Low-Dose Prednisone or Methylprednisolone in Treating Patients With Newly Diagnosed Acute Graft-versus-Host Disease: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Low-Dose Prednisone or Methylprednisolone in Treating Patients With Newly Diagnosed Acute Graft-versus-Host Disease (NCT00929695), a randomized phase 3 study. The primary outcome is Mean Cumulative Prednisone Dose (mg/kg) Over 42 Days From the Start of Treatment.

2. Clinical Question

This was a randomized phase III study of initial treatment. The question concerns dose exposure over a specified follow-up period, rather than whether one strategy improved survival or another clinical outcome. The registry also identifies secondary outcomes concerning toxicity, mortality, malignancy and GVHD progression. Those outcomes provide additional context, but they are distinct questions and do not replace the primary comparison.

3. Trial Design

The registry identifies this as a randomized, parallel-group, open-label phase III study. It records enrollment of 164 and a completed status. The study began in 2009-06, and primary completion was recorded in 2013-07. Fred Hutchinson Cancer Center is listed as sponsor.

The treatment strategies were low-dose and standard-dose prednisone or methylprednisolone. Both could be given once or twice daily; the arm descriptions state that treatment continued in the absence of disease progression or unacceptable toxicity. The registry lists prednisone dose levels separately for Grade IIa and Grade IIb-IV GVHD. These grade-specific descriptions matter when reading the primary outcome comparisons: the sponsor's reported tests compare dose groups within grade categories.

Masking is listed as none. The registry lists prednisone, methylprednisolone and questionnaire administration as interventions. It does not report further detail about randomization procedures, allocation concealment, treatment adherence, dose adjustments, or the exact role of questionnaire administration.

The primary outcome was the mean cumulative prednisone dose in milligrams per kilogram through day 42 after treatment initiation. The registry says this measure was available for 152 of the 164 enrolled patients. It attributes the 12 missing measurements to withdrawal from study (2) and discharge from the Center before day 42 (10). The remainder of the population statement is incomplete, so further details about the analysis population are not reported in the registry record.

The registry lists 12 outcome measures and 5 analyses. It does not report the trial's original sample-size justification or the full statistical analysis plan. In particular, details about the planned treatment of missing data, adjustment for multiple comparisons, or interim analyses are not reported in the registry record.

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: From a total enrollment of 164 patients, the cumulative dose of prednisone at day 42 of treatment was available in 152 patients. The primary outcome was not measured in 12 patients due to withdrawal from study (2), discharge from Center before day 42 of treatment (10). Analysis was not completed in two patients due to an error in stratification.

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: From a total enrollment of 164 patients, 150 were analyzed for the primary endpoint. Excluded from the analysis were 14 patients: 12 withdrew, died or were discharged from the Center early and did not have 42 days of cumulative prednisone dosing and 2 were stratified incorrectly.

6. Primary Endpoint

The registry lists the primary outcome as Mean Cumulative Prednisone Dose (mg/kg) Over 42 Days From the Start of Treatment, measured over At day 42 after initiation of treatment.

The primary endpoint is the mean cumulative prednisone dose, measured in milligrams per kilogram, from the start of therapy through study day 42. It describes total prednisone exposure during the specified treatment period. The registry reports the outcome at day 42 after initiation of treatment and gives a mean and standard deviation for each listed group.

This endpoint is useful for asking whether the treatment strategies led to different amounts of prednisone exposure over the follow-up period. It is not itself a measure of survival, GVHD resolution, infection, or quality of life. Those are separate clinical outcomes, some of which appear among the listed secondary outcomes.

The registry reports the primary outcome for four dose-and-grade groups. For Grade IIa GVHD, the mean cumulative dose was 22.2 mg/kg in the 0.5 mg/kg/d prednisone group and 27.1 mg/kg in the 1.0 mg/kg/d prednisone group. For Grade IIb-IV GVHD, the reported means were 38.4 mg/kg with 1.0 mg/kg/d prednisone and 41.3 mg/kg with 2.0 mg/kg/d prednisone. The corresponding standard deviations describe variation in cumulative dose among participants within each group.

The endpoint is available for 152 patients. The reasons recorded for the other 12 participants are withdrawal from study (2) and discharge from the Center before day 42 (10). How the missing measurements were handled in the sponsor's analysis is not reported in the registry record.

7. Results Posted in the Registry

• Grade IIa GVHD; 1.0 mg/kg/d Prednisone: mean 27.10 versus 22.20 on control; difference 4.90 (95% CI -0.54 to 10.34), standardized effect d = 0.37; z-test p = 0.077.
• Grade IIb-IV GVHD; 1.0 mg/kg/d Prednisone: mean 38.40 versus 22.20 on control; difference 16.20 (95% CI 9.66 to 22.74), standardized effect d = 1.17; z-test p = <0.001.
• Grade IIb-IV GVHD; 2.0 mg/kg/d Prednisone: mean 41.30 versus 22.20 on control; difference 19.10 (95% CI 13.17 to 25.03), standardized effect d = 1.46; z-test p = <0.001.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Prednisone-associated Toxicity as Assessed by Hyperglycemia

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Group A (Low-dose): mean 140.00 versus 142.00 on control; difference -2.00 (95% CI -31.15 to 27.15), standardized effect d = -0.02; z-test p = 0.893.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Group A (Low-dose) and the control. The interval gives the range of effects that remain compatible with the data.

Prednisone-associated Toxicity as Assessed by Invasive Infections (Bacterial, Fungal and Viral)

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Group A (Low-dose): 51.9% (41/79) versus 53.0% (44/83) on control; risk difference -1.1 percentage points (95% CI -16.5 to 14.3), risk ratio 0.98 (0.73–1.31); unadjusted two-proportion z-test p = 0.887.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Group A (Low-dose) and the control. The interval gives the range of effects that remain compatible with the data.

Non-relapse Mortality

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Group A (Low-dose): 15.2% (12/79) versus 15.7% (13/83) on control; risk difference -0.5 percentage points (95% CI -11.6 to 10.7), risk ratio 0.97 (0.47–2.00); unadjusted two-proportion z-test p = 0.934.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Group A (Low-dose) and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes broaden the assessment beyond cumulative prednisone exposure. The registry lists measures of hyperglycemia, invasive infection, myopathy, hypertension, quality of life, non-relapse mortality, recurrent or progressive malignancy, and progression to Grade III-IV acute GVHD. The posted secondary results cover multiple outcome measures, but the available analysis details are not uniform across outcomes.

For the secondary endpoints with numerical recalculations, the overall pattern is one of small estimated differences relative to broad uncertainty: the intervals include the null. This is not evidence that the groups are identical. Rather, the available estimates do not sharply distinguish between a difference and no difference. These outcomes ask separate questions from the primary endpoint, which measures cumulative prednisone dose through day 42.

Secondary findings generally carry less weight than a prespecified primary comparison when many outcomes or tests are considered. Each additional test creates another opportunity for a chance finding. The registry lists 12 outcome measures and 5 analyses, but does not report a multiplicity-adjustment strategy. The unadjusted secondary comparisons should therefore be treated as exploratory evidence, not as conclusive standalone findings.

The registry record does not report enough information to synthesize every secondary endpoint with equal precision, and some of the listed outcomes do not have detailed results in the available record. The secondary outcomes provide context for safety and clinical course but cannot, on their own, resolve the primary treatment question.

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

• Prednisone-associated Toxicity as Assessed by Hyperglycemia
• Prednisone-associated Toxicity as Assessed by Invasive Infections (Bacterial, Fungal and Viral)
• Prednisone-associated Toxicity as Assessed by Myopathy
• Prednisone-associated Toxicity as Assessed by Hypertension
• Prednisone-associated Toxicity as Assessed by Quality of Life
• Non-relapse Mortality
• Recurrent or Progressive Malignancy
• Progression to Grade III-IV Acute GVHD

9. Statistical Interpretation

10. Statistical Methods Explained

For a continuous outcome such as cumulative dose, a group mean gives the average observed value among participants with outcome data. The standard deviation describes how much individual observations vary around that average. It is not the standard error of the mean: the standard deviation concerns participant-to-participant variation, whereas a standard error concerns uncertainty in an estimated quantity.

The sponsor reports two-sided t-tests for comparisons within the two GVHD grade categories. A t-test evaluates whether the observed difference in means is compatible with a null hypothesis of no mean difference, under the assumptions of the test. A two-sided test considers departures in either direction. The reported p-values are 0.08 for Grade IIa and 0.4 for Grade IIb-IV. A p-value is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of a difference.

The comparison estimates recalculated from the posted group summaries are unadjusted mean differences. They compare the 22.2 mg/kg mean in the Grade IIa 0.5 mg/kg/d group with each other listed group. Those pairings are not the sponsor's within-grade comparisons. The group summaries permit a simple unadjusted comparison, but cannot reproduce an analysis that might use additional covariates, stratification, or other model details not reported in the registry record. In this trial, the distinction is especially important because the two comparisons involving Grade IIb-IV also differ in GVHD grade category.

A confidence interval gives a range of values compatible with the data under the method used to construct it. An interval that includes zero for a mean difference indicates that no difference is among the values compatible with that interval; an interval entirely on one side of zero indicates a directional difference under that calculation. The interval level and full method details are not reported in the registry record. These intervals therefore should be read as descriptive recalculations, not as a replacement for the sponsor's analysis.

Why can the recalculated tests differ from the registry tests?

The available summary statistics support comparisons between the selected groups, but the pairing determines the question being tested. The recalculation uses a single reference group, Grade IIa GVHD at 0.5 mg/kg/d, for all three contrasts. The sponsor's reported tests instead compare the two dose groups within Grade IIa and within Grade IIb-IV. A comparison spanning grade categories can reflect both dose and differences between the grade-defined groups. It is therefore not an equivalent test of the randomized dose contrast within a grade category.

The registry reports a discrepancy between the recalculated and sponsor analyses. When a recalculation and the sponsor's own analysis differ, the sponsor's analysis takes precedence for the trial's reported treatment comparison. The recalculation can still illustrate what follows from the posted summary statistics, provided its distinct pairing and unadjusted nature are made clear.

What does over-dispersion mean for count or percentage outcomes?

For binary outcomes, the observed proportion is a summary of how many participants experienced an event among those at risk. The uncertainty around that proportion depends on the number of participants and the event frequency. General statistical context: over-dispersion occurs when the variability in data exceeds what a simple model assumes. If present, it can make model-based standard errors too small and intervals too narrow unless the analysis accounts for the extra variability. The registry record does not report an over-dispersion assessment for these outcomes.

For the binary secondary outcomes, recalculated risk differences compare proportions between groups. A risk difference is expressed as the difference between the two event proportions; its null value is zero. The posted recalculations also include relative measures for some comparisons, but a relative measure has a different null value and interpretation from a risk difference. These measures are unadjusted, and the registry does not report adjusted estimates for the listed secondary comparisons.

What do the sample-size illustrations represent?

The planning illustrations use the observed effect as a starting point and show participant counts for specified fractions of that effect. They are separate from the trial's original design assumptions, which are not reported in the registry record. They do not reconstruct the trial's design or establish what the original investigators expected to observe.

General statistical context: power is the probability that a specified statistical procedure will reject its null hypothesis under a particular assumed effect and model. A power target is meaningful only in relation to those assumptions. A sample-size calculation based on a point estimate from a completed trial is sensitive to how precisely that estimate was measured. The registry's planning assumptions and the distinction from observed power are described with the sample-size figures; they should not be read as an outcome of the trial.

The registry does not report enough detail to assess every element of the analysis. The original sample-size justification, the exact analysis population, missing-data method, treatment of protocol deviations, covariate adjustment, multiplicity strategy and interim-analysis plan are not reported in the registry record. These absences limit independent reproduction and assessment of the sponsor's analysis.

11. Confidence Intervals and Precision

The primary group means are accompanied by standard deviations, which show substantial spread in cumulative prednisone dose within each group. The recalculated mean-difference intervals offer a sense of precision for their particular pairings, but they do not represent the sponsor's within-grade comparisons.

For Grade IIa GVHD, the recalculated difference comparing the 1.0 mg/kg/d group with the 0.5 mg/kg/d group is 4.9 mg/kg. Its interval extends from below zero to 10.34 mg/kg, so the range includes no difference as well as positive differences. The recalculated Grade IIb-IV comparisons against the Grade IIa 0.5 mg/kg/d group are 16.2 mg/kg, with an interval from 9.664 to 22.74, and 19.1 mg/kg, with an interval from 13.17 to 25.03. These cross-grade estimates are not the sponsor's randomized within-grade dose comparisons.

For the sponsor's analyses, the registry reports p-values but does not provide the estimated mean differences or their intervals. The registry record therefore does not allow the precision of those sponsor comparisons to be assessed from the reported analysis results alone. The group means and standard deviations are available, but they do not by themselves disclose whether the sponsor used any additional modeling or adjustment.

More generally, an interval is informative about both direction and uncertainty: narrow intervals constrain the range of plausible effects more than wide intervals. An interval that includes the null does not demonstrate that the true effect is exactly zero; it indicates that the data and calculation do not rule out the null. The interval confidence level is not reported in the registry record. Precision should also be considered separately from clinical importance, for which the registry does not provide a threshold.

12. Planning a Similar Trial: Sample Size

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

The sample-size figures are planning illustrations based on the observed effect, not the trial's original design calculation. All figures are participants per arm; none is a total-event count. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect equal to the stated fraction of the observed effect. The original trial's own sample-size assumptions are not reported in the registry record.

For each effect assumption, the 80% power column gives the number of participants per arm that the planning calculation estimates would be needed for an 80% chance of detecting that assumed effect under those assumptions. The 90% power column gives the corresponding per-arm count for a 90% chance. These are planning targets, not observed outcomes of the trial. This is not observed power.

The rows consider planning for 100%, 75% and 50% of the observed effect. Assuming a smaller effect generally requires more participants to detect it at the same power target. The sample-size figures can therefore rise sharply when planning for a fraction of the observed effect rather than the full observed estimate. The observed estimate itself is uncertain, particularly when based on finite groups, so these calculations should be read as illustrations rather than as precise forecasts.

The calculations also use equal allocation and a normal approximation. If a future study has different allocation, outcome variability, missingness, or analysis methods, its required sample size could differ. The registry does not report the assumptions used for the original study design, so the illustrations should not be mistaken for the rationale behind enrollment of 164.

13. How the Sample Size Is Derived

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

Formula

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

Worked calculation for this trial

1. Observed difference Δ = 4.900; pooled standard deviation σ = 13.193.
2. n = 2 × (1.96 + 0.842)² × 13.193² / 4.900² = 113.8, rounded up to 114 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 <- 4.9                      # observed difference in means
sigma <- 13.192613697815                     # 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(929695)
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 = 114 (80% power), 153 (90% power)
 75% of observed effect: n per arm = 203 (80% power), 271 (90% power)
 50% of observed effect: n per arm = 456 (80% power), 610 (90% power)
power.t.test: 114.8 
Simulated power at n = 114 per arm: 0.794 (20000 replicates)

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

14. Analyses Reported by the Sponsor

The sponsor's reported tests compare the low-dose and standard-dose groups within each GVHD grade category. The registry gives p = 0.08 for the Grade IIa comparison and p = 0.4 for the Grade IIb-IV comparison, using two-sided t-tests.

The recalculated primary comparisons use a different reference: the Grade IIa 0.5 mg/kg/d group is compared with each of the other three groups. The recalculated differences are 4.9 mg/kg, 16.2 mg/kg and 19.1 mg/kg. The latter two contrasts cross GVHD grade categories, so they do not match the sponsor's within-grade tests. The recalculated analysis is also an unadjusted comparison of posted summaries.

Because the analysis pairings differ, the recalculated estimates and p-values should not be treated as reproductions of the sponsor's treatment comparisons. The registry flags a discrepancy, and the sponsor's own analysis takes precedence. The record does not provide enough detail to determine whether the sponsor used additional information or methods beyond the summary statistics.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists 12 outcome measures and 5 analyses. It does not report whether a formal hierarchy of endpoints was used, whether tests were adjusted for multiplicity, or whether a particular secondary outcome was designated as confirmatory. Without that information, the interpretation of multiple unadjusted tests should be cautious: when many hypotheses are tested, some apparently notable results can arise by chance even when there are no corresponding underlying effects.

For the primary outcome, 152 of 164 enrolled patients had a day-42 cumulative prednisone dose available. The record attributes the 12 missing outcomes to withdrawal from study (2) and discharge from the Center before day 42 (10). It does not report whether the missingness was considered related to outcome values, whether imputation was used, or whether sensitivity analyses were conducted. The analysis-population description ends before completing its explanation, so additional details are not reported in the registry record.

The record also does not state whether the sponsor used an intention-to-treat population, a per-protocol population, or another defined analysis set for the reported tests. It does not report how treatment discontinuation, protocol deviations, or missing data were handled in the primary or secondary analyses. Those details matter because exclusions or assumptions about unavailable outcomes can affect estimates and their uncertainty.

Interim analyses and any alpha-spending or other adjustment for repeated looks at the data are not reported in the registry record. Nor are the original sample-size justification and full statistical analysis plan reported. These omissions do not invalidate the results, but they limit independent assessment of how the analyses were planned and whether the reported tests account for all comparisons made.

16. 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

Serious events with the highest arm-level rate: fever/infection (16.0%), graft versus host disease (9.9%), relapse of hematological disease (4.8%), mental status changes (2.4%), abdominal pain (1.2%).

The registry reports serious adverse event counts for 81 participants in the low-dose group and 83 in the standard-dose group. Serious events affected 38 and 30 participants, respectively. These are counts of participants with serious adverse events, not counts of all event episodes. The record does not give death counts in these safety groups.

The listed serious-event terms include fever/infection (16.0%), graft versus host disease (9.9%), relapse of hematological disease (4.8%), mental status changes (2.4%), abdominal pain (1.2%) and palpitations (1.2%). These terms describe the reported serious events; the registry information does not establish that they were caused by study treatment. The record does not provide further details about event severity, timing, attribution, or overlap among participants.

Serious adverse event counts are not interchangeable with the specific secondary outcomes, such as invasive infections or non-relapse mortality. They also do not by themselves provide an adjusted comparison of safety between groups. The figures should be read as descriptive registry counts rather than as proof of a treatment-related safety difference.

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

The primary results describe cumulative prednisone exposure through day 42 in dose groups defined by GVHD grade. The observed mean is higher in each listed standard-dose group than in its corresponding low-dose group. The sponsor's reported tests within grade categories have p-values of 0.08 and 0.4. Those results do not establish a statistically clear difference under the reported tests, but they do not demonstrate that the strategies are equivalent.

The recalculated contrasts answer a different question because they compare the Grade IIa 0.5 mg/kg/d group with all other groups, including groups with different GVHD grades. They are useful only when interpreted as unadjusted comparisons of those particular group summaries. The sponsor's analyses are the relevant reported tests for the within-grade dose comparisons, and the sponsor's own analysis takes precedence when it differs from the recalculation.

The endpoint measures drug exposure, not the full balance of clinical outcomes. The secondary results address some aspects of toxicity and clinical course, but their estimates are imprecise and the registry does not report a multiplicity strategy. No single endpoint should be used to infer a broad overall clinical effect from this record alone.

The registry also leaves important analysis details unreported, including the original sample-size rationale, missing-data methods, analysis populations and any interim-analysis plan. The number of participants with primary outcome data is lower than enrollment, and the effects of missing outcomes cannot be determined from the record. These limitations leave uncertainty about how robustly the analyses address all relevant sources of uncertainty.

Accordingly, these results support a description of the observed doses and the sponsor's reported tests, not a claim that one treatment strategy works or does not work. They do not establish equivalence, non-inferiority, or a clinically important difference. Such conclusions would require defined criteria and supporting analyses that are not reported in the registry record.

18. Why This Trial Matters Statistically

This trial illustrates why a treatment comparison must be read in the context of the groups actually compared. The sponsor's tests compare dose strategies within GVHD grade categories, whereas the recalculated contrasts use one Grade IIa group as a common reference for all other groups. Those are different statistical questions. Comparing across grade categories can confound interpretation of the dose contrast because the groups differ in the grade category as well as the listed dose.

It also shows why a p-value alone is not a complete account of a result. The sponsor reports p-values, but not corresponding treatment-effect estimates or intervals for its within-grade comparisons. The group means and standard deviations provide descriptive context, while the recalculated intervals express uncertainty for different pairings. Keeping those estimands separate prevents the more dramatic cross-grade contrasts from being mistaken for the sponsor's randomized comparisons.

The primary outcome is a continuous measure of cumulative exposure. Its interpretation differs from binary safety outcomes and from outcomes such as mortality or disease progression. A clear statistical description needs to distinguish outcome type, follow-up period, analysis population and comparison group. Here, the primary outcome is available for 152 participants, and the registry gives reasons for the 12 unavailable measurements but does not explain how they were addressed analytically.

Finally, multiple secondary outcomes can broaden understanding while also increasing the risk of chance findings. Because multiplicity procedures and several design details are not reported, the secondary results deserve cautious interpretation. The study is a useful example of how registry results can inform statistical reading while still leaving key questions about precision, missing data and analysis planning unresolved.

19. 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 original sample-size justification or the full statistical analysis plan. The planning illustrations based on observed effects are not a substitute for those design details. The record also does not state the analysis population used for the sponsor's tests or how missing primary outcomes were handled.

The primary outcome is unavailable for 12 participants. The reasons are recorded as withdrawal from study (2) and discharge from the Center before day 42 (10), but the record does not describe whether missingness could be related to treatment or outcome. No imputation method or sensitivity analysis is reported.

The sponsor's reported p-values are not accompanied by treatment-effect estimates or intervals in the registry record. The recalculated estimates use different comparisons, including contrasts between GVHD grade categories, and cannot fill that reporting gap for the sponsor's within-grade tests. Details of any covariate adjustment or stratification beyond the reported t-tests are not provided.

Multiplicity adjustment, interim analyses, randomization procedures, allocation concealment, adherence, and the handling of treatment changes are not reported in the registry record. The record also provides limited detail for some secondary outcomes and does not provide death counts in the serious adverse event summary. These limitations restrict independent assessment of the full statistical and safety evidence.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Low-Dose Prednisone or Methylprednisolone in Treating Patients With Newly Diagnosed Acute Graft-versus-Host Disease — NCT00929695.
• Registry record last updated 2017-08-21; status last verified 2017-07.
• Results first posted 2014-10-09.
• 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.

23. Record Summary

This completed randomized, parallel-group, open-label phase III trial enrolled 164 patients with newly diagnosed acute GVHD. It compared low-dose with standard-dose prednisone or methylprednisolone. The primary endpoint was mean cumulative prednisone dose through day 42; results were available for 152 patients.

The registry reports dose means and standard deviations for four grade-and-dose groups and sponsor-reported two-sided t-tests within Grade IIa and Grade IIb-IV. Recalculated contrasts use different group pairings, so the sponsor's own analysis takes precedence. Secondary outcomes address toxicity and clinical course, while the registry record leaves key details about multiplicity, missing data, sample-size justification and analysis populations unreported.