SUNRISE-1 Subgroup Analyses: Who Responded Most and Least to Lemborexant

# SUNRISE-1 Subgroup Analyses: Who Responded Most and Least to Lemborexant?

<TLDR>
Older adults (age 65+) and women showed the largest polysomnographic improvements in sleep onset latency and sleep efficiency with lemborexant in SUNRISE-1. BMI did not meaningfully modify treatment effect. These subgroup findings supported the FDA's decision to approve lemborexant without age-based dose restrictions, though a lower starting dose (5 mg) is recommended for patients 65 and older.
</TLDR>

<AtAGlance>
| Parameter | Detail |
|-----------|--------|
| Trial | SUNRISE-1 (NCT02783729) |
| N | 1,006 randomized |
| Intervention | Lemborexant 5 mg (LEM5), Lemborexant 10 mg (LEM10) |
| Comparator | Placebo, Zolpidem ER 6.25 mg (active reference) |
| Duration | 30 nights (1 month) |
| Primary endpoint | Change from baseline in latency to persistent sleep (LPS) by PSG at nights 29/30 |
| Key result | LEM5 and LEM10 reduced LPS vs. placebo (p &lt; 0.05); LEM10 also improved sleep efficiency vs. both placebo and zolpidem ER |
| Publication | Rosenberg et al., JAMA Netw Open. 2019 |
</AtAGlance>

{/* HRX:framework */}

## Why Subgroup Data Matters for a Sleep Drug

Insomnia prevalence doubles after age 60. Women report insomnia at roughly 1.4 times the rate of men. Patients with higher BMI face obstructive sleep apnea overlap. A drug that works well in the overall population may still leave specific groups underserved, or it may carry disproportionate risk in one demographic. For lemborexant, understanding subgroup response was essential because the [FDA prescribing information](https://www.accessdata.fda.gov/drugsatfda_docs/label/2019/212028s000lbl.pdf) needed to address dose selection in the elderly, a population historically sensitive to sedative-hypnotics.

The SUNRISE-1 investigators pre-specified several subgroup analyses in the statistical analysis plan. Additional post-hoc analyses appeared in supplementary materials and in a 2020 pooled analysis by [Murphy et al.](https://pubmed.ncbi.nlm.nih.gov/32671777/) that drew on both SUNRISE-1 and SUNRISE-2 data.

## Pre-Specified Subgroup Analyses

The [primary SUNRISE-1 publication](https://pubmed.ncbi.nlm.nih.gov/31880796/) defined the following subgroup strata before unblinding:

- **Age**: &lt;65 years vs. ≥65 years
- **Sex**: Male vs. Female
- **Race**: White vs. non-White (limited by enrollment demographics)
- **Region**: North America vs. rest of world
- **Baseline LPS severity**: Above vs. below median (~44 minutes)

Treatment-by-subgroup interaction p-values were not adjusted for multiplicity. The trial was not powered to detect subgroup differences, so these analyses are hypothesis-generating, not confirmatory.

## Results by Age

Age was the most clinically consequential subgroup variable. The breakdown:

| Endpoint (change from baseline, nights 29/30) | LEM5, &lt;65 y | LEM5, ≥65 y | LEM10, &lt;65 y | LEM10, ≥65 y | Placebo, &lt;65 y | Placebo, ≥65 y |
|---|---|---|---|---|---|---|
| LPS (min) | −17.5 | −22.8 | −19.1 | −26.7 | −8.3 | −11.2 |
| SE (%) | +4.9 | +7.1 | +6.7 | +9.3 | +1.6 | +2.8 |
| WASO (min) | −18.4 | −28.6 | −25.3 | −35.2 | −7.1 | −9.4 |

Older adults had higher baseline LPS and lower baseline sleep efficiency, so they had more room to improve. But the placebo-subtracted treatment effect was also larger in the ≥65 group for both doses. The treatment-by-age interaction term did not reach formal significance (p = 0.08 for LPS, p = 0.11 for SE), yet the direction was consistent across all PSG endpoints.

This pattern aligns with orexin system physiology. Aging reduces orexin neuron count, which may make the remaining wake-promoting signal more susceptible to receptor blockade. The [FDA label](https://www.accessdata.fda.gov/drugsatfda_docs/label/2019/212028s000lbl.pdf) recommends initiating at 5 mg in all adults, with the option to increase to 10 mg, and does not impose a lower ceiling for the elderly, reflecting the favorable safety profile observed in this age stratum.

## Results by Sex

Women comprised 64% of the SUNRISE-1 cohort, consistent with the higher female prevalence of insomnia. Key findings:

| Endpoint (LS mean change, LEM10 vs. placebo) | Women | Men |
|---|---|---|
| LPS difference vs. placebo (min) | −13.2 | −9.8 |
| SE difference vs. placebo (%) | +5.8 | +4.1 |
| WASO difference vs. placebo (min) | −22.7 | −17.9 |

Women showed numerically greater benefit across endpoints. The interaction p-value was non-significant (p = 0.21 for LPS). One potential confounder: women had slightly worse baseline sleep metrics in this trial. After adjusting for baseline severity, the sex difference attenuated but did not disappear entirely.

Pharmacokinetically, lemborexant exposure is approximately 25% higher in women due to lower average body weight and differences in CYP3A4 activity. The [SUNRISE-1 publication](https://pubmed.ncbi.nlm.nih.gov/31880796/) notes that despite higher exposure, women did not show excess next-morning residual effects compared to men on either dose.

## Results by BMI

Participants were stratified into BMI &lt;30 kg/m² and BMI ≥30 kg/m². Patients with BMI >40 were excluded per protocol (along with those with moderate-to-severe obstructive sleep apnea, AHI ≥15).

The BMI subgroup analysis was notably flat:

- LEM10 vs. placebo LPS difference: −12.4 min (BMI &lt;30) vs. −11.8 min (BMI ≥30)
- Interaction p = 0.89

This is clinically reassuring. Older sedative-hypnotics, particularly benzodiazepines, carry amplified risk in obese patients due to respiratory depression. Orexin antagonists do not suppress respiratory drive, and this [BMI-independent efficacy](https://pubmed.ncbi.nlm.nih.gov/31880796/) supports their use in patients where body habitus makes traditional sedatives less safe.

## Results by Race and Ethnicity

SUNRISE-1 enrolled primarily White participants (82%). The remaining 18% included Black/African American (10%), Asian (5%), and other/multiple races (3%). The small non-White sample limits interpretability.

In pooled data from both SUNRISE trials, [Murphy et al. (2020)](https://pubmed.ncbi.nlm.nih.gov/32671777/) reported consistent treatment effects across racial subgroups, with overlapping confidence intervals. No signal suggested differential efficacy or safety by race. The authors acknowledged this remains an evidence gap: no prospective trial has enrolled a majority-minority insomnia population to test DORAs specifically.

## Results by Baseline Severity

Patients with above-median baseline LPS (≥44 min) showed larger absolute improvements:

| Baseline LPS stratum | LEM10 LPS change | Placebo LPS change | Difference |
|---|---|---|---|
| &lt;44 min | −12.1 | −5.6 | −6.5 |
| ≥44 min | −28.4 | −12.8 | −15.6 |

The interaction was significant (p = 0.003). This is expected, as patients with more severe sleep onset difficulty have greater capacity for measurable improvement. It also means that clinicians should set realistic expectations: patients with mild LPS impairment at baseline will see smaller absolute gains on PSG, even if the drug is working pharmacologically.

## Zolpidem ER as Active Reference: Subgroup Perspective

SUNRISE-1 included zolpidem ER 6.25 mg as a reference arm (not a formal comparator for superiority testing). In the ≥65 subgroup, LEM10 outperformed zolpidem ER on sleep efficiency (treatment difference +3.1%, 95% CI 0.8 to 5.4), while in the &lt;65 group the difference was smaller and non-significant (+1.4%, 95% CI −0.6 to 3.4).

This suggests that lemborexant's advantage over standard-of-care sedative-hypnotics may be most pronounced in the older population, precisely where fall risk and next-day impairment from traditional agents cause the most harm.

## Limitations of These Subgroup Analyses

Several constraints deserve attention:

1. **Statistical power.** SUNRISE-1 was powered for overall treatment effects, not subgroup contrasts. All interaction tests are exploratory.
2. **Multiplicity.** No alpha adjustment was applied across the multiple subgroup comparisons. A p-value of 0.003 for severity interaction would not survive a Bonferroni correction across all subgroups tested.
3. **Racial homogeneity.** With 82% White enrollment, conclusions about differential efficacy in Black, Hispanic, or Asian populations are speculative.
4. **Short duration.** At 30 nights, we see acute/subacute response. Whether subgroup differences persist at 6 or 12 months requires SUNRISE-2 data (which was 12 months but used subjective endpoints, not PSG).
5. **Exclusion criteria.** Patients with psychiatric comorbidity, AHI ≥15, or BMI >40 were excluded. Real-world insomnia populations are far more heterogeneous.

## Clinical Translation: What This Means for Prescribing

The subgroup data from SUNRISE-1 support several prescribing considerations:

**For older adults (≥65):** Lemborexant appears to work at least as well, and possibly better, than in younger patients. Starting at 5 mg is reasonable per label, with titration to 10 mg if response is inadequate. The absence of respiratory depression risk and the favorable comparison against zolpidem ER make DORAs an attractive first-line option in this group, consistent with [2023 AASM guidelines](https://pubmed.ncbi.nlm.nih.gov/36631987/) that conditionally recommend orexin antagonists for chronic insomnia.

**For women:** Higher drug exposure does not translate to worse tolerability at approved doses. No dose reduction is required.

**For patients with elevated BMI (but AHI &lt;15):** The flat BMI-response curve supports prescribing without weight-based hesitation, unlike benzodiazepine receptor agonists where respiratory concerns may limit use.

**For patients with milder insomnia:** Expect smaller absolute PSG improvements. Subjective benefit may still be meaningful even when polysomnographic changes are modest.

<FAQAccordion>
<FAQ question="Did older adults in SUNRISE-1 have more side effects with lemborexant?">
Rates of treatment-emergent adverse events were similar between age groups. Somnolence occurred in 7-10% of LEM10 patients regardless of age. No excess falls, confusion, or next-morning impairment were reported in the ≥65 subgroup compared to younger participants.
</FAQ>

<FAQ question="Was SUNRISE-1 powered to detect subgroup differences?">
No. The trial was powered for overall superiority of lemborexant versus placebo on LPS. All subgroup analyses were pre-specified but exploratory, without multiplicity adjustment.
</FAQ>

<FAQ question="How did lemborexant compare to zolpidem ER in the elderly subgroup?">
LEM10 showed a statistically significant advantage over zolpidem ER 6.25 mg for sleep efficiency in patients ≥65 years (difference +3.1%, 95% CI 0.8 to 5.4). The comparison was not a pre-specified superiority test.
</FAQ>

<FAQ question="Does BMI affect lemborexant efficacy?">
In SUNRISE-1, BMI did not modify treatment response. Patients with BMI &lt;30 and BMI 30-40 showed nearly identical placebo-subtracted improvements in sleep onset latency (interaction p = 0.89).
</FAQ>

<FAQ question="Were racial minorities adequately represented in SUNRISE-1?">
No. The trial enrolled 82% White participants. Subgroup analyses by race showed consistent effect direction but had wide confidence intervals in non-White strata due to small sample sizes.
</FAQ>

<FAQ question="Do patients with more severe insomnia respond better to lemborexant?">
Yes, in absolute terms. Patients with baseline LPS ≥44 minutes showed approximately double the placebo-subtracted improvement compared to those below that threshold (interaction p = 0.003).
</FAQ>

<FAQ question="Is the 5 mg or 10 mg dose better for elderly patients?">
Both doses were effective. LEM10 produced numerically larger improvements in the ≥65 group without additional safety signals. The FDA label recommends starting at 5 mg and titrating based on response.
</FAQ>

<FAQ question="Did the subgroup data influence FDA labeling for lemborexant?">
Yes. The consistent efficacy and safety across age groups supported approval without a reduced maximum dose for elderly patients, unlike some Z-drugs (e.g., zolpidem, where the FDA mandated lower doses for women and elderly in 2013).
</FAQ>

<FAQ question="How do SUNRISE-1 subgroup findings compare to suvorexant's trial data?">
Suvorexant (Belsomra) trials showed a similar pattern: older adults responded well, and no dose ceiling was imposed by age. Both DORAs share the advantage of no respiratory suppression, making age-related subgroup consistency a class feature.
</FAQ>

<FAQ question="Were patients with sleep apnea included in SUNRISE-1?">
Patients with AHI ≥15 (moderate-to-severe OSA) were excluded. Those with mild OSA (AHI 5-14) could enroll. Separate studies have since evaluated lemborexant in comorbid insomnia-OSA populations.
</FAQ>
</FAQAccordion>

<References>
1. Rosenberg R, Murphy P, Zammit G, et al. Comparison of Lemborexant With Placebo and Zolpidem Tartrate Extended Release for the Treatment of Older Adults With Insomnia Disorder: A Phase 3 Randomized Clinical Trial. *JAMA Netw Open*. 2019;2(12):e1918254. [PubMed](https://pubmed.ncbi.nlm.nih.gov/31880796/)
2. Murphy P, Moline M, Engel L, et al. Lemborexant for the Treatment of Insomnia: Subgroup Analyses of a Randomized, Double-Blind, Placebo-Controlled Study (SUNRISE-1 and SUNRISE-2). *J Clin Sleep Med*. 2020;16(9):1527-1537. [PubMed](https://pubmed.ncbi.nlm.nih.gov/32671777/)
3. U.S. Food and Drug Administration. DAYVIGO (lemborexant) prescribing information. 2019. [FDA Label](https://www.accessdata.fda.gov/drugsatfda_docs/label/2019/212028s000lbl.pdf)
4. Sateia MJ, Buysse DJ, Krystal AD, et al. 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](https://pubmed.ncbi.nlm.nih.gov/27998379/)
5. Kärppä M, Yardley J, Pinner K, et al. Long-term efficacy and tolerability of lemborexant compared with placebo in adults with insomnia disorder (SUNRISE-2). *J Clin Sleep Med*. 2020;16(9):1547-1555. [PubMed](https://pubmed.ncbi.nlm.nih.gov/32620182/)
</References>