In Session with John W. Winkelman, MD, PhD - Sleep Disorders

Have studies been conducted to indicate whether there are different types of insomnia linked to different psychiatric disorders?

All psychiatric illnesses can, and frequently do, produce insomnia. In fact, the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, lists insomnia as a diagnostic criterion for a number of psychiatric disorders, while for other disorders insomnia is considered a common feature. However, the characteristics of insomnia are not distinct between the different psychiatric illnesses. There have certainly been polysomnographic studies of sleep in individuals with different psychiatric illnesses, although they are difficult to interpret.

Not only do psychiatric disorders produce insomnia, but insomnia is associated with an increased incidence, or new onset of, psychiatric illness. In particular, this has been observed in long-and short-term studies, which have shown an increased risk for the development of new-onset or recurrent depression in people with persistent insomnia. A study by Chang evaluated Johns Hopkins Medical School graduates over 40 years (Chang PP, Ford DE, Mead LA, et al. Am J Epidemiol. 1997;146:105-114). Medical school seniors who reported insomnia had a higher risk of depression 15 years later, and this excess risk continued to increase up to 40 years after graduation.

Is there a cause-and-effect relationship between insomnia and depression, or is it merely an association?

It is unclear whether insomnia is a marker of early depression, or itself predisposes people to depression. Hypercortisolemia, particularly in late afternoon/early evening, has been seen in individuals with chronic primary insomnia. Increased cortisol may indicate a vulnerability to depression, or could be causally related to insomnia.

Since insomnia can be linked to other diseases, should treatment be administered quickly?

Even though we do not have a cause-and-effect relationship established, insomnia should be aggressively treated for a variety of potential reasons, including protection against future psychiatric illnesses. A doctor would not deprive a chronic pain patient of medication just because 5% of the people who take analgesics develop a dependence and tolerance to it. Similarly, insomnia should be treated aggressively, as the majority of those who take hypnotics do not develop dependence or tolerance.

Should we be concerned that patients who rely on medication to sleep over a long period of time will become dependent and abuse the drugs?

Overall, abuse of sleep drugs by insomniacs is rare. Most patients with insomnia are not seeking a high; rather they are seeking adequate treatment for their insomnia. Those with a prior history of substance or alcohol abuse may abuse or develop rapid tolerance to these medications. For patients with addictive tendencies, there are other effective approaches to the treatment of chronic insomnia. For example, cognitive-behavioral therapies have been shown in numerous studies to be as effective as hypnotics in reducing time awake during the night and the amount of time it takes to fall asleep. They work predominantly by restricting time in bed and by reducing dysfunctional sleep-related cognitions.

What differences exist between the benzodiazepines, the sedating antidepressants, and atypical antipsychotics in the treatment of insomnia?

There are very few controlled trials of sedating antidepressants for insomnia and none for atypical antipsychotics or anticonvulsants, which are both being used to treat insomnia. I believe that tolerance can develop to any medication that is sedating. My experience is that people develop a tolerance to alternative agents as often as they do to the benzodiazepine and benzodiazepine receptor agonists.

The short-acting benzodiazepine receptor agonists are generally free of next day carry-over effects. However, because they have a longer half-life (eg, trazodone=5–9 hours; gabapentin=7 hours), sedation is reported more commonly with these agents than with the short-acting benzodiazepines and benzodiazepine receptor agonists.

Should treatment of insomnia be adjusted depending on what the primary disorder is?

Always try to treat the underlying disorder first, whether it is depression, psychosis, thyroid disease, or restless legs syndrome. For patients with substantial insomnia, who have consequences for daytime functioning and quality of life, I start treatment with a benzodiazepine or benzodiazepine receptor agonist, because I believe they work the best. Patients with less substantial adverse consequences can use one of the less reliably sedating agents, such as antidepressants, antipsychotics, or anticonvulsants. In patients with a history of alcohol or substance abuse, I am more likely not to use an agent that works at the benzodiazepine receptor.

I am hesitant to give patients with bipolar disorder antidepressants for sedation. Even though they are taking low doses, there is the risk of inducing a switch to mania. For patients with mood instability or bipolar spectrum disorders, it is important to be aggressive in treating insomnia.

For people who have depression and severe insomnia, it is important to treat both disorders. A sedating agent should be used during the first couple of weeks to provide immediate insomnia relief.

A new drug, eszopiclone, is indicated for long-term use in the treatment of insomnia. Is it safer and more effective than current agents, such as zaleplon, zolpidem, and the benzodiazepines?

No head-to-head studies have been performed comparing eszopiclone to the these other approved medications for sleep. However, a 6-month placebo-controlled study of eszopiclone in the treatment of primary insomnia demonstrated that advantages of this agent over placebo in promoting sleep onset and reductions in time awake during the night were maintained over 6 months of nightly use. Long-term data for the other agents are pretty much absent, except for long-term data on intermittent dosing with zolpidem.

Zaleplon and zolpidem predominantly bind to the a₁ subtype of the g-aminobutyric acid (GABA)_A receptor, which carries amnestic, psychomotor, and anticonvulsant properties. They do not bind to a₂, which carries anxiolytic properties. Zolpidem and zaleplon may lack anxiolytic properties, because of this receptor selectivity.

Some patients who take zolpidem actually develop psychiatric symptoms the next day. Can you explain this behavioral toxicity?

I think that can happen with all benzodiazepines and non-benzodiazepines. Many patients develop disinhibition in the hospital when they are given lorazepam or triazolam for sleep or anxiolysis. In hospitalized medical or psychiatric patients this can occur with the non-benzodiazepines as well. It is rare, but there have been reports of disinhibition, amnesia, and occasionally hallucinations, as potential side effects of any agent in this class.

Is modafinil an effective medication for sleep disorders?

Modafinil is generally effective for sleepiness and fatigue, regardless of etiology. The label has been expanded from daytime sleepiness in narcolepsy to include daytime sleepiness associated with shift work and incompletely treated obstructive sleep apnea. I think that it is effective for fatigue and sleepiness regardless of the cause––narcolepsy, depression, sleep apnea, or even inadequate sleep. However, the modafinil data on mood-elevating properties have not demonstrated improvement in total Hamilton Rating Scale for Depression scores or any components other than energy and fatigue.

As for the negative effects, modafinil can occasionally cause the patient to be jittery and anxious the way a sympathomimetic stimulant can. It can also reduce the efficacy of steroidal contraceptives. The most common side effect is headache, but that is usually transient. Also, at this point in time it is a very expensive medication, but it will become available in a generic form within approximately 18 months.

How often is sleep apnea a cause of psychiatric symptoms?

These disorders coincide more often than would be seen by chance. However, it is not clear to me whether sleep apnea produces a vulnerability for depression. My experience has been that rarely does the treatment of sleep apnea obviate the need for independent treatment of the depression.

Are there any instances when psychiatric symptoms caused by a sleep disorder are resolved when the primary sleep disorder is treated?

People with an acute, severe insomnia can meet criteria for depression: they are agitated, they are not sleeping, they have lost interest in things, they feel hopeless and at times suicidal, their appetite is poor, and they have no energy. However, if you treat their insomnia with an appropriate medication, their “depression” resolves. At times, restless legs syndrome or obstructive sleep apnea have been misdiagnosed as psychiatric disorders, and treatment of these makes the symptoms improve.

Do insomnia patients tend to exaggerate their symptoms, such as how long it takes them to fall asleep?

Insomnia is a diagnosis that is made by self report. When a patient over-estimates sleep difficulties, we call it sleep state misperception, but these people have insomnia. Just because we can record that they have slept in the sleep laboratory, does not necessarily mean that their sleep is refreshing or restorative. Sleep has two components; it has the objective component, measured by an electroencephalogram (EEG), and the subjective component. EEG results, recorded when the patient is asleep do show a difference between insomniacs and people who do not have insomnia.

People with insomnia are not excessively sleepy. Insomnia is a state of hyperarousal in a variety of contexts. The endocrine context has to do with cortisol levels. The cognitive context is characterized in terms of the patient’s ideas about their insomnia, beliefs regarding the consequences of their insomnia, and exaggeration of how much or how little they have slept. The physiological context refers to the whole body metabolic rate, EEG activity, and hypermetabolism by functional imaging. People with insomnia are hyperaroused, but whether that is a marker of their insomnia or a cause of their insomnia is not clear.

Selective serotonin reuptake inhibitors (SSRIs) are commonly used to treat depression, but reportedly produce many side effects, including sleep disturbances, daytime somnolence, and fatigue. Do you think that these effects are clinically significant?

When given to normal volunteers, SSRIs increase the time to fall asleep; they increase the amount of light sleep; and they have profound effects on REM sleep, which contributes to subjective and behavioral toxicity for sleep. SSRIs fragment REM sleep so that there are more awake periods mixed in between the REM periods. They can also disinhibit motor activity, both during REM sleep and non-REM sleep. However, because these agents are so effective in the treatment of depression, and the insomnia of depression is so severe, their net effect on sleep for people with depression is usually beneficial.

In non-REM sleep, effects of SSRIs are seen as an increase in the number of periodic leg movements of sleep, which may end up disturbing sleep. “Twitchiness,” or myoclonic events, as well as longer motor events during REM sleep are observed. SSRIs can actually produce REM sleep behavior disorder, which is a more severe case of disinhibition of REM motor activity.

SSRIs partially reverse the paralysis/atonia that usually occurs during REM sleep. This fragmentation of REM and motor activity in REM may be responsible for patient’s experiences of vivid dreams. Therefore, patients can end up acting out their dreams by thrashing, kicking, or propelling themselves out of bed. I have seen this with the use of several SSRIs, clomipramine, as well as monoamine oxidase inhibitors. Also, SSRIs produce marked excess of eye movements during non-REM sleep, REM sleep, and while awake.

For a patient who is on an antidepressant and has sleep disruption, I recommend lowering the SSRI dose to reduce or eliminate the side effect, and hopefully maintaining the beneficial effect. If that is not possible, one can either consider switching to a nonserotonergic antidepressant or adding a sedative at bedtime to improve sleep quality.

Is REM sleep behavior disorder an early manifestation of Parkinson’s disease?

In approximately 50% of people with REM sleep disorder, the condition seems to be associated with synucleinopathies, which is a class of neurologic disorders that includes Parkinson’s disease. The most common cause of REM sleep behavior disorder is probably antidepressant related. I recommend using shorter benzodiazepines, such as lorazepam or oxazepam, which have less potential daytime consequences than the longer acting ones. This is important because many people with Parkinson’s or other neurologic disease are older and have problems with motor activity and cognitive function.

Are there any other psychotropic drugs with sleep effects that are unique or noteworthy?

In people with restless legs syndrome, an exacerbation may occur with SSRIs, so nonserotonergic agents like bupropion or desipramine may be a better option for these individuals. In addition, any dopaminergic antagonist, such as quetiapine and other atypicals, can produce restless leg syndrome or periodic legs movements of sleep.

Some psychiatric medicines, such as tiagabine and olanzapine, can increase slow-wave sleep. However, it is not clear what the clinical consequences of this increase in slow-wave sleep are; the data for tiagabine and insomnia were negative. I think that these slow-wave sleep findings are very interesting, but need to be supported by clinical consequences.

Are there any common mistakes that general practioners or general psychiatrists make in approaching diagnosis of sleep-related problems or insomnia?

The most common mistake is that the physician does not ask questions about the patient’s sleep. Many physicians have limited time and do not ask questions about sleep because they are concerned that they will not have enough time to make an adequate diagnosis or treatment plan. Also, some of the physicians have never been educated about the appropriate differential diagnoses. The amount of education about sleep in medical school and residency is minimal; possibly only one or two lectures. Thus, physicians tend to avoid the issue because they feel they do not have enough time or knowledge to appropriately address it. The most common mistake is ignoring the issue.

Obstructive sleep apnea is frequently missed by psychiatrists. Many psychiatric medications produce weight gain, which causes vulnerability to sleep apnea. The fatigue and sleepiness are commonly attributed to side effects of medications or to an underlying psychiatric illness, such as apathy, depression, or negative symptoms of schizophrenia. The diagnosis of sleep apnea can be made easily by asking about snoring and looking at the patient’s neck size to see if they have a short, squat neck. Patients who have a neck size of >16.5–17 inches have a 50% chance of having sleep apnea. Sleep study is then recommended.

The diagnosis of restless legs syndrome is commonly missed as well, and the sleep complaints are attributed to primary insomnia or an anxiety or mood disorder.

Similarly, acute insomnia can be misdiagnosed as depression. One can determine if it is insomnia by giving the patient a hypnotic for 3 days and observing whether the symptoms resolve. On the other hand, antidepressant trials may take 1 month, and can worsen insomnia.

Are there some patients who need large doses of benzodiazepines in order for these drugs to be effective?

Patients who require large doses are often those who have taken benzodiazepines in the past and have developed some tolerance to them. When I initiate treatment with a benzodiazepine agent, I tell the patient what the maximum dose is that I feel comfortable prescribing. If the patient develops tolerance and needs an increase in dosage beyond that level, I taper the patient off the medication and use one of the other classes of agents to help them sleep.

Another instance where patients require massive doses of medication is with sleep state misperception. For example, a patient can be taking quetiapine 600 mg, clonazepam 2 mg, zolpidem 20 mg, gabapentin 1,200 mg, and trazodone 150 mg, and still claim they are not sleeping. However, they seem to sleep fine in a sleep laboratory. In this case, there is something wrong with the patient’s subjective appreciation of sleep. I tell these patients that their body is getting enough sleep and that because of the adverse consequences of daytime sedation, that I would like to taper the medication. Many times these patients will have anxiety about the consequences of sleeplessness. This is an instance where bringing them into the sleep laboratory and confirming sleep state misperception can be helpful and reassuring to the patient.

Is there a standard sleep questionnaire that is available?

While there is no standard screening questionnaire for sleep disorders, a helpful sleep assessment is the Pittsburgh Sleep Quality Index (Buysse DJ, Reynolds CF, Monk TH, et al. Psychiatry Res. 1989;28:193-213). PP