Study Comparing Doses Of An Experimental Glucocorticoid Compound To Prednisone And Placebo In Rheumatoid Arthritis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study Comparing Doses Of An Experimental Glucocorticoid Compound To Prednisone And Placebo In Rheumatoid Arthritis (NCT01393639), a randomized phase 2 study. The primary outcome is Percentage of Participants Achieving a 20% Improvement in American College of Rheumatology (ACR) Criteria at Week 8.

2. Clinical Question

The question is about a defined response threshold at a specified assessment time, not about a continuous measure of average change. ACR20 requires improvement in both tender-joint and swollen-joint counts and improvement in at least three of the remaining five ACR core measures. The registry describes those measures as participant pain assessment, participant and physician global assessments of disease activity, self-assessed disability using the HAQ disability index, and C-reactive protein.

The trial also included comparisons with prednisone and a range of secondary assessments over time. The primary endpoint, however, is the Week 8 ACR20 response. The central statistical comparison is the difference in the proportion meeting that response definition between each active-treatment group and placebo. Those comparisons describe the posted data for this trial; they do not by themselves establish how results would generalize to other populations or study settings.

3. Trial Design

This was a randomized, double-blind, parallel-group phase 2 study in participants with rheumatoid arthritis. The registry lists seven groups: four daily doses of PF-04171327, two daily prednisone doses, and placebo. All participants were also to receive background methotrexate. The study medication period lasted eight weeks, followed by a four-week tapering period.

The registry reports enrollment of 323 and a status of completed. The listed study start is 2011-09-27, and primary completion is listed as 2014-06-03. The posted primary outcome is measured at Week 8, before the tapering period is complete. The registry identifies the primary analysis population as participants who were randomized and received at least one dose of randomized study drug; the number for each outcome was the number with observations.

Randomization and double masking are relevant design features because, in general, random allocation aims to balance prognostic factors across treatment groups, while masking can reduce the influence of treatment knowledge on participant reporting, clinical assessment, or trial conduct. The registry does not report enough detail to describe the randomization procedure, allocation concealment, or who specifically was masked. Those details should not be inferred from the broad design labels.

The study assessed both efficacy and safety. Listed secondary outcomes include ACR response thresholds at several assessment times, tender- and swollen-joint counts, and CRP. The registry lists 28 outcome measures and 548 analyses, while 25 secondary outcomes are posted in the record. The original statistical analysis plan, including its approach to multiple testing, is not reported in the registry record. The sample-size justification and any interim analyses are also not reported in the registry record.

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: Participants enrolled who had documented rheumatoid arthritis with a duration of at least 3 months as determined by the investigator using standardly accepted criteria, had received methotrexate for at least 3 months to treat their rheumatoid arthritis, and were free of any signs or symptoms of infection.

Only the first 6 of 7 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set (FAS) included all participants who were randomized to the study and received at least one dose of the randomized study drug (PF 04171327, prednisone, or placebo). N= number of participants with observations.

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.

Only the first 6 of 8 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants Achieving a 20% Improvement in American College of Rheumatology (ACR) Criteria at Week 8, measured over Week 8.

The primary endpoint is the percentage of participants achieving an ACR20 response at Week 8. ACR20 is a composite response definition: a participant must have at least a 20% improvement in tender-joint count and swollen-joint count, as well as at least a 20% improvement in at least three of five other ACR core measures. Those additional measures cover pain, participant and physician global assessments, self-assessed disability, and CRP.

A composite endpoint combines multiple dimensions of a condition into one response classification. In general, this can make the endpoint more clinically informative than a single measure because it requires improvement across several components. It also means that the reported response proportion does not show which components changed for each participant, or the magnitude of change among responders. ACR20 records whether the threshold was met, not the full pattern of symptoms, joint counts, or laboratory results.

The Week 8 time point provides a common assessment point across the randomized groups during the scheduled study-medication period. The endpoint is binary at the participant level: each participant either meets the response definition or does not. The group percentage therefore describes the share with an observed response, and the comparison with placebo expresses the difference between those shares. The registry record does not report the original rationale for selecting this endpoint or the Week 8 assessment as primary.

7. Results Posted in the Registry

• PF-04171327 1 mg: 46.7% (21/45) versus 36.2% (17/47) on control; risk difference 10.5 percentage points (95% CI -9.5 to 30.5), risk ratio 1.29 (0.79–2.11); unadjusted two-proportion z-test p = 0.307.
• PF-04171327 5 mg: 61.7% (29/47) versus 36.2% (17/47) on control; risk difference 25.5 percentage points (95% CI 6.0 to 45.1), risk ratio 1.71 (1.10–2.65); unadjusted two-proportion z-test p = 0.013.
• PF-04171327 10 mg: 68.9% (31/45) versus 36.2% (17/47) on control; risk difference 32.7 percentage points (95% CI 13.4 to 52.0), risk ratio 1.90 (1.24–2.92); unadjusted two-proportion z-test p = 0.002.
• PF-04171327 15 mg: 72.9% (35/48) versus 36.2% (17/47) on control; risk difference 36.7 percentage points (95% CI 18.1 to 55.4), risk ratio 2.02 (1.33–3.06); unadjusted two-proportion z-test p = <0.001.

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 Tender-Joint Counts at Week 12 (Descriptive Statistics)

• 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:
  • PF-04171327 1 mg: mean -6.25 versus -5.59 on control; difference -0.66 (95% CI -3.52 to 2.20), standardized effect d = -0.10; z-test p = 0.651.
  • PF-04171327 5 mg: mean -7.24 versus -5.59 on control; difference -1.65 (95% CI -4.42 to 1.12), standardized effect d = -0.26; z-test p = 0.244.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between PF-04171327 1 mg 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 PF-04171327 5 mg and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Swollen-Joint Counts at Week 12 (Descriptive Statistics)

• 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:
  • PF-04171327 1 mg: mean -6.20 versus -6.17 on control; difference -0.03 (95% CI -2.52 to 2.46), standardized effect d = -0.01; z-test p = 0.981.
  • PF-04171327 5 mg: mean -5.98 versus -6.17 on control; difference 0.19 (95% CI -2.04 to 2.42), standardized effect d = 0.04; z-test p = 0.867.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between PF-04171327 1 mg 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 PF-04171327 5 mg and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline of CRP at Week 12 (Descriptive Statistics)

• 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:
  • PF-04171327 1 mg: mean 0.72 versus -4.07 on control; difference 4.79 (95% CI -6.02 to 15.60), standardized effect d = 0.19; z-test p = 0.385.
  • PF-04171327 5 mg: mean -6.24 versus -4.07 on control; difference -2.17 (95% CI -12.19 to 7.85), standardized effect d = -0.10; z-test p = 0.671.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between PF-04171327 1 mg 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 PF-04171327 5 mg and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes broaden the picture beyond the primary Week 8 ACR20 response. They include other ACR response thresholds, assessed at multiple times, as well as change from baseline in tender-joint counts, swollen-joint counts, and CRP. The registry lists 25 posted secondary outcomes and 28 outcome measures in total. The Week 12 summaries include descriptive statistics for joint counts and CRP, extending observation beyond the primary assessment and into the period following the scheduled tapering phase.

Across these outcome types, the registry presents both binary response measures and continuous changes. They provide different views of disease-related measurements and are not interchangeable with the primary composite response. In broad terms, they can help show whether the primary response pattern is accompanied by changes in other measures, but a consistent statistical conclusion cannot be drawn just from the range of endpoints or the existence of posted descriptive values. The secondary analyses do not replace the prespecified primary question.

Unadjusted secondary results carry less weight because many outcomes and assessment times create many potential comparisons. In general, testing numerous hypotheses without accounting for multiplicity raises the likelihood of chance findings. The registry record does not report the original multiplicity procedure or a hierarchy of secondary hypotheses. For that reason, secondary estimates and nominal p-values should be treated as exploratory descriptions unless their role in the original analysis plan is documented. The posted record does not establish that each secondary analysis was intended as an independent confirmatory test.

Differences in analysis population and available observations can also affect comparisons across outcomes and times. The registry defines the primary Full Analysis Set and notes that outcome counts are based on observations, but full missing-data details for secondary assessments are not reported in the registry record. The secondary results can add context, but they do not by themselves establish a coherent treatment effect across all measured domains.

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

• Percentage of Participants Achieving ACR20 Response at Weeks 2, 4, and 12 (Comparisons to Placebo, and Prednisone 10 mg)
• Percentage of Participants Achieving ACR50 Response at Weeks 2, 4, 6, 8 and 12 (Comparisons to Placebo, and Prednisone 10 mg)
• Percentage of Participants Achieving ACR70 Response at Weeks 2, 4, 6, 8 and 12 (Comparisons to Placebo, and Prednisone 10 mg)
• Change From Baseline in Tender-Joint Counts at Weeks 2, 4, 6, 8 (Comparisons to Placebo, and Prednisone 10 mg)
• Change From Baseline in Tender-Joint Counts at Week 12 (Descriptive Statistics)
• Change From Baseline in Swollen-Joint Counts at Weeks 2, 4, 6, and 8 (Comparisons to Placebo, and Prednisone 10 mg)
• Change From Baseline in Swollen-Joint Counts at Week 12 (Descriptive Statistics)
• Change From Baseline in C Reactive Protein (CRP) at Weeks 2, 4, 6, and 8 (Comparisons to Placebo and Prednisone 10 mg)

9. Statistical Interpretation

10. Statistical Methods Explained

What does the risk difference describe?

The primary endpoint is binary, so the recalculated comparisons use a risk difference: the response proportion in an active group minus the response proportion in placebo. The reported estimate is a simple, unadjusted recalculation from the posted response counts. It is expressed on the proportion scale; a positive value indicates a higher observed response proportion in the named group. No additional adjustment for baseline characteristics or other trial factors is represented in this calculation.

In general, a risk difference is an absolute contrast between two proportions. It answers how far apart the observed proportions are, rather than describing their relative size. The record also includes relative risks and odds ratios in the recalculated comparison data, but the primary interpretation follows the risk difference. The different measures answer related but distinct questions and should not be treated as interchangeable.

What do confidence intervals and p-values add?

A confidence interval describes uncertainty around an estimated comparison. In general, a wider interval indicates less precision, while an interval that includes zero is compatible with no difference at the interval’s stated confidence level. The confidence level for the recalculated primary intervals is not reported in the registry record. A p-value measures how unusual data at least as inconsistent with a no-difference model would be under the assumptions of that test; it is not the probability that a treatment effect is true. The recalculated p-values are unadjusted.

The p-values and intervals should be interpreted together. A small unadjusted p-value does not remove uncertainty about the size of a difference, and a non-small p-value does not prove that the groups are equivalent. The comparison for PF-04171327 1 mg and the comparison for prednisone 5 mg have intervals crossing zero; the corresponding intervals for the other listed active-versus-placebo contrasts do not. These are statistical descriptions of the posted counts, not treatment recommendations.

What is being recalculated?

The primary recalculations use the posted numbers of responders and participants in the compared groups. The count pairs include 17 of 47 for placebo and 21 of 45 for PF-04171327 1 mg; the corresponding pairs are 17 of 47 and 29 of 47 for PF-04171327 5 mg, 17 of 47 and 31 of 45 for PF-04171327 10 mg, and 17 of 47 and 35 of 48 for PF-04171327 15 mg. For prednisone, the posted pairs are 17 of 47 and 23 of 45 for 5 mg, and 17 of 47 and 33 of 46 for 10 mg.

In general, recalculating from summary counts gives a transparent, unadjusted comparison, but it cannot reproduce an analysis that used a different model, covariate adjustment, missing-data handling, or testing procedure. The registry record does not report enough methodological detail about its own primary analysis to establish that it used the same approach. No discrepancy is flagged between the recalculation and the reported material, but the two kinds of results should still be distinguished by their scale and method.

How do the continuous secondary measures differ?

Some posted secondary results are mean changes from baseline in joint counts or CRP. A mean summarizes the average change, while the standard deviation describes variation among participant values. In general, a negative mean change in a count indicates that the measured count was lower at the assessment than at baseline, according to the change-from-baseline convention. A mean difference compares average changes between groups. These summaries are descriptive unless supported by an appropriate inferential analysis.

The registry lists Week 12 descriptive statistics for tender-joint count, swollen-joint count, and CRP. Recalculated comparisons are available for selected PF-04171327 groups against placebo, but these are unadjusted. Their confidence intervals and p-values do not establish that the same methods used in any sponsor analysis were applied. In general, continuous outcomes can be sensitive to skewness, outliers, and the distribution of change scores; the registry record does not report diagnostics or model details for these posted descriptive summaries.

Why does multiplicity matter?

The record contains multiple doses, comparators, outcomes, and assessment times. When many hypotheses are tested, the chance of at least one apparently notable result can increase if tests are interpreted without an adjustment or a prespecified hierarchy. The unadjusted p-values therefore describe each recalculated contrast on its own and should not be read as a family-wise conclusion across the whole trial. The original plan for controlling multiplicity is not reported in the registry record.

Likewise, the registry record does not report interim analyses, the original sample-size justification, or a full statistical analysis plan. These absences prevent reconstruction of the planned decision rules. They do not invalidate the descriptive response percentages, but they limit how confidently the recalculated tests can be connected to the trial’s original confirmatory aims.

11. Confidence Intervals and Precision

The observed response percentages are estimates from groups with outcome denominators ranging from 45 to 48 in the primary comparisons. Estimates based on a limited number of observations can be imprecise: a small change in observed responses could shift the estimated proportion. Confidence intervals make that uncertainty visible by showing a range of values compatible with the data under the interval method.

For the Week 8 risk differences, the intervals for PF-04171327 1 mg and prednisone 5 mg cross zero. Their data are therefore compatible with a range that includes no difference from placebo. For the other active-versus-placebo comparisons, the posted intervals do not cross zero. The size of an estimate and its precision are separate matters: the point estimate is the central calculated contrast, while the interval indicates how uncertain that contrast is.

The confidence level for these recalculated intervals is not reported in the registry record, so it should not be inferred from conventions used elsewhere. The registry also does not report all details needed to determine whether the posted analysis handled missing outcomes, covariates, or multiplicity in a particular way. The primary analysis population is defined as randomized participants who received at least one dose, with the number in each outcome analysis based on participants with observations. This means the outcome denominators are not necessarily the same as enrollment.

Precision is also outcome-specific. Week 12 continuous outcomes have their own group sizes and standard deviations, and their uncertainty is not interchangeable with that of the Week 8 binary endpoint. The smaller sample-size illustrations based on larger observed effects should not be mistaken for evidence that the observed effects are known precisely; such illustrations inherit the uncertainty and possible instability of the estimates used to generate them.

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 in a recalculated comparison, not the trial’s original design calculations. For these binary and continuous outcomes, the listed figures are participants per arm, not total events. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect set at the stated fraction of the observed effect. The original design assumptions are not reported in the registry record.

The 80% power column gives the illustrative number of participants per arm for a design with an 80% chance of detecting the assumed effect under those planning assumptions, if that effect and model assumptions were correct. The 90% power column gives the corresponding planning number for a 90% chance. These are design probabilities under assumptions, not probabilities that the observed result is true. The sample-size figures are not observed power.

Rows that plan for 100% of the observed effect use the full estimated contrast as the target. Rows for 75% or 50% use smaller target effects. Smaller effects are harder to distinguish from random variation, so they require more participants per arm under the same planning framework. This is why the illustrated requirements increase as the target is reduced. The planning curve should be read as reflecting that relationship, not as a forecast of the trial’s actual enrollment needs.

These calculations are especially sensitive to the observed effect used as their starting point. With modest group sizes, an observed difference may be unstable, and a planning illustration based on that estimate can look more optimistic than one based on a smaller target. The figures are useful for understanding how assumed effect size and desired power affect planning; they do not revise or explain the trial’s original design.

13. How the Sample Size Is Derived

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

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.362; treatment proportion p₁ = 0.467; average p̄ = 0.414.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6966 = 1.3653.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6926 = 0.5829.
4. n = (1.3653 + 0.5829)² / (0.1050)² = 344.5, rounded up to 345 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.361702127659574                          # control proportion
p1 <- 0.466666666666667                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(1393639)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 345 (80% power), 461 (90% power)
 75% of observed effect: n per arm = 608 (80% power), 813 (90% power)
 50% of observed effect: n per arm = 1,353 (80% power), 1,810 (90% power)
power.prop.test: 344.5 
Simulated power at n = 345 per arm: 0.792 (20000 replicates)

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

14. Analyses Reported by the Sponsor

The registry includes a set of analyses comparing each PF-04171327 group and each prednisone group with placebo, and selected PF-04171327 groups with prednisone 10 mg. These entries are labeled as mean differences in final values, with 60% confidence intervals. The listed methods and p-values are absent. Because the primary endpoint is a percentage achieving ACR20, those mean-difference entries are not directly expressed on the same scale as the binary response comparisons recalculated from responder counts.

The recalculated primary comparisons are simple, unadjusted risk differences with unadjusted p-values and confidence intervals. They are not a reconstruction of the sponsor’s model. The registry record does not report sufficient detail to determine the population, outcome scale, or statistical method behind the legacy mean-difference entries. No discrepancy is flagged, but the difference in reported scales means the entries should not be treated as direct corroboration of the recalculated risk differences.

The reported confidence intervals are listed as 60%, and no p-values are posted for these entries. The recalculated primary intervals have no confidence level stated in the registry record. Without the model and analysis details, a formal comparison of the two approaches is not possible. The sponsor’s own analysis takes precedence only if a discrepancy is flagged; no such discrepancy is flagged for this trial. That does not make the methods interchangeable or supply missing details about the sponsor analysis.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity is a central interpretive issue because the study includes multiple randomized groups, more than one comparator, several ACR response thresholds, repeated assessment times, and continuous secondary outcomes. If many statistical tests are examined without an adjustment, some may appear notable by chance even when there is no underlying difference. An adjustment, a prespecified testing hierarchy, or a clearly defined primary family of tests can help manage this issue. The registry record does not report which approach, if any, was used for the original analysis.

The recalculated p-values for the primary comparisons are unadjusted. They should be understood as contrast-specific values and not as multiplicity-controlled evidence across all treatment groups and outcomes. The same caution applies to the unadjusted secondary recalculations. The record reports 548 analyses, but does not report a corresponding testing strategy or identify which results were confirmatory versus exploratory. Therefore, the full error rate across the set of analyses cannot be assessed from the registry record.

Missing-data handling is also important. The primary population includes randomized participants who received at least one dose, and the registry says that the analysis number is the number of participants with observations. This is not identical to a statement that every randomized participant contributed an observed Week 8 response. The record does not report the number or reasons for missing primary outcomes, imputation procedures, or sensitivity analyses addressing missingness. For secondary outcomes, the same detailed information is not reported in the registry record.

The randomization and masking labels describe broad design features, but details of allocation procedures, adherence, and analysis implementation are not available in the record. The original sample-size justification, interim analyses, and full statistical analysis plan are also not reported in the registry record. These omissions limit assessment of whether the reported analysis followed prespecified rules and whether the posted estimates are robust to alternative assumptions. They do not change the observed counts, but they constrain confirmatory interpretation.

16. Safety Results

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

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 1% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: coronary artery disease (2.2%), spinal compression fracture (2.2%), intervertebral disc protrusion (2.2%), rheumatoid arthritis (2.2%), bronchitis (2.1%).

The registry lists serious adverse events by group, giving the number at risk and the number affected. The affected counts are 0 of 45 for PF-04171327 1 mg, 1 of 47 for 5 mg, 2 of 45 for 10 mg, and 2 of 48 for 15 mg. For prednisone, the counts are 0 of 45 at 5 mg and 2 of 46 at 10 mg; for placebo, 2 of 47 were affected. These counts describe serious adverse events in the posted safety groups and are not a formal comparison of risk.

Other adverse events are also listed as counts affected: 20 of 45, 18 of 47, 22 of 45, and 16 of 48 across the PF-04171327 groups; 16 of 45 and 19 of 46 for the prednisone groups; and 16 of 47 for placebo. The listed serious-event terms include coronary artery disease, spinal compression fracture, intervertebral disc protrusion, rheumatoid arthritis, bronchitis, and infective cholecystitis, each with the percentages stated in the registry. Death counts are not reported in the registry record.

These summaries do not describe event timing, severity beyond the serious-event classification, exposure duration, or causal attribution. The counts alone cannot establish whether an event was caused by study medication or whether groups differ reliably in safety. Rare-event comparisons can be particularly imprecise, and the registry does not report a statistical safety comparison.

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

The primary result tells readers what share of participants with observed Week 8 data met the ACR20 definition in each randomized group. The posted percentages and recalculated risk differences show the size and direction of the observed contrasts with placebo, while the intervals show uncertainty around those estimates. The comparisons describe this trial’s data; they do not establish a universal effect or guarantee the same response proportions in another setting.

The primary endpoint is a composite threshold. It does not show the detailed contribution of each ACR component, the magnitude of improvement beyond the threshold, or whether every participant experienced improvement in all measured domains. It also does not summarize the full course of outcomes during the four-week tapering period. Secondary results add other measures and assessment times, but the registry does not report enough analysis detail to treat the entire group of outcomes as a single confirmatory body of evidence.

The unadjusted comparisons should not be read as final proof based only on whether a p-value is on one side or another of a threshold. The intervals, group sizes, missing-data information, multiplicity concerns, and absence of the original statistical plan all matter. In general, statistical significance does not measure clinical importance, and a result that is not statistically significant does not prove equivalence. The registry record does not report a clinically important difference threshold for ACR20.

These data also do not determine an individual’s likely response or provide a basis for choosing a treatment. They are a group-level report from a randomized phase 2 study with the stated population, background treatment, schedule, and assessment definitions. The safety counts are descriptive and do not establish causation. Interpretation should remain confined to the outcomes, time points, and methods actually documented in the registry record.

18. Why This Trial Matters Statistically

This trial is statistically informative because it places several doses of an experimental compound alongside two prednisone doses and placebo within a randomized, double-blind, parallel-group design. Multiple groups allow the posted response proportions to be viewed across dose levels and comparators, while the placebo contrasts give a common reference for the primary endpoint. The Week 8 ACR20 definition also makes clear that the main result is a response classification combining joint counts and other core measures.

The record illustrates why point estimates should be read with uncertainty. The observed proportions vary across groups, but the confidence intervals for some contrasts include zero while others do not. A pattern across doses may be useful descriptively, yet it is not itself a formal test of a dose-response trend. The separate contrasts also do not substitute for direct comparisons between active-treatment groups.

The study is also a useful example of the distinction between registry-reported analyses and unadjusted recalculations from posted counts. The primary recalculation has a clear scale and method, but it cannot reproduce an analysis whose model is not described. The record’s many outcomes and analyses raise multiplicity questions, and its missing details about the planned analysis limit reconstruction of the original statistical decision process.

Finally, the sample-size illustrations show how planning depends on assumed effect size and desired power. Since they are based on observed effects and are not the original study design, they should be read as planning scenarios rather than evidence about the power achieved in the completed trial. Together, these features underscore the importance of separating observed results, uncertainty, design planning, and the original analysis plan.

19. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The primary response group sizes are modest, and the outcome denominators vary across arms. The registry defines the analysis population as randomized participants who received at least one dose, with each outcome count based on observations; it does not report detailed reasons for missing primary outcomes or sensitivity analyses for missing data. That limits assessment of whether missingness could have influenced the observed response proportions.

The recalculated comparisons are unadjusted and do not account for multiple primary contrasts or the broader set of secondary tests. The registry record does not report the original multiplicity strategy, complete statistical analysis plan, or original sample-size justification. It also does not report interim analyses. These omissions make it difficult to determine which comparisons were intended to be confirmatory or whether the posted analysis followed a prespecified hierarchy.

The sponsor-reported entries are labeled as mean differences in final values and have 60% confidence intervals, but the method and p-values are not listed. Those entries are on a different reported scale from the recalculated binary risk differences, and the registry does not provide enough context to reconcile them with the primary endpoint.

Generalizability is bounded by the studied population and regimen: participants with rheumatoid arthritis, background methotrexate, the listed treatment groups, and the eight-week medication period followed by tapering. The registry does not provide detailed information about participant characteristics, adherence, or the randomization procedure in the registry record for this analysis. Safety summaries are counts and do not describe timing or causal attribution. Secondary analyses span outcomes and time points, but are not accompanied by a reported multiplicity plan.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Study Comparing Doses Of An Experimental Glucocorticoid Compound To Prednisone And Placebo In Rheumatoid Arthritis — NCT01393639.
• Registry record last updated 2024-09-04; status last verified 2024-08.
• Results first posted 2024-09-04.
• 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

The registry describes a completed randomized, double-blind, parallel-group phase 2 trial in rheumatoid arthritis, with enrollment of 323. Groups received one of four daily PF-04171327 doses, prednisone at 5 mg or 10 mg daily, or placebo, alongside background methotrexate. Study medication was scheduled for eight weeks, followed by a four-week tapering period.

The primary endpoint was the Week 8 percentage achieving ACR20. Posted response values ranged from 37.0 for placebo to 73.0 for PF-04171327 15 mg. Unadjusted recalculated comparisons show positive risk-difference estimates versus placebo for each active group, with intervals crossing zero for PF-04171327 1 mg and prednisone 5 mg. The registry also lists secondary outcomes and descriptive safety counts.

The recalculated comparisons do not replace the sponsor’s analysis. The registry record does not report the original multiplicity plan, sample-size justification, interim analyses, or full statistical analysis plan. The observed results should therefore be read with their uncertainty and the limits of the posted methods in mind.