A Study to Evaluate Rituximab in Combination With Methotrexate in Methotrexate-Naive Patients With Active Rheumatoid Arthritis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Evaluate Rituximab in Combination With Methotrexate in Methotrexate-Naive Patients With Active Rheumatoid Arthritis (NCT00299104), a randomized phase 3 study. The primary outcome is Change From Baseline in Modified Total Sharp Score (mTSS) From Screening at Week 52.

2. Clinical Question

3. Trial Design

This was a randomized, controlled, double-blind, parallel-group, phase 3 study. The registry lists an enrollment of 755 and describes an international, multicenter trial. The brief description characterizes the planned enrollment as approximately 750 patients and says patients would be allocated equally across three groups. The study status is completed.

The three groups were placebo plus methotrexate, rituximab 0.5 g on each of two infusion days plus methotrexate, and rituximab 1.0 g on each of two infusion days plus methotrexate. The methotrexate schedule was shared across groups: treatment began at 7.5 mg per week, with increases of 2.5 mg at intervals of 1–2 weeks toward 15 mg per week by Week 4 and 20 mg per week by Week 8. The active groups could receive subsequent rituximab courses at 24 week intervals. The placebo group’s description allows rituximab treatment from Week 104.

Randomization helps make treatment groups comparable on average, while double masking is intended to limit differences in care or outcome assessment arising from knowledge of assignment. A parallel-group design means participants were assigned to a treatment group rather than each participant receiving multiple study treatments in sequence. These are general features of the design; the registry record does not report the specific randomization procedure, concealment details, or masking procedures.

The primary endpoint was assessed at Week 52 in the modified intent-to-treat population, defined in the record by availability of screening and at least one post-baseline radiographic evaluation. The registry does not report the trial’s original sample-size justification or its 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.

Treatment Period

Reasons for not completing this period:

Safety Follow-Up (SFU) Period

Reasons for not completing this period:

Extended Safety Follow-Up (ESFU) Period

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Modified intent-to-treat population includes patients with a screening and at least one post-baseline radiographic evaluation, grouped as randomized. Linear interpolation/extrapolation used for missing data.

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.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Modified Total Sharp Score (mTSS) From Screening at Week 52, measured over Baseline and week 52.

The primary endpoint was change from screening in modified Total Sharp Score (mTSS) at Week 52. This radiographic score combines an erosion score and a joint-space-narrowing score. The erosion component sums scores from 46 joints and ranges from 0 to 230; the joint-space-narrowing component sums scores from 42 joints and ranges from 0 to 168. The combined score ranges from 0 to 398.

The endpoint is used to quantify structural joint damage on radiographs. A change score of 0 indicates no change, while higher scores indicate worsening joint erosions and narrowing of joint spaces. Because the endpoint is a continuous score, the registry reports group means and standard deviations, which describe average change and variation among participants. It also reports comparisons between each rituximab group and placebo plus methotrexate.

Radiographic structural change is distinct from symptoms, daily functioning, or other clinical outcomes. The primary endpoint therefore addresses one defined aspect of rheumatoid arthritis over the specified follow-up period; it does not by itself describe every outcome that may matter to participants. The record specifies linear interpolation or extrapolation for missing radiographic data in the primary modified intent-to-treat population. The rationale for selecting this endpoint, and whether other endpoints were considered primary during planning, is not reported in the registry record.

7. Results Posted in the Registry

• Rituximab (0.5 g x 2) + Methotrexate: mean 0.65 versus 1.08 on control; difference -0.43 (95% CI -1.01 to 0.15), standardized effect d = -0.14; z-test p = 0.144.
• Rituximab (1.0 g x 2) + Methotrexate: mean 0.36 versus 1.08 on control; difference -0.72 (95% CI -1.26 to -0.18), standardized effect d = -0.24; z-test p = 0.009.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Change From Baseline in Modified Sharp Erosion Score at Week 52

• 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:
  • Rituximab (0.5 g x 2) + Methotrexate: mean 0.45 versus 0.74 on control; difference -0.28 (95% CI -0.59 to 0.02), standardized effect d = -0.17; z-test p = 0.067.
  • Rituximab (1.0 g x 2) + Methotrexate: mean 0.23 versus 0.74 on control; difference -0.51 (95% CI -0.78 to -0.23), standardized effect d = -0.34; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Rituximab (0.5 g x 2) + Methotrexate and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Rituximab (1.0 g x 2) + Methotrexate than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Patients Without Radiographic Progression in Total Erosion Score at Week 52

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Rituximab (0.5 g x 2) + Methotrexate: 59.0% (141/239) versus 54.7% (127/232) on control; risk difference 4.3 percentage points (95% CI -4.7 to 13.2), risk ratio 1.08 (0.92–1.26); unadjusted two-proportion z-test p = 0.351.
  • Rituximab (1.0 g x 2) + Methotrexate: 66.8% (163/244) versus 54.7% (127/232) on control; risk difference 12.1 percentage points (95% CI 3.3 to 20.8), risk ratio 1.22 (1.05–1.41); unadjusted two-proportion z-test p = 0.007.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Rituximab (0.5 g x 2) + Methotrexate and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Rituximab (1.0 g x 2) + Methotrexate than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change From Baseline in Modified Joint Space Narrowing (JSN) Score at Week 52

• 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:
  • Rituximab (0.5 g x 2) + Methotrexate: mean 0.19 versus 0.34 on control; difference -0.15 (95% CI -0.46 to 0.16), standardized effect d = -0.09; z-test p = 0.353.
  • Rituximab (1.0 g x 2) + Methotrexate: mean 0.13 versus 0.34 on control; difference -0.22 (95% CI -0.51 to 0.08), standardized effect d = -0.13; z-test p = 0.159.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Rituximab (0.5 g x 2) + Methotrexate and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Rituximab (1.0 g x 2) + Methotrexate and the control. The interval gives the range of effects that remain compatible with the data.

The posted secondary outcomes broaden the radiographic picture beyond the primary total score. They include erosion and joint-space-narrowing scores, a binary measure of being without radiographic progression, and assessments at more than one follow-up time. Across the numerically detailed secondary outcomes, the posted summaries generally show smaller mean score changes or a higher percentage without progression in the rituximab groups than in the placebo-plus-methotrexate group. The pattern is not equally clear for every dose and measure.

This broad direction is consistent with the primary endpoint’s arm-level pattern, but secondary outcomes do not independently establish the primary finding. They measure related components or alternative definitions of radiographic change, and their statistical analyses differ by outcome type. The registry lists 29 secondary outcomes, while only a subset has detailed arm summaries and comparisons in the record. The evidence available for those listed outcomes should not be generalized to every secondary measure.

Multiple outcomes and pairwise comparisons create a multiplicity concern. If many tests are considered together, some results may appear notable by chance even when there is no underlying difference. The record reports unadjusted recalculations for secondary contrasts and does not report a complete multiplicity adjustment or a prespecified hierarchy for secondary outcomes. Consequently, those recalculated results carry less weight than a primary outcome interpreted under a documented analysis plan. Sponsor analyses are available for some secondary outcomes, but the registry does not report a complete inferential framework for all secondary measures.

The secondary results can therefore be read as supportive context about related radiographic measures, not as a collection of independent confirmations. Their interpretation depends on the analysis method, missing-data handling and multiplicity plan, details of which are not fully reported in the registry record.

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

• Change From Baseline in Modified Sharp Erosion Score at Week 52
• Percentage of Patients Without Radiographic Progression at Week 52
• Percentage of Patients Without Radiographic Progression in Total Erosion Score at Week 52
• Change From Baseline in Modified Joint Space Narrowing (JSN) Score at Week 52
• Change From Baseline in the Modified Total Sharp Score at Week 24
• Change From Baseline in the Total Erosion Score at Week 24
• Change From Baseline in Modified Joint Space Narrowing (JSN) Score at Week 24
• Percentage of Participants Without Radiographic Progression at Week 24

9. Statistical Interpretation

10. Statistical Methods Explained

The outcome is continuous: each participant has a numerical change in mTSS, and the arm summaries describe the distribution of those changes. A mean is the arithmetic average. A standard deviation describes how dispersed individual observations are around that average; it is not the uncertainty in the mean and is not a confidence interval. In the posted results, the mean change and standard deviation are given for each arm.

What does a mean difference represent?

The recalculated mean difference compares an active-treatment arm’s mean with the placebo-plus-methotrexate mean. The comparison is coded as active treatment minus comparator, so negative values indicate a lower mean change in the active-treatment arm. For example, the recalculated differences are -0.433 for the 0.5 g dose group and -0.72 for the 1.0 g dose group. These calculations use unadjusted arm summaries; they do not reproduce a stratified or rank-based analysis. The record flags a discrepancy, so the sponsor’s own analysis takes precedence when describing the trial’s inferential results.

Why are different tests reported?

The sponsor reports a Kruskal-Wallis test for the three-group comparison and Van-Elteren tests for pairwise comparisons. In general, Kruskal-Wallis is a rank-based method for comparing groups and does not directly test equality of arithmetic means. The Van-Elteren method is a stratified rank test; stratification can account for defined strata in the analysis. The registry record does not identify the strata or give the full method specification. These sponsor methods therefore answer questions using a different statistical approach from an unadjusted comparison of means.

A p-value measures how compatible data are with a specified null hypothesis under the assumptions of a particular test. It does not state the probability that the null hypothesis is true, and it does not measure the size or importance of a difference. The reported sponsor analyses include an overall three-group result and pairwise comparisons. The record does not report a complete multiplicity strategy for interpreting these multiple tests, so the posted p-values should not be treated as if a familywise error adjustment were documented.

What is over-dispersion, and does it apply?

Over-dispersion is a general concern for count or binary outcomes when observed variability exceeds what a simple probability model assumes. The primary outcome is a continuous score, not a count or a binary endpoint, so count-model over-dispersion is not the central issue for its mean comparison. The posted standard deviations nevertheless indicate that participant-level changes vary, and the choice of method should account for the outcome’s distribution and analysis design. The registry record does not report diagnostic assessments of the score distribution.

For the recalculated mean differences, the confidence intervals are based on unadjusted summaries and describe uncertainty around those differences under the calculation’s assumptions. The interval for the lower-dose comparison spans zero; the interval for the higher-dose comparison does not. This is not identical to the sponsor’s rank-based testing. Differences in estimand, stratification, missing-data handling, or other details can lead to different inferential results, but the record does not establish which details explain the discrepancy.

How do the secondary outcome methods differ?

The posted secondary outcomes include continuous radiographic scores and a percentage without radiographic progression. For a continuous score, a mean difference summarizes average changes between groups. For a percentage outcome, a risk difference expresses the difference between proportions. Those effect measures have different interpretations and should not be interchanged. The registry reports Cochran-Mantel-Haenszel analyses for the binary secondary outcome and rank-based analyses for some continuous outcomes. In general, the Cochran-Mantel-Haenszel method can form a stratified comparison across categories; the registry record does not specify its strata.

The recalculated secondary comparisons are labeled unadjusted. They are useful as simple summaries of the posted group data, but they do not replace the sponsor’s analysis or account for the full set of outcomes and tests. The registry record lists 30 outcome measures and 20 analyses overall. With several outcomes and comparisons, chance findings become more plausible if many tests are interpreted without adjustment. The multiplicity plan, any hierarchy among secondary outcomes, and any prespecified selection of comparisons are not reported in the registry record.

What is known about the analysis population?

The primary outcome population is explicitly defined as modified intent-to-treat: patients with a screening radiographic evaluation and at least one post-baseline evaluation, grouped as randomized. This retains randomized grouping but conditions inclusion on having radiographic data after baseline. Linear interpolation or extrapolation was used for missing data. In general, interpolation estimates an unobserved value between observed measurements, while extrapolation extends a pattern beyond observed measurements. Both approaches rely on assumptions about the trajectory of the score. The registry record does not report the amount or pattern of missingness, the exact implementation, or sensitivity analyses.

The trial is randomized and double-blind, which are design features intended to support fair comparisons and reduce bias. However, a reader cannot assess all risks of bias from those labels alone. Details of allocation concealment, masking success, analysis-plan prespecification, and handling of deviations are not reported in the registry record.

11. Confidence Intervals and Precision

Precision concerns how narrowly an estimate is determined by the data and analysis. For the unadjusted mean difference between the 0.5 g rituximab group and placebo plus methotrexate, the recalculated 95% confidence interval includes zero. This means the interval is compatible with no mean difference as well as with differences in either direction under the calculation’s assumptions. For the 1.0 g group, the recalculated 95% interval excludes zero and is entirely below zero. The intervals are estimates of uncertainty around group-level differences, not ranges that contain a specified proportion of individual participant outcomes.

The posted arm standard deviations differ across groups. Standard deviations describe participant-level spread, whereas confidence intervals for a mean difference reflect both that spread and the number of observations, along with the calculation method. A larger study can generally estimate a group contrast more precisely when other conditions are similar, but the precision also depends on the variability and effect size. These are general statistical principles, not additional findings about the trial.

The sponsor’s reported analyses are rank-based and do not include effect estimates or confidence intervals in the registry results. Their p-values therefore cannot show the magnitude or precision of the sponsor’s estimated treatment contrasts. The recalculated intervals are simple unadjusted intervals from the posted summary statistics and do not represent a reconstruction of the sponsor’s stratified analysis. Because the registry record flags a discrepancy, sponsor analyses take precedence for the trial’s own inferential interpretation.

Precision is also limited by what the registry reports. The primary endpoint population is defined and the missing-data approach is named, but missingness counts and sensitivity analyses are not reported in the registry record. Without those details, the effect of missing radiographs on the estimate’s uncertainty cannot be fully assessed. Statistical precision should therefore be distinguished from certainty about the result’s robustness, its clinical meaning, or its applicability beyond the studied population and follow-up.

12. Planning a Similar Trial: Sample Size

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

The sample-size figures are planning illustrations based on the observed effect, not the trial’s original design calculation. For continuous outcomes such as the primary mTSS endpoint, the figures refer to participants per arm, not a combined total. The calculation assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each planning row considers detecting the stated fraction of the observed effect.

The 80% power column gives the number of participants per arm that the illustration estimates would be needed for an 80% chance of detecting the assumed effect under those assumptions, if that effect and model assumptions were correct. The 90% power column gives the corresponding requirement for a 90% chance. Power is a property of a planned design under an assumed effect and statistical model; it is not the probability that a particular observed finding is true.

The rows use the full observed effect and smaller fractions of it. As the effect being targeted becomes smaller, more participants are needed to distinguish it from random variation with the same planned power and significance level. For example, the primary endpoint’s planning illustration uses 848 participants per arm for 80% power when planning for 100% of the observed effect, and 3,391 participants per arm when planning for 50% of that effect. These figures should not be confused with the study’s actual enrollment or actual group sizes. They are not observed power.

The registry record does not report the original trial sample-size justification or its planning assumptions. The illustrations should therefore be read as hypothetical design calculations based on the observed data rather than evidence about how the trial was originally sized.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -0.433; pooled standard deviation σ = 3.182.
2. n = 2 × (1.96 + 0.842)² × 3.182² / -0.433² = 847.5, rounded up to 848 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 <- -0.433                      # observed difference in means
sigma <- 3.18164244386402                     # 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(299104)
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 = 848 (80% power), 1,135 (90% power)
 75% of observed effect: n per arm = 1,507 (80% power), 2,018 (90% power)
 50% of observed effect: n per arm = 3,391 (80% power), 4,539 (90% power)
power.t.test: 848.5 
Simulated power at n = 848 per arm: 0.802 (20000 replicates)

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

14. Analyses Reported by the Sponsor

The registry posts sponsor analyses using a Kruskal-Wallis test across the three primary-outcome groups and Van-Elteren tests for pairwise comparisons. The overall reported analysis indicates a difference among the groups. The sponsor’s pairwise analyses do not indicate a clear difference for the 0.5 g rituximab group, while the 1.0 g group’s pairwise result indicates a difference from placebo plus methotrexate.

The recalculated comparisons are simple unadjusted differences in means with confidence intervals and unadjusted p-values. They use the posted arm means, standard deviations and sample sizes; they are not reproductions of the sponsor’s rank-based analyses. For the lower dose, the recalculated interval includes zero; for the higher dose, it excludes zero. The registry record explicitly indicates a discrepancy, so the sponsor’s own analysis takes precedence in describing the trial’s inferential results.

The sponsor’s analysis results do not include effect estimates or confidence intervals in the registry record. This limits direct comparison of effect size and precision across methods. Different methods may test different aspects of the data, and the record does not provide enough detail to identify the reason for the discrepancy. The recalculations can be used to understand the scale and direction of unadjusted mean contrasts, but should not be presented as the sponsor’s result or as a replacement for it.

15. Multiplicity, Missing Data and Analysis Populations

The trial has one designated primary endpoint and a broad set of secondary outcomes. The registry record lists 30 outcome measures and 20 analyses overall, and multiple treatment groups create more than one possible comparison. Testing several outcomes or contrasts increases the chance that at least one result will appear statistically notable by chance when all null hypotheses are true. This is the general multiplicity problem.

The registry reports sponsor analyses for an overall three-group comparison and pairwise comparisons, and the secondary outcome listings include additional tests. It does not report whether the sponsor used a prespecified testing hierarchy, adjusted significance levels, or another multiplicity-control procedure. That information is not reported in the registry record. The unadjusted p-values from recalculations should therefore not be interpreted as if they were adjusted for all endpoints and comparisons.

For the primary endpoint, the modified intent-to-treat population includes randomized patients with screening and at least one post-baseline radiographic evaluation. The record says these participants were grouped as randomized. It also states that linear interpolation or extrapolation was used for missing data. This gives some information about population definition and missing-value handling, but the amount and pattern of missing radiographs, reasons for missingness, and sensitivity analyses are not reported in the registry record.

The original statistical analysis plan, the sponsor’s sample-size justification, interim analyses, and the handling of protocol deviations are not reported in the registry record. Nor does the record fully specify the strata for the Van-Elteren or Cochran-Mantel-Haenszel analyses. These omissions do not erase the posted results, but they limit independent assessment of prespecification, robustness to missing data, and the intended interpretation of multiple tests.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonia (2.4%), urinary tract infection (1.6%), fall (1.2%), infection (1.1%), rheumatoid arthritis (0.9%).

The registry reports serious adverse events for three groups, with separate counts of participants at risk and participants affected. The counts at risk are 249 for placebo plus methotrexate, 348 for rituximab 0.5 g plus methotrexate, and 263 for rituximab 1.0 g plus methotrexate. The corresponding counts affected are 48, 54 and 58. These are group-level registry counts and should be read with their reported denominators; the numbers at risk differ from the primary radiographic analysis group sizes.

The most frequent serious adverse-event terms listed include pneumonia, urinary tract infection, fall, infection, rheumatoid arthritis and road traffic accident, with the registry reporting each term as a percentage. The listed term percentages are not a complete description of all events or their clinical context. The registry record does not report death counts for these groups, and the basis for the adverse-event risk windows and event coding details is not reported in the registry record.

These counts are descriptive. They do not by themselves establish that an event was caused by treatment, and they should not be interpreted as a comparison adjusted for exposure time or participant characteristics. The event counts and the primary radiographic endpoint answer different questions and should be considered separately.

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

The primary endpoint results describe mean radiographic score changes over the stated follow-up in the defined modified intent-to-treat population. The posted means are lower in the rituximab groups than in the placebo-plus-methotrexate group, and the simple unadjusted recalculation gives different levels of statistical evidence for the two dose comparisons. The sponsor’s own analyses take precedence because the registry record flags a discrepancy and uses rank-based methods rather than the recalculated comparison of means.

The result does not mean that every participant in a group had the group’s average change. Individual scores varied, as reflected by the posted standard deviations. A group mean difference also does not describe the distribution of individual responses or identify which participants experienced a particular outcome.

The result concerns one radiographic endpoint at Week 52. It does not, by itself, establish effects on symptoms, physical function, quality of life, long-term outcomes, or other aspects of rheumatoid arthritis. Nor does a difference in a radiographic score alone establish how important that difference is to patients. Clinical importance and statistical evidence are related but distinct questions.

The trial was randomized and double-blind, features that support a comparative interpretation, but the registry record does not contain every detail needed to assess conduct and analysis. In particular, multiplicity procedures, sensitivity analyses for missing data, and the original statistical plan are not reported. The primary analysis population required at least one post-baseline radiographic evaluation, and missing values were handled by linear interpolation or extrapolation; assumptions behind that approach remain relevant.

The most defensible reading is therefore limited: the registry posts a lower average change in mTSS in both rituximab groups than in the comparator group, while the sponsor’s statistical analyses and the simple recalculations are not identical. These results should not be extended beyond the measured endpoint, population, time point, and analyses described in the record.

18. Why This Trial Matters Statistically

This trial illustrates why the endpoint definition and analysis population matter as much as the treatment-group labels. The primary endpoint is a composite radiographic score formed from erosion and joint-space-narrowing components. Its direction is explicit: higher changes represent worsening on the score. That clarity allows readers to interpret the arm means without confusing a score change with a percentage or event rate.

It also demonstrates that a posted group mean and a sponsor’s inferential test need not tell the same statistical story in the same way. The simple recalculations compare arithmetic means using unadjusted summary statistics, whereas the sponsor reports rank-based and stratified analyses. Those methods can differ in their assumptions and target aspects of the data differently. The discrepancy flag makes the distinction especially important: recalculations help explain the summaries, but the sponsor’s analysis is the trial’s own analysis.

The study has three arms and multiple radiographic outcomes, so interpretation requires attention to both pairwise comparisons and multiplicity. A favorable-looking result in one contrast cannot be read in isolation from the broader testing context when the registry does not report a complete multiplicity plan. Similarly, a planning illustration based on the observed effect is not evidence that the trial had a particular probability of detecting its actual result.

Finally, the record shows how analysis populations and missing-data rules shape the meaning of a radiographic endpoint. The primary population required post-baseline imaging, and linear interpolation or extrapolation was used for missing data. Those choices are part of the estimand’s practical interpretation, while the unavailable missingness details and sensitivity analyses limit how fully robustness can be assessed.

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.

The registry record does not report the original sample-size justification or the trial’s design assumptions for power. The planning illustrations based on the observed effect cannot substitute for that information. The full statistical analysis plan, any interim analyses, and the multiplicity strategy are also not reported in the registry record.

The primary analysis population required a screening radiograph and at least one post-baseline evaluation. Although participants were grouped as randomized and missing data were handled by linear interpolation or extrapolation, the number of participants with missing evaluations, reasons for missingness, and sensitivity analyses are not reported. The assumptions behind interpolation or extrapolation therefore cannot be assessed fully from the registry entry.

The sponsor reports rank-based analyses, but the registry does not give effect estimates or confidence intervals for those analyses. The Van-Elteren strata and full specification are not reported. The simple recalculations are unadjusted comparisons of summary statistics and differ from the sponsor’s analysis; the record explicitly indicates a discrepancy, so the sponsor’s own analysis takes precedence.

There are numerous registered outcomes and analyses, but the record does not describe a complete hierarchy or adjustment for multiple testing. This limits interpretation of secondary p-values. Also, the adverse-event groups have different reported numbers at risk from the primary radiographic groups, and the registry does not provide event exposure-time details or deaths counts. These safety summaries are therefore descriptive and cannot be treated as fully adjusted comparative rates.

Finally, the principal result concerns a radiographic score at Week 52 in a defined population. The registry record does not provide enough information to infer effects on other clinical outcomes or to assess long-term radiographic patterns from the primary result alone.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study to Evaluate Rituximab in Combination With Methotrexate in Methotrexate-Naive Patients With Active Rheumatoid Arthritis — NCT00299104.
• Registry record last updated 2017-07-28; status last verified 2017-06.
• Results first posted 2011-08-10.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

23. Record Summary

This completed, randomized, double-blind, phase 3, parallel-group study enrolled 755 participants with active rheumatoid arthritis who were naive to methotrexate. It compared placebo plus methotrexate with two rituximab dose regimens plus methotrexate. The primary endpoint was change in modified Total Sharp Score from screening to Week 52 in a modified intent-to-treat population.

The posted mean mTSS changes were 1.079 with placebo plus methotrexate, 0.646 with rituximab 0.5 g on each of two infusion days plus methotrexate, and 0.359 with rituximab 1.0 g on each of two infusion days plus methotrexate. Sponsor analyses used Kruskal-Wallis and Van-Elteren methods. The registry record flags a discrepancy with the unadjusted recalculations, so the sponsor’s own analysis takes precedence. Missing-data and multiplicity details are incomplete in the registry record.