Clinical Trial of a Nutritional Supplement in Adults With Bipolar Disorder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Clinical Trial of a Nutritional Supplement in Adults With Bipolar Disorder (NCT00109577), a randomized phase 2 study. The primary outcome is Mood as Evaluated by the Overall Bipolarity Index (Composite of the Hamilton Depression Scale and the Young Mania Rating Scale).

2. Clinical Question

The trial was intended to study whether MCN36 was beneficial for bipolar disorder under randomized and fully blinded conditions. The registry summary describes a hypothesis of improved mood stability after 8 weeks in the supplement group relative to placebo. That hypothesis is not itself a result. The posted outcome reports mean changes for both groups and a calculated comparison between them. Those statistics describe the recorded outcome; they do not, by themselves, establish how clinically important the difference may be or whether the finding would apply beyond the enrolled participants.

3. Trial Design

This was a randomized, placebo-controlled, parallel-group phase II trial sponsored by the University of Calgary. The registry lists an enrollment of 40 and a terminated status. The recorded start was 2005-04 and the primary completion was 2009-03. The reason for termination is not reported in the registry record, so the status alone does not explain why the trial ended or whether its planned conduct changed.

Participants were assigned to MCN36 or placebo. Each arm has a reported sample size of 20. The intervention was described as a 36-ingredient micronutrient formulation, primarily vitamins and minerals; the placebo arm received placebo capsules. Both regimens called for 6 capsules three times a day. The allocation ratio and the method used to generate or conceal the random assignment are not reported in the registry record.

Masking is recorded as quadruple, while the identities of the masked parties are not specified. The registry summary describes the comparison as fully blinded, but the operational details of masking and any assessment of whether masking was maintained are not reported in the registry record. The eligible population was limited by residence to the vicinity of Calgary, Alberta, Canada, or the area of San Diego, California. Other recruitment and eligibility details are not reported.

The primary outcome was assessed from baseline to 8 weeks. The record also lists Global Clinical Impressions, an Outcome Questionnaire described as a self-report questionnaire, and the Medical Outcomes Study 36-Item Short Form Health Survey (SF-36) as secondary outcomes. Details of the timing and results for those measures are not reported in the posted results. The trial's original statistical plan, including its intended sample-size justification and any interim analyses, is 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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All participants were included, since the analysis was intent-to-treat with last observation carried forward

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 Mood as Evaluated by the Overall Bipolarity Index (Composite of the Hamilton Depression Scale and the Young Mania Rating Scale), measured over Baseline to 8 weeks.

The primary endpoint is change in mood from baseline to the final visit, with the timeframe identified as baseline to 8 weeks. Mood was assessed using the Overall Bipolarity Index, a composite of the Hamilton Depression Scale and the Young Mania Rating Scale. The score can range from 0 to 103, and higher scores indicate worse symptomatology. The registry describes the analysis population as all participants, using an intent-to-treat approach with last observation carried forward.

A composite measure that brings together depression and mania ratings is relevant to the stated question about mood stability because the trial concerns bipolar disorder and reports an overall mood outcome. The record does not explain the scoring algorithm, how the component scales were combined, or how a change score was calculated. In particular, the direction used to subtract baseline and final scores is not reported. That matters when interpreting the signs of the reported changes: the score direction is known, but the change-score convention is not.

The posted group means are -21.1 for placebo and -18.9 for MCN36. These are mean changes in the score's units, not final scores. The endpoint therefore describes average change within each arm, while the arm comparison describes the difference between those average changes. It does not show how individual participants changed or how the composite's components behaved separately.

7. Results Posted in the Registry

• Micronutrient Formula: mean -18.90 versus -21.10 on control; difference 2.20 (95% CI 0.24 to 4.16), standardized effect d = 0.70; z-test p = 0.028.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: it is a score measured in units, and the registry reports a mean and standard deviation for each arm. A mean represents the arithmetic average of participant-level values. A standard deviation describes how dispersed those values are around the mean. The mean is useful for comparing average changes, while the standard deviation gives information about how much participant outcomes varied within an arm. Neither summary reveals the full distribution or whether a small number of participants had unusually large changes.

What does the arm difference represent?

The posted comparison is a mean difference. It is obtained by subtracting the placebo mean change from the MCN36 mean change: -18.9 minus -21.1 gives the reported estimate of 2.2 units. The estimate therefore expresses the difference between the average changes, not the change for a typical individual and not a difference between final scores. A positive value means the MCN36 mean change is higher than the placebo mean change under that subtraction order. Since the change-score direction is not specified in the registry record, the sign alone does not identify which arm had the more favorable change.

The comparison also reports an unadjusted recalculation from the posted summary statistics. Such a recalculation uses the group summaries rather than participant-level data and does not incorporate adjustments for baseline values, site, or other factors. Whether any such adjustments were part of the trial's own planned analysis is not reported in the registry record. The registry lists no analyses, and no sponsor analysis for this endpoint is described in the posted results. Accordingly, the recalculated comparison should not be presented as a documented analysis specified in advance.

What do the p-value and interval communicate?

The unadjusted recalculated p-value is 0.028. In general, a p-value measures how unusual a result at least as different as the observed result would be under a specified no-difference model and its assumptions. It is not the chance that the result is due to chance, the chance that a treatment is effective, or a measure of clinical importance. The registry record does not report the analysis plan needed to determine whether this was the intended primary test or whether it was selected among multiple analyses.

An interval estimate gives a range of values compatible with the observed data under the method used. Here, both reported limits are positive, but the confidence level is not reported in the registry record. The interval concerns the mean difference, not the range of individual participant changes. Its width reflects uncertainty in estimating the group contrast. A narrower interval would generally indicate more precision under the same analysis, while a wider interval would leave more uncertainty about the effect's size.

Why does variability matter?

General statistical explanation: when participant outcomes vary widely within arms, the mean difference is harder to estimate precisely for a given amount of information. The reported standard deviations are 3.4 in the placebo group and 2.9 in the MCN36 group. These are measures of outcome spread. They should not be confused with standard errors, which describe uncertainty in estimated means. The record contains standard-error values with more than three decimal places, so they are not repeated.

The comparison also includes a standardized effect estimate, but the registry record does not explain how it was intended to be interpreted. Standardization generally expresses a mean difference relative to outcome variability, which can aid comparison across measures with different units. It does not supply a clinical threshold, and it does not remove uncertainty or bias. The unrounded standardized estimate is not repeated because it exceeds the permitted precision for reporting.

What can be inferred from the analysis population?

The registry describes the primary analysis as intent-to-treat with last observation carried forward (LOCF), including all participants. In general, intent-to-treat aims to analyze participants in their randomized groups, preserving the randomized comparison. LOCF is a missing-data procedure that substitutes a participant's last observed outcome for a later missing outcome. This approach treats the last measured value as if it represented the later assessment; that assumption may not hold if outcomes would have changed after the last observation.

The record does not report how many outcomes were missing, when they were missing, why they were missing, or whether sensitivity analyses were conducted. It also does not state how LOCF was applied to participants with no post-baseline measurement. Therefore, the impact of missing data on the posted means and comparison cannot be assessed from the registry record.

General statistical explanation: over-dispersion refers to variability greater than a model expects. It is especially relevant for count outcomes or models with a specified variance structure. The primary endpoint is a continuous score summarized by means and standard deviations, and the registry record does not report an over-dispersion model. No specific concern about over-dispersion can be inferred from these summaries.

10. Confidence Intervals and Precision

The mean-difference estimate is 2.2 units, with the MCN36 mean change minus the placebo mean change as the comparison direction. The reported interval has a positive lower limit and a positive upper limit. This means the interval as posted does not include zero, but the registry record does not state its confidence level. The exact lower limit has more than three decimal places and is not repeated. Without the confidence level and full analysis description, the interval should be read as the posted uncertainty range rather than assigned a conventional confidence interpretation that the record does not document.

An interval is useful because the point estimate is only one estimate of the underlying difference. Values across the interval represent a range of differences compatible with the data under the calculation's assumptions. The positive interval limits support the same direction as the point estimate, but they do not show that the true difference is known precisely. Nor do they establish whether any value in the interval would matter clinically. A threshold for clinical importance is not reported in the registry record.

The reported standard deviations, 3.4 for placebo and 2.9 for MCN36, describe the spread of participant changes. They are not confidence limits and do not show the precision of the group contrast on their own. Precision also depends on the amount of information and the analysis method. There were 20 participants in each arm, but the registry record does not report details such as outcome completeness or an analysis adjusted for baseline.

The interval does not account for uncertainties that are absent from the posted summaries, including the effect of missing observations, any unreported multiplicity, or deviations from the analysis the trial may have intended. The registry record does not give the original statistical plan or identify whether the interval was prespecified. Precision should therefore be judged from the posted estimate and interval alongside these reporting gaps, rather than from the p-value alone.

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 table and curve are planning illustrations based on the observed effect, not a record of the trial's original design calculations. The figures are participants per arm, not total participants or total events. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set at the stated fraction of the observed effect. The original trial's sample-size assumptions are not reported in the registry record. These planning figures are not observed power.

For the row planning for 100% of the observed effect, the illustration gives 33 participants per arm for 80% power and 44 participants per arm for 90% power. For planning for 75% of the observed effect, the corresponding figures are 58 and 78 participants per arm. For planning for 50% of the observed effect, they are 130 and 174 participants per arm.

In general, power is the chance that a study will detect an effect of a specified size under the planning assumptions if that effect is present. The 80% power column means the illustrated sample size is intended to give an 80% chance of detecting the specified effect under those assumptions; the 90% column uses a higher target chance and consequently requires more participants per arm in these illustrations. Power is a design property under assumptions, not a measure of whether the observed result is true.

The curve visualizes the same planning relationship across effect sizes. Smaller effects are harder to distinguish from random variation, so planning for a smaller fraction of the observed effect requires more participants. Because the calculation uses the observed effect as its reference, these numbers should not be read as evidence that the original trial was adequately or inadequately sized. The registry record does not report the trial's own design assumptions or justification.

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

Worked calculation for this trial

1. Observed difference Δ = 2.200; pooled standard deviation σ = 3.160.
2. n = 2 × (1.96 + 0.842)² × 3.160² / 2.200² = 32.4, rounded up to 33 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 <- 2.2                      # observed difference in means
sigma <- 3.15990506186499                     # 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(109577)
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 = 33 (80% power), 44 (90% power)
 75% of observed effect: n per arm = 58 (80% power), 78 (90% power)
 50% of observed effect: n per arm = 130 (80% power), 174 (90% power)
power.t.test: 33.4 
Simulated power at n = 33 per arm: 0.794 (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.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists three secondary outcomes: Global Clinical Impressions, an Outcome Questionnaire described as a self-report questionnaire, and the Medical Outcomes Study 36-Item Short Form Health Survey (SF-36). There are four outcome measures in total, while the record lists no analyses. The posted results concern the primary outcome. Results for the secondary outcomes and the way they were analyzed are not reported in the registry record.

Multiplicity arises when several outcomes, time points, or comparisons are tested. In general, conducting many tests can increase the chance of at least one apparently unusual result, even when no effect exists. Researchers may prespecify one primary test, use an adjustment, or otherwise account for multiple testing. Whether any such approach was planned or used in this trial is not reported in the registry record. The unadjusted recalculated p-value should therefore be read as an unadjusted comparison, not as evidence that multiplicity was addressed.

The primary outcome population is described as all participants under an intent-to-treat approach with LOCF. The number of participants contributing observed outcome data at each assessment, the amount and pattern of missingness, and reasons for missing data are not reported in the registry record. The record also does not describe sensitivity analyses or alternative missing-data methods. This leaves uncertainty about how dependent the reported summary may be on the LOCF assumption.

The registry does not report the trial's statistical analysis plan, the intended primary model, adjustments for baseline or other variables, or interim analyses. It also does not state whether the reported comparison and interval were prespecified. Without those details, it is not possible to evaluate adherence to a planned analysis or to distinguish a planned analysis from a post hoc one. The available summaries support a description of the reported arm means and their unadjusted recalculated contrast, but they do not answer these design and analysis questions.

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 0% reporting threshold.

The posted serious adverse event figures are 0 affected participants among 20 at risk in the placebo arm and 0 among 20 at risk in the MCN36 arm. The registry also reports 0 participants with other adverse events among 20 at risk in each arm. These are counts of affected participants, not counts of all event episodes.

The registry record does not list serious adverse event terms. Death counts are not reported in the registry record. The absence of recorded affected participants in these fields should be described as the posted safety data, not as proof that no adverse events occurred or that the treatments have no safety concerns. The record does not give details about how events were collected, the observation period for safety, severity definitions, or completeness of adverse-event reporting.

These figures are limited to the enrolled trial groups and the registry's recorded categories. They do not establish the frequency of uncommon events or provide a comparison precise enough to rule out differences in risk. They also do not replace a full review of safety information. No medical advice or treatment recommendation follows from these counts.

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

The posted primary outcome results show a mean change of -21.1 units in the placebo group and -18.9 units in the MCN36 group. The mean difference, in the order MCN36 minus placebo, is 2.2 units. The unadjusted recalculation has p = 0.028, and the reported interval limits are positive. These statistics describe an observed difference in average change and the uncertainty around that comparison under its calculation assumptions.

The result does not establish the clinical importance of the difference. The registry record does not report a threshold for a meaningful change, and it does not specify the direction used to calculate change from baseline to final visit. Although higher Overall Bipolarity Index scores indicate worse symptomatology, that score direction alone does not determine how to interpret negative change values when the subtraction convention is unspecified.

The result also does not establish that MCN36 caused the difference. Randomization is a design feature intended to support a fair comparison, but assessment of the actual analysis still depends on trial conduct, missing data, outcome measurement, and adherence to the planned analysis. Relevant details, including the original statistical plan and missing-data sensitivity analyses, are not reported in the registry record. The listed terminated status does not explain why the trial ended.

Nor does the result establish effects in people outside the enrolled population or with different circumstances. The trial was limited to adults with bipolar disorder living in the vicinity of Calgary, Alberta, Canada, or the area of San Diego, California. The registry record does not provide enough detail to assess how broadly the participants represent other populations. A statistically unusual result is not equivalent to a treatment recommendation, and the posted summary does not provide a basis for medical advice.

16. Why This Trial Matters Statistically

This trial illustrates why a group mean difference needs to be interpreted together with its direction, uncertainty, and analysis context. The reported estimate of 2.2 units is a contrast between mean changes, not a direct statement about final symptom levels. Since the record identifies the score direction but not the change-score subtraction convention, even the verbal interpretation of positive and negative changes needs care. That distinction is central to communicating a continuous endpoint accurately.

It also illustrates the difference between a posted group summary and a documented trial analysis. The unadjusted comparison is recalculated from summary statistics, while the registry record does not describe the trial's analysis plan or a sponsor analysis for this outcome. A p-value may appear definitive when read alone, but the missing details about prespecification, multiplicity, and missing observations limit what can be inferred from it.

The planning illustration provides another statistical lesson: the sample size needed depends on the effect chosen for planning. The rows for smaller fractions of the observed effect require more participants per arm. This is a design illustration, not a measure of the trial's observed power and not evidence about the original enrollment rationale, which is not reported in the registry record.

Finally, the record distinguishes outcome results from safety counts and from design details. A randomized, quadruple-masked, parallel-group label describes the stated design, while the posted numerical summaries have their own limits. Clear statistical interpretation keeps those kinds of evidence separate and makes uncertainty and unreported information visible rather than implying stronger conclusions than the record supports.

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.

The registry record leaves important details of the trial's conduct and analysis unspecified. It does not report the reason for termination, the original sample-size justification, the randomization procedure, allocation concealment, or which parties were masked. The registry describes the trial as fully blinded, but does not report whether masking was assessed or maintained.

The primary outcome description does not state the formula used to combine the Hamilton Depression Scale and Young Mania Rating Scale, nor the direction used to calculate change from baseline to final visit. That omission limits interpretation of the negative arm means and the positive mean difference. The record also does not report a threshold for clinical importance or the confidence level for the posted interval.

Although all participants are described as included under intent-to-treat with LOCF, the number and pattern of missing outcomes, reasons for missingness, and sensitivity analyses are not reported in the registry record. The statistical plan, any baseline adjustment, any interim analyses, and the approach to multiple outcomes are also not reported. No secondary-outcome results are described in the posted results.

Finally, the trial enrolled 40 participants and was restricted to adults with bipolar disorder in specified areas around Calgary or San Diego. The record does not provide enough detail to judge representativeness or generalizability. The posted safety fields report no affected participants in the listed categories, but event-collection details and death counts are not reported in the registry record.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Clinical Trial of a Nutritional Supplement in Adults With Bipolar Disorder — NCT00109577.
• Registry record last updated 2012-09-18; status last verified 2012-09.
• Results first posted 2012-09-11.
• 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 describes a randomized, quadruple-masked, parallel-group phase II trial of MCN36 versus placebo in adults with bipolar disorder. Enrollment was 40, with 20 participants reported in each arm. The trial is listed as terminated; the recorded start was 2005-04 and primary completion was 2009-03.

The primary endpoint was change from baseline to 8 weeks on the Overall Bipolarity Index, a composite of the Hamilton Depression Scale and Young Mania Rating Scale. Mean changes were -21.1 for placebo and -18.9 for MCN36. The unadjusted recalculated mean difference was 2.2 units, with p = 0.028. Interpretation is limited by unreported change-score direction, analysis-plan details, missing-data information, and the confidence level of the posted interval.

The registry lists three secondary outcomes but does not report their results. It records 0 participants with serious adverse events and 0 with other adverse events in each arm; death counts are not reported in the registry record.