A Study of Semorinemab in Patients With Moderate Alzheimer's Disease: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Semorinemab in Patients With Moderate Alzheimer's Disease (NCT03828747), a randomized phase 2 study. The primary outcome is Change From Baseline to Last Visit of Double-Blind Treatment Period in Cognitive Function as Measured by the Alzheimer's Disease Assessment Scale, Cognitive Subscale, 11-Item Version (ADAS-Cog11).

2. Clinical Question

The endpoint is a measure of cognitive performance, not a direct measure of every aspect of Alzheimer's disease. The study's secondary outcomes also included CDR-SB and MMSE changes, adverse events, drug concentration, and anti-drug antibodies. The record states that enrollment was 272, but the primary-outcome analysis population has its own eligibility criteria. The registry record does not report how many participants met those criteria or how many contributed data at each endpoint visit.

3. Trial Design

This was a Phase II, multicenter, randomized, double-blind, placebo-controlled, parallel-group study of semorinemab in patients with moderate Alzheimer's disease. The record lists an enrollment of 272 and describes up to two study cohorts. The trial started on 2019-01-25, had primary completion on 2021-07-20, and is marked completed. These dates describe the registry timeline; they do not establish how long any individual participant was followed.

Semorinemab and placebo were administered intravenously during the double-blind treatment period. The study description also includes an optional open-label extension and a safety follow-up period. During the extension, semorinemab was administered to participants in the semorinemab group and to participants originally assigned placebo. Because the extension was open-label and optional, its data should not automatically be treated as a continuation of the blinded randomized comparison. The registry record does not report participation or follow-up counts for the extension.

Masking is recorded as quadruple. That label indicates the study was registered as having multiple masked parties, but the identities of those parties and the procedures used to maintain masking are not reported in the registry record. The allocation method is recorded as randomized, but details such as sequence generation, concealment procedures, and stratification are not reported.

The primary outcome is change in ADAS-Cog11 from baseline to the last visit of the double-blind treatment period. Its stated timeframe is baseline to Week 49 for Cohorts 1 and 2, and baseline to Week 61 for Cohort 2. The wording assigns two timeframes while naming Cohort 2 in both; further details on the cohort-specific analysis schedule are not reported in the registry record. Other registered outcome measures cover cognitive and daily-function measures, safety, pharmacokinetics, and immunogenicity. The registry record does not describe the full protocol or the rationale for the selected outcome schedule.

4. Participant Flow and Accounting

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

Before assignment: A total of 272 participants were enrolled at 49 centers.

Double-Blind Treatment Period

Reasons for not completing this period:

Open-Label Extension

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The modified intent-to-treat population included all randomized participants who received at least one dose of study drug and had at least one baseline and one postbaseline ADAS-Cog11 score.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline to Last Visit of Double-Blind Treatment Period in Cognitive Function as Measured by the Alzheimer's Disease Assessment Scale, Cognitive Subscale, 11-Item Version (ADAS-Cog11), measured over Baseline to Week 49 for Cohorts 1 and 2, and Baseline to Week 61 for Cohort 2.

The primary endpoint measures change in cognitive function using the Alzheimer's Disease Assessment Scale, Cognitive Subscale, 11-Item Version, abbreviated ADAS-Cog11. The scale has 11 items and is intended to quantify areas of cognition often affected in Alzheimer's disease. Its total score ranges from 0-70, with lower scores indicating better cognitive function. Change from baseline compares a participant's later score with that participant's starting score; it is not the same as comparing endpoint scores without accounting for baseline.

The registered outcome is a least-squares mean change from baseline to the last visit of the double-blind treatment period. A least-squares mean is an estimate from a statistical model, often adjusted for factors included in that model. The registry record does not report which adjustment factors or model details were used, nor does it report the resulting group estimates. Therefore, the scale's direction can be explained, but the observed direction or size of the primary treatment contrast cannot be determined from the available record.

ADAS-Cog11 is used because it gives a structured measure of cognitive performance over time. A measured change can help describe how scores evolved in the study population, while a between-group comparison can address whether the changes differed between randomized treatment groups. The record's primary outcome description does not by itself establish that a difference occurred. Nor does a cognitive score capture all clinical experiences or functional outcomes. The registered secondary outcomes include other measures of cognition and daily function, which address related but distinct aspects of the trial.

The outcome's timeframe is stated as baseline to Week 49 for Cohorts 1 and 2, and baseline to Week 61 for Cohort 2. No cohort-specific primary estimates, standard errors, confidence intervals, or significance tests are reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry names a least-squares mean as the primary outcome parameter and standard error as its dispersion measure. It does not give the fitted model, adjustment variables, covariance assumptions, or a numerical estimate. The statistical labels can therefore be interpreted in general terms, but they cannot be used to reconstruct the analysis or its result.

What does a least-squares mean represent?

In general, a least-squares mean is a model-based estimate of an average. It can differ from a simple arithmetic mean because the model may adjust for baseline values or other prespecified factors, and because model-based estimates can account for the structure of repeated measurements. The exact interpretation depends on the model and the variables included. Those details are not reported in the registry record, so it is not possible to say which adjustments were made or how the estimate relates to an unadjusted group average.

The endpoint is a change score: the score at a specified later visit is considered relative to baseline. For ADAS-Cog11, lower scores indicate better cognitive function. A negative change would represent a score moving downward from baseline and a positive change would represent a score moving upward, if the change is defined as later score minus baseline. The registry record does not explicitly state the subtraction convention, and it does not report the change estimates, so no direction should be assigned to the primary result.

What does the standard error describe?

In general, a standard error describes the estimated uncertainty of a statistic, such as a model-based mean. A smaller standard error usually indicates greater precision under the model and analysis assumptions. It does not describe how much individual participants differ from each other; that is a different kind of variability. A confidence interval is often constructed using an estimate and its standard error, but no confidence interval values are reported in the registry record. Without the estimate and uncertainty measure, the primary result's precision cannot be evaluated.

What would a treatment comparison mean?

In general, a randomized parallel-group trial compares outcomes across groups assigned to different treatments. A treatment contrast in this study would concern the difference between the modeled ADAS-Cog11 changes for semorinemab and placebo during the double-blind period. Randomization can support a causal interpretation when allocation is maintained, outcome measurement and follow-up are appropriate, and the analysis handles missing data and protocol deviations suitably. The record establishes randomized allocation and a double-blind, placebo-controlled design, but does not report enough analysis detail to assess all of those conditions. It also does not report a numerical contrast.

The stated timeframe encompasses baseline to Week 49 for Cohorts 1 and 2, and baseline to Week 61 for Cohort 2. The record does not explain the relationship between these scheduled endpoints, nor does it provide separate estimates by cohort or visit. Combining results across cohorts or time points can require a model and a prespecified estimand; those details are not reported. The primary result is therefore not reducible to a single comparison based on the available record.

Why can over-dispersion matter?

In general, over-dispersion occurs when observed variation is greater than a statistical model assumes. It can make standard errors too small if the model understates variability, which in turn can make confidence intervals too narrow and significance tests too optimistic. This concern is relevant to some outcome types, especially counts, but the registered primary outcome is a continuous cognitive score. The record does not identify an over-dispersion assessment or indicate that it was relevant to the primary analysis. It should not be assumed that over-dispersion was present or that a particular correction was used.

What do the planning assumptions mean?

General sample-size illustrations based on an observed effect are planning exercises, not evidence of observed power. The stated assumptions for such illustrations are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect defined as a stated fraction of the observed effect. The effect estimate needed to calculate those illustrations is not reported in the registry record, and the registry record does not report its own sample-size justification. Any planning illustration should therefore be read as conditional on its assumptions, not as a retrospective measure of the trial's ability to detect an effect. Sample sizes calculated this way are planning illustrations, not observed power.

The primary outcome is described as having several categories or time points and is not reduced to a single comparison. That matters because a collection of related comparisons raises questions about which estimate is primary for interpretation and how uncertainty is controlled across analyses. The record lists 14 outcome measures and 8 analyses, but does not report a complete statistical analysis plan. In particular, the model specification, confidence intervals, p-values, missing-data methods, and any multiplicity adjustment are not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates several outcomes, time points, cohorts, or treatment contrasts. In general, making many statistical comparisons can increase the chance of obtaining a nominally unusual result even when there is no underlying treatment difference. A prespecified hierarchy, adjustment method, or clearly defined primary comparison can help govern interpretation. The registry record lists 14 outcome measures and 8 analyses, and the primary outcome timeframe mentions Cohorts 1 and 2 as well as Week 49 and Week 61. It does not report how multiplicity was handled or how these timepoints and cohorts were prioritized for inference.

The record names the modified intent-to-treat population for the primary endpoint: randomized participants who received at least one dose and had at least one baseline and one postbaseline ADAS-Cog11 score. This definition clarifies who was eligible for that analysis, but the number included is not reported in the registry record. It also does not describe an intention-to-treat analysis of all randomized participants, a per-protocol analysis, or how participants with incomplete outcome data were handled.

Missing-data methods can affect estimates when measurements are unavailable. In general, an analysis that uses only observed measurements relies on assumptions about why data are missing; other methods make different assumptions and may produce different estimates. The registry record does not report the amount or pattern of missing ADAS-Cog11 data, reasons for missingness, imputation methods, sensitivity analyses, or assumptions used to address missing outcomes. These omissions limit assessment of how robust the primary estimate would be to incomplete follow-up.

The registry record also does not report the full statistical analysis plan, model details, interim analyses, or a multiplicity strategy. These items should be described as not reported in the registry record, rather than assumed absent from the conduct of the study. In the absence of those details and numerical primary results, the endpoint description cannot substitute for an evaluation of the final inferential analysis.

11. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: fall (2.5%), agitation (2.3%), hip fracture (2.0%), abdominal pain (1.5%), covid-19 (1.5%).

The safety record presents group-specific counts for serious adverse events, other adverse events, and deaths. For Semorinemab (Double Blind), 23 of 135 at risk had serious adverse events, and 78 of 135 had other adverse events; 1 death is reported. For Placebo, the corresponding counts are 22 of 132 and 74 of 132, with 3 deaths. For Semorinemab (OLE), the counts are 37 of 199 and 122 of 199, with 2 deaths. These are separate safety summaries; the counts should not be combined into a total because participants may contribute to more than one event category and the record does not describe overlap.

The listed serious adverse-event terms and percentages are Fall, 2.5%; Agitation, 2.3%; Hip fracture, 2.0%; Abdominal pain, 1.5%; COVID-19, 1.5%; and Pneumonia, 1.5%. The record does not specify the denominator for these percentages in the registry record for this analysis, so they should be read as the reported percentages rather than converted into counts or treated as group-specific rates.

These figures describe events recorded in the safety summaries. They do not establish that an event was caused by semorinemab or placebo, and the death counts do not identify causes or attribution. The open-label extension is a different study period from the blinded randomized comparison and should not be interpreted as a directly comparable treatment arm without further design and analysis information. The registry record does not report that additional context.

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

The registry record establishes that this was a randomized, double-blind, placebo-controlled Phase II study of semorinemab in patients with moderate Alzheimer's disease, with ADAS-Cog11 change designated as the primary outcome. It also records safety counts and lists secondary outcomes related to cognition, daily function, adverse events, drug concentration, and anti-drug antibodies. Those details describe the study's objectives and some posted safety data.

They do not establish the primary efficacy result. Numerical ADAS-Cog11 group estimates, a between-group contrast, confidence intervals, standard errors, and p-values are not reported in the registry record. As a result, the available information cannot show whether cognitive scores changed differently between the randomized groups, how large any such difference was, or how uncertain an estimate might be. No treatment-effect conclusion should be drawn from the endpoint name or trial design alone.

The scale ranges from 0-70, with lower scores indicating better cognitive function, but a scale's direction does not reveal the direction of observed results. Likewise, the enrollment figure of 272 does not tell how many participants contributed to the primary analysis or how complete their follow-up was. The modified intent-to-treat criteria are stated, but the size of that population and the handling of missing scores are not reported.

The serious adverse-event and other adverse-event counts provide a descriptive account for the groups named in the record. They do not, without further information, establish causation or allow a complete comparison of safety. In particular, extension-period data are from an optional open-label period, while the initial period was blinded and placebo-controlled. The record does not report the details necessary to treat the extension as equivalent to the randomized comparison.

Statistical interpretation depends on estimates, uncertainty, analysis populations, and methods, as well as design. Here, the design and outcome definitions are available, but key primary results and methodological details are not reported in the registry record. The appropriate conclusion is therefore limited: the record documents the trial's design and outcome framework and reports specified safety counts, while leaving the primary cognitive treatment comparison unresolved.

13. Why This Trial Matters Statistically

This trial is statistically informative as an example of how design labels and endpoint definitions differ from outcome evidence. Randomization, masking, a parallel-group structure, and placebo control describe how the study was organized. They do not, by themselves, show the size or precision of a treatment contrast. For that, readers need the analyzed estimates and their uncertainty, which are not reported for the primary endpoint in the registry record.

The choice of ADAS-Cog11 makes the endpoint interpretable as change in a defined cognitive scale, with lower scores indicating better function. At the same time, the stated timeframe refers to more than one cohort and visit schedule. This highlights why a statistical analysis needs clear definitions of the target comparison, analysis population, and timepoint. The record gives the endpoint description but not the model and cohort-specific results needed to assess those choices.

The record also illustrates the distinction between an overall enrollment figure and an endpoint-specific population. The modified intent-to-treat definition requires treatment exposure and both baseline and postbaseline measurements. Without the number analyzed and a description of missing-data handling, readers cannot assess how the analysis population relates to enrollment or how incomplete measurement may influence estimates.

Finally, the safety summaries combine a blinded treatment period with a separately identified optional open-label extension. Keeping those phases distinct is important: descriptive counts from an extension do not automatically have the same comparative interpretation as randomized, blinded data. The available information supports careful reading of design and safety summaries, while emphasizing that a complete statistical evaluation needs reported outcome estimates, uncertainty, and analysis methods.

14. 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.

A major limitation of the available record is that it does not report the numerical results for the primary ADAS-Cog11 endpoint. The least-squares mean and standard error are identified as the outcome parameter and dispersion measure, but their values and the treatment contrast are not present. There are no reported confidence intervals or p-values for that comparison. This prevents assessment of both the magnitude and precision of the primary result.

The outcome timeframe is described as baseline to Week 49 for Cohorts 1 and 2, and baseline to Week 61 for Cohort 2. The relationship between those cohort and visit references is not explained in the registry record. There are no cohort-specific estimates or analysis details to clarify how these timepoints were incorporated or which contrast should guide interpretation.

The record defines the modified intent-to-treat population but does not report its size, the number with observations at each visit, or the handling of missing measurements. The statistical model, adjustment variables, assumptions, sensitivity analyses, and multiplicity procedures are also not reported. Thus, readers cannot evaluate whether the analysis choices were prespecified or how sensitive results might be to incomplete data and multiple comparisons.

The safety counts are descriptive and do not include event attribution in the registry record for this analysis. The percentages associated with listed serious adverse-event terms lack a stated denominator here, limiting interpretation of their basis. The optional open-label extension is also distinct from the blinded randomized period, and the record does not provide details that would support a direct comparative interpretation across those phases.

These limitations concern what is reported in the registry record, not necessarily every detail available in other study materials. The registry record does not report the protocol's sample-size justification, interim analyses, or full statistical analysis plan.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: A Study of Semorinemab in Patients With Moderate Alzheimer's Disease — NCT03828747.
• Registry record last updated 2024-09-24; status last verified 2024-08.
• Results first posted 2022-10-03.
• 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.

18. Record Summary

NCT03828747 is a completed Phase II, multicenter, randomized, double-blind, placebo-controlled, parallel-group study of semorinemab in patients with moderate Alzheimer's disease. The registry records enrollment of 272, a start date of 2019-01-25, and primary completion on 2021-07-20. The study description includes a blinded treatment period, an optional open-label extension, and safety follow-up.

The primary outcome is change from baseline in ADAS-Cog11 at the last visit of the double-blind treatment period. Its reported timeframe is baseline to Week 49 for Cohorts 1 and 2, and baseline to Week 61 for Cohort 2. The registry record describes the endpoint and modified intent-to-treat population but does not report the primary numerical estimates or treatment comparison.

Safety summaries report serious adverse events, other adverse events, and deaths for Semorinemab (Double Blind), Placebo, and Semorinemab (OLE). The record also lists 14 outcome measures and 8 analyses. Interpretation of efficacy and comparative safety is limited because primary estimates, analysis details, and several relevant population and missing-data details are not reported in the registry record.