Down Syndrome Memantine Follow-up Study: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Down Syndrome Memantine Follow-up Study (NCT02304302), a randomized phase 2 study. The primary outcome is Efficacy of the Drug Memantine as Assessed by Change in Score on the California Verbal Learning Test-II (CVLT-II) Short Form Total Free Recall.

2. Clinical Question

The study record identifies the trial as a Phase II evaluation. Its stated focus is cognitive ability, especially learning and memory, in adolescents and young adults with Down syndrome. A placebo comparison can help distinguish changes associated with receiving the study medication from changes that may occur over time or in the study setting. Randomization is intended to support a fair comparison of assigned groups, while masking is intended to reduce the influence of expectations on care, assessment, or reporting. These design features do not by themselves establish that the groups remained comparable in every respect or that the result applies beyond the participants who were assessed.

3. Trial Design

The registry describes NCT02304302 as a completed Phase II, multicenter, randomized, quadruple-masked, parallel-group, placebo-controlled study. Its official title specifies a 16-week evaluation of memantine hydrochloride in adolescents and young adults with Down syndrome. The record gives enrollment as 160 and identifies University Hospitals Cleveland Medical Center as the sponsor. The recorded start is 2014-10, and primary completion is listed as 2020-07-22. These registry entries describe the study’s administrative record; they do not by themselves explain why enrollment or follow-up took the recorded course.

Participants were assigned to memantine or placebo. The record describes the trial as quadruple masked, but does not identify in the information presented which specific roles were included in that masking designation. The placebo arm description says the pills looked identical to memantine pills. The posted arm description for memantine also refers to dispensing placebo, so that wording is internally inconsistent and does not clarify the dispensing procedure for the active-treatment arm. The registry record does not report further details about allocation concealment or the practical implementation of masking.

The posted primary-outcome analysis includes 76 participants in the placebo group and 73 in the memantine group. This is an analysis population defined by completion of 16 weeks of treatment and availability of neuropsychological assessments at baseline and after treatment. These outcome-analysis counts should not be confused with the enrollment figure, which describes the overall study record. The adverse-event record uses different at-risk group counts, so those figures also should not be treated as the primary-outcome analysis population.

The primary measure was change in CVLT-II Short Form Total Free Recall score between baseline and 16 weeks from the start of treatment. The registry lists six further named cognitive or safety outcomes, as well as seven secondary outcomes posted in total. The statistical-analysis plan, original sample-size assumptions, missing-data procedures, and any interim analyses are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The analysis population consisted of participants who completed 16 weeks of treatment and had neuropsychological assessments at baseline (T1) and after treatment (T2)

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 Efficacy of the Drug Memantine as Assessed by Change in Score on the California Verbal Learning Test-II (CVLT-II) Short Form Total Free Recall, measured over baseline and 16 weeks from start of treatment.

The primary endpoint is change in score on the California Verbal Learning Test-II (CVLT-II) Short Form Total Free Recall, assessed from baseline to 16 weeks from the start of treatment. The registry describes the test as assessing supraspan word-learning ability and using that performance as an index of episodic verbal long-term memory. The selected measure is the total number of target items correct summed across learning trials 1-4. Its scale ranges from 0 to 36, with higher scores representing better outcomes.

A change score expresses how a participant’s measured score differs after treatment from their baseline score. Comparing the average changes between randomized groups focuses the primary analysis on whether the groups changed differently over the study period, rather than simply comparing their final scores. The use of a verbal-learning measure is consistent with the trial’s stated interest in memory and learning. It does not capture every aspect of cognition, everyday functioning, or quality of life.

The registry says that the primary efficacy measure is focused on episodic memory and records the investigators’ hypothesis that memantine would produce significant improvement on this test. That hypothesis is a rationale for selecting the endpoint, not a result. The interpretation depends on the estimated group contrast and its uncertainty, as well as the analysis population and methods. The endpoint is defined for participants who completed 16 weeks of treatment and had both baseline and post-treatment neuropsychological assessments. How participants without those assessments were handled is not reported in the registry record.

7. Results Posted in the Registry

• Memantine: mean 3.49 versus 3.30 on control; difference 0.19 (95% CI -1.36 to 1.74), standardized effect d = 0.04; z-test p = 0.810.

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.

Efficacy of the Drug Memantine as Assessed by Change in Score on the Paired Associates Learning (PAL) From the Cambridge Neuropsychological Test Automated Battery (CANTAB)

• 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 deviations.
• Result:
  • Memantine: mean 0.67 versus 1.00 on control; difference -0.33 (95% CI -1.55 to 0.89), standardized effect d = -0.09; z-test p = 0.596.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Memantine and the control. The interval gives the range of effects that remain compatible with the data.

Efficacy of the Drug Memantine as Assessed by Change in Score on the Recall of Digits Forward (From the Differential Ability Scales; DAS-II)

• 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 deviations.
• Result:
  • Memantine: mean -0.01 versus 0.03 on control; difference -0.04 (95% CI -0.73 to 0.65), standardized effect d = -0.02; z-test p = 0.910.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Memantine and the control. The interval gives the range of effects that remain compatible with the data.

Efficacy of the Drug Memantine as Assessed by Change in Score on the Pattern Recognition Memory (PRM; Part of the Cambridge Neuropsychological Test Automated Battery -- CANTAB)

• 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 deviations.
• Result:
  • Memantine: mean -0.05 versus 0.45 on control; difference -0.50 (95% CI -1.90 to 0.90), standardized effect d = -0.11; z-test p = 0.484.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Memantine and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes cover several aspects of cognition, including paired associates learning, digit recall, pattern recognition memory, spatial working memory, spatial span, and a Go - No Go task, alongside a QTc safety measure. The registry indicates seven secondary outcomes posted and twelve outcome measures and analyses overall. Detailed numerical results are available in the record for three secondary cognitive measures. Across those three, the reported intervals include no difference between groups. The overall pattern therefore does not provide a clear secondary signal that resolves the uncertainty in the primary comparison.

Secondary outcomes can help assess whether an observed result is consistent across related measures or whether a particular domain appears different. They do not automatically confirm or refute the primary result, especially when they assess distinct abilities and have their own measurement properties. The primary endpoint remains the trial’s stated central efficacy question. The secondary findings should be read as additional context rather than as a substitute primary analysis.

Unadjusted secondary results carry less weight when multiple outcomes are tested. In general, the more statistical comparisons that are examined, the greater the opportunity for chance findings to appear noteworthy. A multiplicity plan can address this through prespecified ordering, adjusted thresholds, or explicit exploratory interpretation. Whether any such approach was used is not reported in the registry record. Without that information, isolated secondary p-values should not be treated as conclusive evidence, and a collection of secondary tests should not be interpreted as if each were the sole planned test.

The record does not give detailed results for every listed secondary measure in the registry record described for this analysis. It would therefore be inappropriate to claim that all secondary outcomes were null, consistent, or clinically unimportant. The cautious conclusion is limited to the reported pattern: the detailed secondary contrasts do not offer a clear, precise separation, and the available information does not establish how multiplicity was handled.

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

• Efficacy of the Drug Memantine as Assessed by Change in Score on the Paired Associates Learning (PAL) From the Cambridge Neuropsychological Test Automated Battery (CANTAB)
• Efficacy of the Drug Memantine as Assessed by Change in Score on the Recall of Digits Forward (From the Differential Ability Scales; DAS-II)
• Efficacy of the Drug Memantine as Assessed by Change in Score on the Pattern Recognition Memory (PRM; Part of the Cambridge Neuropsychological Test Automated Battery -- CANTAB)
• Efficacy of the Drug Memantine as Assessed by Change in Score on the Spatial Working Memory (Part of the Cambridge Neuropsychological Test Automated Battery -- CANTAB)
• Efficacy of the Drug Memantine as Assessed by Change in Score on the Spatial Span (Part of the Cambridge Neuropsychological Test Automated Battery -- CANTAB)
• Efficacy of the Drug Memantine as Assessed by Change in Score on the The Go - No Go Task
• Safety and Tolerability of the Drug Memantine as Assessed by Change in QTc Interval

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous: it is a score, and the registry describes the group summaries as means with standard deviations. For a continuous outcome, a mean represents the average observed change, while a standard deviation describes how spread out individual changes are around that average. The standard deviation is not the same as the uncertainty in the estimated group mean or in the difference between group means. A group can have a substantial spread among participants while still having an estimated mean with a different degree of precision.

What does a mean difference represent?

The reported contrast is the memantine group’s mean change minus the placebo group’s mean change. A positive estimate means the observed average change is numerically higher in the memantine group; a negative estimate means it is numerically lower. For the primary measure, higher scores represent better outcomes, but the contrast alone does not establish that a difference is meaningful or caused by treatment. The primary unadjusted recalculation compares the posted group summaries and does not reproduce the sponsor’s mixed-model analysis.

The recalculated primary estimate is 0.19, with a 95% confidence interval from -1.361 to 1.741. The interval describes a range of values compatible with the data under the assumptions of the estimation procedure. In general, a confidence interval is not a guarantee that the true effect lies within its endpoints, nor does it assign a probability to each possible effect after observing the data. Its width is useful for judging how much uncertainty remains. Here, the interval includes zero and allows values on either side of zero.

How should the p-value be read?

The unadjusted primary p-value is 0.81. In general, a p-value measures how surprising data at least as extreme as those observed would be under a specified null model and its assumptions. It is not the probability that the null hypothesis is true, and it does not quantify the size or importance of an effect. A larger p-value is not proof of no effect. It indicates that the observed data do not provide a strong statistical distinction from the null model by this measure. Interpretation should consider the estimate and confidence interval alongside the p-value.

Why are the sponsor and recalculated estimates not identical?

The registry reports a sponsor analysis using mixed models, with a net mean difference of 0.34, a 95% confidence interval from -0.98 to 1.67, and p = 0.61. The recalculation is a simple, unadjusted analysis from posted summary statistics. Mixed models can account for data structure and use modeling choices that a comparison of group means does not reproduce. The registry record does not provide enough details about model specification to explain the numerical difference fully. The sponsor analysis remains the trial’s reported analysis, while the recalculation offers a separate check based on the summary values.

In general, mixed models can be useful when observations have a structured or repeated-measures relationship. Their estimates depend on the model specification, available observations, and assumptions about the data. The registry gives the analysis label and result but does not report the model terms, covariance structure, adjustment variables, or missing-data assumptions. Those details should not be inferred from the method name alone.

What role do standard deviations play?

The posted standard deviations describe variability in change scores within each outcome group. They help characterize how much individual outcomes varied, but they do not reveal the full distribution of scores or whether the data met assumptions for a particular model. In general, unusually large variation, skewness, or outliers can affect analyses of means. The registry record does not report diagnostic plots or distributional checks for this endpoint.

The primary scale ranges from 0 to 36 and higher scores represent better outcomes. Although that range provides context for the units, it does not supply a minimal clinically important difference. Such a threshold is not reported in the registry record. Statistical evidence about a mean contrast and clinical judgment about what size of score change matters are related but distinct questions.

Why does the analysis population matter?

The registry defines the primary analysis population as participants who completed 16 weeks of treatment and had neuropsychological assessments at baseline and after treatment. This is a completer-and-assessment-based population, not an explicitly described analysis of every randomized participant. In general, restricting an analysis to participants with complete outcome information can affect interpretation if inclusion is related to prognosis, treatment assignment, or outcome. The record does not report how missing outcomes were handled, why assessments were unavailable, or whether sensitivity analyses were conducted.

Randomized assignment supports comparisons between groups at the start of a trial, but the interpretation of an analysis restricted to completed treatment and available assessments also depends on what happened after assignment. The registry record does not report an intention-to-treat analysis or enough detail to determine whether the mixed model included participants with incomplete follow-up. No assumption about those procedures should be made from the posted result alone.

How should secondary analyses be viewed?

The registry lists multiple outcomes, including measures of learning, memory, working memory, response control, and QTc interval. Testing several outcomes creates a multiplicity issue: when many statistical questions are examined, the chance of at least one apparently unusual result can increase. General approaches include prespecifying a hierarchy, adjusting significance thresholds, or treating secondary results as exploratory. The registry record does not report whether any multiplicity procedure or hierarchy was used. Consequently, unadjusted secondary p-values should be interpreted cautiously and should not be treated as independent confirmation of the primary result.

The trial’s original statistical plan, sample-size justification, interim analyses, and detailed model specifications are not reported in the registry record. Their absence limits assessment of whether the posted analyses match the intended plan and how particular modeling choices were made. General explanations of statistical methods do not fill in those study-specific details.

11. Confidence Intervals and Precision

The primary recalculated estimate is a mean difference of 0.19 points, with a 95% confidence interval from -1.361 to 1.741. This interval is wide enough to include values below and above zero. It therefore leaves uncertainty both about the direction of the comparative change and about the size of any difference. The interval is more informative than the point estimate alone because it shows that the observed estimate is not precise enough to rule out a range of alternatives.

The sponsor’s mixed-model result has a mean difference of 0.34 and a 95% confidence interval from -0.98 to 1.67. It, too, spans zero and permits estimates in either direction. The sponsor and recalculated intervals are based on different analyses, so they should not be treated as interchangeable outputs from one model. Both indicate that the posted primary data do not narrowly pin down the comparative effect.

In general, a confidence interval’s width reflects statistical uncertainty under the analysis assumptions. It does not state whether all values within it are equally plausible, and it does not account for every possible source of bias or measurement error. A narrow interval can still be affected by biased methods, while a wide interval can leave a result inconclusive even when the point estimate appears small. The registry does not report a clinically important difference threshold for the CVLT-II outcome. Thus, the interval cannot be classified as excluding effects that would matter clinically.

The outcome analysis includes 76 placebo participants and 73 memantine participants, defined by treatment completion and the required assessments. These counts help contextualize precision, but do not on their own show how representative the analysis is of all randomized participants. The registry record does not report the number or characteristics of participants missing from this endpoint analysis or the methods used to address missing outcomes. Precision should therefore be understood as precision within the reported analysis, not as a complete account of uncertainty about all participants enrolled in the study.

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 are planning calculations based on the observed effect, not the trial’s original design assumptions. They use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The target effect is a stated fraction of the observed effect. For continuous outcomes, the figures are participants per arm; the convention for time-to-event outcomes would be total events, but the primary endpoint is continuous. The registry record does not report the assumptions used in the trial’s original sample-size justification.

The 80% power column describes the number of participants per arm estimated to give an 80% chance of detecting the specified effect under the calculation’s assumptions, if that effect were the true effect and the analysis assumptions held. The 90% power column describes the corresponding planning target for a 90% chance. These are design illustrations, not guarantees about a future study’s results. They depend on the assumed effect and other stated calculation choices.

For the primary outcome, the illustration lists 10,198 participants per arm for planning around 100% of the observed effect at 80% power, and 13,652 participants per arm for 90% power. For 75% of the observed effect, it lists 18,129 participants per arm for 80% power and 24,269 for 90% power. For 50% of the observed effect, it lists 40,789 participants per arm for 80% power and 54,605 for 90% power. The pattern reflects a general statistical principle: smaller effects are harder to distinguish from random variation and require more participants under otherwise similar assumptions.

These values are not observed power and do not show the probability that the completed trial discovered an effect. They are prospective planning illustrations derived from the observed estimate, which itself is uncertain. When the observed effect is small, planning around it can produce very large estimates; those calculations should not be mistaken for a recommendation or for evidence about the effect’s clinical importance.

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 = 0.190; pooled standard deviation σ = 4.843.
2. n = 2 × (1.96 + 0.842)² × 4.843² / 0.190² = 10,197.1, rounded up to 10,198 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 <- 0.19                      # observed difference in means
sigma <- 4.84254634485214                     # 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(2304302)
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 = 10,198 (80% power), 13,652 (90% power)
 75% of observed effect: n per arm = 18,129 (80% power), 24,269 (90% power)
 50% of observed effect: n per arm = 40,789 (80% power), 54,605 (90% power)
power.t.test: 10198.1 
Simulated power at n = 10,198 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 sponsor’s reported primary analysis is a mixed-model analysis with a net mean difference of 0.34, a 95% confidence interval from -0.98 to 1.67, and p = 0.61. The separate recalculation from the posted group summary statistics gives a mean difference of 0.19, a 95% confidence interval from -1.361 to 1.741, and p = 0.81. The registry does not flag a discrepancy between the posted and recalculated results.

The two estimates are not identical because they come from different analytical approaches. The recalculation is a simple, unadjusted comparison of the posted means and standard deviations. The sponsor’s mixed model may incorporate data structure or assumptions that cannot be reproduced from those summary statistics. The registry record does not give the model specification or enough detail to identify precisely which choices account for the difference.

The sponsor’s result is the trial’s reported analysis and should be distinguished from the recalculation rather than replaced by it. Both estimates have confidence intervals that include zero, and neither gives a precise estimate that rules out effects in either direction. The p-values also do not establish that there is no effect. Reading the sponsor’s model and the unadjusted recalculation together makes clear that the broad interpretation is not dependent on treating the simple calculation as the trial’s primary analysis.

The registry does not report enough information to evaluate every modeling assumption, the handling of incomplete outcome data, or the prespecified analysis plan. Those details would be needed for a fuller assessment of why the model was selected and how it was implemented.

15. Multiplicity, Missing Data and Analysis Populations

The trial has a primary cognitive endpoint and multiple secondary outcomes. This raises the general issue of multiplicity: conducting many statistical tests can increase the chance of at least one apparently significant result under a setting where none of the tested effects is real. Prespecification and appropriate interpretation help keep secondary findings in context. The registry record does not report a multiplicity adjustment, a hierarchy for testing outcomes, or whether secondary analyses were designated as confirmatory or exploratory. Unadjusted secondary results should therefore be treated cautiously rather than as stand-alone confirmation.

The registry defines the primary analysis population as participants who completed 16 weeks of treatment and had neuropsychological assessments at baseline and after treatment. That definition gives useful information about who contributed to the primary endpoint analysis. It does not explain how many participants lacked assessments, why they were unavailable, or whether missingness differed between groups. Those details are not reported in the registry record.

The missing-data method is also not reported. In general, methods for incomplete outcomes rely on assumptions about the relationship between missingness and observed or unobserved outcomes. Different assumptions can affect estimates, especially if participants with missing assessments differ systematically from those assessed. No particular method—such as imputation, a complete-case approach beyond the stated analysis-population definition, or a sensitivity analysis—should be attributed to this trial without documentation.

The record does not report whether an intention-to-treat analysis was conducted, how protocol deviations were handled, or whether the sponsor’s mixed model used participants with incomplete follow-up. The available result is consequently best interpreted as the reported analysis among the defined population, with the handling of other randomized participants unresolved. The same caution applies to interim analyses and the original statistical plan: neither is reported in the registry record. This limits evaluation of the analysis choices and of how the posted findings relate to the original plan, but it does not justify assuming that any particular procedure was or was not used.

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 0% reporting threshold.

Most frequent non-serious adverse event terms

The adverse-event record lists 79 participants at risk in the placebo group and 81 in the memantine group. It records zero serious adverse events in each group and zero deaths in each group. These are the posted counts, not a complete description of every possible safety outcome. The record also lists other adverse events for 44 of 79 at-risk participants in the placebo group and 32 of 81 in the memantine group.

Serious adverse events and other adverse events are separate categories in the registry record. A count of zero serious events in the posted groups means none were recorded in those categories for the listed at-risk participants; it does not establish that the treatment is free of serious risk. The record does not provide the serious-event terms, since no top serious terms are listed. The other-event counts indicate that events were recorded, but the information does not describe their nature, severity, duration, relatedness, or consequences.

The safety at-risk counts differ from the primary outcome-analysis counts. These populations should not be conflated: the primary outcome has its own completion-and-assessment definition, while the safety record reports its own group denominators. These descriptive figures do not establish a comparative safety effect or explain whether differences reflect chance, exposure, or other factors. The registry record does not report further details needed for a comprehensive safety assessment.

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

The primary result describes a comparison of average change in a verbal-learning score between memantine and placebo among participants who completed treatment and had the required assessments. The estimated unadjusted mean difference was 0.19, with a 95% confidence interval from -1.361 to 1.741. The sponsor’s reported mixed-model estimate was 0.34, with a 95% confidence interval from -0.98 to 1.67. Both intervals include zero, leaving uncertainty about the direction and size of the comparative change.

This does not mean that the treatments have been shown to be equivalent. Demonstrating equivalence or non-inferiority requires a suitable design, a prespecified margin, and an analysis addressing that question. Such a margin is not reported in the registry record. Nor does a result that does not clearly distinguish a group difference from zero establish that memantine has no effect. It means the posted analysis does not provide a sufficiently clear estimate to rule out a range of possibilities.

The registry does not report a threshold for a clinically important difference on the CVLT-II endpoint. Consequently, it is not possible to determine from the posted information whether particular values within the confidence interval would represent a meaningful benefit or harm. Statistical significance, effect size, precision, and clinical importance are related but separate considerations.

The findings apply most directly to the defined analysis population and measured endpoint. They do not establish effects on every cognitive ability, daily functioning, or outcomes beyond the 16-week assessment period. The secondary measures add context, but the registry does not report whether multiplicity was controlled, and detailed results are not available for every listed measure in the information. Safety counts are also descriptive and should not be treated as proof of absence of risk.

The study therefore contributes evidence about a specific randomized comparison and memory measure, but the result remains uncertain. Decisions about treatment are outside the scope of these statistics; the registry results alone do not provide a complete basis for clinical conclusions.

18. Why This Trial Matters Statistically

This trial illustrates why a randomized comparison, an effect estimate, and its confidence interval should be read together. The trial’s stated hypothesis anticipated better performance on a verbal-learning measure with memantine. The posted primary estimate is close to zero relative to the uncertainty in the interval, and the interval spans values in both directions. That combination is more informative than focusing only on whether a p-value crosses a threshold.

It also demonstrates that a registry can contain both a sponsor-reported model and summary statistics from which a simpler unadjusted comparison can be recalculated. The mixed-model estimate and the recalculation differ, but both intervals include zero. This helps separate two questions: what the trial’s reported analysis concluded numerically, and how a transparent calculation from posted summaries compares. A simple recalculation is useful as a check, but it does not replace the sponsor’s model or resolve missing model details.

The defined primary analysis population matters. Requiring treatment completion and both assessments means the estimate describes participants meeting those criteria. Without information about missing outcomes and their handling, readers cannot fully judge how closely that analyzed group represents everyone randomized. The difference between primary-outcome and safety denominators is another reminder that trial results can involve distinct analysis populations.

Finally, multiple outcomes create interpretive challenges. A collection of cognitive measures may provide broader descriptive information, but unadjusted testing can make isolated findings less reliable as confirmatory evidence. The registry does not report the multiplicity strategy or full original statistical plan. Statistically, the most defensible reading is therefore careful and bounded: describe the measured contrast, acknowledge its uncertainty, and avoid extending it beyond the endpoint and population actually reported.

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.

The registry record does not report the original sample-size justification, so it is not possible to determine whether the planned enrollment was intended to detect a particular effect on the primary endpoint. Planning illustrations based on the observed effect are not a substitute for that original rationale. The record also does not report the statistical-analysis plan, model specification, interim analyses, or a multiplicity strategy. These omissions limit assessment of prespecification and of how the posted analyses were selected.

The primary analysis population is limited to participants who completed 16 weeks of treatment and had assessments at both time points. The registry does not report how many participants lacked primary outcome data, reasons for missingness, or analyses addressing it. This leaves uncertainty about the relationship between the analyzed population and all randomized participants. The record also does not state whether an intention-to-treat analysis was performed.

The primary endpoint is one measure of verbal learning and episodic verbal long-term memory. It does not represent every cognitive domain or everyday function. Although the scale range and score direction are described, a clinically important difference threshold is not reported. That makes it difficult to translate a numerical contrast into a judgment of practical importance.

The registry lists a multicenter trial and quadruple masking, but the information described does not specify the roles masked or provide details on allocation concealment, site effects, adherence, or protocol deviations. The memantine arm description contains wording that refers to dispensing placebo, so it does not clearly document the active-arm dispensing procedure. These details should not be resolved by assumption.

Safety reporting is limited to the posted serious-event, other-event, and death counts. The nature and severity of other events and their relationship to treatment are not described in the registry record considered. Follow-up beyond the 16-week treatment assessment is not reported for the primary outcome. These limitations narrow what can be concluded about the study’s results and their broader applicability.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Down Syndrome Memantine Follow-up Study — NCT02304302.
• Registry record last updated 2023-06-27; status last verified 2022-04.
• Results first posted 2022-04-05.
• 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

NCT02304302 is a completed Phase II multicenter randomized, quadruple-masked, parallel-group placebo-controlled study of memantine in adolescents and young adults with Down syndrome. The registry lists enrollment of 160 and a 16-week treatment evaluation. The primary endpoint was change in CVLT-II Short Form Total Free Recall score from baseline to 16 weeks, with higher scores representing better outcomes.

The primary outcome analysis included 76 placebo participants and 73 memantine participants who completed treatment and had the required assessments. The recalculated mean difference was 0.19, with a 95% confidence interval from -1.361 to 1.741; the sponsor’s mixed-model estimate was 0.34, with a 95% confidence interval from -0.98 to 1.67. Both intervals include zero. The posted findings therefore leave uncertainty about the size and direction of the comparative change.

The record also lists multiple cognitive and safety outcomes, but does not report the multiplicity strategy, missing-data procedures, original sample-size assumptions, or detailed statistical plan. The safety record lists zero serious adverse events and zero deaths in each group. Interpretation should remain limited to the reported measures and populations.