Study Evaluating The Efficacy And Safety Of Bapineuzumab In Alzheimer Disease Patients: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study Evaluating The Efficacy And Safety Of Bapineuzumab In Alzheimer Disease Patients (NCT00667810), a randomized phase 3 study. The primary outcome is The Change From Baseline in the Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog)/11 Total Score at Week 78.

2. Clinical Question

The primary endpoint is a difference in change on a cognitive scale, not a direct measure of every aspect of Alzheimer disease or daily function. The registry lists disability and several biological measures among secondary outcomes, but those address other questions and do not replace the prespecified primary endpoint.

3. Trial Design

This was a phase 3, randomized, multicenter, double-blind, placebo-controlled, parallel-group efficacy and safety trial of bapineuzumab in participants with mild to moderate Alzheimer disease who were apolipoprotein E 4 non-carriers. The listed interventions were bapineuzumab and placebo, with primary efficacy arms for bapineuzumab 0.5 mg/kg, bapineuzumab 1.0 mg/kg, and placebo. The registry describes quadruple masking. The sponsor was Pfizer.

The registry records enrollment of 901 and lists the study status as terminated. The recorded start was 2008-06 and primary completion was 2012-10. The brief description says each participant’s participation would last approximately 1.5 years. The reason for termination, the timing of termination, and the consequences for enrollment or follow-up are not reported in the registry record.

The primary outcome was change from baseline in ADAS-Cog/11 total score at week 78. The stated analysis population was modified intent-to-treat, with eligibility based on receipt of at least one infusion or portion of an infusion and available baseline and post-baseline scores on both ADAS-Cog/11 and DAD. Thus, the primary analysis population is not described as simply every randomized participant.

The record lists 17 outcome measures and 28 analyses, and identifies 15 posted secondary outcomes. The protocol’s intended sample-size assumptions, detailed randomization procedure, full analysis plan, and any interim analysis procedures are not reported in the registry record. The listed enrollment should therefore not be treated as evidence about the original sample-size target or the reasons the study ended.

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: The study originally included bapineuzumab 2.0 mg/kg dose level, which was discontinued on 02 April 2009 based on input from independent safety monitoring committee. It was estimated at the time that about 10 participants received 2.0mg/kg. These participants are not included in the efficacy analyses.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The modified intent-to-treat (mITT) included all randomized participants who received at least one infusion or portion of an infusion of study drug and who had a baseline and at least one post-baseline assessment of the ADAS-Cog/11 total score and Disability Assessment for Dementia (DAD) total 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.

Population: Safety population included all randomized participants who received at least one infusion or portion of an infusion of study drug.

6. Primary Endpoint

The registry lists the primary outcome as The Change From Baseline in the Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog)/11 Total Score at Week 78, measured over 78 weeks.

The primary endpoint was change from baseline in the Alzheimer’s Disease Assessment Scale-Cognitive Subscale, 11-item version (ADAS-Cog/11), at week 78. It is a multi-item measure of cognitive function covering memory, language, and praxis. The listed items include word recall and recognition, naming, orientation, following commands, and comprehension. A trained and certified psychometric rater administered the scale, and the description says that the rater did not have access to information about participants’ adverse events.

The scale ranges from 0 to 70 points, with higher scores indicating greater impairment. A negative change from baseline therefore indicates a decrease in cognitive impairment on this measure. The endpoint is useful for comparing groups on a standardized, structured assessment of cognitive symptoms over the stated follow-up period. It does not, by itself, describe daily functioning, brain measures, or every clinically relevant feature of Alzheimer disease; the registry lists separate secondary outcomes for some of those domains.

The primary results are reported as least-squares means with standard errors. These are model-based group summaries, rather than raw individual score changes. The interpretation of a difference between groups depends on its uncertainty and on the analysis used. In particular, a group’s mean change alone does not establish that its change differs from placebo.

7. Results Posted in the Registry

• Bapineuzumab 0.5 mg/kg: mean 6.05 versus 7.88 on control; difference -1.83 (95% CI -3.70 to 0.04), standardized effect d = -0.16; z-test p = 0.056.
• Bapineuzumab 1.0 mg/kg: mean 8.07 versus 7.88 on control; difference 0.19 (95% CI -1.71 to 2.09), standardized effect d = 0.02; z-test p = 0.845.

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.

The Change From Baseline in Brain Amyloid Burden at Week 71.

• 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:
  • Bapineuzumab 1.0 mg/kg: mean 0.00 versus 0.02 on control; difference -0.02 (95% CI -0.15 to 0.11), standardized effect d = -0.13; z-test p = 0.755.
  • Pooled Bapineuzumab 0.5/1.0 mg/kg: mean -0.01 versus 0.02 on control; difference -0.03 (95% CI -0.14 to 0.08), standardized effect d = -0.19; z-test p = 0.596.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Bapineuzumab 1.0 mg/kg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Pooled Bapineuzumab 0.5/1.0 mg/kg and the control. The interval gives the range of effects that remain compatible with the data.

The Change From Baseline in Phospho-tau Levels in the Cerebrospinal Fluid (CSF) at Week 71.

• 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:
  • Bapineuzumab 0.5 mg/kg: mean -6.62 versus 0.70 on control; difference -7.32 (95% CI -17.00 to 2.36), standardized effect d = -0.42; z-test p = 0.138.
  • Bapineuzumab 1.0 mg/kg: mean -6.35 versus 0.70 on control; difference -7.05 (95% CI -16.47 to 2.37), standardized effect d = -0.40; z-test p = 0.142.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Bapineuzumab 0.5 mg/kg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Bapineuzumab 1.0 mg/kg and the control. The interval gives the range of effects that remain compatible with the data.

The Change From Baseline in Brain Volume at Week 71

• 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:
  • Bapineuzumab 0.5 mg/kg: mean 18.55 versus 17.54 on control; difference 1.01 (95% CI -1.53 to 3.55), standardized effect d = 0.09; z-test p = 0.436.
  • Bapineuzumab 1.0 mg/kg: mean 18.60 versus 17.54 on control; difference 1.06 (95% CI -1.53 to 3.65), standardized effect d = 0.10; z-test p = 0.422.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Bapineuzumab 0.5 mg/kg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Bapineuzumab 1.0 mg/kg and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes cover several domains, including brain amyloid burden, cerebrospinal fluid phospho-tau, brain volume, cognitive and disability trajectories, and time to specified deterioration. The posted results therefore address biological measures and functional or clinical measures alongside the primary cognitive score. Their overall pattern is not a single uniform signal: estimated treatment-placebo differences vary in direction and size, while the reported confidence intervals for the posted comparisons include zero.

The secondary outcomes do not overturn or replace the primary ADAS-Cog/11 result. They address different measures, time points, analysis populations, and in some cases different definitions of change or deterioration. Findings from a biological marker, for example, cannot by themselves establish a difference in cognition or daily function. Likewise, the primary cognitive comparison does not answer every biological or functional question listed in the registry.

The registry lists multiple outcome measures and analyses. When many outcomes, arms, or time points are examined, some results can appear notable by chance even when no underlying effect exists. Unadjusted secondary results therefore carry less inferential weight than a clearly specified primary result, particularly when the registry record does not report the multiplicity strategy, endpoint hierarchy, or which secondary analyses were intended to be confirmatory.

The secondary results are best read as a set of estimates with their own uncertainty, rather than as independent confirmations. Differences in sample availability across outcomes also mean their precision may differ. The record does not report enough detail to determine whether all planned secondary outcomes were analyzed, how missing data were handled for each one, or how the reported analyses relate to the original statistical plan. Conclusions should remain tied to the specific measure and analysis being considered.

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

• The Change From Baseline in Brain Amyloid Burden at Week 71.
• The Change From Baseline in Phospho-tau Levels in the Cerebrospinal Fluid (CSF) at Week 71.
• The Change From Baseline in Brain Volume at Week 71
• Divergence of Effect on the ADAS-Cog/11 Total Scores From Week 39 to Week 78
• Divergence of Effect on the DAD Total Scores From Week 39 to Week 78
• Time to Median Placebo Deterioration on ADAS-Cog/11 Total Score (European Union [EU] Analysis Plan)
• Time to First Clinically Meaningful Deterioration on ADAS-Cog/11 Total Score (United States [US] Analysis Plan)
• Time to Median Placebo Deterioration on DAD Total Score

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is continuous: each participant’s ADAS-Cog/11 score can take values across a scale, and the registry reports group means of change. The posted group summaries are least-squares means and standard errors. A least-squares mean is a model-based estimate of a group mean; it may account for features of the analysis model rather than being a simple arithmetic average of the observed values. The specific full model specification is not reported in the registry record.

A standard error describes the estimated uncertainty in a mean. It is not the same as the spread of individual participants’ scores. Smaller standard errors generally indicate more precise estimates, while larger ones indicate more uncertainty. The group standard errors of 0.71, 0.73, and 0.64 should therefore be read as uncertainty around the corresponding estimated mean changes, not as the range of individual changes.

A mean difference compares the estimated outcome between a treatment arm and placebo. For the primary endpoint, the recalculated comparisons use the placebo mean of 7.88 and the relevant bapineuzumab group mean. A negative difference means the estimated change was lower in the bapineuzumab arm; in the context of this scale, a lower change is directionally consistent with less worsening. That numerical direction alone does not establish a meaningful or reliable group difference.

A confidence interval expresses uncertainty in an estimated difference under the assumptions of the analysis. Intervals that include zero are compatible with no difference, as well as with nonzero differences within the interval. A p-value summarizes how unusual the data would be under a specified null hypothesis and analysis assumptions. It does not measure the probability that the null hypothesis is true, the size or clinical importance of an effect, or the probability that a result will replicate.

The registry reports mixed models analyses for the primary endpoint. Mixed models can use information across repeated observations and account for correlation among measurements from the same participant. Their exact specification, covariance structure, handling of missing observations, and adjustment variables are not reported in the registry record. These details matter because different assumptions and model specifications can yield different estimates or uncertainty.

The recalculated comparisons are unadjusted approximations derived from posted summary statistics, including standard errors. The pooled standard deviations used for standardized effects are marked as approximate. These calculations are useful as an independent, simplified reading of the posted group summaries, but they do not reproduce the sponsor’s model-based analysis. Since the record flags a discrepancy, the sponsor’s own analysis takes precedence when describing the trial’s reported statistical result.

Over-dispersion is a general concern when count data vary more than a simple statistical model expects. It can make standard errors too small if ignored. It is not the central issue for this continuous primary endpoint, but the broader principle applies: uncertainty estimates depend on whether the model captures the data’s variability and dependence. The registry record does not report a detailed diagnostic assessment for the primary model.

General explanation: multiple comparisons arise when analyses test several endpoints, treatment arms, or time points. Even if each test uses the same threshold, the chance of at least one apparently notable result can increase across a family of tests. Adjustment methods or a prespecified hierarchy can address this, but neither the multiplicity strategy nor the full endpoint hierarchy is reported in the registry record. Consequently, secondary results should be treated as supportive or exploratory unless the protocol’s analysis plan clarifies their confirmatory status.

General explanation: the analysis population affects which participants contribute data. The modified intent-to-treat definition includes randomized participants who received study drug and had specified measurements, so it excludes randomized participants who did not meet those criteria. The registry record does not report how many participants were excluded from the primary analysis for each reason, or how missing outcome data were handled in detail. Those omissions limit assessment of possible selection bias and robustness.

11. Confidence Intervals and Precision

The sponsor-reported primary comparison for bapineuzumab 0.5 mg/kg versus placebo is -1.83, with a 95% confidence interval from -3.71 to 0.05. The interval includes zero and extends from a potentially lower change in the bapineuzumab group to a small positive difference. The p-value is 0.057. Taken together, these figures do not establish a clear separation between the groups under the reported analysis, and they leave uncertainty about the magnitude and direction of the difference.

For bapineuzumab 1.0 mg/kg versus placebo, the sponsor-reported difference is 0.19, with a 95% confidence interval from -1.73 to 2.10 and p = 0.848. The interval includes zero and allows for differences in either direction. The point estimate is close to zero, but the interval makes clear that the data do not pinpoint the true difference precisely enough to treat the estimate as proof of no effect or equivalence.

A confidence interval is not a range containing a fixed proportion of individual participant outcomes. It represents uncertainty around a group comparison under the model and sampling assumptions. Its width reflects both the information in the data and the variability modeled by the analysis. An interval that includes zero is not by itself evidence that the groups are identical; it indicates that the analysis does not clearly distinguish the estimated difference from zero at the stated confidence level.

The recalculated intervals and p-values are close to, but not identical with, the sponsor’s results. That difference is expected to matter when interpreting precision: the sponsor’s mixed-model estimates and uncertainty are the trial’s reported analysis, while the recalculation is an unadjusted approximation. The sponsor’s analysis takes precedence. The record does not report a clinically important difference threshold for ADAS-Cog/11, so statistical uncertainty cannot be translated into a claim about clinical importance from these results alone.

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 observed effects, not the trial’s original design calculation. They assume a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The target effect is set to the stated fraction of the observed effect. The original trial’s sample-size assumptions and justification are not reported in the registry record.

For the primary comparisons, the entries are participants per arm, not total participants. The 80% power column gives the estimated per-arm sample size under those assumptions for an 80% chance of detecting the stated effect if that effect were the true one. The 90% power column gives the corresponding per-arm planning figure for a 90% chance. Higher target power requires more participants under the same assumptions.

The rows for 75% and 50% of the observed effect illustrate that smaller effects require more participants to detect, all else equal. This is especially visible when the observed comparison is small: a smaller target difference demands a substantially larger planned sample under the calculation assumptions. These figures are not a recommendation and do not show what the original investigators planned.

They are not observed power. Observed power is not a useful substitute for reading the estimates and confidence intervals: the planning calculations simply translate assumed effects and design conditions into illustrative sample sizes. Their assumptions may not match the actual trial’s design, analysis model, missing-data pattern, or allocation process. These are planning illustrations, not evidence that the trial did or did not have adequate power.

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 = -1.830; pooled standard deviation σ = 11.481.
2. n = 2 × (1.96 + 0.842)² × 11.481² / -1.830² = 617.9, rounded up to 618 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 <- -1.83                      # observed difference in means
sigma <- 11.4809289296546                     # 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(667810)
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 = 618 (80% power), 828 (90% power)
 75% of observed effect: n per arm = 1,099 (80% power), 1,471 (90% power)
 50% of observed effect: n per arm = 2,472 (80% power), 3,309 (90% power)
power.t.test: 618.8 
Simulated power at n = 618 per arm: 0.794 (20000 replicates)

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

14. Analyses Reported by the Sponsor

The registry reports mixed-model comparisons for the primary endpoint. For bapineuzumab 0.5 mg/kg versus placebo, the sponsor reports a mean difference of -1.83, a 95% confidence interval from -3.71 to 0.05, and p = 0.057. For 1.0 mg/kg versus placebo, the reported mean difference is 0.19, the 95% confidence interval is -1.73 to 2.10, and p = 0.848.

The unadjusted recalculations from posted summary statistics are similar in direction and broad interpretation, but they do not match every reported interval limit or p-value. The recalculation is not the sponsor’s mixed-model analysis and relies on approximate variability estimates. The record flags a discrepancy, so the sponsor’s own analysis takes precedence for the trial’s reported inference.

This distinction is important: recalculation can help readers understand how group summaries relate to a simple comparison, but it cannot replace a model that may account for repeated measurements, covariates, or other analysis details. The full mixed-model specification and its assumptions are not reported in the registry record. The recalculated values should therefore be treated as a check on scale and direction, not as a competing definitive analysis.

15. Multiplicity, Missing Data and Analysis Populations

The registry record does not report the statistical plan for controlling multiplicity across the primary and secondary outcomes, treatment comparisons, or time points. It lists 17 outcome measures and 28 analyses, and identifies 15 posted secondary outcomes. Without a reported testing hierarchy or adjustment strategy, secondary p-values should not be interpreted as though each arose from a single isolated test. Multiple testing can increase the chance of apparently strong findings arising by chance.

The primary endpoint has a clear stated time point, week 78, and the registry identifies its analysis population. However, the record does not report whether the primary outcome and analysis were specified before examining the data, whether any protocol amendments changed endpoint definitions, or how missing data were handled in the sponsor’s mixed model. These are not grounds to assume a particular problem, but they constrain independent assessment of the analysis.

The modified intent-to-treat population includes randomized participants who received at least one infusion or part of an infusion and had baseline and at least one post-baseline assessment of both ADAS-Cog/11 and DAD. This differs from an analysis of all randomized participants. The registry does not report the numbers meeting or failing each criterion, the reasons for missing assessments, or whether sensitivity analyses examined alternative assumptions about missing outcomes.

The registry also does not report detailed model specifications, covariate adjustments, covariance assumptions, or a full accounting of deviations from the intended analysis. Such information can affect estimates and standard errors, particularly in longitudinal studies with incomplete follow-up. The posted endpoint estimates remain interpretable as registry results, but the absence of these details limits assessment of robustness and multiplicity-adjusted evidence.

16. Safety Results

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

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: vasogenic cerebral oedema (18.2%), cerebral microhaemorrhage (9.1%), confusional state (9.1%), multiple injuries (0.7%), headache (0.7%).

The registry lists serious adverse event counts for the 0.5 mg/kg, 1.0 mg/kg, placebo, and 2.0 mg/kg groups. The recorded serious adverse event counts are 32 among 267 at risk for 0.5 mg/kg, 34 among 263 at risk for 1.0 mg/kg, 53 among 344 at risk for placebo, and 2 among 11 at risk for 2.0 mg/kg. These are counts as recorded, not calculated comparative risks. The 2.0 mg/kg group appears in the safety record but not among the posted primary efficacy arms.

The registry also lists serious-event terms, including vasogenic cerebral oedema at 18.2%, cerebral microhaemorrhage at 9.1%, confusional state at 9.1%, multiple injuries at 0.7%, headache at 0.7%, and fall at 0.6%. The record does not provide enough context in these entries to treat the listed percentages as comparative treatment effects. Death counts are not reported in the registry record.

Safety counts should be interpreted with attention to the denominators and to the distinction between serious events and other adverse events. They do not by themselves establish whether an event was caused by treatment or whether groups differed reliably. The registry record does not report a complete causal assessment or the detailed exposure-adjusted safety analysis.

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

The primary outcome estimates describe average change in a standardized cognitive measure among the specified modified intent-to-treat population. The sponsor-reported comparisons do not show a clearly separated result from placebo at week 78: the 0.5 mg/kg confidence interval includes zero, and the 1.0 mg/kg interval also includes zero. This means the analyses do not establish a difference at the stated confidence level; it does not prove that the groups were identical.

The estimates also do not establish equivalence, because the registry does not report an equivalence margin or a non-inferiority design for this endpoint. A non-significant or uncertain comparison is not evidence that any potentially important difference has been ruled out. The confidence intervals express uncertainty around each comparison, and the registry does not report a threshold for what change would be clinically important on this scale.

The endpoint is one measure of cognition at one stated time point. It cannot summarize all aspects of Alzheimer disease, daily functioning, biological change, or safety. Secondary measures may add context, but they have distinct meanings and carry additional uncertainty, especially because the record does not report a multiplicity plan. Each result must be interpreted according to its outcome, comparison, and analysis population.

The registry labels the study terminated. That status does not, on its own, explain why the trial ended or establish that termination was caused by efficacy, safety, or any other specific factor. The reason and related decision process are not reported in the registry record. Nor should the listed enrollment be taken as proof of the planned sample size, since the original design justification is not reported.

These results are a statistical description of the posted trial record, not medical advice. They do not support a broad claim about treatment for an individual or about every population with Alzheimer disease. The population was limited to people with mild to moderate disease who were apolipoprotein E 4 non-carriers, and interpretation should remain within that scope.

18. Why This Trial Matters Statistically

This trial illustrates why a group’s average change and a between-group comparison answer different questions. The placebo and bapineuzumab arms all had positive mean changes on the ADAS-Cog/11, while the treatment question depends on how the groups compare and how uncertain that difference is. Focusing only on the direction of one arm’s change would miss the comparator and the interval around the estimated difference.

It also highlights the importance of distinguishing a sponsor’s model-based analysis from a simplified recalculation. The posted summary statistics permit an approximate unadjusted comparison, but a mixed-model analysis may use more of the longitudinal data and depends on modeling choices. Small differences in estimates, confidence limits, and p-values can arise, and the registry flags a discrepancy. The sponsor-reported result is therefore the primary reference for the trial’s stated inference.

The endpoint’s scale gives direction to the estimates: higher ADAS-Cog/11 scores indicate greater impairment, and a negative change indicates a decrease in impairment. Statistical direction, however, is not equivalent to clinical importance. Without a reported threshold for a meaningful difference, the numerical estimates and their uncertainty should not be translated into a clinical judgment.

Finally, the record shows how incomplete reporting affects interpretation. The analysis population is defined, but detailed missing-data methods, multiplicity procedures, and the original sample-size assumptions are not reported. These gaps do not erase the posted results; they define how far the results can be independently evaluated and generalized.

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 identifies a modified intent-to-treat population rather than all randomized participants. It requires receipt of at least part of an infusion and baseline and post-baseline measurements on both ADAS-Cog/11 and DAD. The counts excluded under this definition and the reasons for exclusion are not reported in the registry record, limiting assessment of how the analysis population may differ from the randomized population.

The record does not report the detailed handling of missing observations, sensitivity analyses, or the full mixed-model specification. It also does not report the original sample-size justification, a multiplicity plan, or a clinically important difference threshold for the primary scale. Each omission limits a different part of interpretation: robustness to missingness, reproducibility of the model, confidence in multiple outcome findings, and the clinical meaning of the estimated difference.

The study status is listed as terminated, but the reason and its implications for the conduct or interpretation of the trial are not reported. Enrollment and the posted results should not be used to infer a stopping rule, an interim analysis, or the original target sample size.

The safety record includes a 2.0 mg/kg group with a small at-risk count, while the primary efficacy results concern the 0.5 mg/kg and 1.0 mg/kg arms and placebo. Safety figures across these groups are counts as posted and should not be treated as directly comparable measures of risk without fuller exposure and analysis details. The registry does not report death counts in the safety entries.

Finally, the primary result concerns a particular cognitive score at week 78 in non-carriers with mild to moderate Alzheimer disease. Its scope does not automatically extend to other disease stages, other genetic groups, other endpoints, or other time points.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Study Evaluating The Efficacy And Safety Of Bapineuzumab In Alzheimer Disease Patients — NCT00667810.
• Registry record last updated 2016-01-08; status last verified 2015-12.
• Results first posted 2016-01-08.
• 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

This phase 3 randomized, double-blind, placebo-controlled parallel-group trial evaluated bapineuzumab in people with mild to moderate Alzheimer disease who were apolipoprotein E 4 non-carriers. The registry records enrollment of 901, lists Pfizer as sponsor, and gives the study status as terminated.

The primary endpoint was change from baseline in ADAS-Cog/11 total score at week 78. Sponsor-reported mixed-model comparisons were -1.83 for bapineuzumab 0.5 mg/kg versus placebo and 0.19 for bapineuzumab 1.0 mg/kg versus placebo; both reported 95% confidence intervals include zero. The record also lists secondary biological, cognitive, functional, and safety outcomes.

The registry does not report the reason for termination, the original sample-size justification, a multiplicity strategy, or detailed missing-data methods. The sponsor’s reported analysis takes precedence over the unadjusted recalculations, which are approximate.