Zolpidem Safety Signals and FDA Actions: What Clinicians and Patients Should Know

What has the FDA done to zolpidem?

Zolpidem tartrate is an oral prescription sedative-hypnotic for short-term treatment of insomnia. It is presumed to work by binding the benzodiazepine site of GABA-A receptors that contain the alpha-1 subunit; it is chemically unrelated to benzodiazepines but acts at the same receptor family [3]. Ambien immediate-release tablets were approved on December 16, 1992; Ambien CR extended-release tablets on September 2, 2005; and Edluar sublingual tablets on March 13, 2009. Zolpimist oral spray (approved December 19, 2008) and Intermezzo sublingual tablets (November 23, 2011), the product designed for a middle-of-the-night awakening, are both listed as discontinued in Drugs@FDA. Generic tablets, extended-release tablets and sublingual tablets are approved [4][1].

Two labeling actions changed the drug's dosing and its warnings: the 2013 lowering of the recommended dose, driven by next-morning impairment, and the 2019 boxed warning, driven by serious injuries and deaths during complex sleep behaviors [1][2].

The 2013 dose reduction: what changed and why

On January 10, 2013, the FDA announced it was requiring lower recommended doses for zolpidem products taken at bedtime: from 10 mg to 5 mg for immediate-release products (Ambien, Edluar, Zolpimist) and from 12.5 mg to 6.25 mg for extended-release products (Ambien CR). The agency said blood levels in some patients may be high enough the morning after use to impair activities that require alertness, including driving; that women eliminate zolpidem more slowly than men; and that the risk of next-morning impairment is highest with the extended-release forms. For men, the FDA asked that labeling recommend prescribers consider the lower doses [1].

The current Ambien label reflects those split starting doses: 5 mg for women and 5 or 10 mg for men, and 6.25 mg for women and 6.25 or 12.5 mg for men with Ambien CR, with the label stating the initial doses differ because zolpidem clearance is lower in women [3][5]. The total dose should not exceed 10 mg once daily for immediate-release or 12.5 mg for extended-release; a 5 mg dose that is not effective can be raised, but the higher morning blood levels seen with the top dose increase the risk of next-day driving impairment [3][5]. The FDA stressed that patients can be impaired the next morning even if they feel fully awake [1].

The 2019 boxed warning on complex sleep behaviors

On April 30, 2019, the FDA announced a boxed warning on eszopiclone (Lunesta), zaleplon (Sonata) and all zolpidem products, after rare but serious injuries, including deaths, occurred during sleep behaviors in people who were not fully awake. The agency said these events appear to be more common with those three drugs than with other prescription sleep medicines [2].

Ambien's boxed warning, titled COMPLEX SLEEP BEHAVIORS, states that sleep-walking, sleep-driving and engaging in other activities while not fully awake may occur after use of the drug, that some of these events may result in serious injuries including death, and that the drug should be discontinued immediately if a patient experiences one [3]. The label adds that behaviors such as preparing and eating food, making phone calls or having sex have been reported; that patients usually do not remember these events; and that they can occur after the first or any later dose, at recommended doses, with zolpidem alone and with or without alcohol or other CNS depressants [3]. The Ambien CR label carries the same boxed warning [5].

That 2019 action also added a contraindication. Ambien is contraindicated in patients who have experienced a complex sleep behavior after taking it [3], and the FDA told prescribers not to give eszopiclone, zaleplon or zolpidem to anyone who has had such an episode on any of them [2].

What the case review behind the warning found

A retrospective analysis of the FDA Adverse Event Reporting System from December 16, 1992 through February 27, 2018, plus PubMed and EMBASE, identified 66 cases meeting criteria across the three drugs, four of them from the medical literature. Twenty cases reported death and 46 reported serious injuries. Fatal events included carbon monoxide poisoning, drowning, falls, hypothermia, motor vehicle collisions and apparent completed suicide; nonfatal injuries included accidental overdoses, falls, gunshot wounds, hypothermia, third-degree burns and self-injuries. Twenty-two cases reported an earlier episode of a complex sleep behavior on a Z-drug before the event in the case series [6]. Reports of this kind cannot establish how often these events occur [6].

Next-morning impairment and driving

The label warns of impaired alertness and motor coordination, and says the risk of next-day impairment rises with a higher dose, with fewer than 7 to 8 hours of sleep remaining, with alcohol or other CNS depressants, and with other drugs that raise zolpidem blood levels. Patients, particularly the elderly, are at higher risk of falls [3].

Epidemiologic work points the same way. Using Taiwan's national health insurance database, Yang and colleagues ran a case-crossover study of 12,929 people hospitalized for motor vehicle accidents between 1998 and 2004 and found an adjusted odds ratio of 1.74 (95% CI 1.25 to 2.43) for involvement in a crash after one defined daily dose of zolpidem on the previous day [7]. A review by Gunja reported that drivers taking zopiclone or zolpidem had an increased risk of motor vehicle collisions, over double that of unexposed drivers, and that driving impairment occurred with higher doses of zolpidem [8]. These are associations from observational data, not proof of cause.

Dependence and withdrawal: the Schedule IV question

Zolpidem is classified as a Schedule IV controlled substance under federal regulation [3].

The label warns that use may lead to physical and psychological dependence, with risk rising with dose, with duration of use, and in patients with a history of alcohol or drug abuse. Abrupt discontinuation has produced withdrawal symptoms ranging from mild dysphoria and insomnia to cramps, vomiting, sweating, tremors, convulsions and delirium [3]. In former drug abusers, 40 mg of zolpidem produced effects similar to 20 mg of diazepam, while 10 mg was difficult to distinguish from placebo [3].

The French CEIP pharmacodependence network reviewed reports collected from 1993 to 2002 plus 53 published case reports and concluded that zolpidem has significant dependence and abuse potential; the French drug monograph was changed as a result [9]. A published case report described a 29-year-old woman dependent on 160 mg of zolpidem who had a generalized seizure after sudden withdrawal [10].

Interactions that raise or lower zolpidem levels

An in vitro study projected that CYP3A4 accounts for about 61% of zolpidem's net intrinsic clearance, CYP2C9 for 22% and CYP1A2 for 14% [11]. In the Ambien label, ketoconazole raised zolpidem AUC by 70% and Cmax by 30%, and the label suggests considering a lower zolpidem dose when a potent CYP3A4 inhibitor is given together with it. Itraconazole raised AUC by 34%, sertraline raised Cmax by 43% over 17 days in healthy women, and fluoxetine raised half-life by 17% without an additive effect on psychomotor performance [3].

In the other direction, rifampin cut zolpidem AUC by 73%, Cmax by 58% and half-life by 36%, and the label says that combination is not recommended; St. John's wort may also lower zolpidem levels and is likewise not recommended [3]. Combining zolpidem with benzodiazepines, opioids, tricyclic antidepressants or alcohol increases CNS depression and psychomotor impairment, and opioids add respiratory depression risk [3]. Food slows absorption: after a meal, Cmax fell 25%, AUC fell 15% and time to peak rose from 1.4 to 2.2 hours in 30 healthy men given 10 mg [3].

Who is at higher risk

Older adults. The recommended dose is 5 mg immediate-release or 6.25 mg extended-release, and after a 20 mg dose elderly adults had Cmax 50% higher, half-life 32% longer and AUC 64% higher than young adults [3]. The 2023 AGS Beers Criteria make a strong recommendation, based on moderate-quality evidence, to avoid zolpidem in adults 65 and older because of delirium, falls, fractures and motor vehicle crashes with minimal sleep benefit [12].

Hepatic impairment. Mild to moderate impairment calls for 5 mg immediate-release or 6.25 mg extended-release, and the label says to avoid zolpidem in severe hepatic impairment, where it can contribute to encephalopathy. In cirrhosis, Cmax doubled, AUC rose fivefold and half-life reached 9.9 hours [3].

Compromised breathing. The label asks prescribers to consider respiratory depression risk in patients with sleep apnea, myasthenia gravis, compromised respiration or concurrent opioid use [3].

Pregnancy and children. Zolpidem crosses the placenta; neonates exposed late in the third trimester have had respiratory depression and sedation, some requiring artificial ventilation or intubation, while published data have not shown a clear link with major birth defects [3]. In a trial of 6-to-17-year-olds with ADHD, the drug did not shorten sleep latency and 7% had hallucinations versus 0% on placebo [3][5].

How long treatment is meant to last

Ambien is indicated for short-term treatment of insomnia with difficulty falling asleep, and decreased sleep latency was shown for up to 35 days in controlled trials that ran 4 to 5 weeks. Ambien CR is indicated for difficulty with sleep onset and/or maintenance, with trials up to 3 weeks using polysomnography and 24 weeks using patient-reported measures in adults [3][5]. The label directs use of the lowest effective dose, treatment as short as possible, no long-term use without re-evaluation, and re-evaluation if insomnia persists after 7 to 10 days [3].

The American College of Physicians recommends cognitive behavioral therapy for insomnia as the initial treatment for chronic insomnia in all adults (strong recommendation, moderate-quality evidence), with medication added through shared decision-making when CBT-I alone is unsuccessful (weak recommendation, low-quality evidence) [13].