A Study to Evaluate SAGE-217 in Participants With Bipolar I/II Disorder With a Current Major Depressive Episode: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Evaluate SAGE-217 in Participants With Bipolar I/II Disorder With a Current Major Depressive Episode (NCT03692910), a non-randomized phase 2 study. The primary outcome is Part B: Change From Baseline in the 17-Item Hamilton Depression Rating Scale (HAM-D) Total Score.

2. Clinical Question

The clinical question implicit in the conducted portion was whether depressive symptoms and other listed measures changed during an open-label course of SAGE-217. That is a within-group question: without a concurrent control group, observed change cannot separate a treatment-associated change from changes that might also occur over time, effects of study participation, or other influences. The registry record does not report results for the listed Part A outcomes, so the size or direction of any such changes cannot be described.

The more direct comparative question suggested by the planned Part B was whether HAM-D change differed between SAGE-217 and placebo. That comparison cannot be answered from this trial record because Part B was not conducted. The record's official title describes a two-part study, but the actual completed study represented in the record consists of the open-label portion. The distinction between a planned comparison and a conducted comparison is central to interpreting the clinical question.

3. Trial Design

This was a Phase 2 study listed as completed, with enrollment of 35 participants. Its official title describes an open-label part followed by a double-blind, randomized, placebo-controlled, parallel-group part. The registry record, however, classifies allocation as non-randomized, masking as none, and the model as a single group. The planned second part was not conducted, as stated in the arm descriptions.

In Part A, participants self-administered SAGE-217, 30 mg, orally once daily in the evening from Day 1 to Day 14. Part B was intended to include SAGE-217 and matching placebo arms, also with once-daily evening dosing from Day 1 to Day 14. The sponsor's decision not to conduct Part B means those descriptions represent planned treatment groups, not groups that were enrolled and compared.

The record lists a start date of 2018-08-23 and a primary completion date of 2019-05-21. It identifies bipolar I disorder, bipolar II disorder, and major depressive episode as the conditions. The study summary describes its aims as evaluating safety, tolerability, pharmacokinetics, and efficacy. The primary outcome was designated as Part B change from baseline in the 17-item HAM-D total score, with a timeframe from baseline up to Day 42, but the record says no data were collected or analyzed for that outcome.

These features distinguish the protocol's planned structure from the study that actually generated data. The conducted portion was open-label and single-group, not a randomized comparison. Details about the reasons for stopping before Part B beyond the sponsor's decision, the planned analysis methods, and the sample-size justification are not reported in the registry record.

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: The study was to be conducted in 2 parts: Part A (Open-label) and Part B (Double-blind). Participants received SAGE-217 during the treatment period in Part A. Part B was to be initiated post review of Part A data. However, as per the Sponsor's decision, Part B was not conducted and hence no participants were enrolled in it and no data collection and analysis was done.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Safety Set included all participants who received at least one dose of study drug.

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.

Population: Safety Set included all participants who received at least one dose of study drug.

6. Primary Endpoint

The registry lists the primary outcome as Part B: Change From Baseline in the 17-Item Hamilton Depression Rating Scale (HAM-D) Total Score, measured over Baseline up to Day 42.

The specified primary endpoint was change from baseline in the 17-item Hamilton Depression Rating Scale (HAM-D) total score. The HAM-D was used to rate depression severity in participants identified as experiencing a major depressive episode. Its 17 items cover symptoms including depressed mood, insomnia, guilt, suicidal thoughts, work and activities, anxiety, and somatic symptoms. Some items are scored from 0 to 2 and others from 0 to 4; the total ranges from 0 to 52, with higher scores indicating more depression.

A change score expresses how a participant's endpoint value differs from their baseline value. For this measure, a negative change indicates improvement according to the scale description. In general, a continuous symptom score can convey the magnitude and direction of change, rather than only classifying participants into response categories. That is a general explanation of the measure, not a result from this study.

The endpoint was specified for Part B over baseline up to Day 42. The registry record states that Part B was not conducted and that data for this outcome were not collected or analyzed. Consequently, there is no primary endpoint estimate in the record to interpret, and no comparison of HAM-D change between SAGE-217 and placebo can be made from this outcome.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a change score represent?

In general, a change score is the endpoint measurement minus the baseline measurement for the same participant. It describes movement on a scale over a specified period. For the HAM-D in this study, the scale description says that negative change indicates improvement and that higher total scores indicate more depression. A group summary of individual changes can describe the average or typical movement, depending on the statistic chosen. The registry record does not report the Part B values or the statistic used to summarize them, because that outcome was not collected or analyzed.

Change from baseline is not automatically a treatment effect. In a single-group study, it describes change among participants who received the study treatment, but it does not show what would have happened to comparable participants without that treatment. In general, a randomized concurrent comparison helps estimate the difference between treatment groups by balancing measured and unmeasured influences on average. That comparison was planned for Part B but was not conducted in this trial.

How should uncertainty be read?

In general, a confidence interval gives a range of estimates compatible with the observed data under the method's assumptions. A narrower interval indicates greater precision, while a wider interval indicates greater uncertainty. A confidence interval is not a guarantee that the true value lies within its limits. The registry record does not report a confidence interval for the primary endpoint; no Part B data were collected or analyzed.

A p-value, in general, measures how surprising data at least as inconsistent with a specified null hypothesis would be if that hypothesis and the analysis assumptions were true. It does not measure the probability that the null hypothesis is true, nor does it describe the size or clinical importance of an effect. No primary-outcome p-value is reported in the registry record. With no conducted Part B comparison, there is no reported test of the planned treatment contrast to interpret.

Why does the design determine the estimand?

An estimand is the specific treatment effect a study seeks to quantify, defined by the population, treatments, outcome, and handling of events such as discontinuation or missing measurements. In general, the estimand should align with the question and the analysis. The planned primary outcome identified a population in Part B, a HAM-D change measure, and a timeframe from baseline up to Day 42. However, the record does not provide a completed Part B population or measurements, and details about handling intercurrent events are not reported in the registry record. A precise estimand cannot be reconstructed from the available information.

The conducted Part A offers a different kind of evidence from a controlled comparison. General within-group summaries can describe participants' scores over time, while a randomized contrast is designed to estimate a difference between assigned groups. Part A had no concurrent comparator, and the registry classifies the study as non-randomized, unmasked, and single-group. Those design characteristics limit causal interpretation even if descriptive Part A outcome values were available.

What do the planning assumptions mean?

In general, a sample-size calculation estimates how many participants or events a future study might need under stated assumptions about the target effect, variability, significance level, allocation, and statistical method. The assumptions described for the planning illustration are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the target is expressed as a stated fraction of an observed effect. This is a planning exercise, not an observed result or an account of the original design justification. The registry does not report the study's sample-size justification.

Power is a design property calculated under assumptions about a true effect and a planned analysis. It is not measured after the fact by comparing the observed result with a significance threshold. Accordingly, a planning illustration should not be described as observed power. It also cannot remedy the absence of Part B data or turn the 35-person open-label group into a two-arm trial.

Why can over-dispersion matter?

In general, over-dispersion occurs when outcome variability is greater than a statistical model assumes. If a model understates variability, standard errors may be too small and confidence intervals may be too narrow, making results appear more precise than warranted. Suitable methods can account for extra variability when the outcome and design call for them. This is a general statistical point only: the registry record does not report an analysis model or indicate that over-dispersion was assessed in this study.

Likewise, different outcome types call for different analytic approaches in general. Continuous scale scores, binary response indicators, event counts, and time-to-event outcomes do not all have the same data structure. No detailed statistical analysis plan is reported in the registry record, so the methods that would have been used for the listed outcomes cannot be inferred.

What can be concluded from the available reporting?

The registry names outcomes and describes the HAM-D scale, but naming an endpoint is not the same as reporting an analysis. It lists 23 outcome measures and zero analyses; the primary Part B measure is explicitly stated to have no collected or analyzed data. The registry does not report effect estimates, uncertainty intervals, test results, analysis populations, or missing-data procedures in the information described. Statistical interpretation must therefore focus on the design and reporting limits, not on numerical efficacy conclusions.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 23 outcome measures, including several Part A depression and sleep measures and a Part B HAM-D outcome. Looking at multiple outcomes raises a general multiplicity issue: the more questions tested, the more opportunities there may be for a nominally unusual result to arise by chance. In general, studies may prespecify a primary outcome, organize outcomes hierarchically, or adjust inference for multiple testing. The registry record does not report a multiplicity strategy or a detailed analysis plan, so it is not possible to determine whether or how multiplicity would have been addressed.

The primary outcome was designated for Part B, which was not conducted. The record states that no data were collected or analyzed for it. Although several Part A secondary outcomes are named, their result values are not reported in the registry record. Their presence in the outcome list is not evidence that analyses were performed.

Enrollment was 35, but the number who completed treatment, had a measurement at each time point, or entered any analysis is not reported in the registry record. The analysis populations and participant flow are also not reported. Details of missing-data frequency, reasons for missingness, and any imputation or sensitivity analyses are not reported in the registry record.

In general, missing data can affect interpretation when the chance of a missing measurement relates to participants' health or observed outcomes. Statistical approaches depend on the amount and pattern of missingness and the assumptions considered plausible. Without reporting on the missing-data pattern or analysis methods, those assumptions cannot be assessed. For the primary endpoint, the central issue is more direct: Part B was not conducted, so the planned outcome data do not exist in the registry record.

11. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent non-serious adverse event terms

For Part A, the registry reports 35 participants at risk for serious adverse events and 0 with a serious adverse event. It reports 35 participants at risk for other adverse events and 10 with an event in that category. The death count is 0. These are counts for the open-label SAGE-217 group; they are not a comparison with placebo, because the placebo part was not conducted.

The figures should be read as reported counts, not as proof that an event cannot occur or as a comparative estimate of risk. The registry record does not provide serious adverse-event terms, and no top serious-event terms are listed. It also does not give event-level descriptions or the detail needed to assess severity patterns, timing, duration, or relationship to treatment. The category called other adverse events should not be treated as a specific diagnosis or as a measure of treatment causality.

With no concurrent control group, the reported safety counts cannot distinguish events associated with study participation or treatment from events that might have occurred for other reasons. The record's counts describe what was reported in the study group; they do not by themselves establish a causal safety conclusion.

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

The record establishes that 35 participants were enrolled in an open-label Part A in which SAGE-217 was administered once daily in the evening from Day 1 to Day 14. It also establishes that the planned double-blind, placebo-controlled Part B was not conducted. The primary outcome was assigned to Part B, and the registry explicitly says that data for it were not collected or analyzed.

As a result, the record does not provide a treatment-versus-placebo estimate for change in HAM-D, nor does it provide a primary-outcome estimate for the open-label group. Part A outcome measures are listed, but their results are not reported in the registry record. No conclusion about the amount or direction of symptom change can be drawn from outcome titles alone.

The record also does not establish why Part B was not conducted beyond noting the sponsor's decision, what analysis plan would have governed it, or how the original sample size was justified. Those details are not reported in the registry record. The listed safety counts describe the Part A group and cannot be used as a controlled comparison.

This does not mean that no participants received SAGE-217 or that no safety information was collected: the record describes Part A dosing and reports adverse-event counts. It means that the comparative question implied by the planned placebo-controlled part remains unanswered in this record. The distinction between a study's planned design and its conducted design is essential when interpreting its evidence.

Any planning calculations are not observed power and do not stand in for missing outcome data. Statistical precision for a treatment effect cannot be assessed without a reported estimate and its uncertainty. Overall, the record is informative about the study structure and the absence of the planned primary analysis, but it does not provide a quantitative efficacy comparison.

13. Why This Trial Matters Statistically

This trial is a useful example of why registry design fields and completed study conduct need to be read together. The official title describes a two-part study, including a double-blind randomized placebo-controlled part, while the record classifies the study as non-randomized, unmasked, and single-group. The arm descriptions resolve the apparent mismatch: Part B was planned but not conducted. The completed study therefore should not be interpreted as if a blinded randomized comparison had taken place.

The distinction matters because design determines what an estimate can mean. A single-group study can describe outcomes among participants receiving an intervention, but it cannot provide the same treatment contrast as a concurrent randomized control. Here the primary outcome was attached to the unperformed comparative part, leaving no reported primary result to interpret.

The record also illustrates that outcome listings, enrollment counts, and safety counts are not substitutes for statistical results. It lists 23 outcome measures and 35 enrolled participants, but the primary outcome was not collected or analyzed, and outcome estimates are not reported in the record. The safety figures describe the Part A group alone. Clear reporting about which planned components occurred is therefore necessary before judging what question the data can answer.

Finally, the study shows why planning illustrations must remain distinct from evidence. Sample-size calculations depend on assumptions and can inform future designs, but they do not recover an unobserved comparison or establish the precision of a missing estimate. The statistical lesson is to align every interpretation with the design actually carried out and the results actually reported.

14. 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.

The principal limitation is that the primary outcome was assigned to Part B, which the sponsor did not conduct. The record states that no data were collected or analyzed for that outcome. Therefore the planned placebo-controlled comparison and its primary HAM-D estimate are unavailable.

The conducted portion was open-label and single-group, with no concurrent comparator. This limits what within-group change could establish about treatment effects, even if Part A outcome values were reported. The record classifies allocation as non-randomized and masking as none. It does not describe any method for controlling expectations or other time-related influences.

Although 35 participants were enrolled, the number completing the study or contributing data to particular outcomes is not reported in the registry record. Participant flow, analysis populations, missing-data patterns, and missing-data handling are also not reported. The registry lists secondary outcomes but does not report their numerical results in the information described.

The record does not report a detailed statistical analysis plan, a multiplicity approach, or a sample-size justification. It also does not explain the sponsor's decision not to conduct Part B beyond identifying that decision. These limitations are specific to what is reported for this study and restrict assessment of the planned analysis and the completeness of outcome reporting.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: A Study to Evaluate SAGE-217 in Participants With Bipolar I/II Disorder With a Current Major Depressive Episode — NCT03692910.
• Registry record last updated 2023-12-13; status last verified 2023-11.
• Results first posted 2022-04-20.
• 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.

18. Record Summary

NCT03692910 was a completed Phase 2 study of SAGE-217 in participants with bipolar I or II disorder and a current major depressive episode. The registry reports enrollment of 35 participants and dates of 2018-08-23 for study start and 2019-05-21 for primary completion.

Part A was open-label; participants took SAGE-217, 30 mg, once daily in the evening from Day 1 to Day 14. Part B was planned as double-blind and placebo-controlled but was not conducted. The primary endpoint was Part B change from baseline in 17-item HAM-D total score from baseline up to Day 42; the registry states that it was not collected or analyzed.

The record lists 23 outcome measures and zero analyses. It reports 0 serious adverse events among 35 participants at risk, 10 participants with other adverse events among 35 at risk, and 0 deaths. The reported design and outcome information do not support a randomized treatment comparison.