Study Of The Effects Of Atorvastatin On Cholesterol Levels In Rheumatoid Arthritis Patients Taking CP-690,550: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study Of The Effects Of Atorvastatin On Cholesterol Levels In Rheumatoid Arthritis Patients Taking CP-690,550 (NCT01059864), a randomized phase 2 study. The primary outcome is Percent Change From Baseline (Week 6) in Low Density Lipoprotein-Cholesterol (LDL-C) Level at Week 12.

2. Clinical Question

3. Trial Design

This was a randomized, parallel-group phase 2 study in patients with active rheumatoid arthritis. The registry lists 111 enrolled participants and a completed study. Participants received CP-690,550 throughout the 12-week study, alongside instructions to follow a therapeutic lifestyle diet. At Week 6, they were assigned to receive atorvastatin or a matching placebo for the remainder of the study.

The design therefore has an initial period in which participants were receiving CP-690,550 before the atorvastatin-versus-placebo comparison began. The primary endpoint uses baseline at Week 6 and assessment at Week 12. Its comparison addresses the change over that later period between the two randomized groups. It does not, by itself, estimate the effect of starting CP-690,550, since the registry describes both groups as receiving it.

The registry identifies the study as randomized and lists quadruple masking. It also describes atorvastatin as double-blind relative to its matching placebo. The record does not give the detailed masking roles or describe the allocation procedure beyond randomization. Arm descriptions are blank, so additional operational details about assignment are not reported in the registry record.

The primary analysis was a continuous-outcome comparison of least-squares means, with standard errors reported for each group. The endpoint-specific analysis counts were 46 for CP-690,550 plus atorvastatin and 45 for CP-690,550 plus placebo. These are the participants analyzed for this measure, not a restatement of the trial enrollment. The record does not report the original sample-size justification, the planned analysis model, or whether interim analyses were conducted.

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.

Open-Label Run-In Phase

Reasons for not completing this period:

Double-Blind Treatment Phase

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 treatment groups and received at least one dose of the randomized investigational drug (atorvastatin or placebo). Here 'N' (Number of Participants Analyzed) signifies those participants who were evaluable for this measure.

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 Percent Change From Baseline (Week 6) in Low Density Lipoprotein-Cholesterol (LDL-C) Level at Week 12, measured over Baseline (Week 6), Week 12.

The primary endpoint is percent change in low density lipoprotein cholesterol (LDL-C) from baseline at Week 6 to Week 12. LDL-C is a lipid measure, and expressing its change as a percentage relates the observed change to the participant’s starting level for this comparison period. The endpoint thus focuses on change after the study groups began receiving atorvastatin or placebo, rather than on change across the full 12 weeks.

The registry reports group least-squares means for the percent change and standard errors. A least-squares mean is a model-based estimate of a group mean; it can account for terms included in a statistical model. The exact model and adjustment variables are not reported in the registry record, so the degree and nature of any adjustment cannot be specified from the record. The group estimates are accompanied by a between-group mean difference and a confidence interval.

A continuous endpoint retains information about the magnitude of change rather than reducing outcomes to categories such as improved or not improved. For interpretation, the relevant contrast is the estimated difference between the randomized groups, with the confidence interval indicating the uncertainty around that contrast. The measure is a lipid endpoint; it does not alone establish effects on rheumatoid arthritis activity or other clinical outcomes.

7. Results Posted in the Registry

• CP-690,550 + Atorvastatin: mean -35.34 versus 5.80 on control; difference -41.14 (95% CI -47.40 to -34.88), standardized effect d = -2.70; 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 Low Density Lipoprotein-Cholesterol (LDL-C) at Week 12

• 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 errors.
• Result:
  • CP-690,550 + Atorvastatin: mean -51.49 versus 5.30 on control; difference -56.79 (95% CI -65.12 to -48.46), standardized effect d = -2.80; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for CP-690,550 + Atorvastatin than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes cover additional views of lipid change, lipoprotein characteristics, and rheumatoid arthritis disease activity. The posted secondary LDL-C change result is directionally consistent with the primary percent-change result. That agreement is useful context because it shows a similar pattern on another LDL-C scale, but it is not an independent confirmation: both measures come from the same trial and closely related outcome data.

The registry lists other lipid-profile measures, including particle size and particle levels, as well as disease activity scores using joint counts and inflammatory markers. The record does not post analyses for those other secondary outcomes. Their listing indicates what was considered as an outcome, but does not establish what the results were. The record states that there are 20 secondary outcomes posted, while the analysis listing reports no analyses; the registry’s analysis listing therefore does not provide a complete account of secondary statistical testing.

Secondary results generally carry less weight than a prespecified primary result because they may involve more outcomes and more opportunities for chance findings. The registry reports an unadjusted p-value for the posted comparison, and no multiplicity adjustment is reported. Without a stated analysis plan or multiplicity procedure, the strength of evidence from secondary tests is harder to judge. Selective reporting is also difficult to assess when results are absent for listed outcomes.

Accordingly, the secondary outcome pattern is best treated as supporting context for the primary lipid finding, not as a broad conclusion about every lipid measure or rheumatoid arthritis activity. The registry record does not report the results for the full set of listed secondary measures, and it does not establish whether their analyses were prespecified or adjusted for multiple testing.

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

• Change From Baseline in Low Density Lipoprotein-Cholesterol (LDL-C) at Week 12
• 12-Hours Fasting Lipid Profile
• 12-Hours Fasting Lipid Profile: Particle Size of Lipoproteins
• 12-Hours Fasting Lipid Profile: Level of Lipoprotein Particles
• 12-Hours Fasting Lipid Profile: Level of High Density Lipoprotein Cholesterol (HDL-C) Particles
• Disease Activity Score Using 28-Joint Count and C-Reactive Protein (3 Variables) (DAS28-3 [CRP])
• Disease Activity Score Using 28-Joint Count and C-Reactive Protein (4 Variables) (DAS28-4 [CRP])
• Disease Activity Score Using 28-Joint Count and Erythrocyte Sedimentation Rate (3 Variables) (DAS28-3 [ESR])

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is continuous, so the registry reports means and a mean difference rather than event counts or a ratio of risks. The group summaries are least-squares means, accompanied by standard errors. A standard error describes the estimated sampling variability of a statistic: smaller standard errors generally mean greater precision, while larger ones indicate more uncertainty. It is not the same quantity as the spread of individual participant measurements.

A least-squares mean is generally a model-based group mean. In general, such a mean can adjust for variables or design terms included in the model. The registry does not state which variables or terms were used for this endpoint, so the analysis cannot be reconstructed fully from the posted summary. The contrast between the two least-squares means is reported as a mean difference. Its sign depends on the ordering used: the posted estimate compares the atorvastatin group with the placebo group. A negative estimate therefore indicates a lower estimated percent change in the atorvastatin group on the reported scale.

The confidence interval gives a range of values compatible with the data and the statistical procedure, under its assumptions. It communicates uncertainty around the estimated difference, rather than a range containing a specified proportion of individual participant outcomes. The primary interval is entirely below zero, and the reported p-value is <0.001. In general, a p-value measures how incompatible the observed result is with a specified null model, assuming that model and the analysis assumptions hold. It does not quantify the size or practical importance of the effect, and it is not the probability that the null model is true.

The record also includes a simple unadjusted recalculation based on posted summary statistics. Such a recalculation is an approximation: it does not have access to participant-level observations or the full trial analysis model. In this case, the registry does not flag a discrepancy. Even so, the sponsor’s reported analysis is the trial result to prioritize, particularly because the registry identifies least-squares means while the simple recalculation uses summary values and approximations. A recalculation is useful as a check on scale and direction, not as a replacement for the sponsor’s analysis.

For continuous outcomes, the standard deviation describes participant-level variability, while the standard error describes uncertainty in an estimated mean. The registry reports standard errors for the group estimates. The posted recalculation estimates a pooled standard deviation from those standard errors and the group counts, and labels that estimate approximate. That derived quantity is useful for the planning illustration, but should not be treated as a directly reported participant-level standard deviation or as a feature of the sponsor’s model.

In general, the variance of a count outcome can exceed what a simple Poisson model expects; this is called over-dispersion. A model that ignores over-dispersion can understate uncertainty. That issue is relevant to count data, but the primary endpoint is continuous, not a count. The registry does not describe a count model for the primary analysis, and there is no basis to infer one.

Randomization aims to make treatment groups comparable on average, including with respect to measured and unmeasured characteristics. It does not guarantee that the realized groups are identical, especially in a trial of this size. The record identifies random allocation and a parallel design but does not provide the randomization method, stratification factors, or sequence-generation details. Masking can reduce some forms of bias in treatment delivery and outcome assessment, but the registry does not spell out which roles were masked or how masking was maintained.

The full analysis set is described as including randomized participants who received at least one dose of the randomized investigational drug. For the primary measure, the registry also gives endpoint-specific counts. The exact handling of participants without an evaluable Week 12 measurement, any missing-data assumptions, and any imputation method are not reported in the registry record. This prevents a complete assessment of whether the estimates reflect all randomized participants or rely on additional outcome availability conditions beyond the stated set definition.

The p-value is described as unadjusted. When many outcomes or comparisons are examined, the chance of obtaining at least one apparently noteworthy result by chance can increase. In general, multiplicity procedures can address that issue, but the registry does not report a procedure for these outcomes. The primary endpoint is the stated primary question; secondary findings should therefore be read with more caution, especially when their analyses are not posted.

The record lists 21 outcome measures and 0 reported analyses in its analysis listing, while it posts primary results and secondary outcome results. These record-level counts do not replace the endpoint results themselves. They do indicate that the registry’s structured analysis listing is not a full account of every statistical decision. The original protocol, analysis plan, and sample-size justification are not reported in the registry record, so it is not possible to verify which analyses were prespecified or whether they changed.

11. Confidence Intervals and Precision

The primary difference is -41.14 percentage points, and its 95% confidence interval ranges from -47.4 to -34.88. The interval provides a sense of precision: it shows the range of effect estimates compatible with the reported analysis under its assumptions. Because both limits are below zero, the interval supports a direction of difference favoring a lower estimated percent change in the atorvastatin group on this endpoint. It does not mean that every participant experienced a change within that range.

The width of a confidence interval reflects both the amount of information in the data and the variability in the outcome, as well as the analysis method. A narrower interval would generally indicate a more precise estimate; a wider interval would leave more uncertainty about the magnitude. The interval gives a range of plausible mean differences, not a guarantee that the true effect lies inside it. Its usual repeated-sampling interpretation is conditional on the model and assumptions used to construct it.

The group standard errors, 2.25 and 2.27, describe uncertainty in the estimated group means, not the spread of individual percent changes. The endpoint analysis counts are 46 and 45. The reported interval is more informative about the between-group estimate than considering the group standard errors alone, because it directly addresses the contrast of interest.

The registry does not provide the detailed model specification or missing-data handling, which limits independent assessment of the interval’s assumptions. Nor does the confidence interval address whether the observed difference is clinically important, whether it persists beyond Week 12, or whether the same estimate would be seen in another population. Precision is one part of interpretation; endpoint relevance, study conduct, and completeness of reporting also matter.

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 calculations. For these continuous-outcome comparisons, the entries are participants per arm, not total event counts. The 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: 100%, 75%, or 50%.

The 80% power column indicates the number of participants per arm in the illustration for a design with an 80% chance of detecting the specified effect under the stated assumptions, if that effect and model assumptions are correct. The 90% power column uses a higher target chance of detection and therefore generally requires more participants. These are design targets, not guarantees for a future study.

Smaller effects are harder to distinguish from random variation, so planning for a smaller fraction of the observed effect requires more participants per arm in the illustration. The estimates are based on the effect observed in this study, which may be imprecise or unusually large or small. They do not account for all possible design choices or operational realities, and the registry does not report the assumptions used for the trial’s original design.

This is not observed power. It is a planning illustration under stated assumptions, not a measure of how likely the completed trial was to find its result after the data were collected. The curve should be read as a representation of those planning relationships, not as evidence about the trial’s original recruitment target or its actual probability of success.

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 = -41.140; pooled standard deviation σ = 15.244.
2. n = 2 × (1.96 + 0.842)² × 15.244² / -41.140² = 2.2, rounded up to 3 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 <- -41.14                      # observed difference in means
sigma <- 15.244124674072                     # 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(1059864)
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 = 3 (80% power), 3 (90% power)
 75% of observed effect: n per arm = 4 (80% power), 6 (90% power)
 50% of observed effect: n per arm = 9 (80% power), 12 (90% power)
power.t.test: 3.4 
Simulated power at n = 3 per arm: 0.703 (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. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study examines several outcomes, time points, or comparisons. In general, testing many questions can increase the chance that at least one result appears statistically notable even when none of the underlying effects is real. Researchers may address this using a prespecified hierarchy or a statistical adjustment, but the registry does not report whether such a procedure was planned or used.

The primary endpoint is identified in the record, and its p-value is reported as unadjusted. The registry also lists 20 secondary outcomes and includes a secondary LDL-C result, but results for the other listed measures are not reported. The record gives no multiplicity adjustment or testing hierarchy. As a result, the primary endpoint has a clearer priority than exploratory or secondary findings, while the evidential status of any unadjusted secondary tests is less certain.

The analysis population is described as the full analysis set, including randomized participants who received at least one dose of the randomized investigational drug. The record further notes that the analysis count signifies participants evaluable for the measure, and it gives endpoint-specific analysis counts. It does not explain how participants with missing Week 12 measurements were handled, whether data were imputed, or whether any sensitivity analyses were conducted. These details are important because an analysis limited to participants with evaluable measurements can differ from an approach that retains all randomized participants through a missing-data model.

The registry does not report the trial’s original sample-size justification, the detailed statistical analysis plan, or interim analyses. These omissions do not negate the posted estimates, but they limit the ability to check prespecification, evaluate potential changes in analysis, or determine whether the study was designed around a particular target effect. The record’s structured analysis count is 0 despite posted outcome results, so it should not be taken as a full statistical analysis inventory.

15. Safety Results

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

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 2% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonia (2.1%), pneumonia bacterial (0.9%), arthritis (0.9%).

The registry reports serious adverse events for the CP-690,550 group and for the later treatment groups. In the CP-690,550 group, 2 participants were reported with serious adverse events among 111 at risk. In the CP-690,550 plus atorvastatin group, 0 were reported among 50 at risk; in the CP-690,550 plus placebo group, 1 was reported among 47 at risk. These are participant counts, not event rates computed for comparison.

The record also lists selected serious-event terms with percentages, including pneumonia at 2.1%, bacterial pneumonia at 0.9%, and arthritis at 0.9%. It does not specify in this summary how those term percentages map to the treatment groups, so they should not be interpreted as group-specific estimates. Death counts are not reported in the registry record.

These figures are descriptive. The treatment groups differ in their reported numbers at risk, and the registry does not provide a formal comparative safety analysis or uncertainty intervals for these event counts. Serious adverse event counts alone do not establish that one treatment caused an event, nor do they give a complete account of all safety outcomes. The broader adverse-event definitions, timing, and attribution details are not reported in the registry record.

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

The posted primary result estimates a difference in LDL-C percent change between participants assigned to CP-690,550 plus atorvastatin and those assigned to CP-690,550 plus placebo, from baseline at Week 6 to Week 12. The estimated difference is -41.14 percentage points, with a 95% confidence interval from -47.4 to -34.88 and an unadjusted p-value of <0.001. In this trial, the estimated change was lower in the atorvastatin group than in the placebo group on the reported scale.

The result does not compare CP-690,550 with no CP-690,550 because both groups received CP-690,550. It does not establish that the same estimate applies to other populations, settings, treatment durations, or outcome definitions. The study was phase 2 and assessed the primary measure at Week 12; longer-term outcomes are not reported in the registry record.

The finding concerns LDL-C, not every lipid measure or rheumatoid arthritis disease activity. Although the registry lists several secondary measures, analyses are not posted for most of them. The primary result therefore should not be expanded into a claim about outcomes for which results are unavailable. It also does not by itself establish the clinical importance of the estimated change, since the registry record does not provide a threshold for interpreting clinical importance.

Statistical evidence and clinical importance are different questions. The small reported p-value and interval excluding zero indicate that the observed group difference is difficult to reconcile with a zero difference under the analysis assumptions. They do not establish the probability that a hypothesis is true, nor do they resolve unreported aspects of model specification and missing-data handling. The confidence interval communicates precision for the estimate, but uncertainty remains about how well the result generalizes beyond this study.

The safety counts are descriptive and do not provide a formal comparison of risks. They should not be combined with the lipid result to make a broad treatment judgment. The registry record presents trial findings, not individual medical guidance.

17. Why This Trial Matters Statistically

This trial is statistically informative because its primary contrast is aligned with the timing of treatment assignment. Participants received CP-690,550 before the later atorvastatin-versus-placebo comparison, and the primary measure uses Week 6 as its baseline. The resulting contrast therefore focuses on the additional treatment period rather than mixing the earlier shared treatment period into the primary change measure.

The endpoint also illustrates why both an estimate and its uncertainty matter. The mean difference communicates the magnitude and direction of the observed group contrast, while the confidence interval indicates how precisely that contrast was estimated under the reported analysis. A p-value alone would not communicate the effect’s scale or the range of uncertainty.

The trial also highlights the difference between a posted result and a complete statistical record. The registry gives the primary estimates and a related secondary LDL-C outcome, but it does not report the detailed model, missing-data approach, multiplicity procedure, or original sample-size justification. Those omissions constrain independent evaluation and make it important to distinguish the reported primary finding from less complete secondary evidence.

Finally, the planning illustration demonstrates how sample-size calculations depend on the effect one aims to detect. Planning around a smaller effect requires more participants per arm under the stated assumptions. Because the illustration is derived from the observed effect rather than the trial’s original design, it should not be mistaken for evidence about the study’s design rationale or the probability that the completed study would detect an effect.

18. Limitations

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

The registry record does not provide the detailed statistical model used for the least-squares means, including any adjustment variables or covariance structure. That limits the ability to reproduce the primary analysis from the posted summary statistics. The record also does not report the original sample-size justification or the assumptions used to plan enrollment.

Missing outcome data are another unresolved issue. The registry defines the full analysis set and gives endpoint-specific analysis counts, but it does not explain why some enrolled participants are not included in those endpoint counts, how missing measurements were handled, or whether sensitivity analyses were performed. The impact of missingness therefore cannot be assessed from the record.

The record identifies a primary endpoint and lists multiple secondary outcomes, but it does not report a multiplicity strategy. Most listed secondary results are not posted, and the structured analysis listing does not provide a complete inventory of tests. This limits assessment of selective reporting and the evidential weight of any unadjusted secondary result.

The primary assessment is at Week 12, and longer-term outcomes are not reported in the registry record. The safety summary provides event counts, but not uncertainty intervals or a formal between-group analysis. Finally, arm descriptions are blank and the record does not detail the randomization method or masking roles. These are reporting limitations: they identify what cannot be evaluated from the registry record, rather than proving that a particular bias or problem occurred.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Study Of The Effects Of Atorvastatin On Cholesterol Levels In Rheumatoid Arthritis Patients Taking CP-690,550 — NCT01059864.
• Registry record last updated 2012-12-13; status last verified 2012-11.
• Results first posted 2012-12-13.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

22. Record Summary

NCT01059864 is a completed, randomized, parallel-group phase 2 study in patients with active rheumatoid arthritis. The registry lists quadruple masking and 111 enrolled participants. Participants received CP-690,550 throughout the 12-week study; after 6 weeks, the comparison was atorvastatin versus matching placebo.

The primary endpoint was percent change in LDL-C from baseline at Week 6 to Week 12. The reported mean difference was -41.14 percentage points, with a 95% confidence interval from -47.4 to -34.88 and an unadjusted p-value of <0.001. A secondary LDL-C change result was also posted in mg/dL. The registry does not report analyses for most other listed secondary outcomes.

The record does not report the detailed statistical model, original sample-size justification, missing-data approach, or multiplicity procedure. Serious adverse-event counts are descriptive, and deaths are not reported in the registry record.