An Evaluation of Sarilumab Plus Methotrexate Compared to Etanercept Plus Methotrexate in RA Patients Not Responding to Adalimumab Plus Methotrexate: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for An Evaluation of Sarilumab Plus Methotrexate Compared to Etanercept Plus Methotrexate in RA Patients Not Responding to Adalimumab Plus Methotrexate (NCT01764997), a randomized phase 3 study. The primary outcome is Change From Baseline in Disease Activity Score for 28 Joints - C-Reactive Protein (DAS28-CRP) Score at Week 24.

2. Clinical Question

The clinical question was whether sarilumab combined with methotrexate changed rheumatoid arthritis disease activity differently from etanercept combined with methotrexate in participants who had not responded adequately to adalimumab plus methotrexate. The primary outcome was change from baseline in the Disease Activity Score for 28 joints using C-reactive protein (DAS28-CRP) at Week 24. The registry also describes assessments of signs and symptoms, quality of life, and safety and tolerability.

The record lists 776 participants as enrolled, but also states that the number randomized fell well below target (43 vs. 699). These figures describe different stages of the study: enrollment includes the run-in, whereas the stated randomized count concerns the comparison intended to answer the treatment question. The randomized group sizes listed for the three treatment arms are 17 for etanercept plus methotrexate, 13 for sarilumab 150 mg plus methotrexate, and 13 for sarilumab 200 mg plus methotrexate.

Because the registry says efficacy data were not systematically collected or cleaned and no datasets were created to report, the posted record gives no estimate of the difference in Week 24 change between treatments. The research question is clear, but its efficacy answer is not available in the registry record.

3. Trial Design

This was a Phase 3 randomized, controlled, parallel-group study with quadruple masking. Its official title describes adults with rheumatoid arthritis and an inadequate response to 4 months of treatment with adalimumab and methotrexate. The record lists Sanofi as sponsor and shows a study start of 2013-04 and primary completion of 2015-01. The current status is terminated; the reason for termination is not reported in the registry record.

The design began with an open-label run-in in which adalimumab 40 mg every 2 weeks was added to a stable methotrexate dose for 16 weeks. The registry then lists three randomized treatment groups, each with 24 weeks of treatment added to stable-dose methotrexate: etanercept 50 mg with placebo for sarilumab; sarilumab 150 mg with placebo for etanercept; and sarilumab 200 mg with placebo for etanercept. The descriptions include alternating-week placebo schedules alongside the every-2-weeks schedules. The record identifies quadruple masking but does not specify which roles were masked.

The randomized group counts reported in the adverse-event information are 17 for etanercept plus methotrexate, 13 for sarilumab 150 mg plus methotrexate, and 13 for sarilumab 200 mg plus methotrexate. The primary outcome description says the number randomized fell well below target (43 vs. 699). The 776 enrollment figure should not be interpreted as the size of the randomized comparison: it includes participants recorded in the open-label run-in and other study groups.

The record also lists a sarilumab 150 mg open-label sub-study, in which treatment was planned for 52 weeks with stable-dose methotrexate. Its adverse-event group contains 322 participants at risk. This open-label group is not one of the randomized comparisons described for the primary question. The record reports that efficacy data were not systematically collected or cleaned and that no datasets were created to report. As a result, the intended design can be described, but the posted record does not provide a Week 24 efficacy analysis.

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: Of 776 participants, 365 completed adalimumab run-in phase, of whom 43 non-responders randomized (1:1:1) in double-blind fashion to receive sarilumab 150 mg, sarilumab 200 mg or etanercept 50 mg; 322 responders entered in open label sub-study. 373 participants did not proceed into randomized phase or sub-study and 38 discontinued from run-in phase.

Run-in Period

Reasons for not completing this period:

Study Drug Treatment Period

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: As the number of participants randomized fell well below target (43 vs. 699), the efficacy data were not systematically collected or cleaned and no datasets have been created to report.

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: All participants who included in open label run-in period, randomized treatment and sub-study.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Disease Activity Score for 28 Joints - C-Reactive Protein (DAS28-CRP) Score at Week 24, measured over Baseline, Week 24.

The primary endpoint was change from baseline in the Disease Activity Score for 28 joints using C-reactive protein (DAS28-CRP) at Week 24. DAS28 is a composite measure of rheumatoid arthritis disease activity that incorporates assessment of 28 joints and a blood marker of inflammation, C-reactive protein. A change score expresses how the measured disease-activity score differs at follow-up from the participant’s baseline value.

As general statistical context, a continuous change endpoint can retain information about the size and direction of change rather than reducing each participant to a responder or non-responder category. Comparing changes between randomized groups can estimate the difference in average change, provided the necessary outcome measurements and analysis are available. A confidence interval, when reported, would communicate the range of values compatible with the data and the analysis assumptions; it would also show how precisely the treatment contrast was estimated.

The Week 24 time point gives the endpoint a defined follow-up horizon after randomized treatment. The registry also lists related secondary outcomes, including ACR20, ACR50, and ACR70 response rates at Week 12 and Week 24, achievement of a DAS28-CRP remission score below 2.6 at those time points, and change from baseline in DAS28-CRP at Week 12. These do not substitute for the designated primary endpoint.

No primary endpoint value, between-group estimate, measure of uncertainty, or statistical test is reported. The registry specifically states that efficacy data were not systematically collected or cleaned and no datasets were created to report. Therefore, the intended clinical meaning of the endpoint can be described, but the magnitude and precision of change in this trial cannot be assessed from the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry identifies randomization, quadruple masking, and a parallel-group model. These features describe the intended allocation and follow-up structure, but they do not by themselves provide an analysis method or an efficacy result. The statistical analysis plan, including the model used for the primary endpoint, is not reported in the registry record.

What would a randomized comparison estimate?

As general statistical context, random assignment is intended to make treatment groups comparable on average at the point of allocation, including for factors that are not measured. A parallel-group comparison follows the assigned groups over the same treatment period. The primary analysis would ordinarily quantify how the endpoint differs between assigned groups, but the particular estimand and analysis population for this trial are not reported in the registry record.

For this study, the endpoint is change from baseline in DAS28-CRP at Week 24. A comparison might be expressed as a difference between group-level changes, with a confidence interval and a test of a specified null hypothesis. This is a general description, not a claim about the method used for this trial. The registry gives no estimate, confidence interval, test statistic, or p-value, and no analysis method is reported. The absence of these results means readers cannot assess the estimated treatment contrast or its statistical uncertainty.

Why does the randomized group size matter?

As general statistical context, the amount of information available for an estimate depends in part on how many participants contribute outcome data and on the variability of those data. A smaller randomized sample commonly leads to less precise estimates, although the amount of uncertainty cannot be calculated without outcome data and a specified analysis. Here, the registry states that the number randomized fell well below target (43 vs. 699). It also states that efficacy data were not systematically collected or cleaned and that no datasets were created to report. Thus, the issue is not only that a small randomized group may yield imprecise results; the primary efficacy data needed to estimate the outcome are not reported at all.

The enrolled count of 776 refers to the study overall and should not be substituted for the randomized group count. The record describes an adalimumab run-in and a separate open-label sub-study in addition to randomized treatment arms. Combining those groups would not produce a valid randomized treatment comparison. The randomized arms listed in the adverse-event information have counts of 17, 13, and 13, respectively.

What does masking contribute?

As general statistical context, masking can reduce the opportunity for expectations to influence treatment delivery, assessment, or reporting. The record labels the study quadruple masked and describes placebo components in the randomized regimens. It does not state which groups of people were masked, how masking was maintained, or whether it was assessed. These details are not reported in the registry record. The run-in and sub-study are explicitly described as open label, so those phases should not be treated as masked randomized comparisons.

How should uncertainty be understood?

As general statistical context, a confidence interval describes a range of estimates compatible with the observed data under the stated model and assumptions. Its width reflects precision: wider intervals indicate less precise estimates. A p-value, when reported, addresses how compatible the data are with a specified null model; it does not measure the size or clinical importance of an effect. Neither kind of statistic can be interpreted for the primary endpoint because neither is reported. No estimate is available to accompany an uncertainty measure.

General considerations such as model choice, baseline adjustment, assumptions about residual distributions, and handling of repeated measurements can affect an analysis of a continuous endpoint. The registry record does not identify whether any of these methods were prespecified or used. It also does not report a variance estimate or other outcome distribution information. Therefore, no statement about the adequacy of a particular model, or about over-dispersion in the efficacy outcome, can be made from the posted results. In general, over-dispersion matters when observed variability exceeds what a model assumes, because standard errors and uncertainty estimates may then be unreliable; this is a methodological point, not a finding about this trial.

Can the analysis be reproduced?

No. Reproduction would require outcome data or sufficiently detailed summary statistics, together with the analysis method and population definitions. The registry explicitly states that no datasets were created to report efficacy data. The primary endpoint description contains no numerical result, and the analysis count is zero. The registry record does not report an analysis plan, participant-level data, group summaries, or the handling of protocol deviations for the primary analysis. There is therefore no basis for recalculating or independently checking the primary efficacy comparison.

The safety figures can be read descriptively as reported counts of participants affected in each named group, but they do not supply the missing efficacy information. General statistical methods can explain what a complete comparison would ordinarily communicate, but they cannot recover a treatment effect when the relevant outcome data and analysis are absent from the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The record lists a primary Week 24 change-from-baseline outcome and several secondary efficacy outcomes: ACR20, ACR50, and ACR70 response rates at Week 12 and Week 24; DAS28-CRP remission below 2.6 at Week 12 and Week 24; and change from baseline in DAS28-CRP at Week 12. It does not report how the family of outcomes or time points was handled statistically. A multiplicity procedure, testing hierarchy, or adjustment method is not reported in the registry record.

As general statistical context, assessing multiple outcomes or time points can increase the chance of finding at least one apparently notable result by chance if each is tested separately without a plan to control that risk. A prespecified hierarchy or multiplicity adjustment can address this concern, but no such approach can be attributed to this study from the available record. No efficacy tests are reported, so the effect of any multiplicity approach cannot be evaluated.

Missing-data handling is also not reported. The primary outcome description says efficacy data were not systematically collected or cleaned and no datasets were created to report. The record does not state how many participants had Week 24 measurements, whether participants withdrew from randomized treatment, how missing measurements would have been handled, or whether sensitivity analyses were planned or conducted. These omissions prevent assessment of the potential influence of missing outcomes on an efficacy estimate.

The analysis populations are not reported in the registry record. In particular, the record does not identify an intention-to-treat population, a per-protocol population, or a safety population definition for the efficacy analyses. It also does not describe how participants who discontinued treatment or deviated from the protocol would have been classified. These are distinct decisions that can affect what a treatment estimate represents; no specific approach should be assumed.

11. 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: urinary tract infection (5.9%), hypersensitivity vasculitis (5.9%), colonic abscess (0.3%), diverticulitis (0.3%), escherichia pyelonephritis (0.3%).

The registry reports participants affected by serious adverse events and other adverse events for each named group. In the adalimumab open-label run-in, 25 of 776 participants were reported with serious adverse events and 132 of 776 with other adverse events. In the randomized etanercept-plus-methotrexate group, the corresponding counts were 2 of 17 and 7 of 17. In the sarilumab 150 mg-plus-methotrexate randomized group, they were 0 of 13 and 7 of 13; in the sarilumab 200 mg-plus-methotrexate randomized group, 0 of 13 and 8 of 13. The open-label sarilumab 150 mg sub-study reports 11 of 322 with serious adverse events and 104 of 322 with other adverse events.

These are descriptive counts within groups, not adjusted comparisons of treatments. The randomized groups are small, and the run-in and open-label sub-study are different study settings from the randomized comparison. The registry does not provide a statistical safety comparison or report deaths. It lists serious-event terms with percentages, including urinary tract infection and hypersensitivity vasculitis at 5.9% each, and colonic abscess, diverticulitis, Escherichia pyelonephritis, and bladder cancer at 0.3% each. The denominators and event-counting details for those term percentages are not reported in the registry record. The figures should therefore be read as registry-reported safety summaries, not as proof of a causal treatment effect.

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

The record establishes what investigators intended to study: a comparison of sarilumab plus methotrexate with etanercept plus methotrexate in rheumatoid arthritis after an inadequate response to adalimumab plus methotrexate. It also documents a randomized, quadruple-masked, parallel-group design and identifies Week 24 change in DAS28-CRP as the primary outcome. Those design features make the clinical question explicit.

They do not establish the answer. The registry reports no primary endpoint estimate and no secondary efficacy results. It states that the number randomized fell well below target (43 vs. 699), efficacy data were not systematically collected or cleaned, and no datasets were created to report. Accordingly, the posted record cannot show whether either randomized treatment group had more or less change in disease activity, or how precisely any difference was estimated.

The absence of an efficacy result is not evidence that the treatments had the same effect, nor is it evidence that one was better. No p-value or confidence interval is available, so the record does not support a conclusion about statistical significance or uncertainty around a treatment contrast. The safety counts are separate descriptive information and cannot be used as a proxy for the missing efficacy outcome.

The adverse-event information indicates that the registry recorded affected participants in the run-in, randomized treatment groups, and open-label sub-study. Those group summaries have different contexts and denominators. They should not be read as a controlled estimate of comparative risk. The record also does not report deaths. No causal conclusion about an individual event or a difference between treatments follows from the listed counts alone.

The study’s termination status is recorded, but its reason is not reported in the registry record. The result is therefore best understood as an incomplete efficacy record for a planned randomized comparison. It documents the intended population, interventions, endpoint, and some safety counts, while leaving the central treatment-effect question unanswered in the posted results.

13. Why This Trial Matters Statistically

This trial illustrates why a randomized design and a registered endpoint are not enough to produce an interpretable treatment-effect estimate. The record identifies a clear comparison and a clinically relevant continuous outcome, but also states that the randomized group was far below target (43 vs. 699) and that efficacy data were not systematically collected or cleaned. No efficacy datasets were created to report. The distinction between a well-defined question and an available answer is central to interpreting registry evidence.

The study also demonstrates why enrollment and randomization should not be treated as interchangeable. The record lists 776 enrolled participants, but the primary outcome description points to a much smaller randomized group. The study included a run-in and an open-label sub-study as well as randomized groups. Counting everyone enrolled as if they contributed to a randomized treatment comparison would misrepresent the design and could create a misleading sense of how much information is available for the primary question.

From a statistical perspective, a useful efficacy report needs more than the endpoint name. Readers need an estimate, an indication of its precision, a defined analysis population, and enough information about missing data and analysis methods to understand what the result represents. Those elements are not reported for the primary outcome. Without them, neither statistical significance nor practical importance can be assessed.

The record’s safety counts retain descriptive value, but they answer a different question from the planned efficacy analysis. Their interpretation depends on group context and denominators, and the open-label groups are not equivalent to randomized treatment arms. Clear separation of design, efficacy, and safety evidence helps prevent the available counts from being stretched beyond what they can support.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The principal limitation is that the primary efficacy outcome is not reported. The registry states that efficacy data were not systematically collected or cleaned and no datasets were created to report. There is no Week 24 DAS28-CRP change estimate, treatment contrast, confidence interval, or test result. Consequently, the central randomized comparison cannot be interpreted quantitatively.

The registry says the number randomized fell well below target (43 vs. 699), and lists randomized arm counts of 17, 13, and 13. The reason for the shortfall and the reason for termination are not reported in the registry record. The record lists 776 enrolled participants, but that figure includes study phases and groups beyond the randomized comparison and should not be used as its analysis size.

The analysis populations, statistical methods, missing-data procedures, and handling of protocol deviations are not reported in the registry record. The registry lists multiple secondary outcomes and time points but does not describe multiplicity control or report efficacy analyses for them. No conclusion about the influence of missing data, analysis choices, or selective reporting can be drawn from the information posted.

Safety summaries are descriptive and span the open-label run-in, randomized arms, and open-label sub-study. The registry does not provide a comparative statistical analysis, and deaths are not reported. For the listed serious adverse-event terms, the denominator and counting rules for the percentages are not stated. These limitations mean the safety summaries should not be interpreted as direct randomized estimates of comparative risk.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: An Evaluation of Sarilumab Plus Methotrexate Compared to Etanercept Plus Methotrexate in RA Patients Not Responding to Adalimumab Plus Methotrexate — NCT01764997.
• Registry record last updated 2017-07-27; status last verified 2017-06.
• Results first posted 2017-07-27.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

NCT01764997 was a Phase 3 randomized, quadruple-masked, parallel-group study of sarilumab plus methotrexate versus etanercept plus methotrexate in rheumatoid arthritis after an inadequate response to adalimumab plus methotrexate. The primary outcome was change from baseline in DAS28-CRP at Week 24. The record lists 776 participants enrolled and states that the number randomized fell well below target (43 vs. 699); randomized arm counts were 17, 13, and 13.

The registry reports that efficacy data were not systematically collected or cleaned and no datasets were created to report. No primary or secondary efficacy estimates are posted. Descriptive adverse-event counts are available for the run-in, randomized arms, and open-label sub-study. The study status is terminated, and the reason for termination is not reported in the registry record.