Insomnia is one of the most common comorbidities you’ll see in your depressed and anxious patients (Becker PM and Sattar M, Curr Treat Options Neurol 2009;11(5):349–357). But it is often misunderstood. Over the last several years there has been a shift in how we conceptualize insomnia co-occurring with psychiatric disorders. While the common view is that insomnia is caused by a primary psychiatric or medical condition, it is more accurate to simply say that patients have insomnia and depression at the same time. Insomnia is almost never an isolated problem.

In the 2002 National Health Interview Survey (an in-person structured health interview conducted by the CDC with 35,849 participants with insomnia), only 4.1% of respondents with insomnia reported having no comorbid condition. As compared with normal sleeper cohorts, insomnia was significantly associated with comorbidities such as chronic heart failure (3% comorbid with insomnia vs 0.7% in good sleepers), diabetes (10.8% vs 5.6%), obesity (29.4% vs 20.9%), hypertension (30.3% vs 16.6%), and anxiety or depression (clocking in with a whopping 45.9% in individuals with insomnia vs 9.3% in good sleepers).

The adjusted odds ratio for depression or anxiety comorbid with insomnia is 5.64 (in other words, someone with depression or anxiety is more than five times more likely to suffer from insomnia than someone without) (Pearson N et al, Arch Int Med 2006;166:1775–1782). The bottom line is that for the effective management of depression or anxiety with insomnia, you need to treat them at the same time. Treating depression without addressing comorbid insomnia will not only decrease the effectiveness of treating the depression, but contribute to its relapse (Roth T, Am J Manag Care 2009;15(Suppl):S6–S13).

A useful rule of thumb is that insomnia more commonly precedes a depressive episode, and more commonly follows an episode of anxiety. A large European study of 14,915 people showed that it was more common for a period of insomnia to come before a depression (41%), as opposed to depression preceding insomnia (29%). Similarly, depression relapse tended to be predicted by an insomnia prodrome. In this same study, the opposite pattern was found for anxiety: the anxiety preceded the development of insomnia. These results were replicated in several longitudinal studies (Roehrs T and Roth T, Clinical Cornerstone 2003;5(3):5–12; Ohayon M and Roth T, J Psych Res 2003;37:9–15).

A complete history for all of your patients should include a brief “How is your sleep?” Often this information will be supplied without prompting: “I can’t sleep at all. Can you give me something for that?”

Sure you can. It is important to determine first, however, why your patient can’t sleep. Common potential causes of insomnia that should be on your checklist include:

• Sleep hygiene issues. For example, the patient who drinks super caffeinated drinks so that she can stay up late finishing spreadsheets and answering important phone calls while she watches CNN after her nightly five-mile run—such a patient is unlikely to respond to a simple sleeping pill.


• Sleep apnea.


• Substance abuse.


• Chronic insomnia. A patient who “just can’t get to sleep, no matter how hard he tries,” and who is “afraid it’s going to make him completely useless the next day” is likely to benefit from cognitive behavioral therapy for insomnia (CBT-I; see the interview with Charles Morin in this issue).


• Acute stress-induced insomnia. The patient with an acute, but likely transient bout of insomnia accompanying an event such as a death, birth, moving, or new job might benefit from a short course of hypnotics.


• Insomnia comorbid with a psychiatric disorder. And then there is the patient—often with a mood disorder or anxiety—who just doesn’t sleep well; can’t get to sleep or stay asleep, and who genuinely suffers the next day as a consequence of it. Any of these patients can benefit from CBT-I or at least some components of it, but for some, a sleeping pill is not only an option, it’s an important one. So if your patient is a candidate for a sleeping pill, which one should you use?

Sedating antihistamines. These are popular OTC choices. While diphenhydramine (Benadryl) is the most common antihistamine found in OTC sleep preparations (such as Tylenol PM and Advil PM), you’ll see other antihistamines such as doxylamine in these formulations as well. These medications can be effective, but they are often slow to act, can be associated with next-day “hangover” effects, and your patients may develop tolerance to them. Since these drugs are also muscarinic receptor blockers, you need to watch out for anticholinergic effects (eg, blurred vision, constipation), particularly in your older patients (Neubauer DN and Flaherty KN, Sem Neurol 2009;29(4):340–353). If your patient does respond well to diphenhydramine, recommend the solo preparation rather than the combination with acetaminophen or ibuprofen, which have their own side effects.

Benzodiazepines. Surprisingly to many, only five older benzodiazepines are formally FDA approved for insomnia: flurazepam (Dalmane), temazepam (Restoril), triazolam (Halcion), estazolam (Prosom), and quazepam (Doral). With the exception of temazepam, these medications are no longer commonly prescribed. Instead, modern psychiatrists tend to prescribe benzodiazepines such as diazepam (Valium), alprazolam (Xanax), lorazepam (Ativan), and clonazepam (Klonopin) for insomnia, especially in patients with mood or anxiety disorders (Lader M, Addiction 2011;89(11):1535–1541).

There is no evidence that FDA approval has conferred any hypnotic advantages—all benzodiazepines probably work equally well, though many of the older examples have disadvantages such as very long half-lives or, in the case of short-acting triazolam, troublesome side effects like amnesia.

All benzodiazepines bind non-specifically to the GABA receptor, which leads to side effects such as drowsiness, headache, dizziness, lightheadedness, and difficulty with concentration and memory. Tolerance, dependence, abuse, and withdrawal are well-known occupational hazards of the benzodiazepines (see the September 2011 TCPR for coverage of the tricky use of benzodiazepines in substance abusers).

Non-benzodiazepines. The first non-benzodiazepine hypnotic to appear was zolpidem (Ambien), now available as a generic. A new drug that binds only to certain subtypes of the GABA receptor, it is associated with fewer side effects, faster onset, less potential for abuse, and less next-day hangover (Drugs 1990;40(2):291–313). Other non-benzodiazepines followed zolpidem: zaleplon (Sonata, also available as a generic), eszopiclone (Lunesta, no generic yet), and zolpidem extended-release (Ambien CR, available as a generic). Zolpidem is also available as a rapidly-dissolving sublingual tablet (Edluar) and as an oral spray (Zolpimist); these were developed as faster-acting agents. Zaleplon has the shortest half-life (one hour) of the non-benzodiazepines and is therefore a popular choice for patients who have sleep onset insomnia but can manage to stay asleep once they fall asleep, or for those who awaken in the early morning and need a hypnotic that will be fully metabolized in time for work later in the day.

Zolpidem has a longer half-life (1.5 hours to 2.4 hours), and eszopiclone the longest half-life of all the non-benzodiazepines (5 hours to 7 hours). Eszopiclone’s longer half-life coupled with its slightly slower onset of action makes it a popular choice for patients who can fall asleep, but not stay asleep. Eszopiclone is associated with an unpleasant (sustained) metallic taste, however, experienced by about 40% of the people who try it.

When cost is not an issue, the nonbenzodiazepines are generally the prescription hypnotic of choice. However, during the post-marketing period after zolpidem became available, reports of sleep-eating, sleep-driving, sleep-gambling, sleep-sex and other sleep-related “complex behaviors” emerged. This is because zolpidem and other non-benzodiazepines are associated with sometimes significant anterograde amnesia; people who experienced any of these “complex behaviors” were probably awake when performing them, but do not remember it. You should warn your patients about the possibility of these, and advise them to get into bed and cease normal activities within 15 minutes of taking the pill. Some patients will still unconsciously get up out of bed to engage in complex sleep behavior—such patients may need to be switched to another medication.

Melatonin agonist. The only drug in this class so far is ramelteon (Rozerem). Since it doesn’t bind to GABA, it doesn’t have the troublesome GABA-agonist side effects, and may be a good choice for patients with sleep phase disorders, insomnia associated with shift work or traveling across many time zones, or in patients with substance abuse problems.

Ramelteon may also be a safer choice for older patients (Srinivasan V et al, Adv Ther 2010;27(11):796–813). Ramelteon doesn’t deliver the expected “kick” of a sleeping pill, and some patients do not feel it is as effective as a benzodiazepine or non-benzodiazepine hypnotic. Patients sometimes need to take it continuously for several weeks before noting a benefit. Unlike the benzodiazepines and non-benzodiazepines, which are C-IV scheduled substances, ramelteon is nonscheduled.

Sedating antidepressants and antipsychotics. Low dose tricyclic antidepressants, such as amitriptyline (Elavil), imipramine (Tofranil), and doxepin (Silenor) have long been used as off-label hypnotics. Recently a very low dose (3 mg to 6 mg) formulation of doxepin was approved by the FDA under the trade name Silenor (see TCPR April 2011 for a skeptical review of this agent). While effective, tricyclics can cause the usual slew of anticholinergic side effects, such as constipation and urinary retention, especially in the elderly (Med Lett Drugs Ther 2010;52(1348):79–80).

Other sedating antidepressants have also long been used off label to treat insomnia, such as trazodone (Desyrel) and mirtazapine (Remeron). Trazodone’s long half-life (mean of seven to eight hours) is helpful for keeping patients asleep all night, but can lead to next day sleepiness. Mirtazapine often causes too much weight gain to be useful for the long-term.

Some of the antipsychotics, especially quetiapine (Seroquel) and olanzapine (Zyprexa), are also sedating and are often used off-label to manage insomnia—but given their high expense and risk of sometimes significant weight gain, hyperglycemia, tardive dyskinesia, and EPS, they are best reserved for the toughest cases.