A Study of the Efficacy and Safety of MEDI7352 in Participants With Painful Osteoarthritis of the Knee: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of the Efficacy and Safety of MEDI7352 in Participants With Painful Osteoarthritis of the Knee (NCT04675034), a randomized phase 2 study. The primary outcome is Change From Baseline in Weekly Average of Daily Numerical Rating Scale (NRS) Pain Score to Week 12.

2. Clinical Question

3. Trial Design

This completed Phase IIb study was randomized, quadruple-masked, placebo-controlled, and used a parallel-group design. The registry record lists enrollment of 345 participants. Participants were assigned to one of four MEDI7352 dose levels or placebo. The treatment period lasted 12 weeks, with six subcutaneous doses scheduled once every 2 weeks. The registry record describes the study as a dose-response evaluation of safety and efficacy, with pharmacokinetics, pharmacodynamics, and immunogenicity also among the study aims.

The registry lists the start date as 2020-12-02 and primary completion as 2023-08-16, and records the study status as completed. The primary analysis population was the full analysis set, defined as all randomized participants analyzed according to the intent-to-treat principle, regardless of the treatment actually received. That population definition is important: assignment, rather than adherence or receipt of treatment, determines the group used for the primary endpoint analysis.

The posted primary outcome is change from baseline in weekly average daily NRS pain score through Week 12. The baseline period is Day -7 to Day -1, inclusive, and the reported timeframe extends through Week 12. The record also lists secondary outcomes involving WOMAC pain, physical function, overall score and stiffness, as well as the patient’s global assessment of osteoarthritis. It lists 30 outcome measures and four reported analyses for the primary endpoint.

Although the trial is described as a dose-response study, the registry record does not report the original sample-size justification, detailed randomization procedures, or the planned statistical analysis strategy beyond the information in the posted endpoint description and analyses. Those details should not be inferred from the trial’s design label or from the retrospective planning illustrations.

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 345 participants were randomized, of which 344 participants received at least one dose of study drug.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full analysis set (FAS) included all randomized participants analyzed according to the intent-to-treat principle whereby randomized study treatment will be analyzed regardless of the study treatment actually received.

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: Full analysis set included all randomized participants analyzed according to the Intent-to-treat principle whereby randomized study treatment will be analyzed regardless of the study treatment actually received.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Weekly Average of Daily Numerical Rating Scale (NRS) Pain Score to Week 12, measured over Baseline (Day -7 to Day -1, inclusive) through Week 12.

The primary endpoint measures change in the weekly average of daily pain ratings from baseline through Week 12. Participants rated average pain in the target knee on an 11-point numerical rating scale, where 0 means no pain and 10 means the most severe pain imaginable over the previous 24 hours. Ratings were entered daily, at approximately the same time each morning, through an electronic patient-reported outcome diary.

A change score compares a participant’s follow-up rating with their baseline rating. Negative changes therefore indicate lower pain ratings at follow-up than at baseline. The group mean summarizes the average change, while the standard deviation describes how much individual changes varied around that mean. This is a continuous outcome: it retains information across the rating scale rather than reducing responses to a binary category.

Daily recording can capture changes across the treatment period rather than relying on one isolated pain report. Averaging daily scores over a week also makes the endpoint less dependent on a single day’s rating. The registry description reports a two-step multiple imputation procedure to address missing post-baseline scores, because missing diary entries can otherwise affect the endpoint estimate.

The outcome is clinically relevant to the study question because pain was the target symptom in participants with painful knee osteoarthritis. The endpoint alone, however, does not describe all aspects of the condition or treatment experience. Function, global assessment, safety, and other measures were also listed as outcomes. The registry record does not report a threshold defining a clinically important difference on this scale.

7. Results Posted in the Registry

• MEDl7352 Dose Level 1: mean -2.19 versus -2.35 on control; difference 0.16 (95% CI -0.60 to 0.92), standardized effect d = 0.07; z-test p = 0.681.
• MEDl7352 Dose Level 2: mean -3.00 versus -2.35 on control; difference -0.65 (95% CI -1.44 to 0.14), standardized effect d = -0.28; z-test p = 0.105.
• MEDl7352 Dose Level 3: mean -2.83 versus -2.35 on control; difference -0.48 (95% CI -1.29 to 0.33), standardized effect d = -0.20; z-test p = 0.247.
• MEDl7352 Dose Level 4: mean -2.81 versus -2.35 on control; difference -0.46 (95% CI -1.33 to 0.41), standardized effect d = -0.18; z-test p = 0.301.

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 Western Ontario and McMaster Universities Arthritis Index (WOMAC) Pain Subscale to Week 12

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • MEDl7352 Dose Level 1: mean -1.88 versus -2.62 on control; difference 0.74 (95% CI -0.03 to 1.51), standardized effect d = 0.32; z-test p = 0.059.
  • MEDl7352 Dose Level 2: mean -2.88 versus -2.62 on control; difference -0.26 (95% CI -1.09 to 0.57), standardized effect d = -0.10; z-test p = 0.538.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between MEDl7352 Dose Level 1 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 MEDl7352 Dose Level 2 and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in WOMAC Physical Function Subscale to Week 12

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • MEDl7352 Dose Level 1: mean -1.48 versus -2.14 on control; difference 0.66 (95% CI -0.10 to 1.42), standardized effect d = 0.29; z-test p = 0.088.
  • MEDl7352 Dose Level 2: mean -2.56 versus -2.14 on control; difference -0.42 (95% CI -1.20 to 0.36), standardized effect d = -0.18; z-test p = 0.293.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between MEDl7352 Dose Level 1 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 MEDl7352 Dose Level 2 and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Patient's Global Assessment (PGA) of OA to Week 12

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • MEDl7352 Dose Level 1: mean -0.34 versus -0.57 on control; difference 0.23 (95% CI -0.09 to 0.55), standardized effect d = 0.24; z-test p = 0.160.
  • MEDl7352 Dose Level 2: mean -0.72 versus -0.57 on control; difference -0.15 (95% CI -0.51 to 0.21), standardized effect d = -0.14; z-test p = 0.408.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between MEDl7352 Dose Level 1 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 MEDl7352 Dose Level 2 and the control. The interval gives the range of effects that remain compatible with the data.

The posted secondary outcomes cover pain, physical function, patient global assessment, and other WOMAC domains over time or at Week 12. Across the outcomes with results displayed, the group means generally move in a negative direction from baseline. Comparisons with placebo are mixed: some dose groups have more negative mean changes than placebo, while others have less negative changes. This broad pattern does not provide a uniform dose-by-dose result across measures.

These secondary measures are related to the primary endpoint but are not interchangeable with it. The primary outcome is the weekly average daily pain rating through Week 12; WOMAC pain captures another patient-reported pain measure, while physical function and global assessment address different aspects of the condition. Similarity or difference across measures can offer context, but it cannot substitute for the primary analysis or resolve its uncertainty.

The registry lists many outcome measures and multiple reported primary comparisons. The secondary recalculations shown do not include sponsor-reported analyses, and their unadjusted p-values are exploratory descriptions rather than a unified confirmatory testing strategy. When many outcomes and comparisons are examined, isolated results can arise by chance. The registry record does not report a multiplicity adjustment or testing hierarchy for these secondary outcomes. For that reason, individual unadjusted secondary results carry less weight than a prespecified, multiplicity-controlled analysis would.

The secondary results should be read as supporting context, not as repeated opportunities to select the most favorable comparison. They neither establish a consistent dose-response pattern nor independently confirm the primary result. The registry record lists 29 posted secondary outcomes, but does not report adjusted inferential analyses for the secondary results.

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

• Change From Baseline in Western Ontario and McMaster Universities Arthritis Index (WOMAC) Pain Subscale to Week 12
• Change From Baseline in WOMAC Physical Function Subscale to Week 12
• Change From Baseline in Patient's Global Assessment (PGA) of OA to Week 12
• Change From Baseline in WOMAC Pain Subscale Over Time
• Change From Baseline in WOMAC PF Subscale Over Time
• Change From Baseline in WOMAC Overall Score Over Time
• Change From Baseline in WOMAC Stiffness Scores Over Time
• Change From Baseline in PGA of OA Over Time

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is continuous, so its group summaries are means and standard deviations. The mean is the average change in score for participants in a group. The standard deviation describes variation in participant-level changes around that group average; it is not the uncertainty around the mean itself. A confidence interval for a treatment comparison instead expresses uncertainty about the estimated difference between groups.

What does the sponsor’s ANCOVA estimate represent?

The registry reports ANCOVA analyses with a combined estimate for least-squares mean for each active dose versus placebo. ANCOVA is a regression-based method that can compare groups while accounting for specified covariates. A least-squares mean is a model-based group estimate, rather than simply the arithmetic mean of the posted outcome values. The registry record does not report the covariates, model specification, or further details of the analysis, so those should not be inferred.

Randomization supports comparisons between assigned groups, while adjustment can improve precision or account for variables included in the model. The sponsor-reported analysis is not identical to a calculation based only on each group’s displayed mean and standard deviation. The latter ignores model adjustment and any model-based handling of incomplete observations. In this trial, the registry’s sponsor-reported estimates and the simple unadjusted recalculations differ. Because the discrepancy flag is true, the sponsor’s own analysis takes precedence in describing the primary result.

How should a confidence interval be understood?

A confidence interval gives a range of estimates compatible with the data and the statistical model at the stated confidence level. A narrower interval indicates greater precision; a wider one indicates less precision. The reported 95% confidence intervals describe uncertainty around the sponsor’s estimated contrasts. They are not ranges containing 95% of individual participants’ outcomes, and they do not assign a probability that a particular value is the true effect.

Whether an interval crosses zero is often used to describe whether the data are compatible with no difference under a conventional two-sided test. Here, the reported upper limits for the four sponsor contrasts are positive, while the reported p-values differ in their relation to the usual 0.05 threshold. For two comparisons, the reported p-value is below 0.05 even though the confidence interval extends above zero. The registry record does not explain this mismatch. The interval and p-value should therefore both be quoted, without silently treating them as fully concordant.

Why does missing-data handling matter?

The endpoint description says that a two-step multiple imputation procedure was used to address missing post-baseline scores. Imputation fills in missing values using a statistical procedure that reflects uncertainty about those values. This can preserve randomized participants in an analysis when outcome records are incomplete, but conclusions depend on the assumptions and implementation of the imputation model. The registry record does not report the detailed imputation model, its assumptions, or sensitivity analyses. The full analysis set is defined by randomized assignment, but that definition by itself does not describe how every missing observation was handled.

What do the recalculations add?

The unadjusted recalculations use posted group summaries to express a simple mean difference and its uncertainty. They are useful as a transparent descriptive comparison, but they are not the sponsor’s ANCOVA. They do not incorporate unreported covariate adjustment or necessarily match the sponsor’s population and missing-data procedures. In a trial with multiple arms and model-based reported results, the two approaches can produce different estimates. The recalculations should be treated as a separate, simple analysis, not as a correction to the sponsor’s analysis.

The study has four primary comparisons, one for each dose versus placebo, and it also reports multiple secondary outcomes. When many hypotheses are tested, the chance of at least one apparently notable result can increase. A multiplicity adjustment or testing hierarchy can address that issue, but the registry record does not report whether either was planned or applied. The posted p-values should therefore be understood as reported individual comparisons, not as proof that a family-wise error rate was controlled.

The registry record also does not report an interim analysis plan, the original sample-size justification, the model’s complete covariate set, or a detailed testing hierarchy. The absence of these details limits how far the statistical design can be reconstructed from the posted record. General statistical concepts explain what the methods do; they cannot fill in protocol decisions that are not reported.

11. Confidence Intervals and Precision

The sponsor-reported estimates have 95% confidence intervals that span both negative and positive values. For Dose Level 1, the interval runs from -0.968 to 0.670. For Dose Level 2, it runs from -1.619 to 0.027; for Dose Level 3, from -1.618 to 0.009; and for Dose Level 4, from -1.423 to 0.245. The intervals are broad relative to their point estimates, so the exact size and direction of the underlying contrast remain uncertain.

These intervals include values on either side of zero. In practical terms, the posted data do not pin down a single precise contrast for any dose. The point estimates can be read as the model’s central estimates, but the limits show that materially different values remain compatible with the reported analysis. The registry record does not report a threshold for clinical importance, so the intervals cannot be used to determine whether a particular effect is clinically important.

There is also a reporting tension: the sponsor-reported p-values for Dose Levels 2 and 3 are 0.029 and 0.026, while the corresponding 95% intervals extend slightly above zero. Under the usual matching of a two-sided 0.05 test and a 95% interval, those summaries would ordinarily be expected to align. The registry record does not explain the discrepancy, so both values should be retained as posted and interpreted cautiously.

Precision is distinct from whether a result crosses a conventional significance threshold. A narrow interval can still include values that matter in different ways, while a wide interval can leave substantial uncertainty even when its point estimate appears notable. Here, the confidence intervals communicate that the comparisons are not estimated with enough precision to support fine distinctions among dose levels. They should be considered alongside the difference between the sponsor’s ANCOVA and the separate unadjusted recalculations.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size illustrations translate an observed contrast into hypothetical planning numbers. For these continuous outcomes, the figures are participants per arm, not total events. They are not the number enrolled in the trial and do not represent the trial’s original design calculation. The original sample-size justification and its assumptions are not reported in the registry record.

The illustrations assume a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. They use the stated fraction of the observed effect: 100%, 75%, or 50%. For each assumed effect, the 80% power column gives the number of participants per arm that would yield an 80% chance of detecting that effect under those assumptions if the same planning model applied. The 90% power column gives the corresponding number for a 90% chance.

Smaller effects require more participants to distinguish them from random variation with the same target power. Thus, the planning numbers increase as the assumed effect becomes a smaller fraction of the observed one. The values are especially sensitive to the observed effect used as the starting point: a small observed contrast can produce very large illustrative sample sizes, while an unusually large observed contrast can make a planning estimate look more modest than a future study may require.

These are planning illustrations based on the observed effect, not observed power. They do not establish the probability that this completed study detected an effect, and they should not be read as evidence that any comparison was adequately or inadequately powered. They also do not reproduce the trial’s own design assumptions, which are not reported.

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 = 0.160; pooled standard deviation σ = 2.304.
2. n = 2 × (1.96 + 0.842)² × 2.304² / 0.160² = 3,255.9, rounded up to 3,256 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 <- 0.16                      # observed difference in means
sigma <- 2.30429436053643                     # 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(4675034)
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 = 3,256 (80% power), 4,359 (90% power)
 75% of observed effect: n per arm = 5,789 (80% power), 7,749 (90% power)
 50% of observed effect: n per arm = 13,024 (80% power), 17,436 (90% power)
power.t.test: 3256.9 
Simulated power at n = 3,256 per arm: 0.799 (20000 replicates)

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

14. Analyses Reported by the Sponsor

The sponsor’s own primary analyses are the four ANCOVA comparisons of each MEDI7352 dose level with placebo. They report least-squares mean estimates, 95% confidence intervals, and p-values. Those analyses take precedence over the simple recalculations from posted group means and standard deviations because the two approaches do not agree. The recalculations are unadjusted summaries and cannot replace the sponsor’s model-based analysis.

The posted ANCOVA estimates are -0.15 for Dose Level 1, -0.80 for Dose Level 2, -0.80 for Dose Level 3, and -0.59 for Dose Level 4. Their reported p-values are 0.360, 0.029, 0.026, and 0.083, respectively. The 95% intervals for all four comparisons extend across zero; for Dose Levels 2 and 3, the reported intervals and p-values do not align under the usual two-sided 0.05 interpretation. The registry record does not explain this internal discrepancy.

Accordingly, the sponsor results should be described as the primary reported analysis, while the recalculations remain a distinct descriptive check. Neither should be blended into a newly derived estimate. The registry record does not provide sufficient model details to identify why the two sets of results differ.

15. Multiplicity, Missing Data and Analysis Populations

The registry reports four primary comparisons: each of the four MEDI7352 dose levels versus placebo. It also lists multiple secondary outcomes, covering several symptom and function measures and assessments over time. Testing many comparisons can increase the chance of observing at least one apparently unusual result even when no underlying difference exists. Statistical plans commonly address this through procedures such as a prespecified hierarchy or adjustment of significance levels, but whether those methods were planned or used is not reported in the registry record.

Because the multiplicity strategy is not reported, the individual sponsor p-values should not be treated as though they necessarily control the probability of a false-positive conclusion across all dose comparisons and outcomes. The registry’s primary results include p-values and 95% confidence intervals for each dose-versus-placebo comparison, but no reported adjustment should be assumed. The mismatch between the p-values and confidence intervals for two comparisons adds a separate interpretive concern; the record does not explain it.

The primary endpoint description reports that two-step multiple imputation was used to address missing post-baseline scores. This is relevant because the analysis depends on both observed data and assumptions about missing outcomes. The registry record does not report the imputation model in detail, the assumptions governing missingness, or sensitivity analyses showing how conclusions change under alternative assumptions. Those details are not available for evaluating robustness.

The full analysis set includes all randomized participants and analyzes them according to the intent-to-treat principle, regardless of treatment actually received. This definition supports an analysis based on randomized assignment. It does not, by itself, explain how missing scores were imputed or establish that all randomized participants contributed observed Week 12 measurements.

The registry record also does not report interim analyses, the original sample-size rationale, a detailed statistical analysis plan, or the complete ANCOVA specification. These are distinct omissions: missing information about multiplicity affects how multiple tests should be read, while missing analysis details limit reconstruction of the primary model. Neither gap should be filled by assumption.

16. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: liver function test abnormal (1.5%), anaemia (1.5%), breast cancer (1.5%), gastrointestinal carcinoma (1.5%), nephrolithiasis (1.5%).

The registry safety summaries report serious adverse events and other adverse events by treatment group, with separate denominators for participants at risk. Serious adverse events were reported for 3 participants in Dose Level 1, 5 in Dose Level 2, 3 in Dose Level 3, 1 in Dose Level 4, and 2 in the placebo group. The corresponding at-risk counts were 70, 68, 69, 68, and 69. The record reports zero deaths in every group.

These are descriptive counts. They show how many participants in each group were recorded as affected, but do not establish that a particular event was caused by treatment. The registry also lists serious adverse-event terms, including liver function test abnormal, anaemia, breast cancer, gastrointestinal carcinoma, nephrolithiasis, and Henoch-schonlein purpura, each reported as 1.5%. The posted term frequencies should be interpreted in the context of the study’s group sizes and reporting period.

The number with other adverse events was 51, 45, 54, 51, and 40 across the four dose groups and placebo, respectively. The safety denominators differ slightly from some of the primary endpoint group counts. These summaries do not provide event timing, duration, severity distributions, or causal attribution for each event. The registry record should therefore be read as a report of observed safety counts, not as a complete assessment of comparative risk.

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

The results describe Week 12 changes in pain scores among randomized participants with painful knee osteoarthritis in this Phase IIb study. The sponsor-reported ANCOVA estimates vary across the four MEDI7352 dose comparisons with placebo. The point estimates for Dose Levels 2 and 3 are more negative than those for the other comparisons, but their confidence intervals extend across zero. The p-values and intervals for those comparisons are also not fully concordant in the registry record.

These results do not establish a precise difference for every dose, nor do they establish a dose-response trend. The registry record does not report a formal trend test or a prespecified threshold for clinical importance on the NRS scale. Statistical evidence and clinical importance are related but separate questions: an estimate can be statistically distinguishable from zero without necessarily meeting a meaningful threshold, and the record does not provide such a threshold for this endpoint.

The findings apply to the study’s enrolled population, treatment period, and tested dose groups. They do not by themselves establish what would occur over a longer period, in different populations, or under different treatment conditions. The secondary outcomes provide additional descriptive context but are numerous and lack reported multiplicity adjustment. The adverse-event summaries are counts and do not determine causal relationships.

The simple unadjusted recalculations differ from the sponsor’s ANCOVA results and should not be used to overrule them. At the same time, the sponsor-reported interval and p-value discrepancy should remain visible in interpretation rather than being smoothed over. A careful reading keeps the estimate, its uncertainty, analytic approach, and reporting limitations together.

Nothing in the posted results alone is a medical recommendation. The registry data are a statistical record of a clinical trial, not individualized guidance. Their most defensible use is to describe the measured outcomes and the uncertainty around comparisons, while distinguishing reported analyses from exploratory summaries and avoiding conclusions that require unreported design or analysis details.

18. Why This Trial Matters Statistically

This trial illustrates several central issues in clinical biostatistics. First, a randomized, masked, parallel-group design supports comparison of assigned treatments, but the interpretation still depends on which endpoint is analyzed, how missing observations are handled, and what model is used. The primary outcome is a repeated patient-reported pain measure summarized as change through Week 12, with an imputation procedure reported for missing post-baseline scores.

Second, the study has multiple active dose groups, each compared with placebo, along with many secondary outcomes. That creates an important multiplicity question. Without a reported testing hierarchy or adjustment strategy, individual p-values do not describe the error rate across all comparisons. This matters especially when readers look across several outcomes for a favorable result.

Third, the posted sponsor analysis and the simple recalculations from group summaries differ. The distinction highlights why unadjusted calculations can be useful for transparent description but cannot stand in for a model-based analysis. The registry’s ANCOVA estimates take precedence, while the absence of detailed model information limits independent reconstruction.

Finally, precision and consistency deserve attention alongside point estimates. Confidence intervals show substantial uncertainty, and the reported p-values and intervals for two dose comparisons are not fully aligned. Recognizing these features is more informative than reducing a complex dose-response trial to a single threshold-based label. The record provides a useful example of why statistical interpretation must consider design, estimand, missing data, multiplicity, and reporting coherence together.

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.

Important protocol and analysis details are not reported in the registry record. These include the original sample-size justification, the detailed ANCOVA specification and covariates, the full imputation model and assumptions, and sensitivity analyses for missing data. The record also does not report a formal dose-response trend analysis or the threshold used to define a clinically important difference in pain score.

The reported primary p-values and confidence intervals are not fully concordant for Dose Levels 2 and 3. The registry record does not explain why. The simple recalculations from posted summary statistics also differ from the sponsor’s ANCOVA estimates. Because the model details are unavailable, the cause of that difference cannot be established from the record.

The registry record does not report whether multiplicity was addressed across the four primary comparisons or the many secondary outcomes. Secondary results are therefore best treated as context rather than independent confirmatory evidence. Their breadth also makes selective emphasis on individual comparisons a risk.

Safety summaries are presented as counts and percentages, with no causal attribution in the listed figures. The registry record does not provide full event-level detail such as timing, duration, or severity distributions in the information used for these summaries. Finally, the study’s Week 12 endpoint and treatment period provide a bounded view of outcomes; these results do not describe longer-term effects. These limitations concern what can be inferred from the registry record, not necessarily what information may exist elsewhere.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study of the Efficacy and Safety of MEDI7352 in Participants With Painful Osteoarthritis of the Knee — NCT04675034.
• Registry record last updated 2025-02-10; status last verified 2025-02.
• Results first posted 2025-02-10.
• 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

NCT04675034 is a completed Phase IIb randomized, quadruple-masked, parallel-group, placebo-controlled dose-response study of MEDI7352 in participants with painful osteoarthritis of the knee. The registry lists enrollment of 345 and a 12-week treatment period with six subcutaneous injections once every 2 weeks.

The primary endpoint was change from baseline in weekly average daily NRS pain score through Week 12. The sponsor reported ANCOVA comparisons for four MEDI7352 dose levels versus placebo. Their estimates have 95% confidence intervals extending across zero, and two comparisons have p-values that do not align with the corresponding interval limits under the usual interpretation. The registry record does not explain that discrepancy.

Simple unadjusted recalculations differ from the sponsor analysis, which takes precedence as the reported primary analysis. The record also lists multiple secondary outcomes and descriptive safety counts. Multiplicity strategy, original sample-size justification, and several analysis details are not reported.