A Study of Mavrilimumab Versus Anti Tumor Necrosis Factor in Subjects With Rheumatoid Arthritis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Mavrilimumab Versus Anti Tumor Necrosis Factor in Subjects With Rheumatoid Arthritis (NCT01715896), a randomized phase 2 study. The primary outcome is Percentage of Participants Who Achieved American College of Rheumatology 20 (ACR20) Responses at Day 169.

2. Clinical Question

The central contrast is between two biological treatments given with stable-dose methotrexate in a population with active rheumatoid arthritis and prior inadequate response to specified background or anti-TNF treatments. The endpoint is a composite response: it requires improvement in both swollen and tender joint counts and improvement in at least three of five other ACR core measures. The resulting comparison describes the proportion meeting that response definition at a specified assessment time. It does not, by itself, describe every aspect of disease activity, response duration, safety, or benefit for an individual participant.

The study is described as exploratory. Its primary objective was to explore efficacy of mavrilimumab compared with golimumab in the stated population. The registry record does not report a non-inferiority or equivalence objective, so a result that does not establish a clear difference should not be read as proof that the treatments are equivalent. Interpretation needs to account for the estimated difference, its interval, the sponsor-reported analysis, and the limited size of the analyzed groups.

3. Trial Design

This was a Phase 2 exploratory, randomized, quadruple-masked, parallel-group study. The registry record lists 215 participants enrolled and a completed status. The recorded study start is 2013-03, and the primary completion is 2014-11. The registry record does not report the specific randomization procedure, allocation ratio, or the roles covered by quadruple masking.

One group received mavrilimumab 100 mg subcutaneously every 2 weeks for 24 weeks with stable-dose methotrexate, given orally or parenterally. The comparator group received golimumab 50 mg and placebo matched to mavrilimumab on alternating schedules, also by subcutaneous injection every 2 weeks for 24 weeks, with stable-dose methotrexate. The recorded methotrexate range was 7.5 to 25 mg per week. This schedule means that the active comparator and matching placebo were administered at alternating visits; it should not be interpreted as both active treatments being given together.

The primary endpoint was assessed at Day 169, and its analysis population was the mITT population. For that endpoint, the registry reports 68 participants in the golimumab group and 70 in the mavrilimumab group. These are the endpoint analysis counts, not a description of how all 215 enrolled participants were distributed or followed. The registry record does not report a full account of participant flow, reasons for exclusion from the primary analysis, adherence, or treatment discontinuation.

The entry lists 30 outcome measures and 23 analyses. It also identifies secondary outcomes involving treatment-emergent adverse events, laboratory and vital-sign abnormalities, pulmonary function, dyspnea, disease activity, and ACR response measures. The existence of multiple outcomes is relevant when interpreting isolated nominal statistical findings. The registry record does not report the original sample-size justification, planned interim analyses, or the full statistical analysis plan.

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: Overall, 215 participants were screened, of which 77 participants were considered as screen failures and 138 participants were randomized and completed in the study.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The mITT population analysis set included all participants in the treatment group corresponding to their randomized treatment group.

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: The modified intent to treat (mITT) population analysis set included all participants in the treatment group corresponding to their randomized treatment group.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants Who Achieved American College of Rheumatology 20 (ACR20) Responses at Day 169, measured over Day 169.

The primary endpoint is the percentage of participants who achieved an ACR20 response at Day 169. ACR20 is a composite response criterion: it requires at least 20% improvement in both swollen joint count and tender joint count, together with at least 20% improvement in at least three of five additional core measures. Those measures are participant assessment of pain, participant global assessment of disease activity, physician global assessment of disease activity, self-assessed disability using the Health Assessment Questionnaire disability index, and C-reactive protein. If CRP was missing and erythrocyte sedimentation rate was available, ESR was to be used.

As a general statistical concept, a composite endpoint combines several measurements into one response definition. This can provide a structured way to summarize multiple dimensions of a condition, while making the result dependent on the particular components and thresholds chosen. A participant either meets the endpoint definition or does not; the percentage therefore represents the fraction of the analyzed group meeting all required parts of the criterion.

ACR20 is used as an efficacy outcome because it captures a defined level of improvement across joint counts and other disease-related measures. The registry record’s primary objective was to explore comparative efficacy. The endpoint does not convey the size of improvement for every component, nor does a group-level response percentage guarantee a similar experience for each participant. The record does not report additional details about the endpoint’s clinical interpretation or a threshold for a clinically important between-group difference.

7. Results Posted in the Registry

• Mavrilimumab 100 mg: 61.4% (43/70) versus 66.2% (45/68) on control; risk difference -4.7 percentage points (95% CI -20.8 to 11.3), risk ratio 0.93 (0.72–1.19); unadjusted two-proportion z-test p = 0.562.

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.

Dyspnea Score at Day 169

• 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:
  • Mavrilimumab 100 mg: mean 0.42 versus 0.34 on control; difference 0.08 (95% CI -0.19 to 0.35), standardized effect d = 0.10; z-test p = 0.559.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Mavrilimumab 100 mg and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Continuous American College of Rheumatology (ACRn) Score at Day 169

• 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:
  • Mavrilimumab 100 mg: mean 33.06 versus 40.49 on control; difference -7.43 (95% CI -22.13 to 7.27), standardized effect d = -0.17; z-test p = 0.322.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Mavrilimumab 100 mg and the control. The interval gives the range of effects that remain compatible with the data.

Duration of DAS28 (CRP) Remission at Day 169

• 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:
  • Mavrilimumab 100 mg: mean 69.60 versus 105.00 on control; difference -35.40 (95% CI -35.95 to -34.85), standardized effect d = -38.89; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Mavrilimumab 100 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes cover several kinds of information: safety events and clinical laboratory or vital-sign changes, pulmonary function, dyspnea, disease activity, and other ACR response summaries. This breadth can help describe different aspects of the study, but it also means that results do not all answer the same question or use the same scale. Some are participant counts, some are continuous measurements, and some concern duration or response criteria.

The posted secondary results do not form a single, uniformly clear pattern that can simply be combined into an overall efficacy claim. Some estimates are imprecise, and model-based and unadjusted comparisons may differ. The primary ACR20 result remains the principal efficacy comparison specified in the record. Secondary outcomes can add context, but they cannot turn an inconclusive primary comparison into proof of superiority, equivalence, or non-inferiority.

Several secondary analyses have nominal p-values that may attract attention, including the sponsor-reported time-to-event model result. Such findings are more tentative when many outcomes are examined and no multiplicity strategy is reported. Without a reported testing hierarchy or adjustment, an isolated p-value does not have the same confirmatory standing as a result protected by a prespecified multiplicity plan. The registry record does not report such a plan.

Secondary outcomes may also have smaller analysis groups than the primary outcome, depending on the measurement or event definition. That can affect precision and the representativeness of the analysis population. They should be interpreted with attention to their own denominators, measurement definitions, time points, and analysis methods, rather than as interchangeable evidence. Their main role is to provide exploratory context and identify questions that require confirmation, not to establish a general treatment effect from a collection of separate tests.

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

• Number of Participants With Treatment-Emergent Adverse Events (TEAEs) and Treatment-Emergent Serious Adverse Events (TESAEs)
• Number of Participants With Abnormal Clinical Laboratory Parameters Reported as Treatment-Emergent Adverse Events (TEAEs)
• Number of Participants With Abnormal Vital Signs Reported as Treatment-Emergent Adverse Events (TEAEs)
• Number of Participants With Pulmonary Function Test Values Below Threshold Values Based on Percent Change From Baseline at Day 85 and 169
• Dyspnea Score at Day 169
• Change From Baseline in Disease Activity Score of 28 Joints Using C-Reactive Protein (DAS28 [CRP]) Score at Day 169
• Change From Baseline in Continuous American College of Rheumatology (ACRn) Score at Day 169
• American College of Rheumatology (ACR) Hybrid Score at Day 169

9. Statistical Interpretation

10. Statistical Methods Explained

The primary result is a binary outcome: each participant either met the ACR20 response definition at Day 169 or did not. A group percentage expresses how common that response was among participants included in that endpoint analysis. Comparing the percentages gives a direct summary of the difference between groups, but the estimate is subject to sampling variation because the observed participants are only one possible set of participants from the study population.

What does a risk difference describe?

A risk difference is the response proportion in one group minus the response proportion in the comparison group. It stays on the probability scale, so a value of zero represents no difference in the response proportions. A negative estimate for the mavrilimumab-versus-golimumab contrast indicates a lower observed response proportion in the mavrilimumab group. The recalculated estimate is an unadjusted comparison based on the posted group summaries; it does not account for covariates or model adjustments that might be used in a separate analysis.

Why are there two primary comparisons?

The registry record includes a sponsor-reported logit response model as well as the simple unadjusted re-calculation. A logit model analyzes the log-odds of a binary response and can yield a model-based contrast. The sponsor reports its result as a percent difference, with a 90% interval and a p-value. The re-calculation instead reports a risk difference and a separate interval and p-value. Because the estimands and methods are not identical, the numbers need not match. The registry record does not report further model specification details such as covariates, handling of model assumptions, or all analysis rules.

What does a confidence interval communicate?

In general, a confidence interval describes uncertainty around an estimate under the method used to construct it. A narrower interval indicates greater precision; a wider interval indicates that a broader range of values remains compatible with the data and method. An interval that spans the no-difference value does not demonstrate equivalence. It means the data do not isolate a difference in one direction at the interval’s stated confidence level. The sponsor’s reported primary interval is explicitly 90%; the confidence level for the simple recalculated interval is not reported in the registry record.

How should a p-value be read?

In general, a p-value measures how surprising data at least as incompatible with a specified null hypothesis would be if that hypothesis and the analysis assumptions held. It is not the probability that the null hypothesis is true, nor does it measure the size or importance of a difference. A larger p-value does not prove there is no effect. For this endpoint, the sponsor-reported p-value and the unadjusted recalculated p-value belong to their respective analyses and should not be mixed as if they arose from one model.

Why does over-dispersion matter?

Over-dispersion is a general issue in count or proportion modeling when observed variability is greater than a model’s assumptions allow. If unaccounted for, uncertainty estimates can be too small and statistical evidence can appear stronger than it is. The primary outcome is a binary response summarized as a percentage, but the registry record does not report whether over-dispersion was assessed or how it would have been handled. There is no basis in the registry record to claim that it occurred in this analysis.

What is the role of the analysis population?

The primary analysis is described as using the mITT population. The registry definition says it included all participants in the treatment group corresponding to their randomized treatment group. This indicates that the primary assignment was linked to randomized group, but the registry record does not provide a fuller operational definition or explain how missing Day 169 assessments were handled. Missing-data methods can matter when the outcome is unavailable for some participants, because assumptions about those missing responses can affect the estimate. The record does not report the extent of missing primary endpoint data or the method used to address it.

Why are secondary results more difficult to interpret?

Multiple outcomes create multiple opportunities to observe a nominally unusual result, even when no underlying effect exists. Unless the analysis accounts for this multiplicity, individual p-values across secondary outcomes should be treated cautiously. The registry record lists multiple outcome measures and analyses, but does not report a multiplicity adjustment strategy or a hierarchy controlling which results are confirmatory. General principles therefore support giving the primary endpoint the central role and treating secondary findings as exploratory unless a prespecified testing strategy is reported.

11. Confidence Intervals and Precision

Precision is about how much uncertainty surrounds an estimate, not whether the estimate is important. For the primary endpoint, the posted percentages are relatively close, but the uncertainty intervals are broad enough to include no difference and contrasts in either direction. The re-calculated interval runs from -0.208 to 0.113. The sponsor-reported 90% interval for the model-based percent difference runs from -16.8 to 9.8. These intervals use different analysis scales and methods, so their widths should not be compared as if they were identical quantities.

In general, a confidence interval is produced by a method that would cover the target parameter at its stated rate over repeated samples if the model and assumptions were appropriate. It is not a guarantee that the true value lies inside this particular interval, and it does not mean that every value within the interval is equally plausible. The interval gives a range that helps readers judge how much the estimate is constrained by the data and analysis.

The analyzed groups for the primary outcome contained 68 participants in the golimumab group and 70 in the mavrilimumab group. Those counts help explain why uncertainty remains, but precision also depends on the observed responses and the analysis method. The registry record does not report the original precision target or the effect size the investigators intended to estimate. A non-significant result alongside a wide interval should therefore be read as inconclusive about the direction and size of the difference, not as evidence that the regimens are interchangeable.

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 a reconstruction of the trial’s original design. For the ordinary outcome comparisons, the reported quantities are participants per arm. For a time-to-event comparison, the reported quantity is the total number of events. The assumptions listed for these illustrations are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; each scenario uses a stated fraction of the observed effect as the effect to detect. The registry record does not report the trial’s original sample-size assumptions or justification.

The 80% power column describes the number of participants per arm, or total events for a time-to-event calculation, required under those assumptions to have an 80% chance of detecting the specified effect if that effect were the true one. The 90% power column describes the corresponding planning quantity for a 90% chance. These are design-planning concepts that depend on the assumed effect and statistical assumptions. They do not describe the probability that the trial’s observed result is correct.

Rows planning for a smaller fraction of the observed effect require more participants because a smaller difference is harder to distinguish from random variation. The figures are not observed power. They are conditional planning illustrations and should not be interpreted as proof that the study was adequately or inadequately designed. In particular, the observed effect can be unstable in a small study, so using it as the basis for future planning does not guarantee that the same effect will recur.

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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.662; treatment proportion p₁ = 0.614; average p̄ = 0.638.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6796 = 1.3321.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6788 = 0.5713.
4. n = (1.3321 + 0.5713)² / (-0.0475)² = 1,607.1, rounded up to 1,608 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
p0 <- 0.661764705882353                          # control proportion
p1 <- 0.614285714285714                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(1715896)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 1,608 (80% power), 2,151 (90% power)
 75% of observed effect: n per arm = 2,838 (80% power), 3,798 (90% power)
 50% of observed effect: n per arm = 6,337 (80% power), 8,483 (90% power)
power.prop.test: 1607.1 
Simulated power at n = 1,608 per arm: 0.801 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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

For the primary ACR20 comparison, the sponsor reported a logit response model estimate expressed as a percent difference of -3.5. Its 90% confidence interval was -16.8 to 9.8, and its p-value was 0.666. The simple unadjusted re-calculation gives a risk difference of -0.048, an interval from -0.208 to 0.113, and p = 0.562. These are not duplicate calculations of precisely the same reported quantity: one is a model-based sponsor analysis and the other is an unadjusted re-calculation from the posted summaries.

The two analyses point in the same broad direction for the estimated contrast, and both intervals include no difference. Their p-values and estimates differ because the analyses use different methods and scales. The registry record does not report enough model details to reproduce the sponsor’s analysis, and it does not identify a discrepancy between the posted primary result and the re-calculation. The sponsor result should therefore be read as the registry’s own model-based analysis, while the re-calculation is a transparent, simpler comparison rather than a replacement for the sponsor analysis.

For the secondary continuous ACRn outcome, the registry also reports a repeated-measures model result alongside the unadjusted re-calculation. For duration of DAS28 (CRP) remission, the sponsor reports a Weibull model p-value. The registry record does not give a model estimate or interval for that Weibull analysis. Differences in analysis method and available reporting mean that a recalculated summary should not be mistaken for the full sponsor analysis.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when investigators make several statistical comparisons or evaluate several outcomes. Even if each individual test uses a conventional threshold, conducting many tests increases the chance that at least one result will appear unusual by chance. This is a general statistical concern, not evidence that any particular posted result is false. It affects how much weight should be placed on isolated secondary p-values, especially when the outcomes are related but not identical.

The registry record lists 30 outcome measures and 23 analyses, including one primary outcome and multiple secondary outcomes. It does not report a multiplicity adjustment, a prespecified testing hierarchy, or a gatekeeping strategy for controlling the overall chance of false-positive findings. Therefore, the available record does not establish which secondary comparisons, if any, were intended to be confirmatory after the primary endpoint. Unadjusted secondary findings should be treated as exploratory rather than interpreted as independent confirmation.

The primary analysis population is described as mITT and linked to randomized treatment group. The registry record does not report a complete operational definition of that population, all exclusions, or the amount and pattern of missing primary outcome data. It also does not report the method used to handle missing Day 169 responses, whether sensitivity analyses were performed, or assumptions about participants without an observed endpoint. These details matter because missing outcomes can influence both the estimated response proportions and their uncertainty.

The record likewise does not report whether analyses were adjusted for baseline values, how protocol deviations were handled, or whether interim analyses were planned or performed. The absence of those details does not prove that such procedures were not used; it means they are not reported in the registry record. Interpretation is consequently limited to the outcome definitions, analysis populations, and model descriptions that the record does disclose.

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 1% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: gastroduodenitis (1.5%), peptic ulcer (1.5%), pneumocystis jirovecii pneumonia (1.5%), hepatic enzyme increased (1.5%), lung disorder (1.5%).

The registry reports participants with treatment-emergent serious adverse events and participants with other treatment-emergent adverse events. Serious adverse events were reported for 3 of 68 participants in the golimumab group and 2 of 70 in the mavrilimumab group. Other treatment-emergent adverse events were reported for 29 of 68 participants and 36 of 70 participants, respectively. These are participant counts, not counts of every event episode, and a participant may have experienced more than one event.

The record lists serious-event terms including gastroduodenitis, peptic ulcer, Pneumocystis jirovecii pneumonia, hepatic enzyme increased, lung disorder, and parathyroid tumour benign, with percentages as recorded. Those term-level listings should not be used to infer comparative risk without the corresponding definitions and complete event details. Deaths are not reported in the registry record.

These figures describe the posted safety experience in this study; they do not establish that a treatment caused or prevented an event. Interpretation is limited by the group sizes, the study duration, event rarity, and the absence of a complete narrative of timing, severity, relatedness, and participant-level overlap in the registry record. A simple count comparison cannot replace a fuller safety assessment.

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

The primary result describes the proportion of participants in each randomized group who met the ACR20 response definition at Day 169. The posted percentages were 65.6% for golimumab with alternating placebo and 62.0% for mavrilimumab. The estimated contrast is uncertain: the unadjusted re-calculation interval includes no difference and values on either side, while the sponsor-reported model interval also spans both directions. The result does not provide a precise, decisive separation between the groups.

That does not establish that the treatments are the same. A test that does not show a statistically clear difference is not an equivalence test unless an equivalence margin and suitable design and analysis have been specified. The registry record does not report an equivalence or non-inferiority objective or margin. It is therefore not appropriate to conclude equivalence from the primary result.

Nor does the primary endpoint summarize every relevant outcome. It is a composite threshold at a particular assessment time, and it does not show the trajectory of each component, the duration of response, or the full balance of potential benefits and harms. Secondary outcomes and safety results add context, but their multiplicity and varying analysis methods limit the strength of conclusions from individual findings.

The study was an exploratory Phase 2 comparison with 68 and 70 participants in the primary endpoint groups. Its results can describe the observed response proportions and the uncertainty around the comparison in this study population. They do not establish what would happen in every eligible population or under different treatment schedules. The findings should be understood as evidence from this specific registry-reported study, not as individualized medical guidance or a definitive general conclusion about either treatment.

18. Why This Trial Matters Statistically

This trial is statistically instructive because its central outcome is easy to state—a percentage achieving a defined response—while the comparison still requires careful interpretation. The observed percentages are not the whole result. The estimated contrast, its uncertainty interval, the analysis method, and the population included all affect what can reasonably be concluded.

It also illustrates why non-significance is not equivalence. Both the unadjusted and sponsor-reported primary comparisons leave uncertainty on either side of no difference. That pattern does not prove the groups are identical; it shows that the available analysis does not isolate a difference in one direction. The registry record does not report an equivalence or non-inferiority framework that would support a different interpretation.

The registry contains multiple secondary outcomes with different measurement types. This creates a useful opportunity to examine broader clinical and safety domains, but it also raises multiplicity and interpretation concerns. Without a reported hierarchy or adjustment, isolated secondary findings are exploratory. A collection of individually interesting results should not be treated as a single confirmatory test.

Finally, the sample-size illustrations distinguish planning from inference. They show how assumptions about the effect to detect and desired power influence required enrollment, but they are not a measure of the probability that the trial’s observed finding is true. The original design assumptions are not reported in the registry record. Keeping the planning exercise separate from the observed results is essential for understanding what this Phase 2 comparison can and cannot establish.

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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The primary comparison is based on relatively small analysis groups, and the confidence intervals allow a range of underlying differences. The study is identified as exploratory, so the result should not be treated as a definitive comparative-effectiveness conclusion. The registry record does not report a prespecified clinically important difference, an equivalence margin, or a non-inferiority margin.

Important design and analysis details are not reported in the registry record. These include the original sample-size justification, detailed randomization procedures, the allocation ratio, the roles covered by quadruple masking, the complete analysis plan, and any interim analyses. Missing-outcome counts and handling methods for the primary endpoint are also not reported. These gaps limit independent assessment of the analysis and its assumptions.

The reported primary sponsor analysis is model-based, but the registry record gives only the logit response model label and its result, not a full specification. The recalculation is unadjusted and cannot reproduce or replace that analysis. The record also lists many secondary outcomes, without reporting a multiplicity-control strategy, leaving secondary findings less secure as confirmatory evidence.

The study’s population and regimen are specific: adults aged 18–80 years with moderate-to-severe active rheumatoid arthritis and specified prior inadequate responses, treated for 24 weeks with the described schedules and background methotrexate. The registry record does not establish how well the findings generalize beyond those eligibility and treatment conditions. Safety interpretation is also constrained by the group sizes, duration, and limited event detail in the registry record.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study of Mavrilimumab Versus Anti Tumor Necrosis Factor in Subjects With Rheumatoid Arthritis — NCT01715896.
• Registry record last updated 2016-10-31; status last verified 2016-09.
• Results first posted 2016-10-31.
• 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

NCT01715896 is a completed randomized, quadruple-masked, parallel-group Phase 2 exploratory study of mavrilimumab versus golimumab in adults with rheumatoid arthritis. The registry record lists 215 enrolled participants, a study start of 2013-03, and primary completion in 2014-11. Both regimens were given with stable-dose methotrexate for 24 weeks.

The primary outcome was ACR20 response at Day 169. Posted response percentages were 65.6% for golimumab with alternating placebo and 62.0% for mavrilimumab. The sponsor-reported logit response model and the unadjusted re-calculation both leave uncertainty around the between-group comparison. The registry record does not report an equivalence or non-inferiority objective, original sample-size justification, multiplicity strategy, or detailed missing-data methods.