Scleroderma Lung Study III - Combining Pirfenidone With Mycophenolate: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Scleroderma Lung Study III - Combining Pirfenidone With Mycophenolate (NCT03221257), a randomized phase 2 study. The primary outcome is Percent Predicted Forced Vital Capacity (FVC-%).

2. Clinical Question

3. Trial Design

Scleroderma Lung Study III (SLS III) is a completed phase II, randomized, parallel-group trial of pirfenidone added to mycophenolate for scleroderma-related interstitial lung disease. The registry describes a multi-center, placebo-controlled study. Its masking field says triple masking, while the brief summary describes the study as double-blind; the record does not explain the difference between those descriptions.

The two treatment groups shared background mycophenolate mofetil. One group received pirfenidone as add-on therapy, and the comparator group received placebo as add-on therapy. This design focuses the randomized comparison on the addition of pirfenidone in the setting of MMF. It does not compare MMF with no MMF, and the posted comparison should not be read as an estimate of the effect of either drug used alone.

The registry lists enrollment of 51 participants, with 24 in the placebo-plus-MMF group and 27 in the pirfenidone-plus-MMF group for the primary outcome. The primary endpoint was change from baseline to 18 months in FVC-% and the primary analysis population was modified intention-to-treat. That population included randomized participants who received at least one dose and had at least one post-baseline efficacy assessment.

The record lists a start date of 2017-11-28, primary completion on 2022-03-23, and status as completed. The original sample-size justification, target treatment effect, planned handling of missing outcomes, and original analysis assumptions are not reported in the registry record. Those details should not be inferred from planning illustrations based on the observed result.

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: 88 consented with 31 subsequently deemed ineligible, 6 withdrawals prior to randomization, resulting in 51 participants considered enrolled, eligible and randomized.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The modified intention-to-treat population (m-ITT) is defined as all participants randomized, receiving at least one dose of study medication, and having at least one post-baseline efficacy assessment.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Safety population includes all the randomized participants who received at least one dose of study medication.

6. Primary Endpoint

The registry lists the primary outcome as Percent Predicted Forced Vital Capacity (FVC-%), measured over Baseline to 18 months.

The primary endpoint is change from baseline to month 18 in mean forced vital capacity expressed as a percentage of the predicted value, or FVC-%. The predicted value is adjusted for age, height, gender, and race. FVC measures the volume of air that can be forcibly exhaled after a full inhalation; expressing it relative to a predicted value places observed lung function on a scale that accounts for those characteristics.

Change over time is used to describe how the measure shifts during follow-up, while a comparison between randomized groups estimates the difference in average change associated with assignment to the add-on treatment strategies. The outcome is continuous, so the reported group summaries are means rather than event counts. The primary endpoint uses least-squares means, with standard errors reported for the individual group estimates.

The registry defines the analysis population as randomized participants who received at least one dose and had at least one post-baseline efficacy assessment. This definition matters because it is narrower than simply including every randomized participant: having a post-baseline assessment is required for inclusion in the stated primary population. The registry record does not report the amount or pattern of missing endpoint data, the method used to handle such data, or a clinically important difference threshold for FVC-%.

7. Results Posted in the Registry

• Pirfenidone (PFD) + Mycophenolate (MMF): mean 2.09 versus 2.24 on control; difference -0.15 (95% CI -3.79 to 3.49), standardized effect d = -0.02; z-test p = 0.936.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: each participant has a value on the FVC-% scale, and the analysis compares average change between randomized groups. The registry identifies the primary group summaries as least-squares means and reports a standard error for each group. A standard error describes the uncertainty in an estimated mean. It is not the spread of individual participants' measurements; the latter describes variation among participants.

The sponsor's reported analysis is called a mixed models analysis. In general, mixed models can account for relationships among repeated measurements from the same participant and can use information across measurement occasions when estimating treatment-group patterns. The registry gives the analysis name and the resulting net mean difference and p-value, but does not report the full model specification, covariance structure, covariates, estimation details, or missing-data assumptions. Those choices cannot be reconstructed from the name of the method alone.

The primary population is modified intention-to-treat, defined in the record as all randomized participants who received at least one dose of study medication and had at least one post-baseline efficacy assessment. In general, intention-to-treat analysis aims to preserve randomized comparisons. A modified definition can exclude some randomized participants, and the exact consequences depend on who is excluded and why. The registry does not give participant-level reasons for exclusion or a flow of participants through the primary analysis population.

The unadjusted re-calculation uses the posted group means, standard errors, and sample sizes to estimate a difference between the groups. The result is explicitly approximate because the pooled standard deviation is itself marked as approximate. This calculation does not use participant-level data, adjust for covariates, reproduce the sponsor's mixed model, or account for any features of the original analysis that are not in the registry record. Its value is to provide a transparent comparison derived from the posted summary statistics, not to replace the sponsor's model-based result.

A confidence interval describes a range of values compatible with the estimate under the calculation's assumptions. In general, wider intervals indicate less precision, while narrower intervals indicate more precision. An interval spanning zero means the posted data, under that calculation, do not distinguish the estimated contrast from no difference at the stated confidence level. It does not show that zero is the only plausible value or that the groups are equivalent. The registry does not report a confidence interval for the sponsor's mixed-model estimate, so the approximate interval from the unadjusted calculation should not be attributed to that model.

A p-value summarizes how incompatible the observed data are with a specified null hypothesis under the analysis assumptions. In general, a large p-value is not the probability that the null hypothesis is true, and it does not measure the probability that a result is due to chance. The unadjusted p-value is based on the simple summary-statistic re-calculation; the sponsor p-value is from the mixed-model analysis. They are close but are not identical because they arise from different analyses.

Multiplicity refers to the interpretive issue created by testing or estimating multiple outcomes, time points, or contrasts. In general, repeated testing can increase the chance of finding an apparently notable result when many comparisons are considered. The record lists 23 outcome measures and 14 analyses, but the multiplicity plan, any prespecified hierarchy, and any adjustment procedure are not reported in the registry record. The primary comparison should therefore be distinguished from the wider set of registered outcomes and analyses, without assuming how the trial controlled multiplicity.

Over-dispersion is a general concern when observed variation is greater than a model's assumptions anticipate. In count outcomes it can make standard errors too small if a simple model assumes less variation than the data contain. The primary endpoint is continuous rather than a count, and the registry record does not report an over-dispersion assessment for the primary analysis. More generally, any model's uncertainty depends on whether its assumptions reasonably describe the data; the information posted is not enough to assess every assumption of the mixed model.

The registry does not report the trial's full statistical analysis plan, details of interim analyses, sample-size justification, or original design assumptions. It also does not report a detailed missing-data strategy. These absences limit independent evaluation of design decisions and analysis choices. They are distinct from the planning illustration derived from the observed effect, which is not evidence of the trial's original design or its achieved power.

10. Confidence Intervals and Precision

The simple unadjusted re-calculation gives a mean difference of -0.15 for pirfenidone plus MMF compared with placebo plus MMF, with an approximate 95% confidence interval from -3.795 to 3.495. The interval extends to values on both sides of zero. Its width indicates that the point estimate should not be treated as a precise estimate of the underlying group difference.

In general, a confidence interval represents uncertainty around an estimate under the method's assumptions. A relatively narrow interval would constrain the range of values compatible with the data more tightly; a broad interval leaves more possible values consistent with the calculation. Here, the interval allows for a negative contrast as well as a positive one, so the posted summary statistics do not precisely settle the direction or magnitude of the difference.

The interval is an approximate, unadjusted calculation based on posted summary statistics, not a confidence interval from the sponsor's mixed-model analysis. The registry does not provide confidence limits for the sponsor's reported estimate of -0.14. The two estimates are close, but the approximate interval should not be used as if it were generated by the sponsor's model.

Precision is not the same as clinical importance. The record does not report a threshold for a clinically important difference in FVC-%, so the interval cannot be classified against such a threshold using the registry record alone. Nor does an interval spanning zero demonstrate equivalence: an equivalence conclusion requires a defined margin and an analysis designed to evaluate that margin, neither of which is reported.

11. Planning a Similar Trial: Sample Size

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

The sample-size illustration is based on the observed effect and is not observed power. It assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. These are planning assumptions for the illustration; the trial's original design assumptions and sample-size justification are not reported in the registry record.

Because the primary outcome is continuous, the figures are participants per arm, not total events. The 80% power column gives the number of participants per arm estimated to provide an 80% chance of detecting the specified effect under the illustration's assumptions. The 90% power column gives the per-arm number for a 90% chance under those same assumptions. Higher planned power requires more participants for a given effect size.

For planning at 100% of the observed effect, the illustration gives 30,671 participants per arm for 80% power and 41,060 for 90% power. At 75% of the observed effect, it gives 54,526 per arm for 80% power and 72,994 for 90% power. At 50% of the observed effect, it gives 122,682 per arm for 80% power and 164,237 for 90% power. The pattern reflects a general statistical principle: smaller effects are harder to distinguish from random variation and therefore require more participants under otherwise unchanged assumptions.

These values are planning illustrations calculated from the observed effect, not a retrospective statement about the probability that this completed trial detected an effect. They should not be read as the actual required sample size for the original protocol. The record lists enrollment of 51, and the analysis sizes for the primary outcome were 24 and 27; the illustration's much larger figures follow from its stated effect and assumptions.

12. How the Sample Size Is Derived

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

Formula

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

Worked calculation for this trial

1. Observed difference Δ = -0.150; pooled standard deviation σ = 6.630.
2. n = 2 × (1.96 + 0.842)² × 6.630² / -0.150² = 30,670.5, rounded up to 30,671 participants per arm for 80% power.

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

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -0.15                      # observed difference in means
sigma <- 6.63028967877647                     # 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(3221257)
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 = 30,671 (80% power), 41,060 (90% power)
 75% of observed effect: n per arm = 54,526 (80% power), 72,994 (90% power)
 50% of observed effect: n per arm = 122,682 (80% power), 164,237 (90% power)
power.t.test: 30671.4 
Simulated power at n = 30,671 per arm: 0.803 (20000 replicates)

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

13. Analyses Reported by the Sponsor

The sponsor's reported mixed-model analysis gives a net mean difference of -0.14 for pirfenidone plus MMF compared with placebo plus MMF, with p = 0.933. The registry does not provide confidence limits for this estimate. The sponsor analysis is the trial's own reported analysis and should be kept distinct from the independent summary-statistic re-calculation.

The unadjusted re-calculation gives a mean difference of -0.15, an approximate 95% confidence interval of -3.795 to 3.495, and p = 0.936. The close estimates and p-values indicate that the simple calculation produces a similar descriptive result. It is not a reproduction of the mixed-model analysis: it uses posted summaries and does not incorporate model details that are not reported.

The registry record does not flag a discrepancy between the sponsor analysis and the re-calculation. Accordingly, the sponsor's analysis remains the primary reported inferential result, while the re-calculation provides a transparent, approximate check. The confidence interval from the re-calculation should not be assigned to the sponsor's model-based estimate.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists 23 outcome measures and 14 analyses, but the record does not report a multiplicity strategy. It does not state whether a hierarchy of outcomes was prespecified, whether testing was adjusted for multiple comparisons, or whether any particular analyses were designated exploratory. As a result, the posted primary result can be described as the registered primary comparison, but the handling of the broader collection of measures and analyses cannot be established from the record.

In general, multiplicity matters when several outcomes, time points, or comparisons are examined because repeated testing can affect the interpretation of nominal statistical evidence. That general concern does not establish that the reported analysis was or was not adjusted. The relevant plan is not reported in the registry record, so no adjustment should be assumed.

The primary analysis population is defined as randomized participants who received at least one dose of study medication and had at least one post-baseline efficacy assessment. The record does not report how many randomized participants failed to meet this definition, whether exclusions differed by group, or the reasons for any missing assessments. The amount and pattern of missing primary outcome data are also not reported.

The missing-data method for the sponsor's mixed-model analysis is not reported in the registry record. In general, statistical methods for incomplete repeated measurements rely on assumptions about why measurements are absent and how missingness relates to observed or unobserved outcomes. Without the analysis plan and relevant missingness information, those assumptions cannot be evaluated from the posted summary alone.

The registry record also does not report interim-analysis plans or whether any formal stopping rules were used. It does not provide the original sample-size justification or the trial's full statistical analysis plan. Those omissions limit assessment of prespecification and analytic flexibility; they do not, on their own, show that any particular design or analysis choice was made.

15. Safety Results

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

Deaths

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: colitis (4.2%), covid-19 infection (4.2%), acute chronic hypoxemic respiratory failure (4.2%), right sided chest pain (3.7%), herpes zoster ophthalmicus (3.7%).

The registry reports serious adverse events among 2 of 24 participants in the placebo-plus-MMF group and 4 of 27 participants in the pirfenidone-plus-MMF group. It reports other adverse events among 23 of 24 participants in the placebo-plus-MMF group and 27 of 27 participants in the pirfenidone-plus-MMF group. These are the posted numbers of participants affected, not a complete description of event timing, duration, severity beyond the serious classification, or causal relationship to treatment.

No deaths are reported in either group. The registry lists serious adverse event terms including Colitis, Covid-19 Infection, and Acute Chronic Hypoxemic Respiratory Failure at 4.2% each, and Right Sided Chest Pain, Herpes Zoster Ophthalmicus, and Basal Cell Carcinoma, Nodular at 3.7% each. These terms are listed as the leading serious-event terms in the record; the posted summary does not provide enough detail to infer whether an event was caused by study medication or how individual events were distributed across the groups.

Safety counts should be read alongside the trial's limited enrollment and the registry's level of detail. They describe events recorded in this trial, not the full range or frequency of possible outcomes in other settings. The record does not provide a causal assessment or detailed exposure-adjusted analysis.

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

The primary endpoint comparison describes the average change in FVC-% from baseline to month 18 for two add-on strategies: pirfenidone with background MMF and placebo with background MMF. The posted group estimates are 2.09 and 2.24, respectively, and the sponsor reports a net mean difference of -0.14. The simple unadjusted re-calculation is close to that estimate, but its approximate confidence interval is broad and includes values on both sides of zero.

This pattern means the posted primary result does not precisely distinguish the group contrast from no difference under the reported analyses. It does not prove that there is no difference, and it does not establish that the strategies are equivalent. Equivalence would require a prespecified margin and an analysis directed at that question; such details are not reported in the registry record. The record also does not specify a clinically important difference threshold for FVC-%.

The comparison is about adding pirfenidone to MMF, not about pirfenidone alone, MMF alone, or treatment compared with no background therapy. It is specific to the trial population, the modified intention-to-treat definition, the stated endpoint, and the follow-up through month 18. The record lists other outcome measures, but it does not post secondary outcome results in the registry record.

Statistical interpretation should also keep the sponsor's mixed-model result separate from the approximate unadjusted re-calculation. The sponsor estimate and p-value are the reported trial analysis. The re-calculation is useful as a check from the posted group summaries, but it cannot substitute for the full model or reveal unreported analysis assumptions. Neither result alone establishes a clinical recommendation.

The safety counts are descriptive reports of affected participants and recorded serious events. They do not establish that the study medications caused those events, and they should not be used to infer risks beyond the trial's reported population and follow-up. Overall, the registry offers a result for a specific randomized comparison but leaves important context—including missing-data handling, multiplicity procedures, and clinical-importance thresholds—not reported in the registry record.

17. Why This Trial Matters Statistically

This trial raises a useful statistical question about add-on treatment: when both randomized groups receive the same background MMF, the contrast estimates the incremental difference associated with adding pirfenidone rather than placebo. That interpretation depends on recognizing the shared background therapy. It would be a mistake to describe the comparison as pirfenidone versus MMF or as either drug versus no treatment.

The primary outcome is continuous, and the reported estimates are group means with standard errors. This makes it important to distinguish the estimated average change from uncertainty in that estimate, and to distinguish the variation among participants from the precision of a group mean. The broad approximate interval illustrates why a point estimate close to zero should not be treated as an exact estimate of no difference.

The analysis also demonstrates why a sponsor's model-based result and a summary-statistic re-calculation can be presented side by side without treating them as interchangeable. The sponsor reports a mixed-model estimate and p-value; the independent calculation uses posted group summaries and is approximate. Their proximity is informative, but the registry does not supply the model specification needed to reproduce the sponsor analysis.

Finally, registered outcomes and analyses extend beyond the primary result, while multiplicity and missing-data procedures are not reported in the registry record. That distinction is central to careful interpretation: the stated primary comparison can be described, but its place among the full analysis plan cannot be fully assessed from the posted record. The trial therefore illustrates the value of reporting estimates and uncertainty alongside transparent analysis details.

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 report the full statistical analysis plan, original sample-size justification, design assumptions, or interim-analysis details. It also does not provide the mixed model's specification or confidence limits for the sponsor estimate. These gaps prevent a complete independent reconstruction of the primary analysis.

The unadjusted re-calculation relies on posted summary statistics and an approximate pooled standard deviation. Its confidence interval and p-value are not model-based results from the sponsor's mixed model. The calculation is a descriptive check, and its assumptions may not match the trial's analysis approach.

The primary population excludes randomized participants who did not receive at least one dose or did not have at least one post-baseline efficacy assessment. The number of such participants, their group allocation, and reasons for exclusion are not reported. The record also does not describe the amount or pattern of missing FVC-% measurements or the method used to address missing data.

Although the registry lists 23 outcome measures and 14 analyses, it does not report a multiplicity plan or identify how the many outcomes and analyses were prioritized. No secondary outcome results are posted in the record. The clinical importance threshold for the primary endpoint is also not reported, limiting interpretation of the size of the estimated difference.

Enrollment was 51, with 24 and 27 participants in the primary outcome groups. The approximate confidence interval is broad, so the estimate is imprecise. The adverse-event summary is also limited: it reports affected-participant counts and selected serious-event terms but does not provide detailed causal assessments, timing, or exposure-adjusted comparisons. These limitations are specific to what the registry record reports and should not be filled by assumptions.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Scleroderma Lung Study III - Combining Pirfenidone With Mycophenolate — NCT03221257.
• Registry record last updated 2023-12-15; status last verified 2023-12.
• Results first posted 2023-12-15.
• 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

NCT03221257 is the completed phase II Scleroderma Lung Study III, a randomized parallel-group trial in scleroderma-related interstitial lung disease. It compared pirfenidone added to background mycophenolate with placebo added to background mycophenolate. The registry lists enrollment of 51 participants and identifies change in FVC-% from baseline to month 18 as the primary endpoint.

The sponsor's mixed-model analysis reports a net mean difference of -0.14 and p = 0.933. A simple unadjusted re-calculation gives -0.15, with an approximate 95% confidence interval of -3.795 to 3.495 and p = 0.936. The estimates are near zero, while the re-calculated interval remains broad. The registry record does not report secondary outcome results, the multiplicity plan, missing-data details, or the trial's original sample-size assumptions.