joseph, et al., in 20097 studied 56 OSA patients with treatment for six months. Pre- and post-neuropsychological testing was done after CPAP treatment. They found significant improvement in depression, anxiety, and memory in the regular CPAP treated group.
Zimmerman, et al.,8 studied 58 memory impaired patients with OSA. Memory tests were done pre- CPAP treatment and three months after CPAP. CPAP users were placed into one of three categories based on hours CPAP was used each night. There were 14 poor users who averaged less than two hours of CPAP at night. There were 25 moderate users who used CPAP from two to six hours per night, and 19 optimal users, who used greater than six hours of CPAP per night. They found that patients using CPAP an average of six hours per night were eight times more likely to have memory in the normal range after three months of CPAP. Twenty-one percent of the poor users, forty-four percent of moderate users, and sixty- eight percent of optimal users demonstrated normal memory after three-month of CPAP treatment. Their conclusion, which has been consistently reported in the literature, is that at least six hours per night of CPAP may be needed to produce a reversal of memory impairment in OSA.
Weaver, et al., in 20089 stated that CPAP adherence is defined as greater than four hours of nightly use and that 50-83 percent of OSA patients are non-compliant. This is a real problem, especially in patients with a cognitive disorder that could improve with CPAP treatment.
OSA And Alzheimer's Disease
What do we really know about OSA and Alzheimer's disease? First of all, there does not appear to be any definite evidence that Alzheimer's disease or APOE-4 carriers have a higher incidence of OSA than the normal aging population and non- APOE-4 carriers.10,11 OSA appears to play an important role in many Alzheimer's disease patients. The benefits of CPAP in Alzheimer's tell us what role OSA may play in worsening cognitive function and activities of daily living. Ancoli-Israel, et al., in 2008 12 studied fifty-two women with mild to moderate Alzheimer's and OSA using regular CPAP and placebo CPAP to see if cognitive function improved at the end of three weeks and six weeks by means of neuropsychological testing. All medications were allowed as long as the patient had been on them for at least two months prior to treatment. All patients in the study had AHI (apnea/hypopnea index) of 10 or greater. Twenty-seven patients had normal CPAP use and twenty-five percent had placebo CPAP, and all cases were cognitively equal. During use of the placebo CPAP the AHI did not change. The mean was 27 versus 36 after three weeks of placebo CPAP. During the regular CPAP use the AHI fell from 29.7 to a mean of 6.4. Each group used their CPAP for six hours every night. After three weeks of true CPAP treatment there was significant improvement in neuropsychological scores, especially in episodic verbal learning and memory and some aspects of executive function including cognitive flexibility and mental processing speed.
Chong,13 in 2006 studied daytime sleepiness in Alzheimer's with OSA and using CPAP treatment in 39 patients. Half of the patients received six weeks of regular CPAP or three weeks of placebo CPAP followed by three weeks of regular CPAP. They evaluated daytime sleepiness using the Epworth Sleepiness Scale score. In the regular CPAP patients the Epworth score went from a baseline high of 8.89 to 6.5 after three weeks of regular CPAP. In the placebo CPAP group the baseline Epworth score of 7.68 dropped to 5.4 after three weeks of regular CPAP. The study showed the effectiveness of CPAP in reducing sleepiness in patients who had Alzheimer's and OSA.
Even though CPAP has been shown to improve cognitive function and wakefulness in the normal elderly and in Alzheimer's patients in most studies, the biggest limitation of this treatment in everyday clinical practice is patient compliance. The majority of my patients with OSA, with or without cognitive impairment or dementia, either are not using their CPAP at all or are using it on a very limited basis. They rarely communicate with their sleep physician or receive much training from CPAP equipment suppliers.
As previously stated, Weaver, et al.9 found that 46- 83 percent of all patients are non-compliant with this treatment; they defined CPAP use for four hours as the minimal threshold for adherence and compliance. In their review of the literature and their experience they analyzed many factors in compliance that have very important clinical implications. They are as follows:
- The pattern of CPAP compliance is established within the first week of treatment and predicts long-term use.
- A cause of poor compliance is often refusal of a patient to use a treatment that has a machine and a mask.
- There is no evidence that oral masks or nasal pillows have better compliance than nasal masks.
- CPAP compliance is unrelated to age, sex, marital, or social economics status or level of nocturnal hypoxia.
- Predominately spouse referral for diagnosis and treatment of obstructive sleep apnea leads to poor compliance.
- Increased nasal pressure, if present, can lead to greater than 50 percent chance of rejecting CPAP treatment.
- OSA in patients with few sleep complaints leads to much less compliance. Some characteristics that lead to increased compliance in CPAP use in OSA are as follows:
- Patient is predominately reporting excessive daytime sleepiness.
- Treating increased nasal resistance and/or obstruction with surgery.
- Semi-structured patient interviews after CPAP titration with a sleep technician to improve education and rapport and immediate troubleshooting of problems can improve compliance.
- Improving patient perception of benefits of CPAP in terms of health value (less sleepiness, less fatigue, more alertness and improved memory).14
- Increased education of the spouse or caregiver to provide social support, partner interaction, and improving partner sleep quality.
- New varieties of CPAP equipment, such as smaller machines, bi-level CPAP, autopap, or flexible CPAP designed to lower expiratory pressure or increase patient comfort are being studied to determine their effect on compliance.
Cognitive behavior therapy plus standard information about CPAP early on has resulted in marked improvement in CPAP usage by three hours in the first three months of therapy compared to no cognitive behavioral therapy. 15
In regard to compliance of CPAP use in Alzheimer's disease, many of these suggestions apply, especially in the early stages when the patient has some understanding. However, caregiver education, support and early trouble shooting of any difficulties play an important role in improving compliance in all stages of Alzheimer's disease with OSA.
Ayolon, et al., in 200616 studied 30 patients with mild to moderate Alzheimer's disease and OSA. They studied the adherence to CPAP therapy in these cases using hidden clocks and keeping CPAP pressure within 2cm of water of prescribed pressure. They found the mean number of hours of CPAP use was 4.8. The other factor that stood out and showed improved CPAP compliance was fewer symptoms of depression. The CPAP adherence in the less depressed group was 5.5 hours per night versus 3.7 hours with more depression symptoms. Less depression was felt to improve motivation and quality of life. The paper stated that with a full-time caregiver the presence of mild to moderate dementia should not be a deterrent to using CPAP to treat OSA.
Not all patients with OSA improve totally with CPAP, despite compliance. Sleepiness and cognitive problems may still persist. Roth, et al.,17 studied 391 cases of OSA. Half the group received armodafinil, 250mg a day, and the other half received placebo. They showed that the armodafinil group improved the quality of episodic memory compared to placebo. There was significantly reduced fatigue and sleepiness. CPAP use remained high at seven hours per night, and no appreciable side effects of the medication were found. There were no appreciable side effects or other negative effects on the regular night- time sleep.
Oral Appliance Use in OSA
If the CPAP remains an unsuccessful treatment for OSA due to inability to overcome compliance problems after maximum effort, then the use of an oral appliance should be considered. This subject was extensively reviewed by Fergerson, et al., in 2006.18 They reviewed all the literature since 1995; Only 15 trials met the criteria for good randomized controlled trials, and five were placebo controlled. Most studies used an oral appliance made by a dentist to advance the mandible forward during sleep to allow unobstructed breathing. This works by enlarging the upper airway and by decreasing upper airway collapsibility by improving upper airway muscle tone. Only a few studies discussed an oral appliance that holds the tongue forward. Overall, the review concluded that in mild to severe OSA there was only a 52 percent chance of being able to control the OSA using an oral mandibular appliance. Severe OSA and a very high BMI were more resistant to treatment with an oral apparatus due to more fat and tissue in the nasal pharyngeal structures. On the whole, it was felt that the oral appliances were less effective than CPAP, which has a greater than 80 percent chance of controlling the apnea. The big difference in the oral appliance is the fact that the average compliance rate is 77 percent versus about 20 percent in CPAP. Side effects of the oral appliances were infrequent and included temporal mandibular joint pain, tooth pain, dry mouth, and gum irritation. Tongue repositioning devices only were used in patients with very large tongues, where it made sense to push it down and forward to decrease obstruction.
Surgery for Treatment of OSA
One of the more popular surgical procedures for OSA has been uvulopalatopharyngoplasty, which overall is less successful in treating OSA compared to CPAP or mandibular appliance. Kahn, et al.,1 recently reviewed the Mayo Clinic experience in 63 patients who were treated with surgery. They confirmed that only 33 percent of their patients with OSA achieved the same benefit as an oral appliance and CPAP treatment. They said that surgery is more likely successful in OSA that fails CPAP and oral appliance treatment, and especially those with mild OSA and a normal BMI. High BMIs (obesity) increase adipose tissue in the lateral parapharyngeal fat pads, tongue, and neck, increasing the risk for airway collapse with sleep. Surgery, as described, cannot directly address these excess fat storage sites. Therefore, this would likely lead to a higher apnea/hypopnea index after surgery.
Appropriate to the subject of OSA, deprived sleep, and cognitive impairment is a recent paper by Kang, et al19 showing elevated amyloid beta in brains of mice that were forced to stay awake for the first six hours of a 12-hour sleep cycle. It was also shown that orexin, a substance released by the hypothalamus, leads to wakefulness and amyloid beta release. Blocking orexin receptors by a sub- stance called Almorexant inhibits Abeta plaque accumulation. In human studies, Abeta in CSF also fluctuates with awakening periods. It is not known how much sleep deprivation leads to Abeta changes. The authors suggest that treating sleep disorders aggressively in midlife may be important, since this is the time when amyloid and Alzheimer's pathology usually begins. Almorexant is currently in Phase III clinic trials. It enhances sleep by blocking orexin receptors and appears to reduce Abeta plaque for- mation, suggesting a possible new therapeutic approach for Alzheimer's disease.
OSA with secondary hypoxia and sleep fragmentation is not uncommon. It occurs in many of our patients whether they come for cognitive or other neurological problems or have been diagnosed with dementia. In addition to the standard history and neurological examination, we must be aggressive in obtaining information from the patient, spouse, or other family members about snoring and sleep behavior. The patient who has secondary cognitive impairment or diagnosed dementia often does not recognize their own excessive daytime sleepiness and snoring. With many of our dementia patients, it is not uncommon for a caregiver to tell us that their loved one is sleeping more and more during the day and is frequently up at night. We often tell them that this is a function of the dementia, medication, and more sleep during the day that often leads to less sleep at night. OSA must not be overlooked, and if suspected, a sleep study should be done.
Most studies have shown improved cognitive function, particularly memory, executive function, vigilance, and more wakefulness with the use of CPAP. However, because of poor CPAP compliance, many of these improvements may never be seen, even though it is the most successful treatment for OSA. I believe it is not enough to just order a sleep study and CPAP titration and leave it to the patient and his or her family to deal with the equipment and compliance issues alone. We should contact our patients and families early on to be sure they were trained adequately as to why and how they are using their equipment and where to get help if any problems arise. Aggressive behavioral therapy with freedom to contact someone with CPAP expertise is crucial to compliance.
Remember that the likelihood of long-term good compliance decreases significantly if barriers to CPAP use occur in the first week of therapy. If our patients are not getting expert help from the sleep referral centers or companies supplying the CPAP equipment, we as physicians should contact them and insist on more education and helpful suggestions for those with poor compliance. Their business depends on neurology referrals and poor cooperation should lead to a search for other referral sources. It is a sad day in therapy of cognitive disorders if such a successful therapy as CPAP in OSA is not maximized to the benefit of our patients.
Patients that are maximized but incomplete on CPAP may benefit further from taking Armodafinil, 150-250mg per day. It has been shown to improve wakefulness, vigilance and episodic memory. If CPAP cannot be tolerated for any reason and the best efforts has been made to use it, then an oral mandibular appliance should be tried even though only 52 percent of patients will get satisfactory results despite a 77 percent compliance rate. Palate surgery should be a last resort, since it is least likely to be successful compared to CPAP and an oral appliance. Mild OSA and low BMI patients are more likely to benefit.
Finally, there is early evidence in mice and early data in humans that sleep deprivation can lead to beta amyloid deposition in the brain that can be reduced by inhibition of orexin (a substance leading to wakefulness released by the hypothalamus) receptors. A drug in Phase III clinical trials called almorexant may be the beginning of a new strategy in treatment for Alzheimer's disease.