A Study to Investigate the Safety and Efficacy of ABBV-105 Alone or in Combination With Upadacitinib (ABBV-599 Combination) in Participants With Active Rheumatoid Arthritis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Investigate the Safety and Efficacy of ABBV-105 Alone or in Combination With Upadacitinib (ABBV-599 Combination) in Participants With Active Rheumatoid Arthritis (NCT03682705), a randomized phase 2 study. The primary outcome is Change From Baseline in Disease Activity Score 28 C-reactive Protein [DAS28-CRP]) at Week 12.

2. Clinical Question

The clinical question is about differences in average change on a composite measure of rheumatoid arthritis disease activity over the treatment period. A negative change in DAS28-CRP represents improvement from baseline; the comparison therefore concerns the magnitude of reduction in the combination group relative to placebo, rather than an outcome such as a single symptom or a count of participants reaching remission.

The additional treatment groups allow the registry to display results for ELS alone at several doses and for UPA alone as well as the combination. Those group estimates describe the study arms, but the available sponsor analysis for the primary endpoint is specifically the combination-versus-placebo comparison. Results should be interpreted in the context of the studied population and treatment period, not assumed to apply to people with different disease histories or treatment settings.

3. Trial Design

This was a completed, randomized, quadruple-masked, parallel-group phase 2 trial in participants with active rheumatoid arthritis. The registry lists enrollment of 242. Its stated purpose was to evaluate the safety and efficacy of ELS and ABBV-599, the combination of ELS and UPA, against placebo on a background of conventional synthetic DMARDs, and to define an optimal dose for further development. The listed trial start was 2018-10-08 and the primary completion date was 2020-03-26.

Six groups appear in the posted results: dual placebo; UPA 15 mg with ELS 60 mg; ELS 60 mg, 20 mg, or 5 mg with UPA placebo; and UPA 15 mg with ELS placebo. The active treatments and their placebos were given once daily by mouth for 12 weeks. The trial is described as parallel-group, so participants were assigned to a treatment group rather than being described as switching among treatments. The registry does not report the randomization ratio or the detailed allocation procedure.

Quadruple masking indicates that the registry characterizes the study as masked, but the record does not identify which roles were masked or describe how masking was maintained. Further details about treatment adherence, discontinuation, or use of rescue treatment are not reported in the registry record. The primary endpoint was change from baseline in DAS28-CRP at Week 12. The primary analysis population is defined as the Full Analysis Set, with requirements for randomized treatment, exposure, baseline measurement, and a post-baseline measurement.

Enrollment and endpoint analysis counts are not interchangeable: the record gives the overall enrollment and also gives arm-specific analysis counts for outcomes. Those counts are presented for their respective endpoint populations and should not be combined or treated as a single denominator for every analysis. Detailed explanations for differences in counts are not reported in the registry record. The record also does not report the sample-size justification or the specific design considerations used to choose enrollment.

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: Full Analysis Set: all randomized participants who received at least 1 dose of randomized study drug

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set: all randomized participants who received at least 1 dose of randomized study drug, had non-missing baseline values, and at least one post-baseline value. Baseline is defined as the last non-missing value prior to the first dose of study drug.

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: all randomized participants who received at least 1 dose of randomized study drug

Only the first 6 of 7 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Disease Activity Score 28 C-reactive Protein [DAS28-CRP]) at Week 12, measured over Baseline, Week 12.

The primary endpoint is change from baseline in the Disease Activity Score 28 using C-reactive protein (DAS28-CRP) at Week 12. The registry defines the score as a composite index based on tender joint count and swollen joint count, each assessed out of 28 evaluated joints, the patient’s global assessment of disease activity on a 0-100 mm measure, and high-sensitivity C-reactive protein (hsCRP) in mg/L. Scores range from 0 to approximately 10, with higher scores indicating more disease activity. A negative change from baseline indicates improvement.

A composite endpoint brings multiple aspects of disease activity into one measure. In general, this can provide a broader summary than any single component, but it also means that a change in the composite does not say which component contributed most. The registry does not provide component-level results alongside the primary endpoint estimates, so the relative contributions of joint counts, patient assessment, and hsCRP are not reported in the registry record.

The posted arm values are least-squares means with 90% confidence intervals. Least-squares means are model-based estimates; the registry does not give enough detail to describe the model specification or covariates. The sponsor’s comparison is a least-squares mean difference between the combination and dual-placebo groups. The endpoint therefore addresses average change on the composite scale, rather than the proportion of participants who crossed a particular disease-activity threshold. Remission and low disease activity are reported separately as secondary outcomes.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The statistical quantities in the registry answer related but distinct questions. The arm-specific least-squares means describe estimated average change from baseline for each group. The sponsor-reported least-squares mean difference compares the combination group with dual placebo. The confidence interval conveys uncertainty around that difference, while the p-value reports the result of the stated two-sided test. These should be read together rather than treating any one number as a complete summary.

What does a least-squares mean represent?

In general, a least-squares mean is an estimate from a statistical model, not necessarily the ordinary arithmetic average of the observed measurements. Such estimates can account for terms included in the model. The registry identifies the primary measure as a least-squares mean but does not report the model formula, covariates, or covariance structure. It is therefore not possible to describe which adjustments were made or to reproduce the sponsor’s model from the registry record alone. The arm estimates should be interpreted as model-based results as labeled, not as raw averages.

How should the confidence interval be read?

In general, a confidence interval indicates the range of effect estimates reasonably compatible with the observed data and the statistical procedure, under its assumptions. A narrower interval indicates greater precision, while a wider interval indicates more uncertainty. For this primary comparison, the reported 90% interval for the least-squares mean difference is -2.03 to -0.85. Because the interval lies below zero, its estimates all describe a greater reduction in the combination group on the DAS28-CRP change scale, given the direction of the contrast. The interval does not show the distribution of individual responses and does not establish that each participant experienced a change like the group estimate.

What does the primary p-value say?

The reported primary test is a two-sided t-test, with p < 0.001. In general, a two-sided test considers departures in either direction from a null difference. A small p-value indicates that data at least this incompatible with the null difference would be unusual if the test assumptions and null hypothesis held. It is not an estimate of the probability that the result is due to chance, and it does not quantify clinical importance. The registry does not report the testing hierarchy or whether the primary test was adjusted for any other planned tests.

How do the secondary recalculations differ?

For the posted binary secondary outcomes, the simple comparison estimates are unadjusted recalculations from the event counts and denominators recorded for each arm. A risk difference expresses the difference in the proportion meeting an outcome between the compared groups. In general, its interval communicates uncertainty on the percentage-point scale. These calculations are not the sponsor’s reported analyses: no sponsor-reported analyses are listed for the secondary outcomes included in these comparisons. They also do not incorporate an adjustment for multiple testing. The posted percentages and confidence intervals are useful descriptions, but any inferential reading should account for the absence of a reported multiplicity plan.

Small denominators can make binary percentages imprecise, especially when the number meeting an outcome is low. In general, this is why intervals around percentages or risk differences matter: a point estimate alone can give a misleading impression of certainty. A confidence interval that includes zero for a risk difference is compatible with effects in either direction as well as no difference under the calculation. It should not be interpreted as proof of equivalence or as proof that there is no effect.

What do the planning illustrations mean?

The sample-size illustrations are planning calculations based on the observed effect, not a report of the trial’s planned sample size or observed power. They use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is a stated fraction of the observed effect. In general, a smaller target effect requires a larger planned sample to achieve the same power, while a larger target effect requires fewer participants, all else equal. These illustrations describe hypothetical future planning under stated assumptions. They do not show the probability that the completed trial detected an effect, and they are not evidence that the original study was designed around the observed effect.

Over-dispersion is a general concern for count and binary outcomes when variability is greater than a simple model assumes. In general, failing to allow for extra variability can make standard errors too small and intervals too narrow. The registry does not report an over-dispersion assessment for these outcomes, so no such adjustment can be attributed to the posted recalculations. More broadly, the registry does not report the detailed statistical analysis plan, handling of missing data, or model diagnostics. These omissions limit independent assessment of assumptions, especially for model-based primary results.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary comparison as a two-sided t-test of the least-squares mean difference between UPA 15 mg with ELS 60 mg and dual placebo. The reported difference is -1.44, with a 90% confidence interval from -2.03 to -0.85 and p < 0.001. The comparison is consistent in direction with the two arm-level least-squares mean changes: -2.56 for the combination and -1.12 for dual placebo. Because negative change indicates improvement in DAS28-CRP, the sponsor estimate describes a larger average reduction for the combination group on this endpoint.

The registry does not flag a discrepancy between the reported primary analysis and a recalculation, and the sponsor’s own analysis is the result to use for that primary comparison. The record does not provide the model details needed to independently reproduce its least-squares means or test. The secondary binary-outcome comparisons are different: the posted simple calculations are unadjusted, and the registry lists no sponsor-reported analyses for those comparisons. They should not be presented as if they were sponsor-confirmatory tests or as a recalculation of the primary model.

The sponsor’s primary result is evidence about the specified endpoint, population, contrast, and Week 12 timeframe. It does not by itself settle the interpretation of all treatment arms, all secondary outcomes, longer-term effects, or safety. The statistical plan and its approach to multiplicity are not reported in the registry record.

11. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when an analysis considers several outcomes, treatment groups, timepoints, or statistical contrasts. In general, repeated testing creates more opportunities for apparently positive findings, so confirmatory trials often specify a testing hierarchy or another method for controlling the overall error rate. The registry lists 22 outcome measures and six treatment groups, but it does not report the prespecified multiplicity strategy, a hierarchy of hypotheses, or whether any secondary tests were adjusted. The primary result has a sponsor-reported two-sided p-value; the registry’s available secondary comparisons are unadjusted recalculations. Their p-values, where posted, should therefore be treated as exploratory descriptions rather than as multiplicity-controlled evidence.

The primary analysis population is explicitly defined as randomized participants who received at least one dose, had non-missing baseline values, and had at least one post-baseline value. Baseline is the last non-missing value before the first dose. This makes clear which general eligibility requirements define the Full Analysis Set, but it does not explain how missing Week 12 values were handled, whether any values were imputed, or how participants with incomplete follow-up contributed to the model. Those details are not reported in the registry record. Analysis populations for the secondary outcomes are not specified in the same detail in the available record.

Missing outcomes can affect both the estimated treatment effect and its uncertainty, particularly if the probability of missingness is related to treatment response or adverse events. In general, an analysis relies on assumptions about the relation between observed and unobserved outcomes. Without the missing-data method and its assumptions, the robustness of the estimates to missingness cannot be assessed from the registry summary alone.

The registry does not report the complete statistical analysis plan, sample-size justification, interim analyses, or detailed model specifications. It also does not state whether analyses were conducted according to a prespecified hierarchy or whether adjustments were made beyond the posted description. These omissions do not negate the reported results, but they limit the ability to evaluate the full inferential framework and to reproduce the analysis independently.

12. Safety Results

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

Deaths

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: prostatic specific antigen increased (5.3%), coronary artery disease (2.6%), lumbar radiculopathy (2.6%), cardiac arrest (2.4%), pyelonephritis (2.4%).

The registry reports serious adverse events by treatment group using both the number at risk and the number affected. The recorded serious-event counts are 1 among 19 at risk for dual placebo; 0 among 62 for UPA 15 mg with ELS 60 mg; 0 among 41 for ELS 60 mg with UPA placebo; 2 among 39 for ELS 20 mg with UPA placebo; 3 among 41 for ELS 5 mg with UPA placebo; and 0 among 40 for UPA 15 mg with ELS placebo. These are descriptive counts over the recorded study period, not adjusted comparisons of safety between treatments.

The registry reports deaths of 0 in each group except the ELS 5 mg with UPA placebo group, where it reports 1. It also lists serious-event terms, including prostatic specific antigen increased at 5.3%, coronary artery disease and lumbar radiculopathy at 2.6%, and cardiac arrest, pyelonephritis, and clavicle fracture at 2.4%. The record does not provide enough context in these listed terms to infer comparative risk or causal attribution from the percentages alone.

Rare events and small group counts make safety comparisons imprecise. In general, an absence of recorded serious events in a group does not establish that the risk is zero. These figures should be read as registry-reported observations, not as a complete assessment of all potential harms or as a treatment recommendation. The record does not report detailed event timing or a statistical comparison of serious adverse events.

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

The primary result means that, in the analyzed trial population and over the 12-week treatment period, the sponsor estimated a greater average reduction in DAS28-CRP for UPA 15 mg with ELS 60 mg than for dual placebo on background conventional synthetic DMARDs. The reported least-squares mean difference is -1.44, and its 90% confidence interval is -2.03 to -0.85. The two-sided p-value is < 0.001. These statistics describe the estimated contrast and its uncertainty under the sponsor’s analysis.

The result does not mean that every participant improved by the group average, or that every participant experienced the same outcome. A mean summarizes a group and can conceal variation in individual responses. The registry result also does not identify which DAS28-CRP components drove the mean change, since component-specific results are not reported with the primary endpoint.

The result does not establish the clinical importance of the difference. The registry provides a scale description and direction of improvement but does not report a threshold for a clinically important difference. Statistical evidence against a null difference and clinical importance are distinct questions. Nor does the primary comparison establish that the combination is better than each active treatment group: the sponsor-reported test is for the combination against dual placebo, not a complete set of pairwise tests.

It does not establish longer-term outcomes beyond the reported timeframe, and the record does not provide a basis for generalizing automatically to people outside the trial population. The result is specific to participants with active rheumatoid arthritis and inadequate response or intolerance to biologic DMARDs, treated with background conventional synthetic DMARDs under the studied regimens.

Finally, the primary endpoint cannot by itself determine the balance of benefits and harms. Safety counts are descriptive, and the registry does not report a formal comparative safety analysis in the information summarized. Secondary outcomes offer additional descriptions but have less inferential weight when their analysis and multiplicity handling are not reported. The appropriate interpretation is therefore bounded: a statistically clear estimated difference on the primary disease-activity measure, with limits on what can be concluded about clinical importance, other comparisons, safety, and outcomes beyond the trial period.

14. Why This Trial Matters Statistically

This trial raises several statistical issues that commonly arise in multi-arm clinical studies. It has a primary continuous endpoint, several active regimens, and a number of secondary outcomes that use both continuous and binary definitions. Each type of outcome answers a different question. DAS28-CRP change summarizes average movement on a disease-activity scale, while remission, low disease activity, and response outcomes classify participants according to thresholds or criteria. Keeping these estimands distinct helps prevent a result on one measure from being treated as if it were evidence for every other outcome.

The primary report illustrates why the contrast and analysis method matter. Arm-level estimates describe each group, but the sponsor-reported inferential result is a particular least-squares mean difference between combination treatment and dual placebo. That result is not interchangeable with informal comparisons of the arm estimates or with separately recalculated binary outcomes. The analysis population also matters: the primary estimate is defined for participants meeting the Full Analysis Set criteria rather than simply being identified with overall trial enrollment.

Precision is another central issue. Confidence intervals communicate how much uncertainty remains around estimates, and binary outcomes with limited event counts can be particularly imprecise. A point estimate may look substantial while its interval still permits a wide range of effects. Secondary comparisons also need to be considered in light of multiplicity: without a reported testing plan, nominal secondary findings should not be given the same confirmatory interpretation as the specified primary comparison.

Finally, this record distinguishes a trial’s observed outcomes from planning illustrations derived from an observed effect. Those illustrations can help explain how assumed effect sizes and design assumptions influence hypothetical sample-size requirements, but they are not observed power and do not reveal the original design rationale. The sample-size justification and fuller statistical plan are not reported in the registry record. The case therefore shows why transparent reporting of estimands, model details, missing-data methods, and multiplicity procedures is important for interpreting a trial beyond its headline estimate.

15. Limitations

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

The registry record gives the primary endpoint definition and sponsor comparison but does not report the statistical model details needed to assess or reproduce the least-squares mean estimates. Covariates, model assumptions, diagnostics, and the handling of missing Week 12 observations are not reported in the registry record. These omissions limit independent evaluation of the primary estimate and its sensitivity to analysis choices.

The study has multiple treatment groups and many outcome measures, yet the multiplicity plan and hierarchy of tests are not reported. The available secondary binary comparisons are unadjusted recalculations, with no sponsor-reported analyses listed for those comparisons. They therefore have limited value for confirmatory claims. The registry does not provide numerical results for every listed secondary outcome in the information used for this interpretation, so titles alone cannot support conclusions about those outcomes.

Counts vary by endpoint and treatment group. The primary Full Analysis Set definition is stated, while corresponding detailed population definitions for all secondary outcomes are not reported. The reason for differences between enrollment and analysis counts is also not reported in the registry record. Without additional participant-flow information, it is not possible to determine from these numbers alone why particular endpoint denominators differ.

The treatment period was 12 weeks, which bounds interpretation to that timeframe. Longer-term outcomes are not reported in the registry record. The record also does not report the sample-size justification, detailed randomization procedures, or which study roles were masked. These limitations affect assessment of design and generalizability, but do not change the meaning of the reported primary estimate within its stated population and endpoint.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study to Investigate the Safety and Efficacy of ABBV-105 Alone or in Combination With Upadacitinib (ABBV-599 Combination) in Participants With Active Rheumatoid Arthritis — NCT03682705.
• Registry record last updated 2021-05-03; status last verified 2021-04.
• Results first posted 2021-05-03.
• 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.

19. Record Summary

NCT03682705 is a completed randomized, quadruple-masked, parallel-group phase 2 study of ELS, UPA, their combination, and placebo in active rheumatoid arthritis on background conventional synthetic DMARDs. The registry lists enrollment of 242 and a 12-week treatment period.

The primary endpoint was change from baseline in DAS28-CRP at Week 12. The sponsor-reported comparison of UPA 15 mg with ELS 60 mg against dual placebo was a least-squares mean difference of -1.44 (90% CI -2.03 to -0.85; p < 0.001). Negative change indicates improvement on the disease-activity scale. Secondary outcomes include disease-activity thresholds and response measures, but the available recalculated comparisons are unadjusted and should be interpreted cautiously.

The registry reports serious adverse events and deaths by group, but these counts are descriptive. Detailed multiplicity methods, missing-data handling, model specifications, and the sample-size justification are not reported in the registry record.