Your First Ambien Prescription Predicts Whether You’re Still Taking It in Seven Years

You take it for two weeks after surgery. You take it for a month during a divorce. You take it for six weeks because your doctor’s office was busy and the easiest thing to do was call in a bigger supply so you wouldn’t have to come back. None of those decisions feel like the moment that determines the next seven years of your life. A study published this week in JAMA Network Open says one of them might be exactly that.

The single biggest predictor of whether you’re still taking a Z-drug seven years from now isn’t your diagnosis, your age, or how bad your insomnia was. It’s how many days your very first prescription covered.

The study, led by Jonathan N. Cloughesy at the University of Southern California, followed 129,287 patients prescribed zolpidem (Ambien), eszopiclone (Lunesta), or zaleplon (Sonata) — the class of sedative-hypnotics known as “Z-drugs,” built as a supposedly gentler alternative to benzodiazepines for short-term insomnia — using claims data spanning January 2007 to February 2024. The average initial prescription covered 27 days, already well past the 7-to-10-day window most sleep-medicine guidelines recommend for a first fill. What the researchers found next is the part worth sitting with.

Patients whose first prescription covered 1 to 14 days refilled within the first year 15.6% of the time. Patients whose first prescription covered 35 days or more refilled 49.5% of the time — more than three times the rate, from the same drug, prescribed to comparable patients, differing only in how many pills came in the first bottle. The gap doesn’t close with time. Look out seven years, and patients who started on a short supply were still refilling at 7.9%. Patients who started on 35-plus days were refilling at 18.9% — nearly one in five, still taking a drug that was approved as a short-term bridge, more than half a decade after their doctor wrote the first script. Cloughesy’s own read of the finding, reported by Medical Xpress, was measured: the initial prescription “may play an important role in shaping long-term use.” The data underneath that sentence is less measured. It’s a straight line.

Here’s the layman version of why that line exists, because the mechanism is not mysterious once you say it plainly: a longer first supply gives your body more nights to adapt to falling asleep with help, and adaptation is exactly what makes stopping feel impossible later — not weakness, not a character problem, a nervous system that learned a pattern the same way it learns any other pattern, through repetition. Insomnia itself creates anxiety about not sleeping; a Z-drug interrupts that anxiety fast enough to feel like relief; and the fear of the bad nights returning becomes its own reason to keep refilling, whether or not the original problem — the divorce, the surgery, the busy stretch at work — resolved months ago.

This lands in the middle of a policy conversation that was already moving. In January, the FDA required an updated boxed warning across the entire benzodiazepine class — not Z-drugs specifically, but the closely related sedative-hypnotic family — spelling out abuse, dependence, and withdrawal-seizure risk in language stronger than the prior label carried. And Columbia’s Mailman School of Public Health has been tracking a related, uncomfortable trend: the decade-long decline in benzodiazepine prescribing among older Americans — from 7.2% to 4.7% of adults 56 and older between 2018 and 2022 — has stalled since 2020, with prescribing actually climbing again among patients 75 and older and in long-term care pharmacies. Sedative-hypnotic dependence, in other words, isn’t a problem the system solved and moved past. It’s a problem that’s been quietly rebuilding in the population least equipped to safely withdraw from it — older adults, who face sharply elevated fall and cognitive-impairment risk on this drug class.

Sedative-hypnotic dependence, in other words, isn’t a problem the system solved and moved past.

None of this is a case against Z-drugs existing. Short-term insomnia is real, miserable, and worth treating. What the JAMA Network Open data actually argues for is narrower and more useful than a ban: the size of the very first bottle is a lever clinicians already have their hand on, with a seven-year effect attached to it that most prescribers have no reason to know about unless someone hands them this exact study. If you’re the one holding a 30-day script for something you were told would help you sleep through a hard month, the honest thing this research says isn’t that you did something wrong. It’s that the decision that mattered most happened at the pharmacy counter, before you’d taken a single pill — and it wasn’t yours.