Memantine Augmentation of Lamotrigine Incomplete-Response in Bipolar Depression: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Memantine Augmentation of Lamotrigine Incomplete-Response in Bipolar Depression (NCT00305578), a randomized phase 4 study. The primary outcome is Change in 17-item Hamilton Depression Rating Scale From Baseline to 8 Weeks (Baseline - 8 Wks).

2. Clinical Question

3. Trial Design

This was a randomized, triple-masked, parallel-group Phase 4 trial comparing daily memantine with daily placebo in people with bipolar depression and an incomplete response to lamotrigine. The registry lists Indiana University as the sponsor. The study is marked completed, with a recorded start of 2005-08 and primary completion of 2010-12.

The registry records 29 enrolled participants. The posted outcome data include 15 participants in the placebo arm and 14 in the memantine arm. The parallel design means the posted comparison is between separate groups assigned to the two study arms, rather than a comparison of each participant with themself under both conditions. Random allocation is recorded, but the method used to generate or conceal assignment is not reported in the registry record.

Triple masking indicates that the registry classifies the trial as masked across three roles or groups; which specific roles were masked and how masking was maintained are not reported in the registry record. The intervention descriptions identify daily memantine and daily placebo. Details such as the memantine dose, treatment adherence, permitted co-interventions, and any treatment changes are not reported in the registry record.

The reported primary outcome was change in the 17-item Hamilton Depression Rating Scale from baseline to 8 weeks. The registry describes the scale as measuring depression severity, with higher scores indicating greater severity. The record includes arm means and standard deviations and one mean-difference comparison. It does not report a trial-specific statistical analysis plan, the original sample-size justification, or further operational details of the design.

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

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 Change in 17-item Hamilton Depression Rating Scale From Baseline to 8 Weeks (Baseline - 8 Wks), measured over 8 weeks.

The primary endpoint was change in the total score on the 17-item Hamilton Depression Rating Scale from baseline to 8 weeks. The registry describes the scale as a measure of depression severity, with a total range from 0 - 50 and higher scores signifying greater severity. The outcome was recorded in units on the scale, and the registry reports the mean and standard deviation for each arm.

A change score represents the difference between a participant’s score at the start of the measurement period and the score at its end. The endpoint title specifies Baseline - 8 Wks. Under that direction of subtraction, a positive change corresponds to a lower score at 8 weeks than at baseline, consistent with lower measured severity. The registry does not provide participant-level scores or additional detail about how the change was calculated for participants with missing assessments.

A continuous score can retain information about the degree of change, rather than classifying participants only as improved or not improved. Comparing arm means gives an estimate of the average difference in change between groups. The mean does not show how every participant changed, and it can be affected by variation in individual outcomes. The standard deviations describe the spread of scores around each arm’s mean. The endpoint’s clinical interpretation also depends on context, including the size and precision of the difference and the measurement’s relevance to participants; those considerations cannot be resolved by the mean alone.

7. Results Posted in the Registry

• Memantine: mean 9.00 versus 7.00 on control; difference 2.00 (95% CI -2.76 to 6.76), standardized effect d = 0.31; z-test p = 0.410.

8. Statistical Interpretation

9. Statistical Methods Explained

The reported outcome is continuous: each participant has a numerical change score, and the registry presents the mean and standard deviation in each arm. A mean summarizes the average score in a group. A standard deviation describes how spread out individual scores are around that average. These summaries help describe both the typical direction and the variability of observed change, but they do not show the full distribution or the experience of every participant.

What does the mean difference represent?

The mean difference compares the group averages. In this record, the reported direction is memantine minus placebo: 9.0 minus 7.0 as represented by the registry’s estimate of 2.0 units. This is an unadjusted difference in posted arm means, not a model-based estimate accounting for other participant characteristics. The registry does not report a covariate-adjusted comparison.

Because the outcome is defined as baseline minus the 8-week score, positive change means the recorded score fell over that period. A positive between-arm estimate therefore indicates a larger average baseline-to-week-8 change in the memantine arm within the posted data. It does not by itself establish why the difference arose, whether it reflects a reliable effect, or how much individual outcomes overlapped.

How should the p-value be understood?

A p-value is a measure of compatibility between observed data and a specified statistical model under a null hypothesis, commonly a model in which the groups have no average difference. Smaller values indicate that the data would be less compatible with that null model, given the test’s assumptions. The posted unadjusted p-value is 0.41. It should not be read as the probability that the null hypothesis is true, or as a direct measure of the size or importance of an effect. The specific test, its assumptions, and any trial-specific threshold for interpretation are not reported in the registry record.

Unadjusted means the reported p-value is presented without an adjustment for multiple comparisons. The record has one posted primary outcome and no posted secondary outcomes, but it does not provide a complete account of all analyses considered or performed. The number and nature of any unreported analyses are not reported in the registry record, so the consequences of multiplicity cannot be fully assessed.

Why do confidence intervals matter?

A confidence interval presents a range of values compatible with the data and the method used to calculate it. Its width conveys precision: a narrower interval indicates a more concentrated range of estimates, while a wider interval reflects greater uncertainty. Here, the reported interval for the mean difference runs from -2.761 to 6.761 units. It includes zero and allows for values on both sides of zero. The confidence level and the calculation method are not reported in the registry record.

Confidence intervals are not guarantees that the true value lies within the particular displayed limits. Their interpretation relies on the statistical procedure and its assumptions. A confidence interval also does not determine whether an effect is meaningful in clinical terms. The registry does not report a threshold for a clinically important difference on this scale, so the clinical importance of values within the interval cannot be judged from the record alone.

What does variation contribute?

Individual change scores can differ substantially even when group means are close. The arm standard deviations, 6.0 units and 7.0 units, describe this dispersion. In general, greater variation makes it harder to estimate a group difference precisely when sample size is held fixed. For count outcomes, over-dispersion means that variation exceeds what a chosen count model expects and can make standard errors too small if ignored. That concept is not the basis of this continuous endpoint; the registry reports standard deviations rather than event counts for the primary comparison.

The reported pooled standard deviation is 6.501, and the standardized effect size is 0.308. Standardizing expresses a difference relative to the scale of variation, which can aid comparisons across measures with different units. Such a standardized summary does not replace the original-unit estimate, which is more directly tied to the Hamilton scale. The registry does not report whether or how the standardized value was used in the trial’s own analysis.

What can be inferred from the posted summaries?

The arm means, standard deviations, sample sizes, confidence interval, and p-value allow a reader to describe the posted comparison and its uncertainty. They do not supply individual-level data, a full statistical model, checks of model assumptions, or a complete account of missing observations. The registry record does not report the trial’s original analysis plan, the handling of missing data, analysis populations, or any interim analyses. Those absences limit the ability to reproduce or fully evaluate the inferential analysis.

The trial’s design is randomized and parallel, but the record does not describe the randomization procedure or the details of masking. Randomization is intended to support a fair comparison by balancing factors across arms in expectation; in a small trial, chance differences in participant characteristics may still occur. Whether the posted analysis accounted for baseline score or other prognostic factors is not reported in the registry record. The reported group summaries should therefore be understood as the registry’s posted evidence, rather than as a complete statistical account of every design and analysis decision.

10. Confidence Intervals and Precision

The estimated mean difference is 2.0 units for memantine minus placebo. Its confidence interval runs from -2.761 to 6.761 units. The interval includes zero, indicating that the posted data do not pin down the direction of the average difference with precision. It is compatible with an average difference favoring either arm as well as with no average difference.

The breadth of the interval matters as much as the point estimate. A point estimate is the single value calculated from the observed sample, while the interval communicates uncertainty around that estimate under the method used. The confidence level and interval calculation method are not reported in the registry record, so the interval should be interpreted as posted rather than reconstructed or assigned an unreported confidence level.

The group sizes were 15 for placebo and 14 for memantine. The observed standard deviations were 6.0 and 7.0 units, respectively. With relatively few observations and variation in scores, the estimate has limited precision. The interval does not show that the arms are equivalent; it shows that the observed data leave a range of plausible differences unresolved.

Precision is not the same as clinical importance. The registry does not report a threshold for a clinically important change or a minimally important between-arm difference on the scale. It is therefore not possible to determine from this record alone whether any value within the interval would represent a meaningful difference to participants. The confidence interval should be read together with the trial’s small sample and the absence of detailed reporting on the statistical analysis.

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 planning figures are participants per arm, not total event counts. They illustrate the number of participants per arm under a planning calculation based on the observed effect, not the number enrolled or analyzed in the trial. The trial enrolled 29 participants, with 15 in the placebo arm and 14 in the memantine arm in the posted primary outcome.

The listed 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 as a stated fraction of the observed effect. The registry’s original design assumptions and sample-size justification are not reported in the registry record, so these calculations should not be mistaken for the trial’s own prospective planning.

The 80% power column means the number of participants per arm calculated to give an 80% chance of detecting the specified effect under the stated planning assumptions, if that effect and assumptions held. The 90% power column expresses the corresponding planning target for a 90% chance. For 100% of the observed effect, the figures are 166 participants per arm for 80% power and 223 participants per arm for 90% power. For 75% of the observed effect, they are 295 and 395 participants per arm; for 50% of the observed effect, they are 664 and 889 participants per arm.

Smaller effects require more participants to detect under these assumptions, which is why the planning figures increase as the target effect becomes smaller. These calculations are not observed power and do not describe the probability that the completed trial detected an effect. The curve is a planning aid for relating effect size to sample size under the stated assumptions, not a record of what happened in the trial.

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 = 2.000; pooled standard deviation σ = 6.501.
2. n = 2 × (1.96 + 0.842)² × 6.501² / 2.000² = 165.8, rounded up to 166 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                      # observed difference in means
sigma <- 6.50071221169337                     # 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(305578)
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 = 166 (80% power), 223 (90% power)
 75% of observed effect: n per arm = 295 (80% power), 395 (90% power)
 50% of observed effect: n per arm = 664 (80% power), 889 (90% power)
power.t.test: 166.8 
Simulated power at n = 166 per arm: 0.799 (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 record contains one posted primary outcome and no posted secondary outcomes. It reports one comparison for the primary endpoint, with an unadjusted p-value of 0.41. However, the record does not provide a full analysis plan or account of all outcomes and comparisons that may have been considered. Whether any multiplicity adjustment was planned or used is not reported in the registry record. The unadjusted p-value should therefore be understood as the value posted, without assuming that it reflects a correction for multiple testing.

No analyses are posted in the registry record. This limits the available detail about the statistical procedures, including the test used, any adjustment for baseline measurements, and whether analyses followed the original protocol. The analysis population, such as whether all randomized participants were included in the primary comparison, is not reported in the registry record.

Missing outcome data can affect both the group means and the comparison if participants with missing measurements differ from those with observed measurements. The number of participants with missing 8-week outcome data and the method for handling missing values are not reported in the registry record. The posted arm sample sizes do not by themselves establish whether all randomized participants contributed outcome data or whether the sample counts represent another analysis population.

Interim analyses, stopping rules, and any adjustment related to interim monitoring are not reported in the registry record. Nor are sensitivity analyses or checks of the statistical model assumptions. These omissions do not demonstrate that such procedures were absent; they mean that their use and possible influence on the reported estimate cannot be evaluated from the registry record.

Multiplicity and missing data are separate concerns. Multiplicity relates to the number of statistical questions tested and the chance of finding a small p-value when many comparisons are considered. Missingness concerns incomplete measurements and the assumptions used to analyze them. The posted result can be described, but the available reporting does not support a full audit of either issue.

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 5% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports serious adverse events for 0 of 15 participants in the placebo arm and 0 of 14 participants in the memantine arm. These are the posted counts in the serious adverse event summary. The record lists no serious adverse event terms. The death count is not reported in the registry record.

A count of zero means no serious adverse events were recorded for the listed participants in that arm in this registry summary. It does not establish that the risk is absent or that the two arms have identical safety profiles. With small groups, the posted counts provide limited information about uncommon events, and the registry does not give further detail on event definitions, observation time for safety, or how events were collected.

For other adverse events, the registry reports events affecting 14 of 15 participants in the placebo arm and 11 of 14 participants in the memantine arm. The record does not provide further event details in the figures. These counts should be kept separate from serious adverse events and should not be interpreted as a measure of event severity or causation. The summaries describe what was recorded, not whether an event was caused by study treatment.

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

The result is an estimate of the difference in mean eight-week change on the 17-item Hamilton Depression Rating Scale between the memantine and placebo arms in this randomized trial. The posted means are 9.0 units for memantine and 7.0 units for placebo, with a reported mean difference of 2.0 units. The confidence interval from -2.761 to 6.761 and the unadjusted p-value of 0.41 indicate substantial uncertainty about the direction and size of the difference.

The result does not establish that memantine is more effective, less effective, or equivalent to placebo. An interval that includes zero does not prove there is no difference. It indicates that the observed sample does not distinguish precisely among zero and multiple possible differences under the reported analysis. The registry does not report a threshold for clinical importance, so statistical summaries alone cannot determine whether any of those possible differences would matter to participants.

The result also does not establish that the same estimate would apply to people outside the enrolled trial population. The record describes participants with bipolar depression and an inadequate response to lamotrigine, with a stated minimum lamotrigine dose and treatment duration in the brief summary. Further eligibility details and participant characteristics are not reported in the registry record. Generalizing beyond the described study context would require information not available in the posted record.

The safety summaries report no serious adverse events in either arm, alongside other adverse events affecting participants in both arms. These counts do not show that serious harms cannot occur or that observed events were caused by treatment. Their interpretation is constrained by the small groups and limited event detail.

In short, the record presents a small randomized comparison and a numerical estimate favoring a larger average score change in the memantine arm, but with an interval that allows values on either side of zero. The incomplete reporting of analysis populations, missing data, and trial-specific statistical methods further limits what can be concluded from the registry summaries alone.

16. Why This Trial Matters Statistically

This trial is statistically instructive because it places an estimated group difference beside its uncertainty. The point estimate of 2.0 units is easy to notice, but the confidence interval from -2.761 to 6.761 shows that the observed data leave considerable uncertainty about the direction and size of the average difference. Interpreting the estimate without its interval would give an incomplete account of the evidence.

The trial also illustrates how sample size and outcome variability relate to precision. The arm sizes were 15 and 14, and the standard deviations were 6.0 and 7.0 units. A small study can yield a numerical difference, but usually offers less information for distinguishing among plausible effects than a larger study under otherwise comparable conditions. The planning illustration makes this relationship visible by showing that targeting smaller effects requires more participants per arm under its stated assumptions.

Another useful distinction is between statistical evidence and clinical meaning. The registry supplies an outcome on a depression severity scale, but does not report a threshold for a clinically important difference. A p-value cannot substitute for that context, and a confidence interval does not decide what difference matters to patients or clinicians.

Finally, the record demonstrates why transparent reporting matters. Important details about missing outcomes, analysis populations, interim analyses, and the original statistical plan are not reported in the registry record. Without these details, readers can describe the posted result but cannot fully evaluate every decision that may influence an analysis. Clear reporting of design, outcome handling, and analysis choices is part of interpreting trial evidence, not a separate administrative matter.

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 primary outcome comparison is based on 29 participants overall, with 15 in the placebo arm and 14 in the memantine arm. This is a small sample for estimating a group difference, and the confidence interval is wide enough to include values favoring either arm. The registry does not report participant-level outcomes, so distributional features, outliers, or the influence of individual observations cannot be assessed.

Further trial-specific details are not reported in the registry record. These include the original sample-size justification, the randomization procedure, the precise roles covered by triple masking, the memantine dose, baseline participant characteristics, and the statistical analysis plan. The record also does not state whether the comparison was adjusted for baseline score or other factors.

The number of participants with missing primary outcome measurements, the handling of missing data, and the analysis population are not reported in the registry record. The confidence interval method and confidence level are also not reported. These gaps restrict independent assessment of how the estimate was obtained and whether assumptions were appropriate.

The record reports no secondary outcomes and no analyses, and it does not report a clinical-importance threshold for the scale. Details on adverse event definitions and the death count are not reported in the registry record. These limits mean the posted summaries provide only a partial account of efficacy and safety outcomes; they should be interpreted as descriptive registry results rather than a complete report of all trial methods and findings.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Memantine Augmentation of Lamotrigine Incomplete-Response in Bipolar Depression — NCT00305578.
• Registry record last updated 2016-07-18; status last verified 2016-06.
• Results first posted 2015-11-02.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

21. Record Summary

The registry record describes a completed, randomized, triple-masked, parallel-group Phase 4 trial sponsored by Indiana University. It compared daily memantine with daily placebo in people with bipolar depression and an incomplete response to lamotrigine. The recorded study start was 2005-08, and primary completion was 2010-12. Enrollment was 29.

The posted primary outcome was change in the 17-item Hamilton Depression Rating Scale from baseline to 8 weeks. The placebo mean was 7.0 units, and the memantine mean was 9.0 units. The reported mean difference was 2.0 units, with a confidence interval from -2.761 to 6.761 and an unadjusted p-value of 0.41. Serious adverse events were recorded for 0 participants in each arm. No secondary outcomes or analyses are posted.

The registry record does not report the trial’s original sample-size justification, detailed analysis plan, missing-data methods, or analysis populations. The posted comparison therefore provides an estimate with substantial uncertainty, rather than a complete account of the trial’s statistical methods and results.