TCR: Dr. Winkelman, could you recommend a practical approach for psychiatrists diagnosing sleep disorders?
Dr. Winkelman: I usually base my diagnostic algorithm upon presenting symptoms. A patient will usually present with a chief complaint of insomnia (difficulty falling asleep or staying asleep, or nonrestorative sleep), excessive daytime sleepiness, or unusual behaviors during sleep, which are called "parasomnias." That is where I start, because, like DSM, I think it makes most sense to base your initial approach based on symptoms rather than putative causes. And the fact is that there is no clear one-to-one relationship of any particular sleep disorder or even psychiatric disorder with a particular type of insomnia. For example, the idea that anxiety causes problems falling asleep whereas depression causes early morning awakening is going to get you in trouble.

TCR: What are the most common causes of insomnia in a psychiatric practice?
Dr. Winkelman: The most common cause will be psychiatric illnesses producing insomnia. Most commonly, these would be depression and anxiety disorders, but certainly psychotic disorders can produce serious insomnia. You obviously want to treat the underlying psychiatric illness first. And that is really true for any sleep complaint. However, for many patients with psychiatric illness, the insomnia does not clear up even with adequate treatment of the underlying psychiatric illness; sometimes it even gets worse. And so you need to then go back to your differential diagnosis and see what else could be causing the problem.

“I don't think that the nonbenzos are any less abusable than the benzodiazepines. But both of these classes are rarely abused, except by people with substance abuse histories.”
- John Winkelman, M.D.

TCR: Such as?
Dr. Winkelman: The second major cause of insomnia would be medications. Any of the activating serotonergic, adrenergic, or dopaminergic antidepressants can cause insomnia, or can prevent insomnia from improving even with successful treatment of an underlying psychiatric illness. It is not clear whether taking an SSRI in the morning instead of at night makes a difference, but it is worth a try. You want to try to keep the dose as low as is effective. And make sure that they are not taking their antidepressants inappropriately; for instance, you may have told them to take buproprion "twice a day" and they are taking it all at night. Other medications that can produce sleep disturbance include long-acting stimulants, of which there are many now, and short-acting stimulants like caffeine, which are frequently consumed, particularly by psychiatric patients. Alcohol is a bad actor for sleep, and even small amounts of alcohol--two glasses of wine--can produce nocturnal awakenings. I really challenge patients to stop their alcohol while we are trying to get a handle on their insomnia.

TCR: Do you encourage good sleep hygiene in your patients?
Dr. Winkelman: Sleep hygiene issues are essential. Treating other problems without addressing habits and behaviors that promote good sleep will lead nowhere. You can give as much benzo as you want and, if people are napping all day long, they are not going to sleep at night. I always give my patient a handout of "10 Tips for a Good Night's Sleep," which readers can download for free at: http://www.mass.gov/Eeohhs2/docs/masshealth/pharmacy/10tips_goodnightsleep.pdf.

TCR: What are the most useful sleep hygiene methods in your experience?
Dr. Winkelman: The most effective techniques are sleep restriction and stimulus control. Restricting time in bed ("sleep restriction") is probably the most powerful treatment. I give my insomnia patients a bedtime and a wake time, which are six hours apart. This is a "no pain/no gain" kind of approach. You are using the body's own endogenous sleep drive rather than an exogenous chemical to help with sleep initiation and maintenance. "Stimulus control" is getting up and out of bed when you are awake for more than twenty minutes or so. The idea is to dissociate anxiety and the sleeping environment.

TCR: Can you say something about your approach to Restless Leg Syndrome (RLS)?
Dr. Winkelman: RLS is a common disorder, particularly in the elderly. While it's present in 1-5 % of people under the age of 60, it occurs in 10-20 % of people over the age of 65 or 70. It is a neurological disorder characterized by uncomfortable restless feelings in the legs that are relieved by movement and are most prominent at night.

TCR: How do you treat it?
Dr. Winkelman: First, you want to make sure that there is nothing driving it--in particular, iron deficiency. You want to check a ferritin or stored iron level in everybody with RLS. Anyone who has a ferritin below 40 should be treated with iron. If that is not the issue, then first-line treatments are dopaminergic drugs, primarily Requip (ropinirole) or Mirapex (pramipexole). Requip is started at 0.25 mg QHS for a few days, then increased to 0.5 mg for a few days, then to 1 mg for a few days, and if needed you can go up to 1.5 or 2 mg; the maximum dose is 4 mg. Mirapex is started at 0.25 mg QHS, and increased by 0.25 mg every few days up to a maximum of 1 mg. All doses should be taken about 2 hours before symptoms start.

TCR: They are both equally effective?
Dr. Winkelman: They appear to be. There are no head-to-head studies. The most common side effects with these agents are nausea, sedation, insomnia, and fatigue. You want to avoid these medications--or use them very cautiously--in patients with past or current psychosis. Second-line agents For RLS include opiates, such as low dose oxycodone or codeine, and Neurontin (gabapentin).

TCR: Are any of these treatments going to be FDA-approved?
Dr. Winkelman: Requip will be approved next year, and Mirapex maybe the year after.

TCR: What about sleep apnea as a cause of insomnia?
Dr. Winkelman: Sleep apnea can present as insomnia. Most of those people are not literally sleepless; they fall asleep and then have a series of abnormal or sleep-related breathing events that cause them to wake up briefly.

TCR: Do sleep apnea patients typically believe that they have slept most of the night?
Dr. Winkelman: They may believe that. They may say that they didn't feel refreshed when they wake up, but they do believe that they have slept. But you will also have people saying, "I can't fall asleep," whereas they may actually have fallen asleep, but they haven't recognized it because they haven't entered a deep stage of sleep.

TCR: There continues to be a lack of clarity about whether the non-benzos, such as Ambien (zolpidem) and Sonata (zaleplon), are less abusable than the benzodiazepines. What's your take?
Dr. Winkelman: Anything is abusable. I don't think that the non-benzos are any less abusable than the benzodiazepines. But both of these classes are rarely abused, except by people with substance abuse histories. As long as you use them for the appropriate patients, they are rarely a risk for nonmedical diversion. Rebound insomnia may be reduced in the non-benzos, although it's not clear. Estorra (eszopiclone) has got some great new data demonstrating that there is no tolerance over six months of continuous use.

TCR: Is that different from Ambien, Sonata, or the benzos?
Dr. Winkelman: We have no long-term data on any of these other medications, so we don't know. I am not sure that patients develop any more tolerance to these agents than they do to trazodone, Elavil, Seroquel, or Neurontin. I have seen tolerance to all of those agents; there are no long-term studies on any of them, either.

TCR: So why would we prescribe the more expensive newer alternatives if there is not a huge difference between them and the older agents?
Dr. Winkelman: The main difference is the half-life properties of these agents. Sonata has an ultra-short half-life, Ambien's is short, and Estorra's is going to be medium. Some of the older agents don't have as favorable a half-life profile, so many people will wake up feeling knocked out excessively from drugs like Restoril, Ativan, Klonopin or Dalmane. Halcion, on the other hand, has a short half-life, about equivalent to Ambien, and has an undeserved bad reputation.

TCR: So, for you, the reason to choose the non-benzos would be more based on shorter half-lives?
Dr. Winkelman: More favorable half-life, and when Estorra comes out, more assurance about long-term tolerance issues.

TCR: Do you think this tolerance data is going to allow the insurance companies to start to pay for more than a 14-day supply?
Dr. Winkelman: Yes. I think that there will be no short-term restriction for Estorra, so there will be no basis for them to limit the supply. They will make it a high co-pay, I'm sure, but there will be no basis for restricting it the way they have for the other agents.

TCR: Thank you Dr. Winkelman.