A Study of the Safety and Efficacy of Golimumab (CNTO 148) in Subjects With Active Rheumatoid Arthritis Previously Treated With Biologic Anti-TNFa Agent(s): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of the Safety and Efficacy of Golimumab (CNTO 148) in Subjects With Active Rheumatoid Arthritis Previously Treated With Biologic Anti-TNFa Agent(s) (NCT00299546), a randomized phase 3 study. The primary outcome is American College of Rheumatology (ACR) 20 Response at Week 14..

2. Clinical Question

The trial addressed a population with prior exposure to biologic anti-TNFa treatment, rather than a general rheumatoid arthritis population. That eligibility context is important when considering how directly these results might apply to other groups; the registry record does not describe evidence in people without this treatment history.

The primary comparison concerns a response threshold at a specified assessment time. It is not a comparison of long-term disease course, and the primary outcome does not by itself describe safety or every aspect of disease activity. Secondary outcomes add information on other response definitions and a disease activity measure, but each addresses a different question and should be interpreted on its own terms.

3. Trial Design

This was a phase 3, randomized, parallel-group, quadruple-masked, placebo-controlled trial. The official title describes it as multicenter. The registry record lists enrollment as 461 and the study status as completed. The listed start was 2006-02, and primary completion was 2007-08.

Participants were assigned to placebo, golimumab 50 mg, or golimumab 100 mg. The described injections were subcutaneous and generally scheduled every 4 weeks. The placebo description includes early escape at Week 16 and a transition to golimumab under specified conditions. The 50 mg description also allows early escape and dose adjustment, while the 100 mg description notes later dose adjustment after unblinding. The blinded period for the 100 mg group is described as lasting until the Week 24 database lock. These features mean that the randomized groups describe initial assignments, while the treatment histories could change during follow-up.

The primary outcome was assessed at Week 14. The registry describes its analysis population as randomized participants excluding 1 site. It also states that participants were considered nonresponders if they used prohibited medications or discontinued subcutaneous study agent because of lack of efficacy. Missing ACR components were imputed using Last Observation Carried Forward unless all components were missing; the handling of that latter circumstance is not reported in the registry record.

The record does not report the trial’s original sample-size justification, detailed randomization procedures, the masking roles, or whether interim analyses were planned or conducted. It also does not provide enough detail to reconstruct every treatment transition or analysis decision. The posted primary analysis counts and reported statistical tests should therefore be read alongside these limits in the available design description.

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: Randomized participants (excluding 1 site). Participants considered non-responders if used any prohibited medications or discontinued subcutaneous study agent due to lack of efficacy. Missing ACR components imputed by Last Observation Carried Forward unless all ACR components were missing; in which case considered non-responders.

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 American College of Rheumatology (ACR) 20 Response at Week 14., measured over Week 14.

The primary endpoint is the number of participants with an ACR 20 response at Week 14. The registry defines response as improvement of at least 20% from baseline in both tender and swollen joint counts, together with improvement in at least 3 of 5 assessments: the participant’s pain visual analog scale, the participant’s global assessment of disease activity, the physician’s global assessment of disease activity, the Health Assessment Questionnaire, and C-reactive protein.

This composite definition brings joint findings and other assessments into a single responder classification. A threshold outcome can be useful for comparing how many participants in each group meet a prespecified level of improvement, but it does not show the amount of change for every participant. People just below the threshold and people with much larger changes are both classified as nonresponders and responders, respectively. The responder count therefore complements, rather than replaces, continuous measures and other clinical outcomes.

The endpoint is also tied to a particular assessment time. It describes response at Week 14 under the registry’s stated analysis rules, not the duration of response or the outcome at every later visit. Interpretation should account for the stated nonresponder rules for prohibited medication use and discontinuation due to lack of efficacy, as well as the missing-component imputation procedure. The record’s description of the handling of wholly missing ACR components is incomplete, so that part of the endpoint implementation is not reported in the registry record.

As a general statistical point, a binary endpoint simplifies a complex clinical picture into whether a defined criterion was met. That makes group comparisons direct, but it also makes the result dependent on the chosen threshold and the rules for classifying missing or intercurrent observations.

7. Results Posted in the Registry

• Group 2: Golimumab 50 mg: 34.7% (51/147) versus 18.0% (27/150) on control; risk difference 16.7 percentage points (95% CI 6.8 to 26.5), risk ratio 1.93 (1.28–2.90); unadjusted two-proportion z-test p = 0.001.
• Group 3: Golimumab 100 mg: 36.5% (54/148) versus 18.0% (27/150) on control; risk difference 18.5 percentage points (95% CI 8.6 to 28.4), risk ratio 2.03 (1.36–3.03); unadjusted two-proportion z-test p = <0.001.
• Combined Golimumab: 35.6% (105/295) versus 18.0% (27/150) on control; risk difference 17.6 percentage points (95% CI 9.4 to 25.8), risk ratio 1.98 (1.36–2.88); 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.

American College of Rheumatology (ACR) 50 Response at Week 14

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Group 2: Golimumab 50 mg: 15.0% (22/147) versus 6.7% (10/150) on control; risk difference 8.3 percentage points (95% CI 1.3 to 15.3), risk ratio 2.24 (1.10–4.58); unadjusted two-proportion z-test p = 0.021.
  • Group 3: Golimumab 100 mg: 18.2% (27/148) versus 6.7% (10/150) on control; risk difference 11.6 percentage points (95% CI 4.2 to 19.0), risk ratio 2.74 (1.37–5.45); unadjusted two-proportion z-test p = 0.002.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Group 2: Golimumab 50 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Group 3: Golimumab 100 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Disease Activity Index Score 28 (DAS 28) (Using C-reactive Protein) Response at Week 14

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Group 2: Golimumab 50 mg: 55.8% (82/147) versus 29.3% (44/150) on control; risk difference 26.4 percentage points (95% CI 15.6 to 37.3), risk ratio 1.90 (1.43–2.53); unadjusted two-proportion z-test p = <0.001.
  • Group 3: Golimumab 100 mg: 58.8% (87/148) versus 29.3% (44/150) on control; risk difference 29.5 percentage points (95% CI 18.7 to 40.2), risk ratio 2.00 (1.51–2.66); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Group 2: Golimumab 50 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Group 3: Golimumab 100 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

American College of Rheumatology (ACR) 20 at Week 24

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Group 2: Golimumab 50 mg: 31.3% (46/147) versus 16.0% (24/150) on control; risk difference 15.3 percentage points (95% CI 5.8 to 24.8), risk ratio 1.96 (1.26–3.03); unadjusted two-proportion z-test p = 0.002.
  • Group 3: Golimumab 100 mg: 42.6% (63/148) versus 16.0% (24/150) on control; risk difference 26.6 percentage points (95% CI 16.7 to 36.5), risk ratio 2.66 (1.76–4.02); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Group 2: Golimumab 50 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Group 3: Golimumab 100 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The posted secondary outcomes cover ACR 50 response and DAS 28 response at Week 14, ACR 20 response at Week 24, and the HAQ score at Week 24. Across the binary response outcomes, the broad pattern is that the golimumab groups have more participants recorded as responders than placebo. The secondary results therefore point in a direction consistent with the primary Week 14 ACR 20 comparison, while addressing stricter response criteria, a disease activity measure, or a later assessment time.

These outcomes are related but not interchangeable. ACR 50 uses a different response threshold from ACR 20. DAS 28 response reflects a separate disease activity definition, and the Week 24 measures address a later time point than the primary endpoint. HAQ is listed as a secondary outcome, but the posted record included in this analysis does not give its numerical arm results or comparison estimates. Those results are not reported in the registry record available for this analysis.

The sponsor’s posted secondary analyses include Cochran-Mantel-Haenszel p-values for comparisons against placebo. The recalculated binary comparisons likewise give unadjusted estimates and p-values. These are secondary findings and should not be read as independent confirmations of the primary result: outcomes from the same trial may be correlated, and several comparisons create more opportunities for chance findings.

Multiplicity refers, in general, to the increased chance of obtaining at least one apparently unusual result when many hypotheses are tested. A prespecified adjustment can control that issue, but the registry does not report whether a multiplicity strategy was planned or used for these secondary outcomes. Unadjusted secondary results consequently carry less weight than the primary result when forming an overall statistical interpretation. They are useful for describing consistency and generating questions, but should not be treated as definitive evidence for every outcome or dose.

The registry lists five outcome measures and 15 analyses overall. The specific relationship of all listed analyses to a formal testing hierarchy is not reported in the registry record. Interpret the secondary pattern as supportive context, with attention to the outcome definitions, time points, and absence of reported multiplicity handling.

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

• American College of Rheumatology (ACR) 50 Response at Week 14
• Disease Activity Index Score 28 (DAS 28) (Using C-reactive Protein) Response at Week 14
• American College of Rheumatology (ACR) 20 at Week 24
• Health Assessment Questionnaire (HAQ) Score at Week 24

9. Statistical Interpretation

10. Statistical Methods Explained

The main effect measure used in the recalculations is the risk difference: the response proportion in a golimumab group minus the response proportion in the placebo group. A positive value means the observed response proportion is higher in the golimumab group. The risk difference is expressed as a proportion difference, not as a ratio. It describes the size of the contrast on the absolute scale and depends on the response frequency in both groups.

What does the interval around a risk difference mean?

As a general statistical explanation, a confidence interval expresses uncertainty in an estimate by giving a range of values compatible with the observed data and the method’s assumptions. Narrower intervals indicate greater precision for the effect measure; wider intervals indicate less precision. An interval that includes zero is compatible with no difference under that interval procedure, while an interval entirely on one side of zero is less compatible with a zero difference. The registry-derived recalculations give the interval limits, but the confidence level is not reported in the registry record.

The recalculated values are simple, unadjusted comparisons based on the posted counts. They do not reproduce the full Cochran-Mantel-Haenszel analysis reported by the sponsor. The registry does not give the strata or other analysis details needed to reconstruct that test. The sponsor’s reported analysis and p-values are therefore the appropriate reference for the sponsor’s inferential result, while the recalculations offer a transparent descriptive comparison from the posted summaries.

How should the p-values be read?

In general, a p-value measures how unusual data at least as incompatible with a specified null hypothesis would be if that null model and its assumptions held. It is not the probability that the null hypothesis is true, nor does it quantify the magnitude or importance of a group difference. A small p-value can accompany an effect estimate that is imprecise or modest; effect estimates and intervals are needed to understand magnitude and precision. The registry reports the sponsor’s primary comparisons using the Cochran-Mantel-Haenszel method, and the recalculated comparisons give unadjusted p-values from the posted summaries.

The risk ratio and odds ratio are additional relative measures listed in the recalculations. In general, a risk ratio compares response proportions, while an odds ratio compares response odds. Both are relative measures and can give a different impression of scale from a risk difference. The listed estimates and intervals should be interpreted as accompanying measures, not as independent trials or additional participants. Their values are not necessary to understand the absolute-scale risk-difference comparison.

How do the endpoint rules affect the analysis?

The primary outcome is binary, recorded as the number of participants meeting the ACR 20 criterion. The analysis population excludes 1 site, according to the registry description. Participants who used prohibited medication or discontinued the study agent because of lack of efficacy were classified as nonresponders. For missing ACR components, the registry describes Last Observation Carried Forward unless all ACR components were missing. The further rule for that case is not reported in the registry record. These decisions affect which observations contribute to the endpoint classification and how missing information is represented.

As a general explanation, Last Observation Carried Forward uses an earlier observed value in place of a later missing value. This can make an analysis depend on the assumption that the carried-forward observation is an adequate representation of the missing later measurement. It does not remove uncertainty about missing data. The record does not provide enough detail to evaluate alternative missing-data analyses or the sensitivity of the primary finding to those choices.

The primary comparison is also subject to the design’s treatment changes. Early escape, dose adjustment, and unblinding are mentioned in the arm descriptions. The registry’s primary outcome rules address some events by classifying certain participants as nonresponders, but the available record does not fully describe every intercurrent event or the precise analysis strategy for each one. Results should be understood as applying to the stated endpoint and analysis rules rather than to an unchanging treatment exposure.

As a general statistical point, randomized assignment supports a comparison of groups as assigned, but interpretation still depends on the analysis population and post-assignment events. A registry entry’s brief methods summary may not contain the detail necessary to evaluate all assumptions. The record does not report the original statistical analysis plan, the sample-size justification, or whether interim analyses occurred. Those omissions limit independent assessment of prespecification and the full analytic decision process.

11. Confidence Intervals and Precision

The risk-difference intervals for the primary endpoint are 0.068 to 0.265 for golimumab 50 mg versus placebo, 0.086 to 0.284 for 100 mg versus placebo, and 0.094 to 0.258 for combined golimumab versus placebo. The corresponding estimates are 0.167, 0.185, and 0.176. These intervals describe uncertainty in absolute differences based on the unadjusted recalculations from the posted summary statistics.

Each interval is entirely above zero. This indicates that the recalculated data are not centered on a zero difference and that the range of values covered by each interval is positive. It does not establish that every plausible effect is large, nor does it determine whether an effect is clinically important. Clinical importance requires context beyond a statistical interval, and the registry record does not provide a threshold for interpreting the risk difference as clinically meaningful.

Precision depends in part on the amount of information in the comparison and on the underlying event frequencies. As a general statistical explanation, more information often narrows an interval, while sparse outcomes or substantial variability can leave a wider range of compatible values. The intervals should be considered with the event counts, rather than as stand-alone evidence. The 100 mg and 50 mg intervals overlap; that overlap is not itself a formal test of the difference between doses. A direct dose comparison and its uncertainty are not reported in the registry record.

The interval confidence level is not reported in the registry record, and these recalculated intervals are not the sponsor’s reported Cochran-Mantel-Haenszel intervals. The sponsor posts p-values for the primary comparisons but no corresponding effect estimate or interval in the reported analyses. Accordingly, the recalculated intervals are useful for describing the precision of the simple risk differences, but should not be represented as intervals from the sponsor’s adjusted analysis.

The number of decimal places in an estimate does not remove uncertainty. In general, an interval is more informative than a point estimate alone because it communicates a range of plausible effect sizes under the calculation’s assumptions. The analysis population and missing-data rules remain important when interpreting that range.

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 assumptions. For these binary outcomes, the entries 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, and a normal approximation. The effect to detect is 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.

The 80% power column describes the planned number of participants per arm for a hypothetical design that would have an 80% chance of detecting the specified effect under the stated assumptions, if that effect and those assumptions were correct. The 90% power column describes the corresponding planning target for a higher chance of detecting that same specified effect. These are design calculations, not evidence about the probability that the observed trial result is true.

Each planning row uses a different fraction of the observed effect. Planning for the full observed effect uses a larger assumed difference than planning for 75% or 50% of that effect. Smaller target effects are harder to distinguish from random variation and therefore require more participants per arm under the same assumptions. The curve can be read as a visual representation of that relationship: as the effect targeted for detection becomes smaller, the participant requirement rises.

These figures are illustrations rather than a reconstruction of the trial’s enrollment plan. They use the observed data to set a hypothetical target, and the actual study had three randomized groups and treatment procedures that included early escape and dose changes. The listed calculations assume equal allocation between two arms, so they should not be read as a direct account of the trial’s multiarm design. This is not observed power; it is a planning calculation for hypothetical future designs.

The values are sensitive to the assumptions used, including the effect size chosen, the allocation pattern, the significance threshold, and the approximation. They do not establish that a given number of participants would guarantee a particular result. The original protocol’s sample-size rationale is not reported in the registry record.

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 = 0.180; treatment proportion p₁ = 0.347; average p̄ = 0.263.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6230 = 1.2210.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6117 = 0.5148.
4. n = (1.2210 + 0.5148)² / (0.1669)² = 108.1, rounded up to 109 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 <- 0.18                          # control proportion
p1 <- 0.346938775510204                          # 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(299546)
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 = 109 (80% power), 145 (90% power)
 75% of observed effect: n per arm = 183 (80% power), 245 (90% power)
 50% of observed effect: n per arm = 388 (80% power), 519 (90% power)
power.prop.test: 108.1 
Simulated power at n = 109 per arm: 0.809 (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 reports the sponsor’s primary comparisons using the Cochran-Mantel-Haenszel method. The reported p-values are 0.001 for placebo versus golimumab 50 mg, <0.001 for placebo versus golimumab 100 mg, and <0.001 for placebo versus combined golimumab. The registry does not provide an effect estimate or confidence interval with those reported tests.

The recalculated risk differences use the posted counts and are unadjusted. Their p-values are 0.001 for the 50 mg comparison and <0.001 for the 100 mg and combined comparisons. These agree with the sponsor’s reported p-values at the displayed precision. The discrepancy flag is false, so the registry summaries do not indicate a disagreement requiring resolution. The sponsor’s own analysis takes precedence over the simple recalculations.

The two approaches answer related but not identical statistical questions: the recalculation is a transparent comparison from summary counts, whereas the sponsor reports a Cochran-Mantel-Haenszel test. The registry does not describe the strata or enough details to reproduce that analysis, so the recalculated risk differences should not be presented as adjusted sponsor estimates. The absence of sponsor-reported effect estimates and intervals means the posted record does not allow a direct comparison of effect-size estimates from the two approaches.

The reported p-values do not reveal the full analysis specification, nor do they show whether any multiplicity procedure was used. Such details are not reported in the registry record. The results should therefore be cited with their method and scope in mind, rather than treating matching p-values as proof that the methods are interchangeable.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies a primary outcome and several secondary outcomes, and lists multiple analyses. It reports p-values for the primary comparisons and some secondary comparisons, but does not describe a prespecified hierarchy among hypotheses or state whether multiplicity adjustment was applied. That information is not reported in the registry record. Without it, the family of tests intended to support formal conclusions cannot be reconstructed from the posted results alone.

As a general statistical explanation, testing several outcomes or comparisons increases the chance that at least one result will appear unusual by chance, even when no true effects exist. Methods that account for multiple testing can change how evidence is interpreted. Unadjusted secondary p-values should therefore be regarded as exploratory or supportive unless a suitable analysis plan is documented. The record’s primary endpoint designation helps identify the central question, but does not by itself clarify how secondary tests were handled.

The primary population description states that randomized participants were analyzed, excluding 1 site. Participants were classified as nonresponders if they used prohibited medications or discontinued subcutaneous study agent due to lack of efficacy. For missing ACR components, Last Observation Carried Forward was used unless all components were missing. The rule applied when all components were missing is not reported in the registry record. The record does not provide a fuller account of missingness, reasons for unavailable measurements, or sensitivity analyses.

As a general explanation, missing-data procedures rely on assumptions about how missing observations relate to observed outcomes. Carrying a prior value forward does not make missing information equivalent to observed follow-up and can affect estimates if the assumed course differs from actual outcomes. The available description is not sufficient to judge that assumption or compare it with alternative methods.

The trial descriptions include early escape, dose changes, and later unblinding. These events may affect how a treatment-policy question is framed, but the registry does not report a complete estimand or detailed handling strategy for each event. The original statistical analysis plan, interim-analysis approach, and detailed sample-size justification are also not reported in the registry record. These omissions limit independent evaluation of prespecification, multiplicity control, and robustness to analysis choices.

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 (5.1%), osteoarthritis (4.1%), rheumatoid arthritis (4.1%), sepsis (2.6%), urinary tract infection (2.6%).

The registry presents serious adverse event figures for groups described by golimumab exposure: 50 mg injections only, 100 mg injections only, and both 50 mg and 100 mg injections. The recorded serious adverse event counts are 34 among 98 at risk, 46 among 138 at risk, and 71 among 195 at risk, respectively. These are safety-group descriptions and should not be confused with the randomized placebo and dose groups in the primary efficacy analysis.

The corresponding other adverse event counts are 76 among 98 at risk, 117 among 138 at risk, and 172 among 195 at risk. The registry lists serious adverse event terms including pneumonia at 5.1%, osteoarthritis at 4.1%, rheumatoid arthritis at 4.1%, sepsis at 2.6%, urinary tract infection at 2.6%, and basal cell carcinoma at 2.2%. The available term listing does not provide the denominator or additional detail for each percentage.

These counts describe events recorded in the registry, but do not by themselves establish that treatment caused an event or allow a controlled safety comparison with placebo. The safety groups differ from the randomized efficacy groups, and the record does not provide enough detail to align their exposure periods or fully characterize event ascertainment. Death counts are not reported in the registry record. The figures should be read descriptively, not as a complete assessment of safety.

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

The primary result describes how many participants met a defined ACR 20 response criterion at Week 14 under the registry’s analysis rules. The posted counts and unadjusted risk differences show a higher observed response proportion in each golimumab comparison group than in placebo. The sponsor’s reported Cochran-Mantel-Haenszel p-values are consistent with the recalculated p-values at the displayed precision. The uncertainty intervals describe a range of effect sizes for the unadjusted comparisons, rather than certainty about an individual participant’s outcome.

These findings do not, on their own, establish the clinical importance of the difference. The registry does not provide a threshold for a clinically meaningful risk difference, and a statistical comparison does not answer every question about function, symptoms, durability, or harms. The Week 14 endpoint also does not describe outcomes over the entire period of treatment or determine how results would differ under other treatment schedules.

The result is specific to participants with active rheumatoid arthritis who had previously received a biologic anti-TNFa agent, to the listed treatment options, and to the population and endpoint rules in the record. It should not automatically be generalized to people with different treatment histories or to settings with other protocols. The record describes a randomized trial, but does not include enough information to assess every design and analysis detail.

The secondary outcomes provide additional context, but several tests and outcomes are involved, and multiplicity handling is not reported. The safety results are descriptive and organized by exposure group rather than the primary randomized comparison. Neither the efficacy summary nor the safety counts alone determine an individual treatment decision. They are statistical descriptions of this trial’s posted data, not personalized medical guidance.

Important unanswered questions include the sponsor’s full effect estimates and intervals, the handling of wholly missing ACR components, whether multiplicity was controlled, the original sample-size rationale, and the details of longer-term outcomes. These matters are not reported in the registry record. Their absence should be distinguished from evidence that an analysis or procedure was not performed.

18. Why This Trial Matters Statistically

This trial is statistically informative because it studies a defined group with prior exposure to biologic anti-TNFa treatment and compares two golimumab dose assignments with placebo. The primary endpoint is a clearly described binary response threshold at Week 14, and the registry posts arm counts that make a direct absolute-scale comparison possible. Displaying both counts and recalculated risk differences helps readers distinguish the observed data from the size of the estimated contrast.

The distinction between the sponsor’s Cochran-Mantel-Haenszel tests and unadjusted recalculations is also important. Agreement in the displayed p-values does not make the procedures identical, but it offers a useful consistency check. The sponsor’s reported analysis remains the primary inferential reference, while the recalculations make the scale and uncertainty of the count-based comparisons more transparent.

The study also illustrates why a statistical result depends on more than a p-value. The endpoint’s response threshold, missing-component rules, analysis population, treatment escape provisions, and dose changes all shape what the comparison means. Confidence intervals add information about precision, and the sample-size illustrations show how planned enrollment requirements depend on the effect a study is designed to detect.

At the same time, the record leaves important methodological questions open. The multiplicity strategy, original sample-size justification, full handling of missing components, and detailed sponsor effect estimates are not reported. A careful reading therefore separates what the posted counts and tests support from what cannot be established from the registry entry. That distinction is central to responsible statistical interpretation of clinical trial records.

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’s primary analysis population excludes 1 site, but the reason for exclusion and its potential effect on the results are not reported in the registry record. The endpoint description says participants were considered nonresponders after prohibited medication use or discontinuation due to lack of efficacy. The handling of all-missing ACR components is incomplete, and no sensitivity analysis is reported. These details limit assessment of how robust the primary comparison is to the analysis rules.

The sponsor reports Cochran-Mantel-Haenszel p-values but does not post corresponding effect estimates or intervals in the reported analyses. The recalculated risk differences are unadjusted, and the registry does not provide enough information to reproduce the sponsor’s method or identify its strata. The confidence level for recalculated intervals is not reported in the registry record. The discrepancy flag is false, but agreement in p-values does not fill those reporting gaps.

Multiplicity handling is not reported despite the presence of multiple outcomes and analyses. The record also does not report a detailed statistical analysis plan, the original sample-size justification, or whether interim analyses were conducted. These omissions prevent a full evaluation of prespecification and of how secondary findings should be weighted.

Arm descriptions include early escape, dose adjustment, and unblinding, while the primary endpoint is assessed at Week 14 and another listed endpoint is at Week 24. The available record does not fully describe the analysis strategy for every post-randomization treatment change. The safety data are grouped by golimumab exposure, not the original randomized efficacy groups, and deaths are not reported in the registry record. These safety summaries therefore cannot be used as a complete randomized benefit-harm comparison.

Finally, the posted binary response outcomes do not show the distribution of individual changes or establish durability beyond their stated time points. The record describes longer treatment exposure in the arm descriptions but does not provide the long-term outcome results in the material analyzed. These limitations concern the scope and reporting detail of the available record, not proof that unreported procedures or outcomes were absent from the full study documentation.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study of the Safety and Efficacy of Golimumab (CNTO 148) in Subjects With Active Rheumatoid Arthritis Previously Treated With Biologic Anti-TNFa Agent(s) — NCT00299546.
• Registry record last updated 2014-02-27; status last verified 2014-01.
• Results first posted 2011-09-28.
• 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

NCT00299546 is a completed phase 3 randomized, quadruple-masked, parallel-group, placebo-controlled study of golimumab in active rheumatoid arthritis after prior biologic anti-TNFa treatment. The registry lists enrollment as 461 and describes placebo, golimumab 50 mg, and golimumab 100 mg groups.

The primary endpoint was ACR 20 response at Week 14. The posted counts were 27 among 150 for placebo, 51 among 147 for 50 mg, and 54 among 148 for 100 mg. Sponsor-reported Cochran-Mantel-Haenszel p-values were 0.001, <0.001, and <0.001 for the respective comparisons shown in the record. Unadjusted risk-difference recalculations are positive, with intervals entirely above zero.

The record also lists secondary efficacy outcomes and descriptive adverse event summaries. Multiplicity handling, the original sample-size justification, and some missing-data details are not reported in the registry record. The results should be interpreted in light of the specified endpoint, analysis population, and reporting limits.