Study of the Efficacy and Safety of Various Anti-inflammatory Agents in Participants With Mild Cognitive Impairment or Mild Alzheimer's Disease: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of the Efficacy and Safety of Various Anti-inflammatory Agents in Participants With Mild Cognitive Impairment or Mild Alzheimer's Disease (NCT04795466), a randomized phase 2 study. The primary outcome is Change From Baseline in Cognition as Measured by the Neuropsychological Test Battery (NTB) Z-scores.

2. Clinical Question

The registry describes the broader objective as evaluating anti-inflammatory agents in early Alzheimer’s disease. Because the study terminated early, only a single agent could be studied. The posted comparison is therefore specifically between canakinumab and placebo; it does not provide a comparison among several anti-inflammatory agents. The study enrolled 34 participants overall, while the primary outcome analysis included 10 participants assigned to canakinumab and 16 assigned to placebo. Those analysis counts are smaller than the overall enrollment, so the outcome comparison reflects the participants with available values in the stated analysis population rather than every enrolled participant.

3. Trial Design

This was a phase 2, randomized, placebo-controlled, multicenter platform study. Its design was parallel, meaning participants were assigned to study arms that received canakinumab or placebo. Masking was quadruple. The registry lists 34 participants enrolled and identifies Novartis Pharmaceuticals as the sponsor. The trial status is terminated.

The registry lists a start date of 2021-10-28 and a primary completion date of 2024-03-07. The intervention was canakinumab, given by subcutaneous injection at 150 mg once every 4 weeks for the first 2 doses and 300 mg once every 4 weeks for the subsequent 4 doses. The comparator was matching placebo administered by subcutaneous injection.

Early termination is an important feature of the trial record. The brief summary says that only a single agent could be studied because of early termination. The registry record does not report the reason for termination beyond that statement. The original sample-size justification, planned recruitment targets by arm, and the design assumptions used to determine the trial’s original enrollment are not reported in the registry record.

The primary outcome was change from baseline in cognition measured with NTB Z-scores, assessed at baseline and day 171. The primary outcome population is described as the PD analysis set, with an available value for the outcome measure and no protocol deviations with relevant impact on the PD data. The number of outcome measures is listed as 17, and the registry lists 16 secondary outcomes. The record also includes secondary cognition, neuropsychiatric, and everyday-function measures. Detailed results are available for only some of those secondary measures in the posted results.

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: There was a screening period (Day -60 to Day -8), followed by a baseline period of 7 days (Day -7 to Day -1), before first treatment.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Participants in the PD analysis set with an available value for the outcome measure. The PD (total target) analysis set for biotherapeutic agents included all participants with available (total target) PD data and no protocol deviations with relevant impact on (total target) PD data.

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 in Cognition as Measured by the Neuropsychological Test Battery (NTB) Z-scores, measured over Baseline and day 171.

The primary endpoint measures change in cognition from baseline to day 171 using a composite score from the Neuropsychological Test Battery. The NTB includes tests that assess cognitive domains affected in early Alzheimer’s disease, including memory, executive function, attention, and verbal fluency. In this study, 5 out of 9 NTB components were administered: Rey Auditory Verbal Learning Test immediate and delayed scores, Wechsler Memory Scale Digit Span, Controlled Word Association Test, and Category Fluency Test.

For each component, a raw score was converted to a Z-score, which expresses a score in relation to a mean in standard-deviation units. The total Z-score was derived by averaging the component Z-scores. Change from baseline was calculated as the post-baseline Z-score minus the pre-treatment Z-score. As defined in the registry, zero indicates no cognitive change, a negative value indicates decline, and a positive value indicates improvement.

A composite endpoint can combine information across multiple tests rather than relying on one test alone. General statistical context: composite measures are intended to represent a broader construct, but their interpretation depends on how components are selected, scored, and combined. The registry explains the components and averaging approach but does not report additional details about validation or the rationale for selecting these particular components.

7. Results Posted in the Registry

• Canakinumab: mean 0.23 versus 0.16 on control; difference 0.07 (95% CI -0.22 to 0.36), standardized effect d = 0.19; z-test p = 0.639.

8. Secondary Efficacy Endpoints

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

Change From Baseline in Memory as Measured by the Total Composite NTB Memory Z-score

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Canakinumab: mean 0.46 versus 0.46 on control; difference -0.00 (95% CI -0.34 to 0.34), standardized effect d = -0.00; z-test p = 0.991.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Canakinumab and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Executive Function as Measured by the Total Composite NTB Executive Function Z-score

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Canakinumab: mean 0.11 versus -0.07 on control; difference 0.19 (95% CI -0.20 to 0.57), standardized effect d = 0.38; z-test p = 0.339.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Canakinumab and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Digit Symbol Substitution Test (DSST) Score - 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 errors.
• Result:
  • Canakinumab: mean 1.96 versus 2.45 on control; difference -0.49 (95% CI -3.97 to 2.99), standardized effect d = -0.11; z-test p = 0.783.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Canakinumab and the control. The interval gives the range of effects that remain compatible with the data.

The detailed secondary results cover memory, executive function, and Digit Symbol Substitution Test performance. Across these measures, the posted group differences vary in direction and precision. Their confidence intervals include zero, and the corresponding unadjusted comparisons do not provide a clear separation between the groups. Taken together, this is a mixed and uncertain pattern rather than a consistent secondary signal that resolves the uncertainty in the primary endpoint.

The memory outcome’s reported group means are close, while the executive-function and DSST comparisons have estimates in different directions. Those secondary measures relate to specific cognitive domains or tasks, whereas the primary NTB score is a composite measure. The registry does not report a unified secondary-outcome analysis that combines them into an overall conclusion. Their results therefore complement the primary outcome but do not replace it.

The registry lists 16 secondary outcomes, including measures of memory, executive function, neuropsychiatric symptoms, everyday cognition, and cognitive testing. Numerical results for the other listed secondary outcomes are not reported in the registry record. The three detailed comparisons should not be taken to represent every planned or listed outcome.

Unadjusted secondary results carry less weight when many outcomes are examined. General statistical context: testing multiple outcomes creates more opportunities for apparently notable findings to occur by chance. Methods for controlling multiplicity can address this issue, but the registry record does not report whether a multiplicity adjustment was prespecified or applied. Without that information, secondary p-values are best treated as exploratory rather than as independent confirmation of an effect. The small analysis groups and the breadth of outcomes add further uncertainty.

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

• Change From Baseline in Memory as Measured by the Total Composite NTB Memory Z-score
• Change From Baseline in Executive Function as Measured by the Total Composite NTB Executive Function Z-score
• Change From Baseline in Digit Symbol Substitution Test (DSST) Score - CANTAB
• Change From Baseline in Neuropsychiatric Symptoms as Measured by the Neuropsychiatric Inventory (NPI) Total Score
• Change From Baseline in Neuropsychiatric Symptoms Associated Distress as Measured by the Neuropsychiatric Inventory Caregiver Distress (NPI-D) Score
• Change From Baseline in Mean eNeuropsychiatric at Home Caregiver Assessment Score
• Change From Baseline in Everyday Cognition Scale (ECog) Total Score
• Change From Baseline in eCognitive Testing Scores - SWM Between Errors

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous: each participant’s endpoint is a numerical change in a composite cognition score. For a continuous outcome, the comparison can be expressed as a difference in group means. A positive mean difference indicates a higher mean change in the canakinumab group than in the placebo group, using the registry’s direction for the score. It does not by itself establish the clinical importance of that difference.

What does the reported model estimate?

The registry reports a Mixed Models Analysis with a least-squares mean difference. A mixed model is a statistical approach that can analyze outcome measurements while accounting for features of the data structure, such as repeated measurements within participants. General statistical context: model-based least-squares means are estimates derived from the fitted model, rather than necessarily being simple arithmetic averages of the observed scores. The registry does not report the model specification, included covariates, covariance structure, or detailed assumptions.

The primary comparison is reported as a least-squares mean difference of 0.069. The model result is accompanied by a 90% confidence interval and a p-value. The posted arm summaries also give least-squares means and standard errors. A standard error describes the estimated sampling variability of an estimate; it is not the standard deviation of individual participant scores. Confusing these quantities can make the spread of participant-level outcomes appear smaller or larger than it is.

What is the role of the confidence interval and p-value?

A confidence interval describes uncertainty around an estimated effect under the assumptions of the method used. The primary 90% interval extends from -0.178 to 0.316. Since zero lies within that interval, the result is not a clear distinction between a positive and negative group difference at the interval’s stated confidence level. The p-value reported by the registry is 0.639. General statistical context: a p-value measures how compatible the observed data are with a specified null hypothesis under the analysis assumptions. It is not the probability that the null hypothesis is true, nor is it a measure of the size or importance of an effect.

The re-calculation uses the posted arm-level summaries to form a simple, unadjusted comparison. It gives the same reported point difference, 0.069, and an unadjusted p-value of 0.639. Its estimates rely on summary statistics and an approximate standard deviation. The registry’s mixed-model analysis remains the trial’s own reported analysis, with its model-based interval and p-value. Differences between a model-based analysis and a simple re-calculation can arise because they use different information or assumptions; the registry does not provide enough model detail to attribute the result to any particular feature.

Why does over-dispersion matter?

General statistical context: over-dispersion occurs when the observed variability is greater than a model assumes. If a method understates variability, standard errors may be too small and intervals may be too narrow. If variability is greater than assumed, estimates can be less precise than a simple model suggests. The primary outcome is continuous, and over-dispersion is not identified as a particular issue in the registry record. The more general point is that uncertainty depends on both the estimated difference and the variability in the data.

For continuous outcomes, variation can reflect differences among participants, measurement variability, and other features of the study. Standard errors quantify uncertainty in group estimates, whereas standard deviations describe spread among participant values. The posted arm summaries report standard errors. The simple re-calculation uses an approximate pooled standard deviation derived from those summaries, so it is an approximation and not the sponsor’s full analysis.

How should the analysis population be understood?

The registry identifies the primary population as participants in the PD analysis set with an available outcome value. It further describes the PD set as including participants with available PD data and no protocol deviations with relevant impact on those data. This is the population described for the primary measure; it should not be relabeled as an intention-to-treat population because the registry does not use that description for this endpoint. The primary outcome analysis included 10 participants assigned to canakinumab and 16 assigned to placebo.

The registry record does not report how missing outcome values were handled, whether any imputation was used, or whether sensitivity analyses examined the effect of missing data. It also does not report the details of exclusions from the PD analysis set. These omissions matter because a result among participants with available outcome values can differ from a result based on all randomized participants if missingness or exclusions are related to outcomes or treatment assignment.

General statistical context: randomization helps balance groups in expectation, but it cannot remove uncertainty in a small realized sample. With limited numbers, chance differences in baseline characteristics or outcome availability can influence estimates. The registry identifies randomized allocation and masking, but detailed information about baseline balance and the analytic handling of such differences is not reported in the registry record.

What should be inferred from statistical significance?

The p-value and interval are parts of the statistical evidence, not stand-alone judgments about clinical relevance. The primary result has a confidence interval that includes zero and an unadjusted recalculation with a p-value of 0.639. Those features indicate substantial uncertainty in the estimated contrast. They do not demonstrate equivalence, because no equivalence margin is reported. They also do not establish that the groups are identical. The registry’s reported analysis is described as a superiority analysis; the registry record does not report a non-inferiority or equivalence framework.

11. Confidence Intervals and Precision

The primary estimate is 0.069, with a registry-reported 90% confidence interval from -0.178 to 0.316. This interval is broad in relation to the point estimate and includes zero. It therefore leaves open both a negative difference and a positive difference, as well as values near no difference. The data do not locate the group contrast precisely.

General statistical context: a confidence interval is a range produced by a statistical method that reflects uncertainty in an estimate under specified assumptions. In repeated use of the same method under its assumptions, intervals at a stated confidence level would cover the true parameter at that rate. A particular interval does not assign a probability to the true effect after the data have been observed. The 90% level is the level reported in the registry’s analysis.

The primary analysis included 10 participants assigned to canakinumab and 16 assigned to placebo. Limited analysis counts mean that estimates can be sensitive to the observed outcomes and that uncertainty can remain substantial. The interval’s width is a direct reminder that the point estimate should not be treated as a precise value. A confidence interval including zero also does not prove the absence of a difference; it indicates that zero is among the values compatible with the reported interval.

Precision concerns the range of plausible effects, while clinical importance concerns whether effects are meaningful in context. The registry does not report a threshold for a clinically important difference in this Z-score. Without such a threshold, the interval can describe statistical uncertainty but cannot by itself establish the practical importance of the values it contains.

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 planning figures are illustrations based on the observed effect, not the trial’s original enrollment plan. For these continuous outcomes, the figures are participants per arm, assuming equal allocation between two arms. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. Time-to-event planning figures, if present, would instead count total events using the Schoenfeld method; the outcomes are not time-to-event outcomes.

The 80% power column describes the planned number of participants per arm under those assumptions for a design with an 80% chance of detecting the specified effect at the stated significance level, if that effect is the true effect and the assumptions hold. The 90% power column expresses the analogous planning target for 90% power. These are design probabilities for hypothetical future studies, not probabilities that the current result is correct.

For each effect scenario, the planning numbers use a fraction of the observed effect: 100%, 75%, or 50%. A smaller target effect is harder to distinguish from random variation, so the planning illustration requires more participants per arm as the target effect becomes smaller. This relationship is a general feature of study planning, not evidence that the smaller effect is present.

The observed effects come from a small analysis, so planning numbers that use them can be unstable and should not be treated as definitive enrollment recommendations. The trial’s own sample-size justification and original design assumptions are not reported in the registry record. These planning figures are not observed power: they are forward-looking illustrations based on stated assumptions, not a retrospective measure of the trial’s chance of success.

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 = 0.069; pooled standard deviation σ = 0.364.
2. n = 2 × (1.96 + 0.842)² × 0.364² / 0.069² = 436.4, rounded up to 437 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 <- 0.069                      # observed difference in means
sigma <- 0.363816574663662                     # 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(4795466)
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 = 437 (80% power), 585 (90% power)
 75% of observed effect: n per arm = 776 (80% power), 1,039 (90% power)
 50% of observed effect: n per arm = 1,746 (80% power), 2,337 (90% power)
power.t.test: 437.4 
Simulated power at n = 437 per arm: 0.798 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The registry’s own primary analysis used a Mixed Models Analysis and reported a least-squares mean difference of 0.069, a 90% confidence interval from -0.178 to 0.316, and a p-value of 0.639. This is the trial’s reported analysis and should be distinguished from the simple re-calculation based on posted arm summaries.

The re-calculation also gives a mean difference of 0.069 and an unadjusted p-value of 0.639. It is an approximation using summary statistics, not a re-fit of the sponsor’s mixed model. The registry does not report enough detail about the model specification to reproduce the sponsor’s analysis independently. The registry record does not flag a discrepancy between the reported and recalculated results.

The sponsor’s analysis and the simple re-calculation should not be treated as interchangeable methods. The sponsor’s result reflects the model it used, while the re-calculation provides a basic, unadjusted check using the posted group summaries. Their close point estimates support a similar broad reading of the primary contrast, but the model-based result is the official trial analysis recorded in the registry. Neither analysis removes the uncertainty indicated by the confidence interval or the limited number of participants in the outcome analysis.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists 16 secondary outcomes in addition to the primary cognitive outcome. The posted results include detailed analyses for memory, executive function, and DSST, as well as the primary endpoint. The registry record does not report whether the trial prespecified a hierarchy among outcomes, adjusted significance levels for multiple comparisons, or controlled the family-wise error rate or false discovery rate. These details are important for interpreting secondary p-values because a larger set of tests increases the chance of finding an apparently unusual result by chance.

The primary analysis is identified as a superiority analysis. The registry does not report an equivalence or non-inferiority margin, and the primary endpoint should not be interpreted as an equivalence test. The registry record also does not report a detailed statistical analysis plan, interim analyses, or a stopping rule. The fact that the study was terminated early is stated, but further details about the decision process are not reported in the registry record.

The primary outcome population is the PD analysis set with an available value for the outcome measure, and the record describes the PD set in relation to available PD data and relevant protocol deviations. This differs from an explicit statement that all randomized participants were included. The primary outcome analysis counts were 10 for canakinumab and 16 for placebo, while overall enrollment was 34.

The registry record does not report the number or reasons for missing primary outcome values, the handling of missing data, or any sensitivity analyses for missingness. It also does not report the treatment of protocol deviations beyond the population description. General statistical context: if outcome availability is related to treatment assignment or participants’ outcomes, analyses limited to those with observed values can be subject to bias. The available record does not establish whether that concern affected this result, so the direction or size of any such effect cannot be inferred.

16. Safety Results

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

Deaths

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: covid-19 (6.2%), streptococcal infection (6.2%), transitional cell carcinoma (6.2%), metabolic encephalopathy (5.6%).

The adverse-event summary reports serious adverse events in 2 of 16 participants at risk in the canakinumab group and 1 of 18 in the placebo group. Across the trial, 3 of 34 participants at risk were reported with serious adverse events. The record lists no deaths in either group or overall. It also reports other adverse events in 13 of 16 participants at risk in the canakinumab group and 16 of 18 in the placebo group.

The serious-event terms listed include COVID-19, streptococcal infection, transitional cell carcinoma, and metabolic encephalopathy. The registry reports 6.2% for each of the first three listed terms and 5.6% for metabolic encephalopathy. These term-level figures do not by themselves establish whether events were caused by study treatment. The registry record does not report a causal attribution for these terms in the information summarized.

These are descriptive counts from a small trial. They should not be interpreted as a stable estimate of comparative safety, and the counts do not establish that either treatment caused or prevented an event. The registry summary does not provide enough detail to assess the timing, duration, or clinical course of individual events.

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

The primary result is an estimate of the difference between canakinumab and placebo in change in a composite cognition score from baseline to day 171 among participants included in the stated analysis population. The registry’s mixed-model estimate is 0.069, with a 90% confidence interval from -0.178 to 0.316. This interval includes zero and permits effects in either direction. The estimate is uncertain, not a precise measure of a treatment effect.

The reported result does not establish that the groups are identical. A result that does not clearly separate the groups is different from evidence of equivalence, which would require a specified margin and an analysis designed to assess whether the difference lies within that margin. The registry record does not report an equivalence margin. Nor does the result establish that a clinically important difference is impossible; the registry does not give a threshold for such a difference.

The arm-level estimates and endpoint definition describe how the composite scores changed on average. Positive change corresponds to improvement under the registry’s score definition, and negative change corresponds to decline. However, the treatment question concerns the contrast between randomized groups, not whether an individual arm’s mean is positive. The registry result should not be translated into medical advice or a claim that a treatment works or does not work.

Interpretation is also limited by the early termination and the small analysis groups. The overall enrollment was 34, while 26 participants contributed to the primary outcome analysis. The reason for termination beyond the statement that only one agent could be studied is not reported in the registry record. The available results can describe the observed comparison, but they do not settle what a larger or differently designed trial might find.

The secondary outcomes broaden the range of cognitive and other measures considered, but their unadjusted nature and the number of listed outcomes make them exploratory. They do not overturn the uncertainty in the primary analysis. In statistical terms, the evidence in this record is best characterized as an imprecise estimate from a small, early-terminated study, rather than a definitive conclusion about comparative effects.

18. Why This Trial Matters Statistically

This trial is statistically informative because it illustrates the difference between a point estimate and the uncertainty around it. The primary estimate is 0.069, but the registry’s 90% confidence interval extends from -0.178 to 0.316. Reading only the point estimate would hide how much uncertainty remains. The interval provides essential context: the study does not precisely distinguish among a range of plausible group differences.

The trial also shows why outcome population counts matter. Although 34 participants were enrolled, the primary outcome analysis included 10 assigned to canakinumab and 16 assigned to placebo. The analysis population is described using availability of PD data and relevant protocol deviations. These details affect what population the estimate describes and how broadly it can be generalized.

Early termination and limited enrollment matter for interpretation, especially in a study with multiple outcomes. A small number of participants provides limited information, while the registry lists a range of secondary outcomes. In that setting, secondary findings need careful interpretation, particularly when adjustment for multiple testing is not reported. The record’s results are useful for understanding the observed data and the uncertainty, but they do not resolve questions that would require a larger and fully documented analysis.

Finally, comparing the sponsor’s mixed-model result with a summary-statistic re-calculation highlights that analytical methods can differ. The simple re-calculation is a check, not a replacement for the trial’s reported model. Distinguishing those roles helps readers avoid treating an approximate calculation as if it were the prespecified or complete analysis.

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 most direct limitation is the small number of participants contributing to the primary outcome analysis: 10 in the canakinumab group and 16 in the placebo group. The overall enrollment was 34, but the primary analysis was limited to participants in the described PD analysis set with an available value. The registry record does not report why the remaining enrolled participants were not in that analysis, or how missing outcome values were handled.

The trial terminated early, and its brief summary states that only a single agent could be studied. The reason for termination, the timing and decision process, and any implications for the analysis are not reported in the registry record. The study’s original sample-size justification and the assumptions used for its original design are also not reported. Consequently, the planning illustrations based on the observed effect should not be mistaken for a description of the trial’s original design.

The registry identifies a mixed-model method but does not report the model specification, covariates, covariance structure, or other details needed for independent reproduction. The record also does not report how missing data were handled or whether sensitivity analyses were conducted. Further, although several secondary outcomes are listed, detailed results are available for only a subset in the posted results; numerical results for the other listed measures are not reported in the registry record.

Multiplicity procedures are not reported, limiting the weight that can be placed on exploratory secondary comparisons. The registry does not provide a clinically important difference threshold for the primary Z-score, so statistical uncertainty cannot be translated into a threshold-based judgment about practical importance. Finally, the posted adverse-event counts are descriptive and small; they do not establish causation or provide a precise comparative safety estimate.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Study of the Efficacy and Safety of Various Anti-inflammatory Agents in Participants With Mild Cognitive Impairment or Mild Alzheimer's Disease — NCT04795466.
• Registry record last updated 2025-10-16; status last verified 2025-10.
• Results first posted 2025-05-06.
• 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

NCT04795466 is a randomized, quadruple-masked, parallel, placebo-controlled phase 2 platform study sponsored by Novartis Pharmaceuticals. It enrolled 34 participants and is listed as terminated. The primary comparison was canakinumab versus matching placebo in participants with mild cognitive impairment due to Alzheimer’s disease or mild Alzheimer’s disease.

The primary endpoint was change from baseline to day 171 in NTB Z-score. The registry’s mixed-model analysis reported a least-squares mean difference of 0.069, with a 90% confidence interval from -0.178 to 0.316. The primary outcome analysis included 10 canakinumab participants and 16 placebo participants. The estimate is imprecise, and the interval includes zero.

The registry reports secondary cognitive and other outcomes, along with adverse-event counts. Early termination, the small analysis population, and unreported details about missing-data handling, multiplicity, and the model specification limit how much can be concluded from the posted results.