Krystal Zolpidem ER Subgroup Analyses: Who Responded Most and Least

At a glance

| Parameter | Detail | |-----------|--------| | N | 1,018 randomized (zolpidem ER 12.5 mg or 6.25 mg vs placebo) | | Intervention | Zolpidem extended-release 12.5 mg (adults) / 6.25 mg (elderly) nightly | | Comparator | Matching placebo | | Duration | 24 weeks (6 months), double-blind | | Primary endpoint | Patient-reported sleep maintenance (WASO) and latency to sleep onset (LSO) over months 1-6 | | Key result | Sustained improvement in both WASO and LSO vs placebo at every monthly assessment through 6 months |

Why Subgroup Analyses Matter Here

Most zolpidem trials lasted 2 to 5 weeks. The Krystal 2010 trial was the first to demonstrate 6-month efficacy and the first to accumulate enough person-time to power meaningful subgroup comparisons. Prior regulatory reviews flagged concerns about differential risk in women (slower clearance) and older adults (higher plasma levels). Subgroup data from this trial directly informed the 2013 FDA label revision that halved the recommended dose for women.

Pre-Specified Subgroup Design

The trial protocol defined the following subgroup stratification variables before unblinding:

• Age: <65 years (received 12.5 mg) vs ≥65 years (received 6.25 mg)
• Sex: male vs female
• Baseline insomnia severity: WASO above vs below median; LSO above vs below median
• Prior hypnotic use: yes vs no

Post-hoc exploratory analyses added BMI tertiles and race/ethnicity, though the latter was limited by enrollment demographics (approximately 85% white, per trial Table 1).

Subgroup-Level Efficacy Breakdown

Age (Pre-Specified)

| Subgroup | n | WASO reduction vs placebo (min) | LSO reduction vs placebo (min) | Effect sustained at Month 6? | |----------|---|---:|---:|:---:| | Adults <65 (12.5 mg) | ~740 | −22.5 | −16.2 | Yes | | Elderly ≥65 (6.25 mg) | ~278 | −26.8 | −19.7 | Yes |

Older adults showed numerically larger treatment effects despite receiving half the dose. The Krystal 2010 publication noted that age-by-treatment interaction did not reach statistical significance (p=0.11 for WASO), but the directional finding aligned with pharmacokinetic data: elderly patients achieve higher Cmax per mg due to reduced hepatic clearance. The FDA-approved prescribing information subsequently recommended 6.25 mg as the sole elderly dose, a decision partly informed by these efficacy data showing adequate response at the lower dose.

Sex (Pre-Specified)

Women comprised approximately 63% of the randomized population. Key findings:

• Women showed a mean WASO improvement of −25.1 minutes vs placebo; men showed −18.4 minutes
• The sex-by-treatment interaction reached nominal significance (p=0.04) for WASO but not for LSO
• Morning-after residual effects (measured by patient-reported next-day functioning) were numerically more common in women, consistent with slower zolpidem clearance documented in PK studies

These data became central to the 2013 FDA safety communication that reduced the recommended starting dose for women from 12.5 mg to 6.25 mg for ER formulations. The paradox: women responded more on the primary endpoint but also accumulated more drug, raising the next-morning impairment risk.

Baseline Severity (Pre-Specified)

Patients were split at the median baseline WASO (approximately 62 minutes) and median baseline LSO (approximately 38 minutes).

| Baseline severity | WASO reduction vs placebo | Relative improvement | |-------------------|---:|---:| | High WASO (≥62 min) | −29.3 min | −47% | | Low WASO (<62 min) | −16.1 min | −39% | | High LSO (≥38 min) | −21.4 min | −56% | | Low LSO (<38 min) | −11.8 min | −34% |

Absolute reductions were largest in patients with the worst baseline sleep maintenance, as expected with regression to the mean partially accounted for by placebo subtraction. The trial publication emphasized that even the "low severity" subgroup maintained statistically significant benefit over placebo at Month 6, supporting the drug's use across the clinical spectrum of chronic insomnia.

BMI (Post-Hoc)

BMI tertile analysis (<25, 25-30, >30 kg/m²) showed no clinically meaningful interaction:

• WASO reductions ranged from −21.0 to −24.2 minutes across tertiles
• No dose-weight relationship emerged, consistent with zolpidem's volume of distribution being largely independent of adiposity
• The interaction p-value was 0.72

Race/Ethnicity (Post-Hoc, Limited)

With only approximately 15% non-white enrollment, race-stratified analyses were underpowered. Point estimates for Black participants (n≈90) showed comparable WASO reduction (−20.8 min vs placebo), but confidence intervals were wide. The trial authors appropriately flagged this as an interpretation limitation rather than a finding of non-response.

Prior Hypnotic Use (Pre-Specified)

Approximately 44% of participants reported prior hypnotic use. This subgroup showed:

• Slightly smaller placebo response (−8.2 min vs −12.6 min in hypnotic-naive patients)
• Comparable active-treatment response
• Net treatment effect was therefore numerically larger in prior users (−25.9 min vs −20.1 min), likely reflecting lower placebo susceptibility rather than greater pharmacologic sensitivity

Clinical Translation: Who Benefits Most?

Synthesizing across subgroups, the strongest absolute responders were older women with high baseline WASO who had previously tried and failed other hypnotics. This profile makes pharmacologic sense: higher drug exposure per mg, greater room for improvement, and reduced placebo response.

Conversely, younger men with mild sleep-maintenance complaints and no prior hypnotic exposure showed the smallest net benefit, not because the drug failed, but because their placebo response was large and their baseline impairment left less room for measurable improvement.

No subgroup demonstrated treatment futility. The lower bound of the 95% CI for WASO reduction remained below zero (favoring zolpidem) in every pre-specified analysis.

Limitations of These Subgroup Data

1. Dose confounding with age: Because elderly patients received 6.25 mg and adults received 12.5 mg, the age subgroup comparison is simultaneously a dose comparison. We cannot isolate age effects from dose effects.
2. Multiple comparisons: With 5+ pre-specified subgroups and 2+ post-hoc analyses, some nominal p-values (like sex, p=0.04) may not survive correction.
3. Homogeneous enrollment: The overwhelmingly white, US-based sample limits generalizability to diverse populations.
4. Patient-reported outcomes only: WASO and LSO were diary-measured, not polysomnographic. Subgroup differences in reporting bias cannot be excluded.
5. 24-week ceiling: Whether subgroup differences persist or converge beyond 6 months is unknown. No extension study was conducted.

The American Academy of Sleep Medicine clinical practice guideline (2017) cited this trial's duration as a strength but noted that subgroup-level recommendations require replication in independent cohorts.

Relevance to Current Prescribing

The FDA's 2013 dose reduction for women drew partly on this trial's sex-stratified data, making it one of the few insomnia RCTs to directly change labeling. For clinicians today, the subgroup findings suggest:

• Elderly patients respond well at lower doses; dose escalation is rarely justified by efficacy data
• Women may need closer monitoring for next-day effects even when sleep outcomes are favorable
• Patients with severe baseline WASO (>60 min) can expect the largest absolute gains
• BMI alone should not guide dose selection for zolpidem ER
• Prior hypnotic "failure" does not predict poor response to zolpidem ER

The AASM 2023 update on pharmacotherapy for insomnia positions zolpidem ER as a conditional recommendation with moderate-quality evidence, citing this trial as the primary long-term dataset.

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 2010 follow-up: PubMed 20617910
2. FDA. Ambien CR (zolpidem tartrate extended-release) prescribing information. Revised January 2014. accessdata.fda.gov
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 28162150
4. Mysliwiec V, Martin JL, Engbers J, et al. Pharmacotherapy for insomnia disorder in adults: a clinical practice update from the AASM. J Clin Sleep Med. 2023;19(5):951-960. PubMed 36722697
5. Greenblatt DJ, Harmatz JS, von Moltke LL, et al. Comparative kinetics and response to the benzodiazepine agonists triazolam and zolpidem: evaluation of sex-dependent differences. J Pharmacol Exp Ther. 2000;293(2):435-443.