Scleroderma: Cyclophosphamide or Transplantation: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Scleroderma: Cyclophosphamide or Transplantation (NCT00114530), a randomized phase 2, phase 3 study. The primary outcome is Global Rank Composite Score (GRCS) (Month 54, ITT).

2. Clinical Question

The registry lists this as a randomized, open-label, parallel-group study, with 75 participants enrolled. Its brief description frames the study as a comparison of the potential benefits of transplantation and high-dose monthly cyclophosphamide in severe scleroderma. The statistical question is narrower than a general question about treatment: it concerns the groups’ relative ordering under a specified hierarchy of outcomes at a stated follow-up time. The record does not report a broader claim about every manifestation of systemic sclerosis, nor does the composite itself provide a single clinical measurement that can be interpreted independently of its components.

3. Trial Design

The registry describes a randomized, open-label, parallel-group trial and lists both phase II and phase III. It is titled as a multicenter study, and its status is completed. Enrollment was 75. The primary completion date is listed as 2016-04, and the start date as 2005-06. The registry does not report further details about how the random sequence was generated, how treatment allocation was concealed, or the timing and conduct of enrollment at individual centers.

Participants were assigned to mHSCT or intravenous cyclophosphamide. The mHSCT arm included collection of participants’ hematopoietic stem cells, high-dose chemotherapy and radiation, and autologous stem cell transplantation. The official title names total body irradiation, cyclophosphamide, ATGAM, and auto-CD34+HPC. The cyclophosphamide arm received high-dose intravenous cyclophosphamide, described as 12 monthly pulses. The arm descriptions do not provide a complete account of all treatment procedures in the record available for this analysis.

Because masking is listed as none, the study was open label: participants and relevant study personnel were not blinded to assigned treatment. The registry does not report whether outcome assessment or adjudication was blinded. That distinction matters because knowledge of assignment can affect care, reporting, and some assessments, while it does not by itself establish that any particular measurement was biased.

The primary endpoint was assessed at 54 months after randomization in the intention-to-treat population. The record also lists analyses at other time points and in per-protocol populations among secondary outcomes. The analysis population definitions and numbers vary by outcome. Details of the protocol’s planned statistical methods, sample-size justification, interim analyses, and procedures for departures from assigned treatment are not reported in the registry record. The available information therefore describes the broad design and endpoint structure, but not every operational or analytical choice.

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: Intention to treat (ITT). The ITT population includes all randomized participants.

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: Randomized intent to treat

6. Primary Endpoint

The registry lists the primary outcome as Global Rank Composite Score (GRCS) (Month 54, ITT), measured over 54 Months Post-Randomization.

The primary endpoint is the Global Rank Composite Score (GRCS) at 54 months after randomization, analyzed in the intention-to-treat population. It is defined as a sum of subject-pair comparison scores. Rather than measuring one symptom or organ system, it compares participants using a hierarchy: death, event-free survival, forced vital capacity, Health Assessment Questionnaire–Disability Index, and Modified Rodnan Skin Score.

The hierarchy gives priority to survival-related outcomes. Participants who were alive ranked higher than those who died, and participants who survived event-free ranked higher than those with an event-free survival failure. The registry defines event-free survival failure to include death and specified respiratory, renal, or cardiac events. For the lower components, change from baseline is classified in ordinal categories such as improvement, stability, or worsening. The record describes thresholds for those categories, including ±10% change in FVC percent predicted, ±0.4 change in HAQ-DI, and ±25% change in mRSS.

The rationale for a composite of this kind is to represent several important disease-related outcomes in a single ordering, while allowing more consequential outcomes to take precedence. That can be useful when a condition has multiple manifestations and no single outcome captures the full comparison of interest. The trade-off is interpretability: a GRCS value is not itself a direct measure of disease activity or severity. Its meaning depends on the hierarchy, the pairwise scoring procedure, and the mix of component outcomes contributing to the ranking. The registry reports the median and full range for each group, alongside a Wilcoxon (Mann-Whitney) comparison. It does not report a confidence interval for the primary group comparison.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the primary rank test compare?

The Wilcoxon (Mann-Whitney) test is a rank-based method for comparing observations from two independent groups. In general, it asks whether values from one group tend to be higher or lower in the combined ordering than values from the other group. It does not directly test whether the group means differ. Its interpretation is particularly relevant for the GRCS because that endpoint is itself built from ordered participant comparisons rather than being a conventional continuous measurement.

A rank test does not require the outcome to follow a normal distribution in the way a standard two-sample t test does. However, a result should not automatically be translated into a difference in medians alone. Such an interpretation depends on the shapes of the group distributions and other assumptions. The registry reports group medians and full ranges, but it does not report distribution plots or a confidence interval for the primary rank comparison. The p-value therefore provides a test result, not a complete description of effect magnitude and precision.

What does a p-value tell the reader?

In general, a p-value describes how unusual data at least as incompatible with a specified null hypothesis would be if that null hypothesis and the test assumptions were true. The primary p-value is 0.013. It is not the probability that the treatment groups are truly identical, nor is it the probability that the result occurred “by chance.” It also does not measure clinical importance. Interpretation should consider the endpoint’s construction, the amount of information in the trial, the uncertainty around the effect, and the analysis plan.

The registry labels the primary analysis as a superiority comparison. That label describes the testing framework, not a guarantee that every relevant outcome favors one arm. The reported score is a composite hierarchy, so the test concerns the rank-based composite result. A result for this endpoint cannot identify which component drove the ordering without component-specific results and analyses.

Why is the composite hierarchical?

In general, a hierarchical composite can put outcomes considered more consequential ahead of others. A participant’s comparison can be decided by survival or event-free survival before the comparison reaches lung function, disability, or skin score. This may preserve a priority ordering across different outcome types, but it also means the resulting score is not measured on a single familiar clinical scale. The GRCS is defined as a sum of subject-pair comparison scores, and the registry describes it as an analytic tool that accounts for multiple disease manifestations simultaneously. It specifically cautions that it does not measure clinical disease activity or severity.

The underlying components do not all share the same measurement properties. Death and specified event-free survival failures are event outcomes; FVC, HAQ-DI, and mRSS are assessed through changes from baseline and classified using ordinal categories. Consequently, a composite difference may reflect multiple types of information. The registry does not report component contributions to the primary comparison, so the relative influence of those outcomes cannot be determined from the posted primary summary alone.

How should the secondary binary analyses be read?

The registry reports Fisher exact tests for treatment-related mortality. Fisher’s exact test is a method for comparing categorical outcomes between groups, often used when counts are small. It evaluates the evidence for an association between group and event status under its null hypothesis. The reported intention-to-treat p-value is 0.48, and the per-protocol p-value is 0.49. These results do not prove that mortality risks are equal; with few events, tests can have limited ability to distinguish differences.

The posted results also include simple unadjusted recalculations expressed as risk differences with confidence intervals. In general, a risk difference expresses the absolute contrast in event proportions between groups, while a confidence interval describes a range of effects compatible with the data and method used. Those recalculations are not the same analysis as the registry’s Fisher exact test. They should be distinguished from the sponsor’s reported inference rather than treated as a replacement for it.

What do planning illustrations mean?

In general, a sample-size calculation asks how many participants would be needed under specified assumptions to have a chosen probability of detecting an effect at a stated significance level. The planning illustrations associated with these results use the observed effect as a reference, a two-sided significance level of 0.05, equal allocation, and a normal approximation; the effect to detect is a stated fraction of the observed effect. For time-to-event outcomes, the illustrations concern the total number of events; for other outcomes, they concern participants per arm.

These are planning illustrations, not observed power and not evidence that the trial was designed around those assumptions. The registry’s own sample-size justification is not reported. Calculations based on an observed effect can be sensitive to the uncertainty in that estimate, so they should not be read as independent confirmation of the primary result or as a substitute for the original design rationale.

What analytical details remain unclear?

The record lists 41 outcome measures and 64 analyses, with 40 secondary outcomes posted. It does not report the complete prespecified statistical analysis plan, the handling of missing outcome data, any interim analysis plan, or any adjustment for multiple testing. Those omissions limit the ability to determine how closely the posted analyses track the original protocol or how the chance of false-positive findings across multiple tests was addressed. General statistical interpretation must therefore separate the reported tests from assumptions or procedures that are not documented in the registry record.

10. Analyses Reported by the Sponsor

The sponsor’s posted primary analysis is the Wilcoxon (Mann-Whitney) comparison of the 54-month intention-to-treat GRCS, with p = 0.013. The registry reports the group medians and ranges but does not report a separate recalculated primary effect estimate or confidence interval. The primary interpretation should therefore remain tied to the sponsor’s rank-based analysis and the endpoint’s hierarchical construction.

For treatment-related mortality, the sponsor reports Fisher exact tests. The intention-to-treat analysis has p = 0.48, and the per-protocol analysis has p = 0.49. The posted unadjusted recalculations express the comparison as risk differences with confidence intervals and report p-values of 0.295 for intention to treat and 0.306 for per protocol. These are not identical inferential procedures: the sponsor’s tests are Fisher exact tests, while the recalculations are simple unadjusted comparisons of the posted summary statistics. Differences between their p-values should not be treated as a contradiction or as evidence that one method has uncovered a definitive result.

The registry record does not give enough detail to reconstruct a full analysis plan or to establish which estimand and inferential approach were prespecified for every secondary comparison. The sponsor’s reported analyses remain the official posted analyses; the recalculations are supplementary descriptions of the summary data, not replacements for the sponsor’s methods. Neither analysis makes sparse event counts precise, and neither supports a claim of equivalence between groups.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 41 outcome measures, 64 analyses, and 40 posted secondary outcomes, in addition to the primary outcome. A broad outcome set can be useful for examining different manifestations and follow-up points, but it creates a multiplicity issue: when many statistical tests are conducted, some results can appear noteworthy even if there is no underlying effect. In general, a prespecified strategy may control the interpretation of multiple comparisons through adjustment, a hierarchy of testing, or a clear distinction between confirmatory and exploratory outcomes.

An adjustment for multiplicity, a formal testing hierarchy beyond the endpoint’s own clinical ranking, or a strategy for classifying secondary analyses is not reported in the registry record. Therefore, the reported secondary p-values should not automatically be treated as if each were an independent confirmatory test. The registry’s primary outcome is identifiable, but the available record does not show how the broader set of comparisons was handled in the original statistical plan.

The primary analysis uses the intention-to-treat population, defined in the record as all randomized participants. This generally preserves the comparison created by random assignment, even when participants do not follow the assigned treatment exactly. The registry also lists per-protocol analyses for secondary outcomes. Per-protocol populations can answer a different question because they include participants selected according to adherence or other protocol criteria; the criteria and their application are not reported in the record available for these results.

Missing-data procedures are not reported in the registry record. The available information does not explain how missing endpoint measurements, incomplete follow-up, or participants without an evaluable outcome were handled. It also does not state whether any sensitivity analyses assessed the effect of missing data. These details matter because missingness can affect both the number of participants contributing to an analysis and the interpretation of the comparison.

The record does not report an interim analysis plan, stopping rules, or the protocol’s sample-size justification. Their absence from the registry record does not establish that none existed; it means these features cannot be assessed from the posted information. The planning illustrations calculated from observed effects are not evidence of the trial’s original design assumptions.

12. 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: pneumonia (14.7%), lymphopenia (11.8%), pulmonary function test decreased (10.8%), respiratory failure (10.8%), pneumonitis (8.8%).

The registry reports serious adverse events affecting 25 of 34 participants at risk in the mHSCT group and 19 of 37 in the cyclophosphamide group. Other, non-serious adverse events affected 33 of 34 participants at risk in the mHSCT group and 37 of 37 in the cyclophosphamide group. These counts describe participants with events, not the total number of event episodes. A participant may experience more than one event, and the counts alone do not describe severity, duration, timing, or causality.

The registry also reports 7 deaths in the mHSCT group and 14 in the cyclophosphamide group in its adverse-event summary. Those figures should not be equated with the treatment-related mortality endpoint, which is separately reported as a participant count at specified follow-up. The record does not establish that every death in the adverse-event summary was treatment-related or explain the relationship between those deaths and the narrower endpoint.

The listed serious adverse event terms include pneumonia, lymphopenia, decreased pulmonary function tests, respiratory failure, pneumonitis, and congestive cardiac failure, each with a registry-reported percentage. The record does not specify the denominator or group attribution for each listed percentage in the available summary, so those percentages should not be assigned to a treatment arm. Counts of affected participants are useful descriptive safety information, but without event-level detail and a prespecified comparative safety analysis, they do not establish the cause of an event or provide a complete estimate of comparative risk.

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

The primary result describes a statistically detectable difference in a rank-based composite comparison at 54 months, with group medians of 17.0 for mHSCT and -6.0 for cyclophosphamide and a reported Wilcoxon (Mann-Whitney) p-value of 0.013. The endpoint orders participants according to a hierarchy of survival, event-free survival, lung function, disability, and skin score. The result is therefore about this particular composite ranking in the intention-to-treat population.

It does not mean that the GRCS is a direct measure of disease activity or severity; the registry explicitly says it is not. It does not establish that every component favored the same group, identify which component contributed most, or provide a stand-alone estimate of change in lung function, disability, or skin score. It also does not show how much difference would be clinically important. The record does not report a primary confidence interval or a clinical threshold for interpreting the score.

The statistical result should not be confused with proof that one treatment is universally preferable, or that the observed result will apply to people outside the study’s severe systemic sclerosis population. The registry does not provide eligibility details sufficient to characterize that population further, nor does it report enough information to assess the full range of treatment procedures, adherence, missing outcomes, or subgroup effects.

The secondary analyses and safety counts add context but have limitations. Treatment-related mortality analyses include few events and do not demonstrate equivalence. Many secondary outcomes and analyses are listed, while a multiplicity strategy is not reported. Adverse-event counts describe participants affected, but do not alone show causation or fully characterize the balance of benefits and harms.

In general, a trial result is one contribution to a larger evidence base. The registry record supports a careful statement about the posted primary rank comparison and its statistical test, not a broad clinical recommendation. Interpretation requires attention to the composite’s construction, the uncertainty and missing details, and the distinction between statistical evidence and clinical importance.

14. Why This Trial Matters Statistically

This trial is statistically notable because its primary outcome is not a single conventional clinical measurement. The GRCS combines ordered comparisons across survival, event-free survival, lung function, disability, and skin score. That approach addresses the challenge of evaluating a disease with multiple manifestations, while making interpretation more dependent on the hierarchy and scoring rules than a simple outcome such as a measurement at one time point.

The endpoint also illustrates a broader methodological trade-off. A composite can capture several dimensions and prioritize more serious outcomes, but a single summary can obscure which component drives the group comparison. Here, the registry provides a median and full range for the composite and a rank-based test, but does not report component contributions or a primary confidence interval. Statistical significance can therefore be identified without a complete account of effect magnitude, precision, or clinical relevance.

The design offers another useful teaching point: randomized assignment supports a treatment-group comparison, while open-label conduct and missing details about outcome assessment mean some aspects of implementation cannot be evaluated from the registry record. The primary analysis is intention to treat, while the record also lists per-protocol analyses among secondary outcomes. These populations answer related but not identical questions.

Finally, the registry lists many outcomes and analyses, illustrating why multiplicity and transparent analysis plans matter. A clearly identified primary endpoint helps orient interpretation, but a large set of secondary tests needs context about prespecification and adjustment. The record’s lack of detail on multiplicity, missing data, and sample-size justification limits how confidently readers can reconstruct the original inferential strategy. These features make the trial informative not only for its posted result, but also for understanding how endpoint design and registry reporting shape the interpretation of clinical evidence.

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 registry record leaves important design and analysis details unreported. It does not include the full eligibility criteria, the randomization procedure, allocation concealment details, or whether outcome assessors were blinded. The study is open label, but the record does not provide enough detail to evaluate the potential influence of treatment awareness on individual outcome assessments.

The primary composite is complex. Although its hierarchy and major components are described, the primary result does not include a confidence interval or a component-by-component account of how participants contributed to the overall comparison. The score is explicitly not a measure of clinical activity or severity, and the record does not report a threshold for clinical importance. Those limits make it difficult to translate the rank result into a familiar clinical magnitude.

Missing-data handling, protocol deviations, adherence, and sensitivity analyses are not reported in the registry record. The per-protocol population criteria are also not described in the available information. As a result, the robustness of comparisons to incomplete follow-up or the population restrictions used in per-protocol analyses cannot be assessed.

The registry lists many outcomes and analyses, but a multiplicity strategy is not reported. The distinction between prespecified and exploratory secondary analyses is therefore unclear. The original sample-size justification and interim analysis plan are also not reported; planning illustrations based on the observed effect do not resolve those gaps.

Safety summaries provide counts of affected participants and selected event terms, but the available information does not give enough context to attribute every listed percentage by treatment group, establish causality, or understand the timing and recurrence of events. The adverse-event death counts and the narrower treatment-related mortality endpoint should not be conflated. These limitations do not invalidate the posted results, but they constrain the conclusions that can be drawn from the registry record alone.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Scleroderma: Cyclophosphamide or Transplantation — NCT00114530.
• Registry record last updated 2023-04-12; status last verified 2023-03.
• Results first posted 2017-07-14.
• 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

NCT00114530 is a completed randomized, open-label, parallel-group study of mHSCT versus high-dose intravenous cyclophosphamide in severe systemic sclerosis. The registry lists phase II and phase III, enrollment of 75, and a multicenter study in the official title. The primary outcome is the intention-to-treat GRCS at 54 months after randomization, a ranked composite of survival, event-free survival, forced vital capacity, disability, and skin score.

The registry reports group medians of 17.0 for mHSCT and -6.0 for cyclophosphamide, with a Wilcoxon (Mann-Whitney) p-value of 0.013. The GRCS is not a direct measure of clinical disease activity or severity. Secondary analyses cover additional outcomes, time points, and populations, while safety summaries report serious and other adverse events, deaths, and selected serious event terms.

The record does not report the full statistical analysis plan, missing-data procedures, multiplicity strategy, primary confidence interval, or original sample-size justification. These omissions limit assessment of precision and the broader inferential context.