Extracorporeal Photopheresis (ECP) With Methoxsalen for Chronic Graft Versus Host Disease (cGVHD): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Extracorporeal Photopheresis (ECP) With Methoxsalen for Chronic Graft Versus Host Disease (cGVHD) (NCT01380535), a randomized early/phase 1 study. The primary outcome is Number of Participants With an Overall Response at Week 28.

2. Clinical Question

The trial addresses whether adding ECP with methoxsalen to standard care is associated with a different observed response count than standard care alone. The result is a count of participants meeting the overall-response definition, not a measure of how large an individual participant’s response was. The posted outcome does not separately report partial and complete responses, so their contributions to the overall count cannot be distinguished.

The registry describes this as a randomized controlled study in people with moderate to severe cGVHD. Its reported results can describe the observed outcome in the study groups, but the interpretation depends on the endpoint definition, the number of outcomes recorded, and the uncertainty around the comparison. The response assessment was blinded to treatment assignment according to the registry description; the registry record does not identify which study staff were masked or give further details of that process.

3. Trial Design

This was a randomized, single-masked, parallel-group study of ECP therapy with UVADEX for moderate to severe chronic graft-versus-host disease. The registry lists the phase as early phase 1, an enrollment of 60, and a completed status. The study began in 2011-11, and primary completion was recorded as 2015-03. The sponsor was Mallinckrodt.

One group received methoxsalen administered via ECP in addition to standard of care. The comparison group received standard of care only. The registry does not specify the components of standard care, how random allocation was generated or concealed, or who was masked under the single-masking designation. It does state that study staff assessing the primary response did not know treatment assignment.

The primary outcome was the number of participants with an overall response at Week 28. The registry identifies the outcome population as intention to treat. The posted group results are 20 of 29 for ECP with methoxsalen plus standard of care and 14 of 24 for standard of care. Those are the outcome counts and denominators reported for this measure; the registry record does not explain how they relate to the enrollment figure or describe any exclusions from the outcome analysis.

The record lists one outcome measure and one analysis, with no secondary outcomes posted. It does not report the trial’s original sample-size justification, the intended analysis plan, or whether interim analyses were planned or conducted. Consequently, the enrollment figure alone does not establish what effect the study was designed to detect or how the final analysis was intended to handle missing outcome data.

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: All participants enrolled were included in the safety analysis set

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intention to treat population

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: All participants enrolled and included in the safety analysis set

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants With an Overall Response at Week 28, measured over Week 28.

The primary endpoint counts participants with an overall response at Week 28. A response includes either a partial response or a complete response, as assessed by study staff who did not know the participant’s treatment assignment. The registry identifies the analysis population as intention to treat.

A binary endpoint such as this classifies each participant according to whether the defined outcome occurred. The reported result therefore gives the number who met the response definition and the number in the outcome denominator for each group. It does not show the degree of improvement among responders, the separate counts of partial and complete responses, or how response changed over time. Those details are not reported in the registry record.

Using a common response definition and a stated assessment time makes it possible to compare the groups on the same outcome. Blinding the assessors to treatment assignment may reduce the risk that knowledge of treatment influences classification, although the registry record does not provide further information about assessment procedures. The endpoint is a count of participants with the event, not a count of all response episodes or a time-to-event measure.

The Week 28 result is the registry’s primary reported outcome. Its interpretation should be read together with the estimated group difference and confidence interval: the counts alone do not indicate how precisely the underlying difference has been estimated or whether the observed contrast is compatible with no difference.

7. Results Posted in the Registry

• Standard of Care (SoC): 58.3% (14/24) versus 69.0% (20/29) on control; risk difference -10.6 percentage points (95% CI -36.6 to 15.3), risk ratio 0.85 (0.56–1.28); unadjusted two-proportion z-test p = 0.422.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a binary endpoint mean?

A binary endpoint records whether a defined event occurred for each participant. For this trial, the event is an overall response at Week 28, with partial response and complete response both counted as responses. The group results are counts of participants with that event, paired with the reported denominators. A binary result does not retain finer distinctions in response status once participants have been classified, and the registry does not post the separate partial-response and complete-response counts.

What does the risk difference describe?

The risk difference compares the response proportions in the two groups by subtraction. The posted comparison is oriented as standard care minus ECP with methoxsalen plus standard care. Accordingly, a negative estimate means the observed response proportion was lower in the standard-care group. The estimate of -0.106 is the reported difference on the proportion scale; it is not a count of participants. Its confidence interval runs from -0.366 to 0.153, covering negative and positive differences as well as zero.

A risk difference is an absolute measure: it describes the gap between the group proportions. It can be useful because it keeps the comparison tied to the event proportions observed in the groups. Its meaning depends on the specified direction of subtraction, so stating which group is subtracted from which is essential.

How do relative measures differ?

The risk ratio compares the response proportions as a ratio, and the odds ratio compares the odds of response. For standard care relative to ECP plus standard care, the posted risk ratio is 0.846, with a confidence interval from 0.557 to 1.284; the posted odds ratio is 0.63, with a confidence interval from 0.203 to 1.951. A ratio of 1 represents no relative difference. Both intervals include that reference value. These measures are not interchangeable with the risk difference: each places the comparison on a different scale, and the odds ratio is not itself a difference in proportions.

What does a p-value tell the reader?

A p-value describes how compatible the observed data are with a specified null model, under the assumptions of the test. It is not the probability that the null hypothesis is true, nor does it measure the clinical importance of an effect. The summary-based unadjusted recalculation reports p = 0.422. The registry reports a Fisher Exact p-value of 0.373. The registry record does not give enough procedural detail to explain the difference between these values, so they should be identified by their source rather than treated as the same calculation.

Fisher’s exact test is a method for comparing categorical outcomes using the observed counts. In general, exact methods evaluate possible arrangements of those counts under a null model rather than relying on a large-sample approximation. The posted registry analysis names this method, but does not report additional test settings or an effect estimate and confidence interval from that analysis.

Why does a confidence interval matter?

A confidence interval conveys uncertainty around an estimated effect. It shows a range of values that are compatible with the data and the interval procedure, rather than only a single point estimate. Here, the risk-difference interval extends from -0.366 to 0.153 and includes zero. That breadth indicates that the data do not pin down the contrast tightly. In general, a wider interval means less precision, while a narrower interval means greater precision, subject to the method and assumptions used.

The registry record does not state the confidence level or fully specify the procedure used for the posted interval. The interval should therefore be interpreted using the information actually reported, without assigning an unreported confidence level or assuming an unreported adjustment.

What is over-dispersion, and does it apply?

In general, over-dispersion means that observed variability is greater than a statistical model expects. It can matter for outcomes modeled as counts, because a model that assumes too little variability can produce uncertainty estimates that are too narrow. This primary endpoint is described as a binary participant count, with no dispersion statistic reported. The registry record does not report an over-dispersion analysis, so there is no basis to claim that over-dispersion was assessed or affected this result.

What analysis population is identified?

The primary endpoint is labeled as using the intention-to-treat population. In general, an intention-to-treat analysis aims to retain participants in the groups to which they were randomized, regardless of subsequent treatment changes. The label alone does not disclose how missing outcomes, withdrawals, or other data issues were handled. The registry record does not report those details, nor does it explain the relationship between the outcome denominators and the enrollment figure.

10. Confidence Intervals and Precision

The estimated risk difference is -0.106 for standard care minus ECP with methoxsalen plus standard of care. Its confidence interval extends from -0.366 to 0.153. Because the interval includes zero, it is compatible with no difference as well as with a difference favoring either group under the stated direction. The point estimate is only one value within a substantially broader range of plausible contrasts.

The relative-effect intervals tell a similar story of uncertainty. The risk ratio is 0.846, with a confidence interval from 0.557 to 1.284. The odds ratio is 0.63, with a confidence interval from 0.203 to 1.951. Each interval includes the value representing no relative difference. The interval endpoints also allow effects in different directions, so the estimates do not precisely settle the direction or magnitude of the group contrast.

Confidence intervals are useful because they show how much uncertainty remains around a point estimate. A result with an interval spanning a broad range should not be read as a precise estimate merely because the point estimate is printed with several decimal places. The observed counts were 20 of 29 and 14 of 24, and those data support only the precision conveyed by the posted intervals.

The registry record does not state the confidence level or all details of the interval method. No unreported level should be assumed. The recalculated intervals are unadjusted summaries; the registry’s Fisher Exact analysis reports a p-value but no corresponding interval in the posted analysis. Precision is therefore best described from the reported estimates and interval limits, not inferred from a single significance test.

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 figures are planning illustrations based on the observed effect, not the trial’s original design assumptions. They give participants per arm for 80% power and participants per arm for 90% power. They are not total events: this endpoint is a participant outcome, and the illustration reports participants in each arm. The registry record does not report the trial’s original sample-size justification or design assumptions.

The illustration assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It considers detecting the stated fraction of the observed effect. Planning for 100% of the observed effect gives 321 participants per arm for 80% power and 429 participants per arm for 90% power. Planning for 75% of the observed effect gives 561 participants per arm for 80% power and 750 participants per arm for 90% power. Planning for 50% of the observed effect gives 1,240 participants per arm for 80% power and 1,660 participants per arm for 90% power.

In general, the 80% column describes a design with an 80% chance of detecting the specified effect under the stated assumptions if that effect is the true one; the 90% column expresses the corresponding chance at the higher target. Smaller effects are harder to detect and therefore require more participants under the same planning framework. These are prospective planning illustrations, not observed power for this completed study.

The curve should be read as a visual representation of the same planning relationship, not as a record of enrollment or a new trial result. The registry record does not report how the original enrollment target was chosen, so the illustration should not be substituted for the trial’s own design rationale.

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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.690; treatment proportion p₁ = 0.583; average p̄ = 0.636.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6802 = 1.3333.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6761 = 0.5690.
4. n = (1.3333 + 0.5690)² / (-0.1063)² = 320.1, rounded up to 321 participants per arm for 80% power.

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

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.689655172413793                          # control proportion
p1 <- 0.583333333333333                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(1380535)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 321 (80% power), 429 (90% power)
 75% of observed effect: n per arm = 561 (80% power), 750 (90% power)
 50% of observed effect: n per arm = 1,240 (80% power), 1,660 (90% power)
power.prop.test: 320.1 
Simulated power at n = 321 per arm: 0.802 (20000 replicates)

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

13. Analyses Reported by the Sponsor

The registry reports the sponsor’s analysis as a Fisher Exact test, with p = 0.373. It does not post an effect estimate, confidence interval, or additional procedural details for that analysis. The summary-based unadjusted recalculation reports a risk difference of -0.106, with a confidence interval from -0.366 to 0.153, and p = 0.422. The comparison is oriented as standard care minus ECP with methoxsalen plus standard of care.

The reported p-values are not identical. The registry record does not give enough information to determine precisely why they differ, including the implementation details needed to reproduce the sponsor’s test from the posted record. They should therefore remain clearly distinguished: the Fisher Exact result is the registry-reported analysis, while the other figures are unadjusted recalculations from the posted group summaries. The record does not flag a discrepancy.

The sponsor’s p-value does not replace the effect estimate or its uncertainty. A test result alone cannot show the size or direction of the contrast, and the recalculated estimate and interval help describe those features. Conversely, a recalculation from summary counts does not establish that it reproduces every element of the sponsor’s analysis. The original statistical analysis plan and further test details are not reported in the registry record.

14. Multiplicity, Missing Data and Analysis Populations

The registry posts one primary outcome measure and one analysis, with no secondary outcome results posted. It does not report whether the investigators planned to test additional outcomes, subgroups, or multiple versions of the primary analysis. The multiplicity strategy, if any, is therefore not reported in the registry record. With no posted secondary results, there is no basis to assess consistency across other outcomes or to determine whether any adjustment for multiple testing was used.

Missing-data handling is also not described. The primary outcome is labeled as using the intention-to-treat population, but the record does not say how participants without a Week 28 response assessment were treated, whether any outcomes were imputed, or whether a sensitivity analysis was performed. The outcome denominators are 29 for ECP with methoxsalen plus standard of care and 24 for standard of care; the registry does not explain the difference between those denominators and the stated enrollment.

The record does not report interim analyses, stopping rules, or a prespecified statistical analysis plan. Those details should not be inferred from the randomized design or from the reported p-value. It also does not provide subgroup analyses or additional population definitions that would clarify the extent to which the posted analysis followed an intention-to-treat principle in practice.

These omissions matter because analysis choices can affect which participants contribute to a result and how uncertainty is calculated. The posted result remains interpretable as the reported Week 28 group counts and comparison, but the registry record does not provide enough information to assess all potential sources of analytic flexibility or missing-outcome bias.

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: sepsis (7.1%), myopathy (7.1%), cardiac failure (3.6%), thyroiditis subacute (3.6%), graft versus host disease in liver (3.6%).

The registry reports serious adverse events among 28 participants at risk in each group. Serious events affected 8 participants in the ECP methoxsalen plus standard-of-care group and 9 in the standard-of-care group. Other adverse events affected 26 participants and 24 participants, respectively. The record also lists deaths as 4 in the ECP group and 0 in the standard-of-care group.

These are posted counts, not causal assessments. The registry record does not specify whether the deaths were judged related to treatment, nor does it provide enough detail to interpret the timing or causes of those deaths. The adverse-event at-risk counts are not the same as the Week 28 primary-outcome denominators, and the record does not explain the relationship between them.

The serious-event terms listed include Sepsis and Myopathy, each at 7.1%; Cardiac Failure, Thyroiditis subacute, Graft versus host disease in liver, and Bronchopulmonary aspergillosis, each at 3.6%. The registry excerpt does not identify these terms by treatment group or state whether they refer to participants or event occurrences. They should not be used to infer comparative risks or treatment causality without that context.

Counts and at-risk denominators provide a descriptive safety summary, but they do not by themselves establish whether an event was caused by treatment. The registry record does not provide a detailed safety analysis or further attribution information.

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

The posted results describe the Week 28 overall-response outcome in a randomized comparison of ECP with methoxsalen plus standard of care against standard of care alone. The observed counts were 20 of 29 and 14 of 24, respectively. The summary-based risk difference is negative when expressed as standard care minus ECP, while the reported confidence interval includes zero and extends to values in either direction.

This means the observed group counts favor a higher response proportion in the ECP group, but the estimate is imprecise. The interval and the relative-effect intervals do not establish a definitive direction or size for the underlying difference. The Fisher Exact p-value reported by the registry is 0.373, while the unadjusted recalculation reports p = 0.422; neither value describes clinical importance or eliminates the uncertainty shown by the effect estimates.

The result does not establish that one treatment works or does not work. It is a comparison of the reported outcomes in this study, under the registry’s endpoint definition and analysis population label. The record does not give the original sample-size rationale, detail how missing outcomes were handled, or provide secondary outcomes that might broaden interpretation. It also does not describe the specific standard-of-care treatments.

Safety figures are likewise descriptive. The posted serious-event and death counts should not be interpreted as proof of treatment causation, and the record does not attribute the deaths. In general, clinical decisions require consideration of evidence beyond a single trial result; these posted data alone are not medical advice.

The most defensible reading is therefore bounded: the registry reports an observed Week 28 response contrast, with a point estimate favoring the ECP group in the stated direction and substantial uncertainty around that estimate. The record does not resolve the true magnitude of the difference or provide all details needed to evaluate the analysis and safety findings fully.

17. Why This Trial Matters Statistically

This trial illustrates why an effect estimate and its uncertainty should be read alongside the raw outcome counts. The counts communicate what was observed in each group, while the risk difference gives the direction and scale of the contrast. The confidence interval makes clear that the estimate is not precise enough to rule out materially different underlying effects.

It also demonstrates why the direction of a comparison needs to be explicit. The posted risk difference is standard care minus ECP with methoxsalen plus standard of care, so a negative value corresponds to a lower observed response proportion in the standard-care group. Reading the sign without checking the subtraction order could reverse the interpretation.

The record offers both a sponsor-reported Fisher Exact p-value and an unadjusted recalculation from the posted summary statistics. Keeping those results distinct is important: an analysis based on summary data is not necessarily an exact reproduction of the registry analysis. The record does not provide the extra details needed to reconcile their values, and a p-value should not take the place of the estimated effect and its interval.

Finally, the sample-size illustration shows the planning consequences of targeting smaller effects under stated assumptions. It is separate from the original trial design, whose assumptions are not reported, and it is not evidence about the study’s observed power. This distinction helps prevent retrospective planning calculations from being mistaken for a trial’s prespecified justification or for a verdict on its findings.

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 primary result is limited to the overall-response definition at Week 28. The registry does not report the separate numbers of partial and complete responses, the timing of responses, or outcomes at other follow-up times. No secondary outcome results are posted, so the primary finding cannot be compared with additional efficacy measures in this record.

The outcome denominators differ from the stated enrollment, but the registry does not explain the difference or report the handling of missing Week 28 assessments. Although the outcome is labeled intention to treat, the record does not describe exclusions, imputation, or sensitivity analyses. This limits assessment of how fully the analysis retained the randomized groups.

The trial is described as early phase 1, but the registry does not report its original sample-size rationale, statistical analysis plan, interim-analysis procedures, or allocation-concealment methods. It also does not identify the person or staff role covered by single masking, beyond stating that response assessors did not know assignment. The standard-of-care regimen is not specified, limiting detail about the comparison treatment.

The sponsor’s Fisher Exact p-value differs from the summary-based unadjusted recalculation, and the registry record lacks the procedural details needed to reconcile them. The safety record gives affected-participant counts and serious-event terms, but does not provide detailed causal attribution; the listed serious-event terms are not assigned to treatment groups in the registry entries. The death counts are reported without causes or treatment-relatedness information. These limitations constrain interpretation of both the efficacy and safety summaries.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Extracorporeal Photopheresis (ECP) With Methoxsalen for Chronic Graft Versus Host Disease (cGVHD) — NCT01380535.
• Registry record last updated 2020-02-19; status last verified 2020-02.
• Results first posted 2019-10-29.
• 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

NCT01380535 is a completed randomized, single-masked, parallel-group study sponsored by Mallinckrodt. The registry lists the phase as early phase 1, enrollment as 60, study start as 2011-11, and primary completion as 2015-03. It compared ECP with methoxsalen plus standard of care with standard of care alone in participants with moderate to severe chronic graft-versus-host disease.

The primary endpoint was the number of participants with an overall response at Week 28, defined as partial or complete response and assessed by staff unaware of treatment assignment. The posted counts are 20 of 29 in the ECP group and 14 of 24 in the standard-of-care group. The recalculated risk difference is -0.106, with a confidence interval from -0.366 to 0.153; the registry reports a Fisher Exact p-value of 0.373.

No secondary outcome results are posted. Missing-data procedures, original sample-size assumptions, and detailed statistical-plan information are not reported in the registry record. Safety results include serious adverse events, other adverse events, and deaths, without sufficient detail to establish causation.