A Phase III Study of Eszopiclone in Patients With Insomnia (Study SEP 190-150): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Phase III Study of Eszopiclone in Patients With Insomnia (Study SEP 190-150) (NCT00770692), a randomized phase 3 study. The primary outcome is Incidence of Adverse Events.

2. Clinical Question

3. Trial Design

This was a Phase III, randomized, double-masked, parallel-group study sponsored by Eisai Co., Ltd. The registry record lists insomnia as the condition and 369 participants as the enrollment. The study began in 2008-10, had primary completion in 2010-05, and is marked completed. The brief summary describes a long-term safety evaluation of eszopiclone, with doses assigned by age category.

The four listed groups were elderly participants receiving 1 mg or 2 mg, and non-elderly participants receiving 2 mg or 3 mg. The arm descriptions contain no additional details. The available pairwise calculations use the elderly 1 mg group as the reference for all three comparisons. As a result, the two comparisons involving non-elderly participants also contrast age categories. They should not be read as clean dose-only comparisons.

The primary endpoint covered up to 25 weeks, specified as a 24-week treatment period and a 1-week follow-up. The primary measure was the percentage of participants with adverse events in the safety analysis set. Four secondary outcomes are named: mean change from baseline in sleep latency, wake time after sleep onset, total sleep time, and total number of awakenings. Results for those outcomes are not reported in the registry record.

The randomization method, allocation ratio, operational details of masking, eligibility criteria, and age threshold are not reported in the registry record. The original sample-size justification and detailed statistical plan are also not reported. Accordingly, the available description supports a summary of the study structure and posted safety results, but not a reconstruction of all design decisions.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: 181 non-elderly \& 188 elderly participants were enrolled in the screening period 1 week prior to the first dose. Among these, 20 non-elderly \& 24 elderly participants discontinued during the screening period. 161 non-elderly and 164 elderly participants enrolled. 1 elderly participant enrolled for treatment did not receive treatment.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Safety analysis set: All 161 non-elderly participants who were enrolled in the treatment period were included. All 164 elderly patients who were enrolled in the treatment period were included. The participant who was excluded from the efficacy analysis set was included in the safety analysis set because the participant had evaluable safety data.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Incidence of Adverse Events, measured over Up to 25 weeks (24 weeks treatment period & 1 week follow-up).

The primary endpoint is the incidence of adverse events, expressed as a percentage of participants. The registry definition divides the number of participants with adverse events by the number analyzed in the safety analysis set, then multiplies by 100. An adverse event could be an unwanted disease, symptom, sign, or abnormality in laboratory parameters in a participant who received study drug. The definition explicitly does not require a causal relationship with the study drug.

Investigators or subinvestigators evaluated and recorded adverse events in the case report form. Events occurring after study treatment began were recorded irrespective of whether they were considered related to the drug or study procedures; follow-up data were also recorded. This broad capture is relevant to a safety endpoint: it describes events observed during the specified period rather than events proven to have been caused by treatment.

The timeframe was up to 25 weeks, comprising a 24-week treatment period and a 1-week follow-up. Reporting the proportion of participants with at least one event provides a participant-level summary of how commonly adverse events were recorded in each group. It does not convey event duration, severity distribution, recurrence, or the number of events per participant. Those details, beyond the serious-event counts and terms listed separately, are not reported in the registry record.

7. Results Posted in the Registry

• Eszopiclone 2 mg- Elderly: 79.5% (66/83) versus 81.5% (66/81) on control; risk difference -2.0 percentage points (95% CI -14.1 to 10.2), risk ratio 0.98 (0.84–1.13); unadjusted two-proportion z-test p = 0.751.
• Eszopiclone 2 mg- Non-elderly: 82.1% (69/84) versus 81.5% (66/81) on control; risk difference 0.7 percentage points (95% CI -11.1 to 12.4), risk ratio 1.01 (0.87–1.16); unadjusted two-proportion z-test p = 0.912.
• Eszopiclone 3 mg- Non-elderly: 87.0% (67/77) versus 81.5% (66/81) on control; risk difference 5.5 percentage points (95% CI -5.8 to 16.8), risk ratio 1.07 (0.93–1.22); unadjusted two-proportion z-test p = 0.341.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the endpoint estimate?

The primary measure is a participant-level proportion: the number of people in a group who had an adverse event divided by the number analyzed in that group's safety analysis set. The registry reports this as a percentage. This kind of endpoint answers how common it was for participants to experience at least one recorded event during the observation period. It is different from counting every event, since one participant with several events still contributes as a participant with an event to the incidence measure.

The registry definition is intentionally broad. An adverse event does not necessarily have a causal relationship with study drug. Thus, the percentage is an account of events recorded after treatment began, not a percentage of events demonstrated to have been caused by eszopiclone. The investigator or subinvestigator evaluated and recorded events, including those judged unrelated to treatment or study procedures.

How should a risk difference be read?

A risk difference subtracts the event proportion in the reference group from the event proportion in the comparison group. A negative estimate means the observed proportion was lower in the comparison group; a positive estimate means it was higher. The recalculated estimates are -0.020 for elderly participants receiving 2 mg, 0.007 for non-elderly participants receiving 2 mg, and 0.055 for non-elderly participants receiving 3 mg, each relative to elderly participants receiving 1 mg.

The confidence interval expresses uncertainty around an estimated contrast. In repeated samples under the assumptions of the method, confidence intervals constructed in this way would contain the underlying contrast at the stated confidence level at the corresponding long-run rate. For these comparisons, each reported interval includes the no-difference value. That means the estimates do not distinguish clearly between no contrast and a range of positive or negative differences. An interval is not a probability statement that the true value lies within that particular interval.

What do the p-values say?

The reported p-values are unadjusted: 0.751 for elderly 2 mg, 0.912 for non-elderly 2 mg, and 0.341 for non-elderly 3 mg, each compared with the elderly 1 mg group. A p-value measures how compatible the observed data are with a specified no-difference model, assuming that model and the calculation's assumptions. It is not the probability that the no-difference model is true, and it does not measure the size or clinical importance of a difference. These p-values should be read alongside the estimates and their intervals rather than as standalone conclusions.

The estimates and p-values are unadjusted recalculations from the posted summary statistics. The registry record does not report the sponsor's prespecified comparison method, whether covariates were included, or a statistical model for the primary endpoint. It also does not report an adjustment for making several comparisons. Therefore, the recalculated tests should not be mistaken for a complete account of the trial's original inferential analysis.

Why does the comparison structure matter?

The reference group is elderly participants receiving 1 mg. The comparison with elderly participants receiving 2 mg holds the broad age category constant, while the comparisons with non-elderly participants change both age category and dose. As a result, the latter estimates cannot isolate a dose contrast from an age-group contrast. Randomization is listed for the study, but the registry record does not describe allocation within age strata or provide details needed to evaluate how the four groups were formed.

Even within the elderly groups, the reported contrast is an unadjusted comparison of observed adverse-event proportions. The registry record does not report adjustment for baseline characteristics, handling of missing endpoint data, or a prespecified analysis population beyond the description of the safety analysis set. The estimates are thus best interpreted as simple contrasts of posted group summaries.

Why is a broad safety endpoint useful, and what does it omit?

For a safety evaluation, a broad adverse-event definition can capture a wide range of unwanted occurrences during treatment and follow-up, including events not judged drug-related. That supports a descriptive account of what was recorded among participants exposed to study treatment. At the same time, a single incidence percentage does not distinguish mild from serious events, related from unrelated events, or isolated events from recurring events. The registry separately lists serious-event counts and selected terms, but the primary percentage alone cannot answer those questions.

The endpoint timeframe is up to 25 weeks, consisting of treatment and follow-up. It does not characterize events outside that period. The registry record does not report longer-term ascertainment or details of event adjudication. The statistical interpretation should therefore remain tied to the recorded endpoint, its stated period, and the populations described in the record.

10. Confidence Intervals and Precision

The confidence intervals around the recalculated risk differences are broad relative to their point estimates. For the comparison of elderly participants receiving 2 mg with the elderly 1 mg reference group, the estimate is -0.020 and the 95% confidence interval runs from -0.141 to 0.102. For non-elderly participants receiving 2 mg, the estimate is 0.007 and the interval runs from -0.111 to 0.124. For non-elderly participants receiving 3 mg, the estimate is 0.055 and the interval runs from -0.058 to 0.168.

All three intervals include the no-difference value and extend to both negative and positive contrasts. The data therefore leave uncertainty about both the direction and size of the differences. The point estimates should not be treated as exact effects; the intervals show that the underlying contrasts may differ meaningfully from those estimates in either direction.

Each group in these comparisons has a denominator between 77 and 84 participants. With group sizes of this scale, uncertainty in proportions can remain substantial, particularly when observed event proportions are high and close to one another. General statistical explanation: a confidence interval is a range of values compatible with the data under the calculation's assumptions; a wider interval indicates less precision than a narrower interval. The interval does not prove equality when it includes the no-difference value.

The posted percentages and recalculated contrasts also answer different presentation needs. The arm-level percentages describe each group, while the risk differences quantify the contrast with the reference group. Neither resolves the age-and-dose entanglement in the non-elderly comparisons. The registry record does not report an adjusted estimate or a prespecified precision target for the primary endpoint.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size figures are planning illustrations based on the observed effect, not the original trial's design justification. For this binary participant-level outcome, the figures are participants per arm, not total events. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each scenario uses the stated fraction of the observed effect: 100%, 75%, or 50%.

The 80% power column gives the illustrative number of participants per arm for a design that would have an 80% chance of detecting the chosen effect under those assumptions, if that effect were the true one. The 90% power column gives the corresponding per-arm planning number for a 90% chance. These are prospective design quantities, not evidence about the probability that the trial's findings are correct. This is not observed power.

The displayed scenarios make the effect-size dependence visible. Planning for 75% or 50% of the observed effect requires more participants per arm than planning for 100% of the observed effect. Smaller effects are harder to distinguish from sampling variability, so they require larger planned groups under otherwise comparable assumptions. The required figures differ substantially among the three comparisons because their observed contrasts differ in size.

The registry record does not report the trial's own sample-size assumptions or justification. The illustrative figures should therefore be kept separate from the enrollment of 369 and from the groups actually analyzed. They do not show that the completed study was designed to achieve a particular power for its primary endpoint, and they do not replace the uncertainty conveyed by the reported confidence intervals.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.815; treatment proportion p₁ = 0.795; average p̄ = 0.805.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.5603 = 1.0982.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5601 = 0.4714.
4. n = (1.0982 + 0.4714)² / (-0.0196)² = 6,391.0, rounded up to 6,392 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.814814814814815                          # control proportion
p1 <- 0.795180722891566                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(770692)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 6,392 (80% power), 8,556 (90% power)
 75% of observed effect: n per arm = 11,254 (80% power), 15,066 (90% power)
 50% of observed effect: n per arm = 25,076 (80% power), 33,570 (90% power)
power.prop.test: 6391 
Simulated power at n = 6,392 per arm: 0.797 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

The posted pairwise calculations make three comparisons with the elderly 1 mg reference group. When several comparisons are considered, the chance of at least one apparently unusual result can increase if each is judged separately at the same significance threshold. Methods such as prespecifying a primary contrast or adjusting for multiple comparisons can address that issue, but the registry record does not report whether any such approach was planned or used.

The displayed p-values are unadjusted recalculations: 0.751, 0.912, and 0.341. They should not be interpreted as multiplicity-adjusted results. The registry record does not report a prespecified statistical analysis plan, the original model for the primary endpoint, or an inferential hierarchy among the four groups and the listed secondary outcomes. The original sample-size justification and any interim-analysis plan are not reported in the registry record.

The safety analysis set is described as including all 161 non-elderly participants and all 164 elderly participants enrolled in the treatment period. The text also begins to explain why a participant excluded from the efficacy analysis set was included in the safety set, but that explanation is incomplete in the registry record. Detailed inclusion and exclusion rules for the safety population, and the denominators used for each posted arm-level result, are not fully described beyond the posted group counts.

Missing-data procedures are not reported in the registry record. It is not stated how participants with incomplete follow-up, missing event information, or deviations from treatment were handled in the primary analysis. Nor does the record describe sensitivity analyses for missing outcomes. The primary endpoint's event definition includes data collected during follow-up, but the available information does not establish that all participants had complete observation through the full period.

Four secondary outcomes are named, but their results are not reported in the registry record. The number and structure of analyses planned for those outcomes, and whether multiplicity was addressed across primary and secondary endpoints, are also not reported. These omissions limit assessment of selective testing and of how the primary endpoint fit within the broader analysis plan.

14. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 2% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: clavicle fracture (1.3%), completed suicide (1.3%), type 1 diabetes mellitus (1.2%), major depressive disorder (1.2%), sick sinus syndrome (1.2%).

The serious adverse-event counts are listed alongside the number at risk in each group. In non-elderly participants, 2 of 84 in the 2 mg group and 2 of 77 in the 3 mg group had serious adverse events. In elderly participants, 3 of 81 in the 1 mg group and 3 of 83 in the 2 mg group had serious adverse events. These are counts of participants, as represented in the registry record, and are descriptive rather than adjusted comparisons.

The registry also lists selected serious-event terms with percentages: Clavicle fracture, 1.3%; Completed suicide, 1.3%; Type 1 diabetes mellitus, 1.2%; Major depressive disorder, 1.2%; Sick sinus syndrome, 1.2%; and Loss of consciousness, 1.2%. The denominator and group-specific distribution for these listed percentages are not reported in the registry record, so they should not be used to infer a rate within any particular arm.

Death counts are not reported in the registry record. Serious events and the broad primary adverse-event endpoint are distinct summaries; serious-event counts do not explain the full pattern of adverse events in each group.

15. What the Result Does, and Does Not, Mean

The results describe the frequency of recorded adverse events in four age-and-dose groups during a period of up to 25 weeks. The posted arm-level percentages range from 79.5% to 87.0%. The recalculated unadjusted risk differences relative to elderly participants receiving 1 mg are close to zero for the elderly 2 mg and non-elderly 2 mg comparisons, while the non-elderly 3 mg estimate is positive. The confidence intervals for all three comparisons include the no-difference value and allow for contrasts in either direction.

These findings do not establish that the groups have equivalent adverse-event incidence. A non-small p-value is not proof of no difference, and an interval spanning no difference can still include values that may matter. The results also do not show that any event was caused by treatment: the registry definition includes events regardless of causal relationship. They do not establish a dose-response pattern because the comparisons involving non-elderly groups also differ in age category.

The serious adverse-event counts are small in each group, but the registry record does not provide a powered, adjusted comparison of serious events. The selected serious-event terms are not accompanied by a reported denominator or arm-level allocation. They should be read as registry listings, not as evidence that one group had a greater risk of a particular event.

The record does not report the original statistical plan, missing-data handling, or results for the four secondary outcomes. Consequently, the available information cannot settle questions about the planned inferential analysis, long-term outcomes beyond the stated timeframe, or changes in sleep measures. The statistical meaning is limited to the posted adverse-event summaries and the unadjusted recalculated contrasts.

16. Why This Trial Matters Statistically

This trial illustrates why safety results need to be read with both the event definition and the comparison structure in mind. The primary endpoint counts participants with adverse events whether or not investigators considered the events related to study drug. That broad measure is useful for describing what was recorded during treatment and follow-up, but it is not the same as an estimate of treatment-caused harm.

It also shows how similar-looking percentages can coexist with substantial uncertainty in between-group contrasts. The group percentages are all relatively high, while the recalculated confidence intervals remain wide enough to include positive and negative differences. Statistical interpretation therefore depends on the estimates and their precision, not simply on which group has the largest displayed percentage or whether a p-value crosses a conventional threshold.

The age-specific dosing structure is important. Comparing elderly participants receiving 1 mg with elderly participants receiving 2 mg is different from comparing either elderly group with a non-elderly group. In the latter comparisons, age category changes along with dose. Without an analysis that separates those features, a contrast cannot be attributed to dose alone.

Finally, registry results can contain important limits on reproducibility and interpretation. Here, the original sample-size justification, detailed statistical plan, missing-data procedures, and secondary-outcome results are not reported in the registry record. Recognizing what the posted summaries can support—and what they leave unresolved—is central to sound statistical reading.

17. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The posted record leaves several trial-specific questions unresolved. It identifies the study as randomized and double-masked, but the method of sequence generation, concealment of allocation, allocation ratio, and operational details of masking are not reported in the registry record. Eligibility criteria and the age threshold separating elderly from non-elderly participants are also not reported.

The four intervention groups are defined by age category and dose, but the reported pairwise calculations use the elderly 1 mg group as the common reference. The comparisons with non-elderly groups therefore combine differences in age category and dose. The registry record does not provide a separate comparison structure that would isolate dose effects across age groups or an adjusted analysis accounting for participant characteristics.

The primary endpoint is broad and covers up to 25 weeks. Although the adverse-event definition and collection approach are described, the record does not report event severity distributions, participant-level event histories, recurrence, or detailed judgments of causality. The listed serious-event terms lack a reported denominator and group allocation, and death counts are not reported in the registry record.

The record does not report how missing or incomplete safety follow-up was handled, whether all participants contributed complete observation time, or whether sensitivity analyses were performed. The description of the safety population is partly truncated. The original sample-size justification, prespecified statistical plan, multiplicity handling, and interim-analysis plan are not reported in the registry record. Results for the four named secondary outcomes are also not reported.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: A Phase III Study of Eszopiclone in Patients With Insomnia (Study SEP 190-150) — NCT00770692.
• Registry record last updated 2012-11-22; status last verified 2012-10.
• Results first posted 2012-11-22.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

21. Record Summary

The registry record describes a completed, randomized, double-masked, parallel Phase III study of eszopiclone in participants with insomnia. Enrollment was 369. The study began in 2008-10 and had primary completion in 2010-05. Elderly participants received 1 mg or 2 mg, and non-elderly participants received 2 mg or 3 mg.

The primary endpoint was the percentage of participants with adverse events in the safety analysis set, through up to 25 weeks. Posted arm-level percentages ranged from 79.5% to 87.0%. Three unadjusted recalculated comparisons use elderly participants receiving 1 mg as the reference; their confidence intervals all include the no-difference value. Comparisons involving non-elderly groups also differ by age category.

The record lists four secondary sleep outcomes but does not report their results. The original statistical plan, missing-data procedures, and sample-size justification are not reported in the registry record.