Study of the Efficacy of a Fixed-dose Regimen of Cariprazine Compared to Placebo for Treatment of the Depressive Episode in Participants With Bipolar I Disorder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of the Efficacy of a Fixed-dose Regimen of Cariprazine Compared to Placebo for Treatment of the Depressive Episode in Participants With Bipolar I Disorder (NCT02670538), a randomized phase 3 study. The primary outcome is Change From Baseline in Montgomery-Åsberg Depression Rating Scale (MADRS).

2. Clinical Question

3. Trial Design

This was a Phase 3, randomized, double-blind, placebo-controlled, parallel-group, multicenter trial evaluating cariprazine for depressive episodes in participants with bipolar I disorder. The registry lists quadruple masking and random allocation. It reports enrollment of 493 and a completed status. The sponsor was Forest Laboratories. The listed study start was 2016-03-31 and primary completion was 2018-01-18.

There were three assigned groups: placebo, cariprazine 1.5 mg, and cariprazine 3.0 mg. The 1.5 mg regimen continued for 6 weeks. In the 3.0 mg group, participants began at 1.5 mg per day for 2 weeks and then received 3.0 mg per day for 4 weeks. Placebo was given for 6 weeks. The arm descriptions state that capsules were taken orally once per day, following a screening/washout period of 7 to 14 days.

The primary endpoint was change from baseline in MADRS score from Baseline (Week 0) to Week 6. The registry describes an analysis using a mixed-effects model for repeated measures, with treatment group, pooled study center, and visit as fixed factors; baseline as a covariate; and treatment-group-by-visit and baseline-by-visit interactions. The primary analysis population is described as participants in the Safety Population with at least one postbaseline MADRS total-score assessment. The registry also lists change in CGI-S score over the same timeframe as a secondary outcome.

The registry record does not report the original sample-size justification or the assumptions used to plan enrollment. It does not report details of any interim analyses. Those design elements should not be inferred from the observed endpoint counts or from planning illustrations based on the observed effects.

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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT Population consisted of all participants in Safety Population who had at least 1 postbaseline assessment of MADRS total score.

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: Randomized Population consisted of all participants in the Screened Population who were randomized to a treatment group in the study.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Montgomery-Åsberg Depression Rating Scale (MADRS), measured over Baseline (Week 0) to Week 6.

The primary endpoint was change from baseline in total score on the Montgomery-Åsberg Depression Rating Scale (MADRS), assessed from Baseline (Week 0) to Week 6. The registry describes MADRS as a clinician-rated measure of depressive symptoms during the preceding week. It covers feelings of sadness, lassitude, pessimism, inner tension, suicidality, reduced sleep or appetite, difficulty concentrating, and lack of interest.

Each of the scale's 10 items is scored from 0 to 6, with higher item scores indicating greater symptom severity. The total score ranges from 0 (best) to 60 (worst). Because the endpoint is a change score, the direction matters: a negative change from baseline indicates improvement. A comparison of change between a treatment group and placebo estimates how the groups differed over the study period; it is not the same as comparing their final scores alone.

MADRS is used in this trial to quantify depressive symptom severity in a consistent way across participants and visits. A clinician-rated scale can capture several symptom dimensions in a single total score, while repeated assessments allow the analysis to use measurements over time. The registry's model accounts for visit and specified interactions rather than relying only on a single unadjusted endpoint comparison. The endpoint does not by itself describe every aspect of the disorder, functioning, tolerability, or longer-term course.

7. Results Posted in the Registry

• Cariprazine 1.5 mg: mean -14.80 versus -12.40 on control; difference -2.40 (95% CI -4.49 to -0.31), standardized effect d = -0.25; z-test p = 0.025.
• Cariprazine 3.0 mg: mean -14.10 versus -12.40 on control; difference -1.70 (95% CI -3.82 to 0.42), standardized effect d = -0.18; z-test p = 0.116.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Change From Baseline in Clinical Global Impressions-Severity (CGI-S) Score

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Cariprazine 1.5 mg: mean -1.50 versus -1.20 on control; difference -0.30 (95% CI -0.55 to -0.05), standardized effect d = -0.26; z-test p = 0.018.
  • Cariprazine 3.0 mg: mean -1.40 versus -1.20 on control; difference -0.20 (95% CI -0.45 to 0.05), standardized effect d = -0.18; z-test p = 0.116.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Cariprazine 1.5 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Cariprazine 3.0 mg and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists change from baseline in CGI-S score as a secondary outcome over Baseline (Week 0) to Week 6. The posted group estimates follow the same general direction as the primary endpoint: the active-group changes are more negative than the placebo change. The registry-reported contrasts and the simple recalculations likewise show a broadly similar directional pattern. This directional consistency can be considered alongside the primary result, but it does not make the secondary outcome an independent replication of the primary endpoint.

Secondary outcomes help describe additional dimensions of the study's findings. CGI-S addresses clinical global severity, while MADRS is a clinician-rated measure with specified symptom items. Agreement in direction can add context, but the measures are not interchangeable. They differ in what is assessed, and their model-based estimates have their own uncertainty. The registry record lists one posted secondary outcome and reports analyses for the treatment comparisons, but it does not state a multiplicity adjustment for interpreting those results.

When several outcomes or comparisons are tested, the chance of observing at least one apparently notable result can increase if every test is interpreted as though it were the only test. Adjustment methods can address this, but the registry record does not report whether such a procedure was specified or used for the secondary analyses. Therefore, unadjusted secondary p-values carry less confirmatory weight than a result protected by a prespecified multiplicity strategy. They are better interpreted as supportive or exploratory evidence, with their estimates and uncertainty considered rather than p-values alone.

The secondary findings should not be used to resolve the discrepancy between the sponsor's primary analysis and the recalculation, nor to claim that a particular primary result is confirmed. The outcomes contribute descriptive context, while the primary MADRS endpoint remains the central question identified in the registry. Clinical importance thresholds for CGI-S are not reported in the registry record.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Change From Baseline in Clinical Global Impressions-Severity (CGI-S) Score

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous: MADRS is recorded as a numeric score, and the posted arm summaries are least-squares means with standard errors. The treatment contrasts compare estimated changes in each cariprazine group with the placebo estimate. Since lower change values are more negative and the registry states that negative change indicates improvement, the sign of a contrast needs to be interpreted in that direction. A negative treatment-minus-placebo contrast indicates a larger mean reduction in the treatment group according to this coding.

Why use a repeated-measures model?

The registry states that the primary endpoint was analyzed with a mixed-effects model for repeated measures (MMRM). This type of model is designed for outcomes assessed at multiple visits. Rather than treating every visit as unrelated, it models the repeated outcome information while including terms specified in the analysis description. The listed fixed factors were treatment group, pooled study center, and visit; baseline was included as a covariate. The model also included treatment-group-by-visit and baseline-by-visit interactions.

Including visit lets estimated outcomes vary across assessment times. A treatment-group-by-visit interaction allows the difference between groups to vary by visit, rather than requiring a single shared difference across all visits. Including baseline as a covariate accounts for the starting score in the model, while the baseline-by-visit interaction allows its relationship with the outcome to vary over visits. These are general explanations of the model terms; the registry record does not report all implementation details, such as the covariance structure or the specific handling of every incomplete measurement.

What do least-squares means represent?

A least-squares mean is a model-based estimate, not necessarily the raw arithmetic average of observed participant scores. It is estimated within the stated model, which incorporates the included factors and covariates. The arm estimates are reported with standard errors, a measure of the estimated sampling uncertainty of each model-based mean. Standard errors are not the same as standard deviations: the former describe uncertainty in an estimate, while the latter describe variation among individual observations.

The reported analyses use contrast t-tests for the least-squares mean differences. A contrast expresses a model-based comparison between two groups. The registry reports a point estimate, a 95% confidence interval, and a p-value for each primary comparison. The confidence interval gives a range of values compatible with the data and the analysis under the model's assumptions. A p-value describes how unusual the data would be under a specified null hypothesis, assuming the analysis model and its assumptions; it is not the probability that the null hypothesis is true.

How do the recalculated comparisons differ?

The recalculated comparisons use the posted least-squares means and standard errors to make simple, unadjusted comparisons. The pooled standard deviations used in the calculations are approximate, as indicated by the calculation notes. These comparisons are not a reconstruction of the full MMRM or the sponsor's contrast procedure. They do not incorporate the full model structure, including its visit-specific information and covariate handling. Their role is to offer a transparent rough comparison from the posted summaries, not to replace the reported analysis.

In this case the recalculated and registry-reported contrast values and p-values are not identical. That is an important distinction when interpreting the numbers: the sponsor's model-based results are the trial's reported analysis, while the recalculations are secondary checks based on summary statistics. The registry record does not explain the reason for the discrepancy. Accordingly, the sponsor's own analysis takes precedence, and the recalculated results should not be presented as a correction to it.

Why does over-dispersion matter in general?

For a continuous outcome, greater variability among participants generally makes an estimated group difference less precise when other features are held fixed. In count or event outcomes, over-dispersion means that the variation is larger than a simple model expects; failing to account for it can make uncertainty appear too small. The primary endpoint is continuous, and its reported standard errors provide model-based uncertainty for the least-squares means. No separate over-dispersion assessment is reported for this endpoint, so it should not be assumed to have been part of the analysis.

Interpretation also depends on the analysis population. The primary endpoint population includes people in the Safety Population who had at least one postbaseline MADRS assessment. Thus, the analysis population is defined partly by availability of a follow-up score. The registry does not provide a detailed accounting of missing assessments or state how departures from the model's missing-data assumptions were assessed. The MMRM description alone is not enough to establish that missingness could not affect the estimate.

11. Confidence Intervals and Precision

Precision concerns how narrowly the data locate an estimated contrast. The registry-reported 95% confidence interval for the cariprazine 1.5 mg contrast runs from -4.6 to -0.4. The interval is entirely negative, consistent with a larger reduction in MADRS in that group than in placebo under the stated contrast direction. The 95% confidence interval for cariprazine 3.0 mg runs from -3.9 to 0.4. It includes zero, so the posted result is compatible with no difference as well as with negative and positive contrast values within that interval.

A confidence interval is not a guarantee that the true value lies within the particular calculated interval. In general, it is a procedure that, under its assumptions and repeated use, produces intervals that cover the true parameter at the stated rate. It is also not a range of individual participant outcomes. The interval describes uncertainty about a group-level contrast, not the spread of scores among participants.

The recalculated intervals differ from the sponsor's reported intervals: -4.493 to -0.307 for the 1.5 mg contrast and -3.821 to 0.421 for the 3.0 mg contrast. These use an approximate, unadjusted calculation from posted summary statistics. The discrepancy means the recalculated intervals should not displace the registry-reported model-based intervals when describing the trial's own analysis.

Precision should be considered alongside the effect estimate and the measurement scale. The registry record does not provide a threshold for a clinically important MADRS difference. Therefore, the intervals can describe statistical uncertainty but cannot, on their own, establish whether the estimated differences matter to participants. Nor do these intervals characterize effects beyond the Week 6 endpoint.

12. 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 illustrations translate an observed effect into hypothetical planning scenarios. For the continuous MADRS outcome, the entries are participants per arm, not total participants or total events. The calculation 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 as the effect to detect. The registry record does not report whether those assumptions match the original trial design; these are planning illustrations based on observed results.

The 80% power column gives the planned participants per arm for a hypothetical study to have an 80% chance of detecting the specified effect under the stated assumptions, if that effect and the assumptions were correct. The 90% power column gives the corresponding planning count for a 90% chance. Power is a property of a proposed design under assumptions; it is not a measure of whether the observed trial result is true.

For each comparison, planning for 75% of the observed effect requires more participants per arm than planning for the full observed effect, and planning for 50% requires still more. This follows the general statistical principle that smaller effects are harder to distinguish from random variation, so more information is needed. The values should be read as conditional planning scenarios, not as guarantees that a future trial would achieve the stated power.

This is not observed power. Observed results and their uncertainty are described by the estimated contrasts and confidence intervals; a post hoc power calculation based on the observed effect does not add independent evidence about those results. The original design rationale, including its planned target effect and sample-size justification, is not reported in the registry record.

13. 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, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -2.400; pooled standard deviation σ = 9.624.
2. n = 2 × (1.96 + 0.842)² × 9.624² / -2.400² = 252.4, rounded up to 253 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
delta <- -2.4                      # observed difference in means
sigma <- 9.62426277985631                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^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 (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(2670538)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 253 (80% power), 338 (90% power)
 75% of observed effect: n per arm = 449 (80% power), 601 (90% power)
 50% of observed effect: n per arm = 1,010 (80% power), 1,352 (90% power)
power.t.test: 253.4 
Simulated power at n = 253 per arm: 0.800 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-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.

14. Analyses Reported by the Sponsor

The registry reports model-based contrast t-tests for the primary endpoint. For cariprazine 1.5 mg, the reported least-squares mean difference is -2.5, with a 95% confidence interval from -4.6 to -0.4 and p = 0.042. For cariprazine 3.0 mg, the reported difference is -1.8, with a 95% confidence interval from -3.9 to 0.4 and p = 0.105.

The recalculation uses the posted arm-level least-squares means and standard errors to form simple unadjusted comparisons. It produces -2.4 for the 1.5 mg contrast and -1.7 for the 3.0 mg contrast, with corresponding p-values of 0.025 and 0.116. The values are close in direction and broad pattern, but the estimates and inferential results do not match exactly. The recalculation is approximate and does not reproduce the full repeated-measures model.

The registry record does not explain the discrepancy. It could reflect differences between the model-based analysis and the summary-statistic calculation, but the exact explanation is not reported and should not be guessed. The sponsor's own analysis takes precedence as the reported analysis of the trial. The recalculated results are best treated as a transparent, unadjusted check rather than an alternative official result.

15. Multiplicity, Missing Data and Analysis Populations

The registry record gives the primary endpoint and identifies a secondary CGI-S outcome, and it reports treatment comparisons for both. It does not report a multiplicity plan, such as a hierarchy of hypotheses or a method for adjusting inference across outcomes and treatment comparisons. Without that information, the reported p-values should be read as the values attached to the individual analyses, not as evidence that family-wise error across all analyses was controlled.

Multiplicity matters because testing several hypotheses creates more opportunities for a small p-value to occur by chance. The issue is not that any particular result is necessarily false; rather, confirmatory interpretation depends on which outcomes and comparisons were prespecified and how their joint testing was handled. The registry record does not report enough detail to determine that strategy. Secondary findings therefore warrant more caution, especially when interpreted as confirmatory claims.

The primary analysis population is defined as all participants in the Safety Population who had at least one postbaseline MADRS total-score assessment. The registry does not provide further detail about the number or pattern of missing endpoint measurements, reasons for missingness, or sensitivity analyses addressing missing data. It describes an MMRM, but the exact missing-data assumptions and implementation details are not reported in the registry record. A model choice alone does not show that missing-data concerns have been eliminated.

The registry record also does not state whether the primary analysis population differs from a broader randomized population for reasons other than the stated assessment requirement, or provide a participant-level account of exclusions from the endpoint analysis. The arm-level endpoint counts and the enrollment figure describe different aspects of the record and should not be treated as interchangeable. No unreported analysis population or missing-data procedure should be inferred from those counts.

These gaps limit how fully the analysis can be audited from the registry record. They do not negate the posted estimates, but they make the prespecified testing strategy and the robustness of the estimates to incomplete outcome data difficult to assess.

16. 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 serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: abortion (1.0%), non-cardiac chest pain (0.6%), lumbar vertebral fracture (0.6%), bipolar disorder (0.6%), mania (0.6%).

The safety figures report serious adverse events and other adverse events by group, with the number at risk and the number affected. For serious adverse events, the registry lists 5 affected among 165 at risk in placebo, 1 among 167 at risk in the cariprazine 1.5 mg group, and 1 among 158 at risk in the cariprazine 3.0 mg group. The deaths column reports 0 for each group.

For other adverse events, the registry lists 36 affected among 165 at risk in placebo, 45 among 167 at risk in the 1.5 mg group, and 45 among 158 at risk in the 3.0 mg group. These categories should be kept distinct: the serious-event counts are not the same measure as the other adverse-event counts. The registry also lists serious-event terms, including Abortion, Non-cardiac chest pain, Lumbar vertebral fracture, Bipolar disorder, Mania, and Substance abuse.

These counts describe reported events in the registry record; they do not establish that a treatment caused or prevented an event. The figures alone do not show event timing, duration, recurrence, clinical context, or whether the groups differed statistically. The registry record does not report a comparative safety analysis in the information summarized by these counts. Safety interpretation therefore requires more detail than a simple comparison of affected counts.

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

The primary endpoint captures change in clinician-rated depressive symptom severity from baseline to Week 6. The registry-reported estimates show negative changes in all groups, with more negative estimated changes in both cariprazine groups than in placebo. The reported contrast for the 1.5 mg regimen has a confidence interval that is entirely negative, while the interval for the 3.0 mg regimen includes zero. Those statements describe the posted model-based estimates and their uncertainty; they are not a blanket conclusion about treatment effectiveness.

The result does not establish that the observed difference is large enough to matter to participants. The registry record does not report a threshold for clinical importance on MADRS. Statistical compatibility with a difference and clinical importance are separate questions. The estimates also apply to the specific analysis population, endpoint definition, and Week 6 timeframe described in the registry, not automatically to other populations, outcomes, or durations.

The data do not establish what would happen after the endpoint period, whether any change would persist, or how the findings relate to longer-term outcomes. Nor do the available summaries show how individual participant experiences varied around the group means. A mean contrast can conceal variation in response, and the registry summaries do not establish which individuals improved or how consistently they did so.

The primary result should be interpreted alongside its model, confidence intervals, analysis population, and the discrepancy between the registry-reported analysis and the simple recalculation. The sponsor's reported analysis takes precedence, while the recalculation illustrates the general direction using posted summary statistics. The secondary outcome and safety counts add context but do not resolve questions about clinical importance, missing data, or multiplicity.

Nothing in these results alone supports an individual treatment decision. They are trial-level statistical findings and do not replace clinical judgment or evidence from other studies. The appropriate reading is specific: the registry reports estimated group changes and contrasts over the stated trial period, with uncertainty and reporting limitations that should remain visible.

18. Why This Trial Matters Statistically

This trial illustrates several recurring issues in clinical biostatistics. First, the endpoint is a change score whose sign has a defined clinical interpretation: negative values indicate improvement. Stating that direction clearly prevents a common mistake in reading contrasts. A negative treatment-minus-placebo difference indicates a greater reduction in the symptom score, not worsening simply because the number itself is negative.

Second, the registry reports model-based least-squares means from a repeated-measures analysis, while a simple recalculation from arm summaries produces slightly different estimates and p-values. This contrast shows why a summary-statistic check can be useful for orientation but cannot substitute for the trial's specified model. The model uses visit and baseline information as described in the registry, and the recalculation does not reproduce all of that structure. When results disagree, the registry-reported sponsor analysis remains the primary reported result.

Third, precision and interpretation are inseparable. One reported confidence interval excludes zero and the other includes zero, but neither interval establishes clinical importance. The absence of a stated importance threshold means statistical evidence cannot be translated directly into a judgment about the size of benefit that matters.

Finally, registry reporting has limits. The record gives an analysis population definition, but does not fully describe missing-data handling, multiplicity, or the original sample-size rationale. The safety counts are also descriptive rather than a complete comparative analysis. Recognizing what is reported and what is not helps prevent a precise-looking estimate from being treated as a complete account of the evidence.

19. 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.

Several trial-specific limitations affect interpretation of the posted results. The registry does not report the original sample-size justification, the assumptions behind the planned enrollment, or whether interim analyses occurred. The planning illustrations based on observed effects cannot fill that gap because they are not descriptions of the trial's original design.

The primary analysis population requires at least one postbaseline MADRS assessment. The registry does not report a detailed account of missing endpoint data, the reasons for missingness, or sensitivity analyses. Although an MMRM is specified, its covariance structure and detailed missing-data implementation are not reported. This limits assessment of how robust the estimates are to incomplete follow-up.

The record does not report a multiplicity strategy for the primary comparisons and secondary outcome analyses. It also does not give a clinical-importance threshold for MADRS or CGI-S, so the size of the estimated differences cannot be classified as clinically important from these registry entries alone. The discrepancy between the sponsor's model-based primary contrasts and the summary-statistic recalculations is not explained; the sponsor's analysis takes precedence, but the reason for the difference cannot be established from the registry record.

The safety information consists of affected counts and listed serious-event terms. It does not provide enough detail in these summaries to determine event timing, duration, recurrence, attribution, or the statistical uncertainty of between-group differences. The stated treatment period and endpoint also constrain interpretation to the period recorded; longer-term outcomes are not reported in the registry record.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Study of the Efficacy of a Fixed-dose Regimen of Cariprazine Compared to Placebo for Treatment of the Depressive Episode in Participants With Bipolar I Disorder — NCT02670538.
• Registry record last updated 2019-03-12; status last verified 2019-02.
• Results first posted 2019-02-05.
• 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.

23. Record Summary

NCT02670538 is a completed Phase 3 randomized, double-blind, placebo-controlled, parallel-group, multicenter trial sponsored by Forest Laboratories. It enrolled participants with bipolar I disorder and depressive episodes and compared fixed-dose cariprazine regimens with placebo.

The primary endpoint was change in MADRS score from Baseline (Week 0) to Week 6, analyzed with an MMRM in the stated ITT Population. The registry reports model-based contrasts for each active group versus placebo. Simple unadjusted recalculations from the posted summaries differ from those contrasts, so the sponsor's reported analysis takes precedence. The record also lists CGI-S change as a secondary outcome and provides group-level adverse-event counts.

The findings are estimates over the specified endpoint period and should be read with their confidence intervals. The registry does not report the original sample-size rationale, multiplicity strategy, or detailed missing-data assessments, and it does not provide a clinical-importance threshold for the scales.