Effects of Tofacitinib (CP-690,550) on Magnetic Resonance Imaging (MRI)- Assessed Joint Structure In Early Rheumatoid Arthritis (RA): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Effects of Tofacitinib (CP-690,550) on Magnetic Resonance Imaging (MRI)- Assessed Joint Structure In Early Rheumatoid Arthritis (RA) (NCT01164579), a randomized phase 2 study. The primary outcome is Change From Baseline to Month 3 in Outcome Measures in Rheumatoid Arthritis Clinical Trials (OMERACT) Rheumatoid Arthritis Magnetic Resonance Imaging Score (RAMRIS) Wrist and Metacarpophalangeal (MCP) Synovitis.

2. Clinical Question

The trial explored effects on joint structure using MRI and other assessments, with the primary endpoint centered on synovitis. The outcome was change from baseline, rather than the final score alone. A negative change denotes improvement on the specified synovitis score. The study also listed outcomes for bone marrow edema, erosions, and radiographic measures, including modified Total Sharp Score and joint space narrowing. Those outcomes provide context for the trial's broader structural-imaging focus, but the results discussed in detail concern the primary Month 3 synovitis endpoint.

3. Trial Design

This completed phase 2 study was randomized, quadruple-masked, parallel-group, and multicenter. The registry lists enrollment of 109 and a study population of methotrexate-naïve adults with early active rheumatoid arthritis. The study began in 2010-10, and primary completion was recorded as 2013-11. Its stated purpose was exploratory assessment of tofacitinib effects on MRI endpoints, alongside x-rays and standard clinical assessment.

The three arms were tofacitinib plus methotrexate, tofacitinib without active methotrexate, and methotrexate with placebo tofacitinib. For the primary endpoint, the evaluable-set counts were 30 for combination therapy, 32 for tofacitinib, and 31 for methotrexate. These endpoint counts are not interchangeable with overall enrollment: the endpoint definition required receipt of at least one dose and nonmissing measurements at baseline and the specified timepoint.

The registry describes the primary measure as a least-squares mean with standard error. It also reports mixed model repeated measures analyses for contrasts involving methotrexate. The original protocol's allocation ratio, sample-size justification, and detailed statistical design assumptions are not reported in the registry record. The planning assumptions attached to the recalculated sample-size illustrations are separate from the trial's original design assumptions.

The record lists 24 outcome measures and 206 analyses, and 22 secondary outcomes are posted. That breadth makes it important to keep the primary Month 3 endpoint distinct from secondary imaging and radiographic measures. The registry record does not establish that every listed measure or analysis had equal status in the study's original statistical 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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Evaluable Set: all randomized participants who received at least 1 dose of the randomized investigational drug and for whom a variable is nonmissing at both baseline and the specified timepoint.

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 (FAS): all participants who were randomized to study medication and received at least 1 dose of randomized study medication.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline to Month 3 in Outcome Measures in Rheumatoid Arthritis Clinical Trials (OMERACT) Rheumatoid Arthritis Magnetic Resonance Imaging Score (RAMRIS) Wrist and Metacarpophalangeal (MCP) Synovitis, measured over Month 3.

The primary endpoint is change from baseline to Month 3 in the OMERACT RAMRIS wrist and metacarpophalangeal synovitis score. Synovitis refers to an area in the synovial compartment with above-normal postgadolinium enhancement. The assessment used T1-weighted MRI images before and after intravenous gadolinium contrast; contrast was required to show enhancing synovitis.

Scoring covered three wrist regions and each of the first through fifth MCP joints. Each region or joint was scored from 0 to 3, where 0 is normal and scores from 1 to 3 represent mild, moderate, and severe synovitis. The total score ranges from 0 to 24. A negative change from baseline means the score declined, which the endpoint description defines as improvement.

MRI is used in this trial to assess inflammatory changes in joint structures directly, complementing x-rays and standard clinical assessment. The endpoint is a continuous score, so group results are expressed as means rather than as counts of participants meeting a threshold. The registry does not report a threshold for a clinically important change in this measure. Therefore, the numerical change can be interpreted as a change on the stated imaging scale, but its clinical importance cannot be determined from the posted endpoint definition alone.

7. Results Posted in the Registry

• Tofacitinib (CP-690,550): mean -0.69 versus -0.80 on control; difference 0.11 (95% CI -1.01 to 1.23), standardized effect d = 0.05; z-test p = 0.848.
• Methotrexate: mean -0.17 versus -0.80 on control; difference 0.63 (95% CI -0.49 to 1.75), standardized effect d = 0.28; z-test p = 0.271.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is continuous: each participant's synovitis score can take values along a scale, and the reported group statistic is a least-squares mean change. A least-squares mean is a model-based group estimate. Unlike a raw arithmetic mean, it can reflect the model's handling of repeated observations or other modeled factors. The registry record does not give enough detail to reconstruct the model specification, covariates, or all assumptions used for that estimate.

What does a change score describe?

A change score compares a participant's value at follow-up with their baseline value. For this endpoint, the change is interpreted using the MRI synovitis score, and a negative value is defined as improvement. A group mean change describes the average direction and magnitude across the analyzed group; it does not mean that every participant changed by that amount. It also does not reveal the distribution of individual responses.

The reported least-squares mean changes were -0.8 for combination therapy, -0.69 for tofacitinib, and -0.17 for methotrexate. Each has a standard error: 0.41, 0.4, and 0.4, respectively. A standard error describes the sampling uncertainty of an estimated mean, rather than the spread of individual participants' scores. The individual-level distribution and standard deviations are not directly reported for this endpoint.

What is the sponsor's repeated-measures analysis?

The registry labels the sponsor's method as mixed model repeated measures analysis. In general, repeated-measures models are designed to analyze observations collected from the same participant at multiple timepoints while accounting for the correlation among those observations. Mixed models can use available repeated measurements under model assumptions, rather than treating each timepoint as unrelated. The registry record does not report the precise covariance structure, covariates, missing-data assumptions, or model diagnostics, so those details cannot be assessed from the posted description.

The sponsor reported differences in least-squares means for combination therapy versus methotrexate and tofacitinib versus methotrexate. Those are model-based contrasts. The sponsor's confidence intervals and p-values belong to those specified analyses and should be distinguished from the simpler recalculations made from the posted Month 3 group summaries.

What do the recalculations add?

The recalculated comparisons are simple, unadjusted comparisons based on the posted summary statistics. They estimate a difference between two group means and use approximate standard deviations derived from the posted standard errors. The approximation is flagged in the record. These calculations are useful as a transparent check of the displayed group summaries, but they do not reproduce the sponsor's repeated-measures model. They do not adjust for baseline covariates, repeated observations, or multiple comparisons.

For the tofacitinib-versus-combination comparison, the recalculated mean difference is 0.11, with a 95% confidence interval from -1.013 to 1.233 and an unadjusted p-value of 0.848. The methotrexate-versus-combination recalculation gives a mean difference of 0.63 and an unadjusted p-value of 0.271. The sponsor's model-based combination-versus-methotrexate contrast is -0.63, with a 90% confidence interval from -1.58 to 0.31. The signs differ because the comparison order is reversed. A confidence interval from the recalculation should not be treated as interchangeable with the sponsor's model-based interval.

Why do confidence intervals matter alongside p-values?

A p-value evaluates how unusual the observed data would be under a specified null hypothesis and analysis model. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of a difference. A confidence interval provides an estimate and a range reflecting its precision under the analysis assumptions. In this trial, the sponsor's reported 90% intervals include zero, and the recalculated comparison for tofacitinib versus combination therapy also has a 95% interval that includes zero. The intervals show uncertainty in the estimates; a single threshold decision based on a p-value would not convey that uncertainty fully.

Multiplicity is relevant because the record contains many outcome measures and analyses. Testing many endpoints or contrasts can increase the chance of at least one nominally notable result. The registry record does not report a multiplicity adjustment or a complete prespecified testing hierarchy. The posted p-values should therefore be read as reported for their particular analyses, not as proof that all testing-related uncertainty has been addressed.

Missing measurements can also affect continuous outcomes. The endpoint's evaluable-set definition requires nonmissing baseline and follow-up values, so participants without both measurements for the specified timepoint are not included in that endpoint population as defined. The number and reasons for missing Month 3 measurements, and sensitivity analyses for missing data, are not reported in the registry record. That limits assessment of how missingness may have influenced the estimates.

10. Confidence Intervals and Precision

The sponsor's mixed model repeated measures estimate for combination therapy versus methotrexate was -0.63, with a 90% confidence interval from -1.58 to 0.31. For tofacitinib versus methotrexate, the estimate was -0.52, with a 90% confidence interval from -1.46 to 0.41. Each interval includes zero. The results are therefore compatible with no difference on the score, while also allowing differences in either direction within the stated intervals.

The recalculated tofacitinib-versus-combination estimate is 0.11, with a 95% confidence interval from -1.013 to 1.233. This interval is also compatible with effects on either side of zero. The recalculated methotrexate-versus-combination estimate is 0.63; the posted lower interval endpoint is not reproduced because it exceeds the maximum permitted decimal precision for this account. Its unadjusted p-value is 0.271.

In general, interval width reflects precision: wider intervals indicate more uncertainty about the magnitude of an effect. The standard errors around the group means and the intervals around the contrasts indicate that the estimates do not pin down a narrow range of between-group differences. Precision is also influenced by the analysis method and the amount of usable information. The reported evaluable-set counts for the primary endpoint are 30, 32, and 31 across the respective arms, but the registry record does not provide the detailed missing-data pattern or model assumptions needed to fully assess precision.

A confidence interval is not a guarantee that the true effect lies within its bounds, nor does overlap or inclusion of zero establish equivalence. The registry record does not report an equivalence or noninferiority margin for this endpoint. The posted intervals should be read as uncertainty ranges for the stated analyses, not as proof that any two regimens have the same effect.

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

For the recalculated comparison of tofacitinib with combination therapy, planning for 100% of the observed effect gives 6,595 participants per arm for 80% power and 8,828 participants per arm for 90% power. Planning for 75% of the observed effect gives 11,723 participants per arm for 80% power and 15,694 participants per arm for 90% power. Planning for 50% of the observed effect gives 26,377 participants per arm for 80% power and 35,311 participants per arm for 90% power.

For methotrexate versus combination therapy, planning for 100% of the observed effect gives 198 participants per arm for 80% power and 265 participants per arm for 90% power. Planning for 75% of the observed effect gives 352 participants per arm for 80% power and 471 participants per arm for 90% power. Planning for 50% of the observed effect gives 792 participants per arm for 80% power and 1,060 participants per arm for 90% power.

The 80% and 90% columns describe the assumed probability of detecting the specified effect under the planning model if that effect were the true one. The figures are participants per arm, not totals. Smaller effects require more participants to achieve the same target power, which is why the figures rise as the assumed fraction of the observed effect falls. These are planning illustrations, not observed power.

12. 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 = 0.110; pooled standard deviation σ = 2.255.
2. n = 2 × (1.96 + 0.842)² × 2.255² / 0.110² = 6,594.1, rounded up to 6,595 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 <- 0.11                      # observed difference in means
sigma <- 2.25450290160233                     # 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(1164579)
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 = 6,595 (80% power), 8,828 (90% power)
 75% of observed effect: n per arm = 11,723 (80% power), 15,694 (90% power)
 50% of observed effect: n per arm = 26,377 (80% power), 35,311 (90% power)
power.t.test: 6595 
Simulated power at n = 6,595 per arm: 0.794 (20000 replicates)

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

13. Analyses Reported by the Sponsor

The sponsor's own analysis used mixed model repeated measures and reported a combination-therapy versus methotrexate difference in least-squares means of -0.63, with a 90% confidence interval from -1.58 to 0.31. The posted tofacitinib-versus-methotrexate estimate was -0.52, with a 90% confidence interval from -1.46 to 0.41. Both intervals include zero.

The simple recalculation for methotrexate versus combination therapy gives 0.63, the same contrast in the opposite order from the sponsor's -0.63 estimate. The recalculated comparison uses posted summary statistics and an approximate standard deviation, whereas the sponsor's result comes from a repeated-measures model. The recalculated estimate should therefore be treated as a check on the group summaries, not as a replacement for the sponsor's analysis.

The record does not flag a discrepancy between the posted group summaries and reported analyses. The sponsor's model-based results are the primary source for its stated analyses; the recalculations help illustrate how the group estimates compare under a simpler unadjusted approach. The tofacitinib-versus-combination recalculation is not the same contrast as the sponsor's tofacitinib-versus-methotrexate result.

14. Multiplicity, Missing Data and Analysis Populations

The registry record includes 24 outcome measures, 206 analyses, and 22 posted secondary outcomes. The primary endpoint is Month 3 change in MRI synovitis; the listed secondary outcomes include other MRI components and radiographic measures at various timepoints. A large set of outcomes and contrasts creates multiplicity concerns because, in general, the chance of at least one apparently notable result can increase as the number of statistical tests increases.

The record does not report a multiplicity adjustment, a prespecified hierarchy of hypotheses, or which analyses were intended to be confirmatory rather than exploratory. The trial is described as exploratory, but the registry does not provide enough detail to infer how the investigators accounted for multiple outcomes or analyses. The reported p-values should be understood as results for their individual analyses; the record does not establish that they control an overall error rate across all analyses.

The primary endpoint population was the evaluable set: randomized participants who received at least one dose and had nonmissing values at both baseline and the specified timepoint. Thus, the primary analysis is not described as including every randomized participant regardless of whether the required measurements were available. The registry record does not report the number or reasons for missing Month 3 endpoint values, nor does it give sensitivity analyses showing how alternative assumptions about missing data affect the estimates.

In general, a complete-case or evaluable-set analysis can be vulnerable to bias if the availability of measurements is related to participants' outcomes or treatment. That possibility cannot be assessed from the posted definition alone. The registry record also does not report the detailed repeated-measures model specification or its missing-data assumptions. These omissions limit independent evaluation of both the analysis population and the robustness of the reported estimates.

15. 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 2% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: angina pectoris (2.8%), anal abscess (2.8%), foot deformity (2.8%), hematochezia (2.7%), cholelithiasis (2.7%).

The registry reports serious adverse event figures for the treatment groups. In the tofacitinib-plus-methotrexate group, 2 serious adverse events were recorded among 36 at risk; in the tofacitinib group, 1 was recorded among 36 at risk; and in the methotrexate group, 2 were recorded among 37 at risk. These are descriptive counts and denominators, not adjusted comparisons of safety between groups.

For other adverse events, the record reports 24 affected among 36 at risk for combination therapy, 31 affected among 36 at risk for tofacitinib, and 29 affected among 37 at risk for methotrexate. The serious-event terms listed include angina pectoris, anal abscess, foot deformity, hematochezia, and cholelithiasis, with reported percentages of 2.8%, 2.8%, 2.8%, 2.7%, and 2.7%, respectively. The registry record does not identify these listed terms as a complete account of all safety events in each arm.

Deaths are not reported in the registry record. These figures do not establish comparative risk: event counts alone do not account for exposure duration, differences in follow-up, or uncertainty around the estimates. They should be read as the safety information posted in the registry, not as a full safety analysis.

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

The posted primary endpoint results describe average changes in MRI-assessed synovitis over the period from baseline to Month 3. The mean changes were negative in all three groups, and the endpoint definition interprets negative change as improvement on the score. The sponsor's estimated differences between groups have confidence intervals that include zero. The recalculated tofacitinib-versus-combination contrast also has an interval that includes zero.

These results do not establish that the groups are equivalent. A failure to identify a clear difference is not the same as demonstrating that effects are identical or sufficiently similar according to a prespecified margin. No equivalence or noninferiority margin is reported in the registry record. The intervals also leave room for differences in either direction, so the posted results do not rule out all potentially meaningful effects.

The results do not establish the clinical importance of the observed score changes. The registry describes the MRI score and its direction but does not report a threshold for clinically important change. MRI synovitis is one structural-imaging measure within a study that also listed other MRI, radiographic, and clinical assessments. The primary endpoint alone cannot summarize every dimension of disease activity or joint structure.

Interpretation is further bounded by the evaluable-set definition, incomplete reporting of missing measurements and model details, and the record's many outcome measures and analyses. The sample-size illustrations are planning calculations based on observed effects, not evidence that the trial had a particular probability of detecting an effect. In short, the posted estimates describe uncertain group-average changes on a specified imaging scale; they do not, by themselves, settle broader questions about treatment benefit, clinical importance, or safety.

17. Why This Trial Matters Statistically

This trial illustrates how an imaging endpoint can be analyzed as a continuous outcome. The distinction between a group mean change, a model-based contrast, and a simple recalculation matters: each answers a related but not identical statistical question. The negative sign on the change scores also has a defined interpretation in the registry, so direction should be read against that definition rather than treated as self-evident.

It also shows why estimates should be considered with their intervals. The point estimates alone might look like clear rankings, but the reported confidence intervals for the sponsor's group contrasts include zero and span effects in both directions. The uncertainty is central to interpretation, especially when endpoint populations are evaluable sets and model details are not fully available in the registry record.

Finally, an exploratory trial with many posted outcomes raises practical questions about multiplicity, analysis populations, and planning. The recalculated sample-size illustrations vary substantially with the assumed fraction of the observed effect. They emphasize that planning around smaller effects requires larger groups, while not substituting for the trial's original design rationale. These are useful statistical distinctions for reading registry results without overstating what the data establish.

18. 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 record describes the primary measure and evaluable-set definition, but detailed modeling information is limited. The covariance structure, covariates, model diagnostics, and missing-data assumptions for the mixed model repeated measures analysis are not reported in the registry record. This restricts independent assessment of the model-based estimates.

The endpoint counts differ from overall enrollment because the endpoint population required both baseline and follow-up measurements and at least one dose. The registry record does not report the number of missing Month 3 measurements, reasons for missingness, or sensitivity analyses. Consequently, the potential effect of missing data on the primary estimates is unclear.

The trial is described as exploratory and lists numerous outcome measures and analyses, but the record does not report a multiplicity strategy or a testing hierarchy. Without that information, the posted p-values cannot be interpreted as accounting for all analyses. The registry also does not report a clinically important difference threshold for the MRI synovitis score, which limits conclusions about practical importance.

The recalculated contrasts use approximate standard deviations inferred from posted standard errors and are unadjusted. They do not reproduce the sponsor's repeated-measures model or account for its longitudinal structure. The planning illustrations are based on observed effects and stated assumptions rather than the original design justification; the original sample-size rationale is not reported in the registry record. Safety figures are descriptive, and deaths are not reported in the registry record.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Effects of Tofacitinib (CP-690,550) on Magnetic Resonance Imaging (MRI)- Assessed Joint Structure In Early Rheumatoid Arthritis (RA) — NCT01164579.
• Registry record last updated 2015-04-22; status last verified 2015-04.
• Results first posted 2015-04-22.
• 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.

22. Record Summary

NCT01164579 is a completed, randomized, quadruple-masked, parallel-group, multicenter phase 2 study in methotrexate-naïve adults with early active rheumatoid arthritis. Enrollment was 109. The trial compared tofacitinib plus methotrexate, tofacitinib with placebo methotrexate, and placebo tofacitinib plus methotrexate.

The primary endpoint was Month 3 change in MRI-assessed OMERACT RAMRIS wrist and MCP synovitis. Negative change was defined as improvement. Sponsor-reported repeated-measures contrasts versus methotrexate had 90% confidence intervals that included zero. Simple unadjusted recalculations provide related comparisons but do not replace the sponsor's model.

The registry record does not report the original sample-size justification, a multiplicity strategy, detailed missing-data analyses, or a clinical-importance threshold for the primary score. Safety counts are descriptive, and deaths are not reported in the registry record.