12- Week Open Label Treatment of Refractory Bipolar Depression: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for 12- Week Open Label Treatment of Refractory Bipolar Depression (NCT00223496), a na phase 4 study. The primary outcome is Primary Measure:Reduction in Depression Symptoms.

2. Clinical Question

The registry also states an objective of assessing changes in symptomatology and function, as well as the tolerability and safety of adding aripiprazole to divalproex. The primary measure is a response definition, not a comparison between two treatment groups. Consequently, even a reported response count would describe outcomes among participants receiving the combination; by itself it would not show how outcomes would have differed under another treatment or without the added drug.

The study is described as open-label, so the registry does not describe masking of participants or study personnel. It is also a single-group study, with allocation listed as not applicable. These design features are central to interpreting the question: the record describes an assessment of outcomes during treatment, not a controlled estimate of the treatment's effect relative to a concurrent control.

3. Trial Design

The registry identifies NCT00223496 as a Phase 4, completed, open-label, single-group study titled “12- Week Open Label Treatment of Refractory Bipolar Depression.” Its official title specifies a combination of Depakote ER and aripiprazole. The listed condition is bipolar disorder, and the reported enrollment is 32.

The intervention entry names aripiprazole. The arm description says it was given open-label as an adjunct to divalproex, with aripiprazole doses ranging from 5-30mg QD and divalproex doses ranging from 500-2500mg QD. The study's stated treatment period was 12 weeks; the primary measure's timeframe is described as up to 12 weeks. The registry lists a start date of 2004-09 and a primary completion date of 2009-09.

There is no separate control arm in the design described by the registry. Allocation is listed as not applicable, masking as none, and the model as single group. The record therefore does not describe random assignment or a blinded comparison. A single-group design can describe observed outcomes over treatment, but it cannot isolate the effect of treatment from changes that might occur for other reasons, including the course of the condition, concurrent care, or expectations. Whether these factors were assessed or addressed is not reported in the registry record.

The registry lists one primary outcome measure and no analysis entries. It describes the primary efficacy criterion as a combined response based on MADRS and CGI-S. Although the record states that symptomatology, function, tolerability, and safety were study objectives, further outcome details and analysis procedures are not reported in the registry record. The completed status indicates the study's recorded status; it does not, by itself, establish how many enrolled participants completed treatment or contributed to each measure.

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: Subjects who entered study in a bipolar depressed state were started on divalproex ER. Those still depressed, according to depression rating scales, after visit 4 were given aripiprazole in addition to divalproex ER.

Overall Study

Reasons for not completing this period:

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: 32 subjects consented, 19 subjects were baselined

6. Primary Endpoint

The registry lists the primary outcome as Primary Measure:Reduction in Depression Symptoms, measured over up to 12 weeks.

The primary measure is the number of participants meeting a combined definition of depression response during up to 12 weeks. The definition requires a 50% reduction on the MADRS together with improvement of 1 or 2 on the CGI-S. The MADRS is described as ranging from 0-60, with higher values indicating more depression. The CGI-S is described as ranging from 0-7, with higher values indicating more severe overall bipolar symptoms.

Using both scales makes the stated response criterion depend on improvement in depression symptoms and on a clinician-rated change in overall bipolar symptom severity. Requiring both conditions is more specific than counting improvement on either scale alone: a participant must satisfy both parts of the registry definition to be classified as a responder. This is a general explanation of how a combined endpoint works; the record does not report whether the two components were assessed by independent raters or how borderline or unavailable assessments were handled.

The endpoint is expressed as a participant count, not as a mean change in scale score. A response count or proportion can be clinically interpretable, but its meaning depends on the denominator, the completeness of assessments, and the exact implementation of the criterion. The registry record does not report a primary response count, proportion, uncertainty interval, or analysis result. It therefore describes what was intended to be measured, but does not provide the posted primary efficacy result needed to assess how often the criterion was met.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The study design determines which statistical conclusions are supported. The registry describes one open-label group and no randomized comparator. General statistical explanation: a single-group analysis can estimate or describe an outcome among participants with recorded data, but it cannot make a randomized between-group estimate. Without a comparison group, observed changes cannot be separated from other changes over time or from features of participant selection and care. The registry record does not report an analysis intended to address those alternative explanations.

What does the primary response definition mean?

The primary measure requires both a 50% reduction on MADRS and improvement of 1 or 2 on CGI-S. This is a binary classification at the participant level: the stated criterion is either met or not met. General explanation: a binary endpoint is often summarized by the number or proportion of participants with the defined event. The interpretation depends on a clear definition of the event and a clear denominator. Here, the definition is stated, but the primary response count and denominator are not reported in the registry record.

The measure combines a change threshold on one scale with an improvement criterion on another. That joint requirement can capture more than a change on either scale alone, but it also means that a participant who meets only one component does not meet the full response definition. The registry does not report component-specific results, the timing of each assessment within the stated timeframe, or how assessments that were missing or not classifiable were handled.

What can uncertainty estimates tell a reader?

General explanation: a confidence interval describes the statistical uncertainty around an estimate under the assumptions of the analysis. For a binary outcome, an interval around a response proportion can indicate how compatible the observed data are with a range of underlying response rates. It does not account automatically for bias from an unblinded, uncontrolled design, nor does it correct for missing data unless the analysis explicitly addresses them. The registry record does not report a response estimate or confidence interval for this endpoint.

A significance test, if one had been reported, would answer a question specified by the statistical model and null hypothesis; it would not by itself establish that a change was caused by treatment or that the result was important. No test, p-value, or hypothesis specification is reported in the registry record. The record also does not state whether results were adjusted for baseline measures or other variables.

How does the single-group design affect comparisons?

General explanation: random allocation to concurrent groups helps balance measured and unmeasured characteristics on average, allowing a treatment contrast to be estimated more credibly. In this study, the registry lists allocation as not applicable and identifies a single-group model. There is no randomized contrast to interpret. Changes from an earlier assessment, if measured, could be described, but the registry does not provide such results or a baseline-to-follow-up analysis plan.

The study is open-label, with masking listed as none. General explanation: when treatment is known, expectations or knowledge of treatment can influence subjective assessments, participant reporting, and study conduct. This is particularly relevant to rating scales that involve symptom judgments. The record does not report steps taken to limit such influences, so no such procedures should be assumed.

What do the planning illustrations represent?

The registry describes planning illustrations as based on an observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It says the effect to detect is a stated fraction of the observed effect. For time-to-event illustrations, the stated quantity is total events; for other illustrations, it is participants per arm. General explanation: these calculations ask what sample size a hypothetical design might require under specified assumptions. They are not observed results, do not show the power actually achieved by this trial, and do not change the single-group design described in the registry. The record does not report the trial's own sample-size justification.

Because the primary endpoint result is absent from the registry record, no effect estimate can be assessed against those planning assumptions. Nor can any planning illustration substitute for the missing outcome data. Sample-size calculations are conditional on assumptions about the effect and statistical model; they do not supply evidence about the treatment outcome.

What can be concluded from the safety counts?

The safety listing distinguishes participants at risk from participants affected, and the serious and other categories have different denominators. General explanation: an event count is not interchangeable with a participant count, because one participant may experience more than one event. The record describes affected participants, not a detailed event total or event-by-event chronology. The relationship between these denominators and the reported enrollment is not stated. No adjustment, comparison, or inference about causality is supported by these counts alone.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Aripiprazole Plus Divalalproex ER")
n      <- c(16)
events <- c(16)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Aripiprazole Plus Divalalproex ER  16  16  100.0  79.4  100.0  80.6  100.0  80.5  99.9

11. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure, with a response definition requiring both a MADRS threshold and a CGI-S improvement criterion. That is one stated combined endpoint, but the record does not explain whether the two components were also analyzed separately or whether any additional efficacy outcomes were tested. It does not report a multiplicity strategy, such as how statistical error across multiple tests would be handled. No adjustment should be assumed.

The registry states that symptomatology and function, tolerability, and safety were objectives, but it does not provide a full outcome inventory or analysis plan. It lists one outcome measure and no analysis entries. The record therefore does not establish which outcomes were analyzed, which analyses were prespecified, or how any selection among analyses was handled. Details of interim analyses are not reported in the registry record.

Missing-data methods and analysis populations are also not reported in the registry record. The enrollment figure does not show how many participants had an evaluable primary endpoint, completed the treatment period, or contributed to each adverse-event denominator. The record does not describe exclusions, withdrawals, imputation, or sensitivity analyses. General explanation: missing outcome data can affect both the denominator and the interpretation of a response estimate, particularly if the availability of data is related to participants' outcomes. Without information on completeness and handling, the direction or size of any such effect cannot be assessed.

The registry record also does not report whether analyses used all enrolled participants or a subset defined by treatment exposure or endpoint availability. The completed status is not a substitute for participant-flow information. These omissions limit assessment of the analysis population and of how robust any reported estimate would be to missing data; in this record, the primary outcome estimate itself is not reported.

12. 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 5% reporting threshold.

Most frequent non-serious adverse event terms

The registry lists 0 participants affected by serious adverse events among 19 at risk. In the other adverse-event category, it lists 6 participants affected among 16 at risk. These are participant counts, not a detailed tally of every event. The record does not report the relationship between either denominator and the enrollment of 32, or whether the categories include overlapping participants.

No serious adverse-event terms are listed. The death count is not reported in the registry record. The absence of listed serious-event terms does not provide a description of event types, timing, or possible causes. The record also does not provide event-level details or an explanation of how participants were counted across categories.

General explanation: safety counts describe what was recorded in the population represented by each denominator. They do not, on their own, establish whether an event was caused by a treatment, how risks compare with another treatment, or whether all relevant events were captured. Because there is no control group described and the denominators differ, these counts should be read descriptively rather than as a comparative estimate.

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

The registry record establishes the study's stated question and design: an open-label, single-group assessment of aripiprazole added to divalproex for refractory bipolar depression, with a combined symptom-response measure intended for assessment during up to 12 weeks. It also reports enrollment of 32 and provides limited adverse-event participant counts.

The primary result needed to answer how many participants met the response definition is not reported in the registry record. There is no reported response count, denominator, estimate of uncertainty, or analysis result. This absence does not mean that no participants responded, and it does not establish that an outcome was not collected. It means that the registry record does not report the result.

Even if the response outcome were reported, this design would describe outcomes in a treated group rather than establish a causal treatment effect. The single-group structure supplies no concurrent control for comparison, and the open-label design does not mask treatment. General explanation: an observed change during treatment may have more than one explanation, so a change over time alone cannot isolate the contribution of treatment.

The safety counts are also limited. They report affected participants within two categories, with different numbers at risk, but the registry does not explain the denominators or list serious-event terms. They cannot establish comparative safety or attribution, and they do not give a complete account of event timing or severity.

Accordingly, the record supports a description of the study's planned endpoint, treatment approach, and reported safety counts, but not an estimate of primary efficacy or a controlled treatment comparison. Important details about analysis populations, missing data, and statistical methods are not reported in the registry record. Those gaps constrain what can be concluded from the posted information.

14. Why This Trial Matters Statistically

This study illustrates why design and reporting details are essential to statistical interpretation. The endpoint is clearly framed as a participant-level response requiring both a specified reduction on MADRS and improvement on CGI-S. Yet without a reported response count and denominator, the endpoint definition alone cannot establish how frequently that criterion was met.

The single-group, open-label design also draws a clear boundary around possible conclusions. It can describe outcomes recorded among participants receiving the combination, but it does not provide a randomized comparator for estimating a treatment contrast. General statistical explanation: a treatment effect is a comparison between outcomes under different treatment conditions; a treated-group outcome alone does not supply that comparison.

The record further highlights the importance of participant flow. Enrollment is reported as 32, while the serious and other adverse-event categories have different numbers at risk. Without an explanation of how those denominators were formed, readers cannot determine their relationship to all enrolled participants or to one another. Similarly, without missing-data and analysis-population details, the scope of any endpoint result would be difficult to assess.

Finally, planning calculations and observed outcomes answer different questions. The registry describes hypothetical two-arm sample-size assumptions, while the study itself is described as a single group. Such illustrations can be useful for understanding planning under assumptions, but they are not evidence of an observed treatment effect, achieved power, or a valid randomized comparison in this study. The absent primary result makes this distinction especially important.

15. 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 most direct limitation of the registry record is that it does not report the primary outcome result. Although the endpoint is defined, the response count, denominator, and uncertainty are not reported. This prevents an assessment of the observed efficacy outcome described in the study objective.

The design is a single-group, open-label study without a separate comparator. As a result, the record does not support a randomized estimate of treatment effect. The absence of masking may also matter for symptom-rating assessments, but procedures intended to address this are not reported in the registry record.

The relationship between the reported enrollment of 32 and the adverse-event denominators of 19 and 16 is not reported. Participant completion, withdrawals, the endpoint analysis population, and missing-data handling are also not reported. These omissions prevent a clear account of who contributed to each measure and how incomplete records may have affected interpretation.

The registry describes tolerability and safety as objectives, but the safety information presented is limited to participant counts in serious and other categories. Serious-event terms are not listed, and the death count is not reported in the registry record. The record does not provide event timing, severity details beyond the category, or attribution information.

The registry does not report a statistical analysis plan, multiplicity handling, an interim-analysis plan, or a study-specific sample-size justification. It also does not report outcome-specific analysis methods or whether any results were adjusted. These are limits on what can be assessed from the record, not evidence that a particular method was or was not used outside what the registry reports.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: 12- Week Open Label Treatment of Refractory Bipolar Depression — NCT00223496.
• Registry record last updated 2017-05-16; status last verified 2017-04.
• Results first posted 2013-05-21.
• 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.

19. Record Summary

NCT00223496 is recorded as a completed Phase 4, open-label, single-group study of aripiprazole added to divalproex for refractory bipolar depression. The registry reports enrollment of 32 and describes treatment for 12 weeks, with aripiprazole doses of 5-30mg QD and divalproex doses of 500-2500mg QD.

The intended primary measure is the number of participants meeting a combined response definition based on a 50% reduction on MADRS and improvement of 1 or 2 on CGI-S, assessed up to 12 weeks. The primary result and analysis are not reported in the registry record. Safety listings report 0 participants affected by serious adverse events among 19 at risk and 6 affected in the other category among 16 at risk; serious-event terms are not listed, and the death count is not reported.

The record does not describe a concurrent comparator, detailed analysis methods, missing-data handling, or participant flow. Its information supports understanding the intended endpoint and design, but does not provide a primary efficacy estimate or a controlled comparison.