Zolpidem: the legal and safety record

January 2013: the dose for women was cut in half

This is a dosing-science story, and it is worth telling straight. Driving-simulation and laboratory data submitted to FDA showed that zolpidem blood levels above roughly 50 ng/mL appear capable of impairing driving. In pharmacokinetic trials of about 250 men and 250 women taking 10 mg immediate-release zolpidem, about 15% of women and 3% of men still exceeded that level 8 hours after dosing — morning-commute time. On extended-release 12.5 mg it was about 33% of women and 25% of men [1].

The mechanism is unglamorous: women eliminate zolpidem from the body more slowly than men [1][2]. So FDA required the manufacturers to halve the recommended dose for women — 10 mg to 5 mg for immediate-release products, 12.5 mg to 6.25 mg for extended-release — and to tell prescribers to consider the lower dose for men too, since “in many men, the 5 mg dose provides sufficient efficacy” [1]. The label changes were approved in April 2013, with the required text spelled out in the approval letter [3], and they are in the current label today: 5 mg for women, 5 or 10 mg for men [5]. Twenty years of one-size-fits-all dosing, revised by pharmacokinetics — not by a lawsuit.

April 2019: the boxed warning, in FDA’s own numbers

In April 2019, FDA added its most prominent warning — a boxed warning — to zolpidem, eszopiclone, and zaleplon for complex sleep behaviors: sleepwalking, sleep-driving, and engaging in other activities while not fully awake. It also added a contraindication, its strongest warning: anyone who has had such an episode on one of these drugs should not take any of them again [4]. The events can occur at the lowest recommended doses, after a single dose, and with or without alcohol [4].

The evidence base deserves to be stated exactly, because it is both genuinely grave and genuinely small. FDA identified 66 cases over roughly 26 years (1992–2018) in which complex sleep behaviors led to serious injury or death: 20 deaths and 46 serious non-fatal injuries — falls with intracranial hemorrhages and fractures, burns, near-drownings, hypothermia with loss of limb, fatal motor-vehicle collisions, gunshot wounds, apparent suicides. Zolpidem was the drug in 61 of the 66. Patients usually did not remember the events [4].

Now the denominator, from the same document: an estimated 26.6 million US zolpidem prescriptions and 5.1 million patients in 2018 alone [4]. FDA’s wording is “rare but serious,” and its own caveat cuts the other way: the 66 counts only reports FDA received or found in the literature, “so there may be additional cases about which we are unaware” [4]. A page that quotes the 20 deaths without the 26.6 million prescriptions is amplifying; one that quotes the denominator without the deaths is minimizing. You get both.

The courtroom shelf is nearly bare — and that is the finding

For most drugs in this series the legal section is the headline: gabapentin has a $430 million guilty plea behind its psychiatric use [8]. Zolpidem has nothing comparable. What the record does show: in August 2008, Sanofi-Aventis US received a subpoena from the HHS Office of Inspector General and the US Attorney’s Office in San Francisco over Ambien and Ambien CR sales and marketing practices — an investigation of possible false Medicare and Medicaid claims. In November 2010, the federal court in the Northern District of California dismissed the related False Claims Act case without prejudice, and no resolution, payment, or charge ever followed in the public record [7].

So the honest sentence is: no DOJ plea, settlement, or corporate resolution naming zolpidem appears in the record we checked — an investigation opened, went quiet, and closed without a case. We publish that as a verified negative rather than dressing an insinuation up as a finding.

What we excluded, on purpose. Ambien has a body of criminal-case folklore — the “Ambien defense,” where a defendant says the drug explains conduct they do not remember. Those cases are about the people who took the drug, not about the company that sold it; no court in them decided anything about the manufacturer. They do not belong on a corporate legal record, and importing their drama here would make this page worse at its actual job. The pharmacology they gesture at is in the FDA documents above, with numbers attached.

What is not in the record

• No Class I recall (within openFDA’s ~2012-onward coverage window). Ten recall records, eight Class II and two Class III, for packaging, labeling, and manufacturing-practice problems — a repackager label mixup, CGMP deviations, failed dissolution specs — nothing about the molecule itself [6].
• No adjudicated manufacturer-liability verdict over complex sleep behaviors in the record we checked. The 2019 boxed warning is a regulatory instrument, not a court finding of liability, and reading it as either an indictment or an exoneration of anyone is overreach — in either direction.
• No off-label marketing plea. Unlike gabapentin, zolpidem’s prescription volume sits on an actual insomnia approval (December 1992), and its label, approval history, and reported adverse events are on our zolpidem page, caveats first.

Common questions

Does zolpidem (Ambien) have a black box warning?

Yes, since April 2019. FDA required a boxed warning — its most prominent label warning — for complex sleep behaviors: sleepwalking, sleep-driving, and doing other activities while not fully awake, some resulting in serious injuries or death. FDA also added a contraindication: anyone who has ever had an episode of complex sleep behavior on zolpidem, eszopiclone, or zaleplon should not take these drugs again.

Why is the recommended dose lower for women?

Pharmacokinetics, not politics. Women eliminate zolpidem more slowly than men, so morning blood levels run higher. Driving-simulation and laboratory data put the impairment threshold near 50 ng/mL, and in FDA’s trials about 15% of women versus 3% of men still exceeded that level 8 hours after a 10 mg immediate-release dose. In January 2013 FDA required the recommended dose for women be halved: 5 mg immediate-release, 6.25 mg extended-release.

How common are the sleepwalking and sleep-driving events?

FDA’s own numbers, both directions: 66 cases of complex sleep behaviors causing serious injury or death reported over roughly 26 years (1992–2018) — 20 fatal, 46 non-fatal — against an estimated 26.6 million US zolpidem prescriptions and 5.1 million patients in 2018 alone. FDA calls the events “rare but serious” and also notes the count includes only reports it received or found in the literature, so there may be cases it does not know about. Both halves of that sentence are the honest answer.

Was there ever a big Ambien settlement, like the Neurontin one?

Not that the record shows. Sanofi-Aventis disclosed in its SEC filings that it received a federal subpoena in August 2008 over Ambien and Ambien CR sales and marketing practices, and that in November 2010 the related False Claims Act case was dismissed without prejudice. No DOJ plea, settlement, or corporate resolution naming zolpidem appears in the DOJ archives we searched — nothing comparable to the $430 million Warner-Lambert guilty plea over Neurontin. A verified negative, published as one.

What about crimes people say they committed on Ambien?

Deliberately excluded. The so-called “Ambien defense” cases are criminal prosecutions of people who took the drug — they are about individual defendants, not about the manufacturer’s conduct, and no court in them adjudicated anything about the company. Mixing them into a corporate legal record would inflate the page while informing no one. The pharmacology those cases gesture at is covered, precisely, by the FDA documents above.

Have there been serious zolpidem recalls?

No Class I recall — FDA’s most serious category — appears in the openFDA enforcement database (which is reliable only from roughly 2012 onward, so this is a statement about the covered window, not the drug’s whole history). Ten records exist, eight Class II and two Class III, for packaging, labeling, and manufacturing-practice problems rather than anything about the molecule.

Should I stop taking zolpidem because of this page?

No — nothing here is a reason to change treatment on your own, and the label documents withdrawal signs and symptoms after rapid dose reduction or abrupt discontinuation. Even when FDA cut the women’s dose in 2013, its instruction to patients on the old dose was to keep taking it as prescribed until they had talked to their prescriber. The one exception is FDA’s own: if you experience an episode of complex sleep behavior — doing things while not fully awake, or not remembering what you did — stop the medicine and contact your health care professional right away.

Sources

Last verified 2026-09-08. Corrections change this date.