Oral Pirfenidone for the Pulmonary Fibrosis of Hermansky-Pudlak Syndrome: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Oral Pirfenidone for the Pulmonary Fibrosis of Hermansky-Pudlak Syndrome (NCT00001596), a randomized phase 2 study. The primary outcome is Change in Forced Vital Capacity (36 Months).

2. Clinical Question

3. Trial Design

This was a phase 2, randomized, parallel-group clinical trial comparing oral pirfenidone with placebo for pulmonary fibrosis associated with Hermansky-Pudlak Syndrome. The registry record lists quadruple masking. It does not describe the masking procedures in detail or identify which participants, clinicians, assessors, or other roles were masked. The allocation ratio is not reported in the registry record.

The recorded enrollment was 35. The brief summary describes a plan to select up to 40 patients, which is a planned maximum rather than the recorded enrollment. The study status is completed. Its recorded start was 2005-09, and primary completion was 2009-09. The sponsor is listed as William Gahl, M.D.

Participants assigned to pirfenidone received 801 mg, consisting of 3 pills of 267 mg each, three times daily. Participants assigned to placebo received 3 pills three times daily. The registry lists 11 outcome measures and no reported analyses. The primary endpoint was change in forced vital capacity at 36 months, measured from baseline. The outcome population is described as the Intent to Treat population for whom baseline and 36-month data were available.

The brief summary states that the study would stop if one therapy proved more effective than the other. The registry record does not report the criteria for such a decision, whether an interim analysis was planned or conducted, or the analysis plan used to assess efficacy. It also does not report the sample-size justification. The design labels establish randomization, masking, and parallel groups, but do not by themselves supply those missing procedural details.

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: Participants from the Intent to Treat population for whom data was available at baseline and 36 months.

5. Baseline Characteristics

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

6. Primary Endpoint

The registry lists the primary outcome as Change in Forced Vital Capacity (36 Months), measured over Measured at baseline and 36 months.

The primary endpoint was change from baseline in forced vital capacity at 36 months, reported as a percentage of predicted volume. Forced vital capacity, or FVC, is the volume of air a person can forcibly blow out after taking a full breath. The predicted volume is calculated using height, age, sex, and weight, according to the endpoint description. Expressing the measure as a percentage of predicted volume places the observed lung-function measurement in relation to a value predicted from those characteristics.

Change from baseline focuses on how the measurement differs from its starting value over the specified follow-up period. FVC is used in this trial as a measure of lung function, the principal outcome identified in the brief summary. The primary measurement was specified at 36 months; a separate secondary outcome assessed FVC change at 12 months. Those time points address different follow-up durations and should not be treated as interchangeable.

The record reports a mean change and standard deviation for each group with data available at baseline and 36 months. It does not provide a confidence interval or a reported test for the primary comparison. The available group summaries describe the participants contributing endpoint data, not necessarily everyone enrolled or randomized. The registry record does not explain why endpoint data were unavailable for other participants.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary measure is continuous: it records a quantity on a numerical scale, and the registry reports a mean and standard deviation for each group. A mean describes the group’s average observed change. A standard deviation describes the spread of individual measurements around that group average; it is not the uncertainty interval for the mean and is not itself a comparison between groups.

What does a mean difference represent?

For a two-group continuous outcome, a mean difference expresses the average in one group relative to the average in the comparator group. Its sign depends on the order used for subtraction. Interpretation also depends on the measure’s scale and direction. The primary record does not give a mean difference for the 36-month outcome, so no such estimate should be inferred from the two arm summaries. In particular, an unreported contrast is not the same as a contrast of zero.

The secondary 12-month comparisons include mean differences and confidence intervals. These comparisons are labelled as recalculated, meaning they are simple unadjusted calculations from the posted summary statistics rather than reported sponsor analyses. The registry lists no reported analyses for these outcomes. The record indicates no discrepancy between those recalculations and a sponsor analysis; however, there is no reported sponsor analysis to compare with them. Such calculations do not recover details of an original model that are absent from the record.

How should standard deviations and confidence intervals be read?

As a general statistical concept, a standard deviation describes variation among observations, while a standard error describes uncertainty in an estimated quantity such as a mean. A confidence interval uses that uncertainty to give a range of values compatible with the data under the analysis assumptions. It is not a guarantee that the true effect lies in the interval, nor does a wide interval identify which value is correct. Larger variation and limited information generally make estimates less precise.

The 36-month primary summaries include standard deviations but no confidence interval for a between-group contrast. The reported reason is that values were missing or not estimable for a two-arm comparison. The 12-month secondary estimates do include intervals. Because those intervals include zero, they do not rule out no mean difference under the stated comparison; they also encompass effects in either direction. This is why a point estimate should not be read on its own as a definitive result.

What does an unadjusted p-value say?

A p-value is a measure of how compatible the observed data are with a specified null model, conditional on the statistical assumptions. It is not the probability that the null hypothesis is true, and it does not measure clinical importance. The p-values reported for the secondary recalculations are unadjusted: the registry record does not indicate that they account for testing multiple outcomes. The record does not report a p-value for the primary 36-month comparison. It would therefore be inappropriate to present a primary significance test that is not in the record.

For repeated or multiple outcome testing, a general concern is that looking across many results creates more opportunities for apparently notable findings by chance. Methods that control multiplicity adjust the interpretation or testing procedure across a family of outcomes. Whether such a method was planned or used is not reported in the registry record. Secondary p-values should consequently be treated as unadjusted results, not as confirmatory evidence.

Why can over-dispersion matter?

General explanation: over-dispersion refers to variability greater than a statistical model expects. It can make standard errors too small and confidence intervals too narrow if the model fails to account for that extra variation. This issue is most directly relevant to count or event models, rather than the mean-based continuous summaries reported for the primary outcome. The record does not report an over-dispersion analysis, and there is no basis to assert that it affected this trial’s continuous endpoint estimates.

What do the planning illustrations mean?

The planning illustrations use the observed secondary effect as an input and show participants per arm for 80% or 90% power at the stated two-sided significance level of 0.05, with equal allocation and a normal approximation. Other rows use a stated fraction of the observed effect. These are hypothetical planning calculations, not observed power and not the trial’s original sample-size justification. The sample-size assumptions explicitly say the calculations are not the trial’s own design. The original design justification is not reported in the registry record.

General explanation: planning sample sizes depend on assumptions about the effect to detect, outcome variability, allocation, and the chosen error rate. Changing an assumed effect changes the planning result. A calculation based on an effect observed in a small trial can be unstable, so it should not be treated as a reliable estimate of the effect that a future study will necessarily see. The illustrations describe assumptions, not evidence that the present trial was adequately or inadequately powered.

How do available-data populations affect interpretation?

The primary population is described using an Intent to Treat label but is restricted to participants with data available at baseline and 36 months. The record does not explain how this available-data condition was implemented, whether imputation was used, or whether the analysis retained all randomized participants through another method. In general, missing outcome data can affect an estimate when the probability of having an observed measurement is related to the outcome or to group assignment. The appropriate interpretation depends on the missing-data process and analysis method, neither of which is detailed.

These methods distinctions matter because design labels and arm summaries answer different questions. Randomization concerns assignment; masking concerns knowledge of assignment; descriptive statistics summarize observed data; and a comparison estimate quantifies the contrast under an analysis method. The registry record identifies the design and gives some descriptive results, but does not report enough primary-outcome analysis detail to quantify a 36-month treatment contrast.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Pirfenidone", "Placebo")
n   <- c(11, 3)
m   <- c(-23.52, -20.93)
sp  <- c(24.32, 25.59)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Pirfenidone  11  -23.520  24.320  -39.858  -7.182
CARM  Placebo  3  -20.930  25.590  -84.499  42.639

11. Multiplicity, Missing Data and Analysis Populations

The record identifies a primary outcome and lists secondary outcomes, but does not report a multiplicity strategy. It does not state whether the secondary outcomes were organized into a testing hierarchy, whether any adjustment was made for testing more than one outcome, or how results were to be interpreted across time points. This matters because multiple testing can increase the chance of finding a result that appears unusual when many comparisons are considered. The secondary p-values available in the posted results are unadjusted.

The registry identifies 8 outcome measures, and numeric results are not posted for all listed outcomes. The record does not explain whether unreported outcomes were not measured, were measured but not analyzed, or were analyzed but not entered in the registry. No analyses are listed as reported analyses. The statistical analysis plan, including any prespecified covariates or model choices, is not reported in the registry record.

The primary outcome population is described as the Intent to Treat population with data available at baseline and 36 months. The record does not give a participant flow, reasons for missing measurements, or the missing-data method. It also does not state whether a sensitivity analysis assessed the effect of missing data. The endpoint summaries therefore describe the available measurements, while the extent and consequences of unavailable measurements remain unclear.

The reported primary values are based on 11 participants in the pirfenidone group and 3 in the placebo group. The registry enrollment is 35, but the record does not explain how participants moved from enrollment to the endpoint population or why these endpoint counts differ. It would not be justified to infer reasons for missingness or assume that the available cases represent all randomized participants.

The brief summary says the study would stop if one therapy proved more effective. Operational criteria, interim-analysis methods, and any alpha-spending or other decision procedure are not reported in the registry record. The summary alone does not establish that an interim analysis occurred or how early stopping was assessed. The sample-size justification and original statistical plan are also not reported, limiting assessment of the intended inferential framework.

12. 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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: chest pain (8.7%), respiratory failure (8.7%), blood creatinine phosphokinase increased (8.3%), deep vein thrombosis (4.3%), haematochezia (4.3%).

The registry records serious adverse events for participants at risk in each group. In the pirfenidone group, 23 were recorded as at risk and 6 as affected. In the placebo group, 12 were recorded as at risk and 2 as affected. These are the registry’s group-level counts. They should be read with their denominators and not as a complete account of the type, timing, duration, or severity of every event.

The record also lists other adverse events: 18 affected among 23 at risk in the pirfenidone group, and 9 affected among 12 at risk in the placebo group. These counts are distinct registry categories; the record does not provide enough detail to determine how categories overlap or to reconstruct individual event histories. Death counts are not reported in the registry record.

The listed serious-event terms and percentages are chest pain, 8.7%; respiratory failure, 8.7%; blood creatinine phosphokinase increased, 8.3%; deep vein thrombosis, 4.3%; and haematochezia, 4.3%. The record does not assign these listed percentages to a particular treatment group in the information reported, so they should not be interpreted as group-specific comparative rates.

These figures are descriptive safety records, not a randomized comparison adjusted for exposure or follow-up. The registry does not report event-level timing, exposure-adjusted rates, or a statistical comparison of adverse events. The counts alone cannot establish whether a difference is attributable to treatment, and they do not describe the full clinical context of each event.

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

The central statistical limitation is that the registry provides no estimable two-group comparison for the primary outcome, change in FVC at 36 months. It reports group means and standard deviations among participants with available measurements, but not a mean difference, confidence interval, or p-value. Those summaries describe observed data; they do not quantify the uncertainty around a comparative effect.

The secondary results add information about selected outcomes at 12 months, but they are unadjusted recalculations and have broad intervals that include zero. They do not replace the primary comparison, and they do not constitute a complete analysis of all registered secondary outcomes. The absence of a multiplicity plan, missing-data explanation, and reported statistical analysis plan leaves important aspects of interpretation unresolved.

The randomized, quadruple-masked, parallel design is relevant context for understanding the trial. It does not eliminate the need for a transparent analysis of available outcome data. Likewise, the registry’s completed status does not mean that every planned outcome has a posted numerical result or that all analysis details are available.

These results do not establish that pirfenidone was superior to placebo, that the groups were equivalent, or that there was no difference. Nor do the registry summaries establish a clinical benefit or lack of benefit. They show the posted arm-level measurements and some imprecise secondary comparisons, while leaving the primary comparative estimate unavailable.

In general, a treatment-effect conclusion requires an estimate, its uncertainty, and a clear account of the population and analysis. Here, the missing primary contrast and limited details mean that the statistical evidence cannot be fully assessed from the registry record alone. The appropriate reading is constrained to what is reported: descriptive primary endpoint values, partial secondary results, and adverse-event counts without a detailed comparative analysis.

14. Why This Trial Matters Statistically

This trial is statistically informative because it illustrates the difference between a registered primary question and the results that can be interpreted from available data. The primary question concerns a randomized comparison of FVC change at 36 months. Yet the registry reports only group summaries for participants with available measurements and states that a two-arm comparison is not estimable. A trial can therefore have a defined primary endpoint and a completed status while still lacking a usable comparative estimate in the registry record.

The endpoint counts also show why denominators matter. The reported primary summaries come from 11 participants in the pirfenidone group and 3 in the placebo group, whereas the recorded enrollment is 35. The record does not explain this difference. Without participant flow and missing-data methods, readers cannot determine how the available endpoint population relates to the enrolled population or how that affects the result.

The record also distinguishes the primary endpoint from multiple secondary outcomes. Some 12-month secondary comparisons can be recalculated, but the results are imprecise and unadjusted, and numeric results are not posted for all listed outcomes. This is a useful example of why isolated p-values or point estimates should not substitute for a prespecified primary analysis or an account of multiplicity.

Finally, the planning illustrations are separate from the trial’s original design justification and do not describe power achieved. They show how planning quantities depend on assumptions, including the effect chosen for detection. Taken together, the record highlights the value of reporting complete outcome data, analysis methods, missing-data handling, and uncertainty around the primary comparison.

15. 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 most direct limitation is the absence of an estimable primary comparison. The 36-month FVC means are reported for 11 pirfenidone participants and 3 placebo participants with available data, but no primary effect estimate or confidence interval is available. The small endpoint groups make the arm-level summaries particularly limited as descriptions of a comparative effect, while the record offers no analysis that resolves this limitation.

The registry does not report why endpoint measurements are unavailable, provide participant flow, or describe missing-data handling. The primary population is characterized as Intent to Treat participants with baseline and 36-month data, but the operational details of that restriction are not given. It is not possible to determine from the record whether missingness is related to outcomes, assignment, or other factors.

There are also gaps in outcome reporting. The record lists 8 outcome measures, while numeric secondary results are posted for only three outcomes at 12 months. It does not explain the status of every other listed outcome. No statistical analysis plan, sample-size justification, multiplicity procedure, or detailed interim-analysis method is reported. The brief summary’s statement about stopping if one therapy proved more effective does not specify how that decision would be made.

The secondary comparisons are unadjusted recalculations from posted summary statistics, not reported sponsor analyses. They may not reproduce any original model or adjustment that is not documented. The record does not provide enough information to judge whether other analysis choices, such as covariate adjustment, were considered or used.

Finally, the safety data are aggregate counts, with no detailed event timing or group-specific attribution for the listed serious-event terms and percentages. Death counts are not reported. These limitations constrain both efficacy and safety interpretation; they do not, by themselves, indicate what unreported results or analyses would have shown.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Oral Pirfenidone for the Pulmonary Fibrosis of Hermansky-Pudlak Syndrome — NCT00001596.
• Registry record last updated 2017-10-16; status last verified 2017-09.
• Results first posted 2012-01-26.
• 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.

19. Record Summary

The registry record describes a completed, randomized, quadruple-masked, parallel phase 2 trial of oral pirfenidone versus placebo for pulmonary fibrosis associated with Hermansky-Pudlak Syndrome. Recorded enrollment was 35. Pirfenidone was given as 801 mg, three times daily; placebo was given as 3 pills, three times daily.

The primary endpoint was change in FVC from baseline at 36 months, expressed as a percentage of predicted volume. Arm-level means and standard deviations are reported for participants with available measurements, but the record says a two-arm comparison is not estimable. Numeric secondary comparisons are posted for three outcomes at 12 months; the record lists 7 secondary outcomes overall.

The registry lists no reported analyses and does not report the primary analysis plan, sample-size justification, multiplicity approach, or missing-data method. Safety summaries give serious and other adverse-event counts by group, while death counts are not reported in the registry record. These gaps limit interpretation of the comparative efficacy and safety results.