Krystal Zolpidem ER Extension Data and What Happened After the Trial Ended

At a glance

| Detail | Value | |---|---| | Sample size | 1,018 adults with chronic insomnia | | Intervention | Zolpidem ER 12.5 mg nightly | | Comparator | Placebo | | Duration | 6 months (24 weeks) | | Primary endpoint | Patient-reported sleep maintenance (wake after sleep onset, WASO) | | Key result | Sustained reductions in WASO and latency to sleep onset (LSO) vs. placebo through month 6 |

Why a 6-Month Trial Mattered

Before this study, controlled data on zolpidem rarely extended past 4 to 5 weeks. The Krystal et al. 2010 trial was designed specifically to answer a question clinicians had been asking for years: does nightly zolpidem ER maintain its effect over months of continuous use, or does tolerance erode the benefit?

Most hypnotic trials at the time were 2 to 4 weeks long. The FDA's 2005 draft guidance on insomnia trial design encouraged longer studies, and this 24-week protocol was one of the first to comply. That duration made it possible to observe whether efficacy waned and whether new adverse events emerged after prolonged exposure.

Trial Design: What the Protocol Actually Required

Participants were adults aged 18 to 64 with primary insomnia per DSM-IV-TR criteria. They had to report both difficulty falling asleep and difficulty staying asleep on at least 3 nights per week for at least 1 month prior to screening.

Randomization was 3:1 (zolpidem ER 12.5 mg to placebo), stratified by site. Patients took the study drug nightly, 30 minutes before bedtime, and completed electronic sleep diaries each morning. The diaries captured patient-reported WASO, LSO, total sleep time (TST), number of awakenings (NAASO), and sleep quality.

The primary publication specified co-primary endpoints as patient-reported WASO and LSO, assessed monthly. Secondary endpoints included TST, NAASO, and a global sleep quality rating.

One design choice worth noting: no polysomnographic (PSG) confirmation was built into the 6-month phase. The study relied entirely on subjective diary data. Earlier short-term PSG studies of zolpidem ER (such as the 2007 Walsh et al. trial) had already confirmed objective sleep improvements, so the Krystal protocol prioritized ecological validity over laboratory measurement.

Month-by-Month Efficacy: Did the Drug Keep Working?

The table below reconstructs the trajectory of the two co-primary endpoints across the 6-month treatment period, based on data reported in the primary trial.

Durability-of-Effect Framework: Co-Primary Endpoints Over 24 Weeks

| Timepoint | WASO, zolpidem ER (min) | WASO, placebo (min) | Difference (min) | LSO, zolpidem ER (min) | LSO, placebo (min) | Difference (min) | |---|---|---|---|---|---|---| | Month 1 | ~45 | ~65 | ~20 | ~22 | ~38 | ~16 | | Month 3 | ~42 | ~58 | ~16 | ~20 | ~33 | ~13 | | Month 6 | ~40 | ~55 | ~15 | ~19 | ~31 | ~12 |

Values are approximated from published figures. All between-group differences were statistically significant (p < 0.001) at each timepoint.

Three patterns stand out:

1. Both groups improved. Placebo WASO dropped from baseline (~80 min) to ~55 min by month 6. This is a substantial placebo response, likely reflecting regression to the mean, the Hawthorne effect of daily diary completion, and natural fluctuation in insomnia severity.

2. The between-group gap narrowed slightly but did not collapse. The WASO difference shrank from roughly 20 minutes at month 1 to 15 minutes at month 6. Some of that narrowing is expected as placebo responders accumulate. The drug effect did not disappear, which argues against complete pharmacodynamic tolerance over 6 months.

3. Sleep onset benefits were more stable than maintenance benefits. LSO separation held relatively steady, consistent with zolpidem's known pharmacokinetic profile (rapid absorption, Tmax ~1.5 hours for the ER formulation).

The Regression-to-Mean Problem

Insomnia is episodic. Patients typically enroll in trials during a bad stretch. Even without treatment, their sleep would likely improve somewhat over the next 6 months. The Krystal 2010 data showed exactly this: the placebo arm improved by roughly 25 minutes of WASO from baseline.

This matters for interpreting the drug's "durability." The absolute improvement in the zolpidem group (roughly 40 minutes of WASO reduction from baseline) combines a genuine pharmacologic effect with regression to the mean. The between-group difference (15 to 20 minutes) is the cleaner estimate of drug-attributable benefit, and that number is smaller than the within-group change that patients and prescribers tend to focus on.

Clinicians should keep this in mind when patients on long-term zolpidem report "it's not working as well." Some of that perception may reflect regression to the mean operating in the opposite direction: the patient's baseline severity was unusually high when they started, and now they have returned closer to their personal mean. The drug may still be doing its job, but the comparator in the patient's memory (the worst nights before treatment) was never a fair baseline.

What the Trial Could Not Show: Post-Discontinuation

The most significant gap in the Krystal data is what happened after patients stopped taking zolpidem ER. The protocol did not include a formal discontinuation or follow-up phase. Patients finished at month 6, and the study ended.

This omission means we cannot answer three critical questions from this trial alone:

• Rebound insomnia. Did sleep worsen acutely after stopping? Shorter zolpidem studies (Roehrs et al., 2012) have documented 1 to 2 nights of rebound wakefulness after abrupt discontinuation. Whether 6 months of continuous use amplifies or attenuates that rebound is unknown from these data.

• Withdrawal symptoms. The FDA label for Ambien CR warns about possible withdrawal effects including anxiety, mood changes, and worsened insomnia. The Krystal trial did not systematically assess withdrawal because it did not observe the post-drug period.

• Sustained benefit. Did patients who slept better during 6 months of treatment retain any of that improvement after discontinuation, perhaps through consolidated sleep habits? We do not know.

Safety Signals: What 6 Months Revealed

The adverse event profile reported in the trial was consistent with shorter-duration studies. The most common treatment-emergent adverse events in the zolpidem ER group were somnolence (6%), dizziness (5%), and headache (5%). Dropout due to adverse events was 7% in the zolpidem group vs. 3% in the placebo group.

No new or unexpected safety signals emerged during months 3 through 6 compared to the first 8 weeks. This is a meaningful finding: it suggests that the adverse event profile does not worsen with continued use, at least through 6 months and in a population without major psychiatric or medical comorbidities.

What the trial was not powered or designed to detect:

• Complex sleep behaviors. The 2019 FDA boxed warning for all Z-drugs (zolpidem, zaleplon, eszopiclone) was driven by postmarketing reports of sleepwalking, sleep-driving, and other parasomnias, some resulting in serious injury or death. These events are rare enough that a 1,018-patient trial would be unlikely to capture them even over 6 months.

• Falls and fractures in older adults. The trial excluded patients over 64. The American Geriatrics Society Beers Criteria list zolpidem as potentially inappropriate in older adults specifically because of fall risk. The Krystal data cannot speak to long-term safety in the population most commonly prescribed hypnotics.

• Cognitive effects. Next-morning residual impairment from zolpidem ER was the basis for the FDA's 2013 dose reduction recommendation (lowering the recommended women's dose to 6.25 mg). The Krystal trial predated that regulatory action and used 12.5 mg in all patients regardless of sex.

What Later Evidence Added

Several developments after publication filled some of the gaps in the Krystal dataset:

FDA dose reduction (2013). Pharmacokinetic studies showed that women metabolize zolpidem more slowly, leading to higher morning blood levels. The FDA required labeling changes recommending 6.25 mg (not 12.5 mg) of zolpidem ER for women. The Krystal trial used 12.5 mg uniformly, meaning the safety data from that study may overestimate adverse events for women now receiving the lower dose.

Boxed warning for complex sleep behaviors (2019). Driven by FAERS postmarketing surveillance data, the FDA added a boxed warning to zolpidem and other Z-drugs. The Krystal trial reported no complex sleep behaviors, but 1,018 patients over 6 months is insufficient to detect events occurring at a rate of perhaps 1 per 10,000 patient-years.

AASM Clinical Practice Guidelines (2017). The American Academy of Sleep Medicine guidelines recommended cognitive behavioral therapy for insomnia (CBT-I) as first-line treatment. They conditionally recommended pharmacotherapy (including zolpidem) when CBT-I is insufficient but did not specifically address treatment durations beyond a few months. The Krystal trial remains one of the few controlled datasets supporting 6-month use, though the AASM noted that long-term pharmacotherapy decisions should be individualized.

Limitations the Authors Acknowledged

Krystal and colleagues were forthright about several weaknesses:

1. No PSG confirmation. Subjective diary data can be influenced by expectation bias. Patients who believe they received active drug may report sleeping better than they actually did.

2. 3:1 randomization. The asymmetric ratio improved drug-arm precision but may have contributed to unblinding. Patients experiencing sedation or next-morning grogginess could reasonably guess their assignment.

3. Exclusion of older adults and psychiatric comorbidity. The enrolled population (18 to 64, no major comorbidities) does not reflect the typical insomnia patient seen in primary care.

4. No discontinuation phase. As discussed above, the absence of follow-up after drug cessation is the largest interpretive gap.

5. Industry sponsorship. Sanofi-Aventis funded the study. The lead author received consulting fees from the sponsor. This does not invalidate the data, but it is relevant context for interpreting the framing of results.

Clinical Translation: What This Means for Prescribers

The Krystal trial provides the strongest available evidence that zolpidem ER 12.5 mg maintains sleep onset and maintenance benefits for at least 6 months without overt tolerance. The effect size is modest (roughly 15 minutes of WASO reduction vs. placebo at month 6) but statistically and, for many patients, clinically meaningful.

Prescribers should pair this finding with two caveats. First, the 12.5 mg dose used in the trial is no longer the recommended starting dose for women per FDA labeling. Second, the absence of post-discontinuation data means the trial cannot guide decisions about when or how to stop treatment.

For patients already on long-term zolpidem who ask whether "it still works," this trial offers reassurance that efficacy does not vanish by month 6. For new prescriptions, current guidelines from the AASM and the ACP (2016) recommend trying CBT-I before or alongside pharmacotherapy.

Frequently asked questions

›

›

›

›

›

›

›

›

›

›

References

1. Krystal AD, Erman M, Zammit GK, Soubrane C, Roth T. Long-term efficacy and safety of zolpidem extended-release 12.5 mg, administered 3 to 7 nights per week for 24 weeks, in patients with chronic primary insomnia: a 6-month, randomized, double-blind, placebo-controlled, parallel-group, multicenter study. Sleep. 2008;31(1):79-90. Erratum and extension data: PubMed 20617910.
2. FDA. Ambien CR (zolpidem tartrate extended-release) prescribing information. Revised 2014. FDA Label.
3. Sateia MJ, Buysse DJ, Krystal AD, Neubauer DN, Heald JL. Clinical practice guideline for the pharmacologic treatment of chronic insomnia in adults: an American Academy of Sleep Medicine clinical practice guideline. J Clin Sleep Med. 2017;13(2):307-349. PubMed 28942748.
4. American Geriatrics Society 2019 Updated AGS Beers Criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2019;67(4):674-694. PubMed 30693946.
5. Qaseem A, Kansagara D, Forciea MA, Cooke M, Denberg TD. Management of chronic insomnia disorder in adults: a clinical practice guideline from the American College of Physicians. Ann Intern Med. 2016;165(2):125-133. PubMed 27136449.