Combination Chemotherapy With or Without Rituximab in Treating Participants With Stage III-IV Classic Hodgkin Lymphoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Combination Chemotherapy With or Without Rituximab in Treating Participants With Stage III-IV Classic Hodgkin Lymphoma (NCT00654732), a randomized phase 2 study. The primary outcome is Event-free Survival (EFS) Rate.

2. Clinical Question

3. Trial Design

This was a randomized phase II, parallel-group trial with no masking. It compared ABVD plus rituximab (RABVD) with ABVD alone in participants with advanced-stage classical Hodgkin lymphoma and poor-risk features. The registry lists a total enrollment of 58 and identifies M.D. Anderson Cancer Center as the sponsor. The record is marked completed. The primary completion date is 2018-09-05, and the start date is 2008-03-19.

In the RABVD arm, rituximab was administered intravenously over 7 hours on days 1, 8, 15, and 22 of course 1 and on days 1 and 8 of course 2. Both arms received doxorubicin hydrochloride, bleomycin, vinblastine, and dacarbazine intravenously over 1 hour on days 1 and 15. The arm description says ABVD repeats every 4 weeks, but the available description ends before the full course schedule is stated. Further treatment-duration and treatment-modification details are not reported in the registry record.

The primary endpoint was labeled Event-free Survival (EFS) Rate, with a stated timeframe from the start of study treatment up to 3 years. The record describes a Kaplan–Meier estimate, a log-rank comparison, and use of a Cox proportional hazards model to include multiple covariates. It also mentions logistic regression for assessing prognostic factors and response rate. Those descriptions identify intended analytic approaches, but the record does not post results from those analyses. No secondary outcomes are posted in the registry record.

The primary results list 22 participants for RABVD and 26 for ABVD. The adverse-event summary uses at-risk denominators of 26 and 32, respectively. The registry record does not explain the differences between these denominators and the overall enrollment or the endpoint denominators. It also does not report participant flow, analysis populations, or reasons for any 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.

Overall Study

Reasons for not completing this period:

5. Baseline Characteristics

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

6. Primary Endpoint

The registry lists the primary outcome as Event-free Survival (EFS) Rate, measured over From the start of study treatment up to 3 years.

Event-free survival is a time-to-event concept. In general, it concerns the time until a prespecified event occurs, with participants who have not had an event by the end of follow-up handled as censored observations. The specific events included in this trial’s EFS definition are not reported in the registry record. The stated timeframe runs from the start of study treatment up to 3 years.

The registry describes estimating EFS with the Kaplan–Meier method. In general, Kaplan–Meier estimation uses the observed event times and censoring information to estimate the proportion remaining event-free over time. A log-rank test can compare the event-time distributions between groups, while a Cox proportional hazards model can examine group differences while incorporating covariates. These methods are useful for time-to-event outcomes because they can use information from participants with different follow-up durations.

The posted primary outcome is labeled “Event-free Survival (EFS) Rate,” but the result type is a count of participants with an event. It reports 17 of 22 in the RABVD arm and 20 of 26 in the ABVD arm. These counts are not Kaplan–Meier estimates at the stated 3-year timeframe, and they do not convey when events occurred or how follow-up and censoring were handled. The registry record does not post a 3-year EFS estimate, event-time curves, or results of the stated log-rank or Cox analyses. Consequently, the count-based comparison and the intended time-to-event endpoint should not be treated as interchangeable summaries.

7. Results Posted in the Registry

• ABVD: 76.9% (20/26) versus 77.3% (17/22) on control; risk difference -0.3 percentage points (95% CI -24.2 to 23.5), risk ratio 1.00 (0.73–1.36); unadjusted two-proportion z-test p = 0.977.

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes a randomized, parallel-group phase II trial without masking. Randomization is intended, in general, to make treatment groups comparable at assignment with respect to measured and unmeasured prognostic characteristics. It does not guarantee balance in a particular small trial. Open-label means that masking was not used; the registry does not describe whether any outcome assessment was separately masked.

What does the endpoint description imply?

The primary endpoint is labeled Event-free Survival (EFS) Rate and has a timeframe from the start of study treatment up to 3 years. The registry states that EFS would be estimated using Kaplan–Meier methods and that a log-rank test would compare the groups’ time-to-event distributions. In general, a time-to-event analysis uses both whether an event occurs and when it occurs. Participants who have not experienced the event by their last observation can contribute follow-up information through that time, even though their event time is not observed.

The Cox proportional hazards model is described as a way to include multiple covariates in the time-to-event analysis. In general, this model estimates associations between covariates and the event rate over follow-up, under a proportional-hazards assumption. The registry record does not report which covariates were included, whether that assumption was assessed, or any Cox model results. Logistic regression is also mentioned for assessing effects of prognostic factors on response rate. The response outcome, covariates, and logistic-regression results are not reported in the registry record.

What do the posted numbers represent?

Despite the time-to-event endpoint description, the posted primary outcome is a count of participants with an event. It gives counts and denominators for the two arms, with no posted event times, censoring details, or Kaplan–Meier estimates. The simple risk difference, risk ratio, odds ratio, and p-value are unadjusted summaries of the reported counts. They do not adjust for prognostic factors or follow-up duration. The registry record does not post a separate analysis result that would permit a direct review of the stated log-rank or Cox analysis.

The risk difference is expressed as the proportion in the comparison group minus the proportion in the reference group. A value near zero describes little difference in the observed proportions, while the confidence interval expresses uncertainty around that estimate. The risk ratio compares the two proportions as a ratio, and the odds ratio compares the odds of the event. These measures answer related but different questions; neither replaces the time-to-event estimand described for EFS.

The reported p-value is unadjusted. In general, a p-value measures how unusual data at least as incompatible with a specified null model would be if that model and the test assumptions were true. It does not measure effect size or precision, and it is not the probability that the null hypothesis is correct. The registry record does not report a multiplicity adjustment or a prespecified hierarchy of tests.

Why can time-to-event results differ from counts?

General explanation: a count comparison treats participants as having or not having an event within the data summarized, while a survival analysis uses the timing of events and accounts for varying observation periods through censoring methods. A participant followed for a shorter time may not have the same opportunity for an event to be observed as someone followed longer. The registry’s stated endpoint timeframe is up to 3 years, but the posted count results do not clarify whether each participant had complete follow-up to that horizon or how censoring was handled. The time-to-event estimate cannot be reconstructed from the counts alone.

Over-dispersion is mainly a concern for count models when the variability in counts exceeds what the assumed model allows. It is not a central adjustment to the posted two-group binary risk comparison. General explanation: when over-dispersion is present in a count model, standard errors from a model that assumes less variability can be too small, making uncertainty appear narrower than it is. The registry does not report a count-regression analysis or an over-dispersion assessment.

What can be inferred from the comparisons?

The unadjusted risk difference is -0.004, with a 95% confidence interval from -0.242 to 0.235. The interval covers zero and allows effects in both directions. Its breadth reflects limited precision in these posted counts. The risk ratio and odds ratio intervals are also broad, and both include their respective no-association values. The p-value is 0.977. Taken together, these summaries do not identify a precise treatment-group difference. They also do not demonstrate equivalence or rule out meaningful differences.

The original sample-size justification, randomization details, covariate plan, interim-analysis plan, missing-data handling, analysis-population definitions, and protocol-specific multiplicity strategy are not reported in the registry record. The planning illustration accompanying the observed effect uses separate stated assumptions; it should not be mistaken for the trial’s original design calculations.

10. Confidence Intervals and Precision

The unadjusted risk difference for ABVD compared with RABVD is -0.004, with a 95% confidence interval from -0.242 to 0.235. The point estimate is close to zero, but the interval is broad and includes zero. It is compatible with either a lower or a higher event proportion for ABVD under the comparison as defined. The interval describes statistical uncertainty around the estimate; it does not assign a probability to a specific true value.

The risk ratio is 0.996, with a 95% confidence interval from 0.731 to 1.356. The odds ratio is 0.980, with a 95% confidence interval from 0.254 to 3.787. Both intervals include the no-difference value for their respective scales and span a broad range. The point estimates are close to one, but that proximity should not be read as proof of equivalence when the intervals remain wide.

Precision is constrained by the endpoint denominators of 22 and 26 and by the fact that the posted primary results are counts rather than time-to-event estimates. The confidence intervals do not incorporate the missing event timing or clarify whether follow-up was complete through the stated timeframe. No Kaplan–Meier confidence interval or model-based uncertainty estimate is posted in the registry record.

General explanation: a confidence interval describes a range of values reasonably compatible with the data under the statistical method used. A wider interval indicates less precision; a narrower interval indicates more precision. Interpretation also depends on whether the method matches the outcome and analysis plan. Here, the count-based intervals should be understood as unadjusted summaries of the posted counts, not as intervals around a 3-year Kaplan–Meier EFS estimate.

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 planning figures are participants per arm for 80% power and participants per arm for 90% power. They are participant counts per arm, not total event counts. The listed planning scenarios use 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. For each scenario, the figures report the number of participants per arm associated with the stated power level.

The assumptions for this illustration are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is the stated fraction of the observed effect. These figures are planning illustrations computed from the observed effect, not the trial’s own original design assumptions; those original assumptions are not reported in the registry record.

In plain language, 80% power means that, if the assumed effect and planning assumptions were correct, a study of the stated size would have an 80% chance of detecting that effect at the specified significance level. The 90% power column represents a higher target chance under the same assumptions. The figures are not a retrospective measure of the chance that the completed trial detected an effect. This is not observed power.

The number of participants required rises as the effect being targeted becomes smaller, because smaller differences are harder to distinguish from random variation. The large planning figures reflect the very small observed difference used as the starting point for these calculations. They should not be interpreted as a recommended trial size without considering the intended clinical question, outcome definition, analysis method, and assumptions. The registry record does not report the original sample-size justification or protocol-specific target effect.

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 = 0.773; treatment proportion p₁ = 0.769; average p̄ = 0.771.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.5943 = 1.1647.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5943 = 0.5001.
4. n = (1.1647 + 0.5001)² / (-0.0035)² = 226,717.6, rounded up to 226,718 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 <- 0.772727272727273                          # control proportion
p1 <- 0.769230769230769                          # 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(654732)
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 = 226,718 (80% power), 303,511 (90% power)
 75% of observed effect: n per arm = 402,514 (80% power), 538,851 (90% power)
 50% of observed effect: n per arm = 904,438 (80% power), 1,210,785 (90% power)
power.prop.test: 226717.6 
Simulated power at n = 226,718 per arm: 0.797 (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. Multiplicity, Missing Data and Analysis Populations

The registry record lists one primary outcome and no posted secondary outcomes. It describes more than one analytic approach: Kaplan–Meier estimation and a log-rank test for the time-to-event endpoint, a Cox model with multiple covariates, and logistic regression for prognostic factors and response rate. The record does not report which analyses were ultimately conducted or their results. It also does not report a multiplicity strategy, such as a hierarchy of hypotheses or adjustment for multiple testing. Without that information, it is not possible to determine how the analyses were intended to control the chance of false-positive findings across multiple comparisons.

The registry record does not define analysis populations, such as whether the primary analysis included all randomized participants or only participants with outcome data. The total enrollment is 58, while the primary outcome denominators are 22 and 26. Safety at-risk denominators are 26 and 32. The record does not explain these differences, report participant flow, or state whether any participants were excluded from particular analyses.

Missing-data methods are not reported. In general, missing outcome information can affect estimates if the probability of missingness is related to treatment, prognosis, or the outcome itself. For a time-to-event endpoint, censoring rules and follow-up completeness are especially relevant. The posted counts do not provide information about censoring, loss to follow-up, or handling of participants whose event status was unknown.

The original protocol’s statistical analysis plan, interim analyses, allocation procedures, and sample-size justification are not reported in the registry record. The stated methodological descriptions are not enough to establish the exact analysis population, covariate selection, or missing-data assumptions used in practice. These omissions limit assessment of whether the posted counts represent the intended primary analysis and whether the stated time-to-event methods were carried out as planned.

14. 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

Serious events with the highest arm-level rate: neutropennia (50.0%), infection (15.4%), neuropathy (9.4%), anemia (6.2%), thrombocytopenia (6.2%).

The serious adverse-event summary reports 13 affected participants in each arm. The at-risk denominators are 26 for RABVD and 32 for ABVD. It reports 5 deaths for RABVD and 6 for ABVD. The summary also records zero affected participants for the listed “other” adverse events in both groups. These are descriptive counts; the registry record does not clarify the relationship between the safety denominators and the endpoint denominators or provide further detail on timing, duration, severity grading, or attribution.

The listed serious-event terms and percentages are Neutropennia, 50.0%; Infection, 15.4%; Neuropathy, 9.4%; Anemia, 6.2%; Thrombocytopenia, 6.2%; and Respiratory Complicaton, 6.2%. The term spellings are reproduced as recorded. The listing does not identify an arm for each term, and the record does not describe how these percentages were calculated. They should not be interpreted as arm-specific rates.

These figures do not establish that an event was caused by a particular treatment. The registry record does not provide event-level narratives or a comparative statistical analysis of serious adverse events. They are a limited descriptive summary rather than a complete account of safety outcomes.

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

The posted count comparison describes similar observed proportions in the two arms: 17 of 22 participants in the RABVD arm and 20 of 26 in the ABVD arm are recorded as participants with an event. The unadjusted risk difference is -0.004, with a 95% confidence interval from -0.242 to 0.235, and the p-value is 0.977. The point estimate is close to no difference, while the interval is broad enough to include effects in either direction. These results do not establish that the groups are equivalent or that a clinically important difference has been ruled out.

The registry describes a time-to-event primary endpoint through 3 years and names Kaplan–Meier estimation, a log-rank test, and Cox regression. However, the posted primary results are counts. They do not provide event timing, a Kaplan–Meier estimate, a log-rank result, or a Cox model result. The count-based figures therefore cannot answer the full time-to-event question as described in the endpoint record.

The record also leaves uncertainty about which participants contributed to each result. Overall enrollment is 58, the primary endpoint denominators are 22 and 26, and the safety denominators are 26 and 32. Participant flow and analysis-population definitions are not reported. In addition, the event definition, censoring details, missing-data approach, and original analysis plan are not reported in the registry record.

These statistics describe what the posted registry results contain and how precisely they estimate a count-based difference. They do not determine whether either treatment is appropriate for an individual, and they do not establish that either regimen works or does not work. A clinical interpretation would require the full outcome definitions, time-to-event results, follow-up information, and broader evidence beyond the posted counts.

16. Why This Trial Matters Statistically

This trial illustrates a common interpretive challenge: the primary endpoint is described as a time-to-event outcome, but the posted result is a count. Time-to-event methods use event timing and censoring, while a count-based comparison summarizes event status without showing when events occurred. The distinction matters because a count cannot reproduce a Kaplan–Meier estimate or a log-rank comparison. Here, the registry describes those methods but does not post their results.

The record also demonstrates why an effect estimate should be read together with its confidence interval. The risk difference is close to zero, but its 95% confidence interval extends from -0.242 to 0.235. The relative-measure intervals are also broad. A near-zero point estimate by itself can suggest more certainty than the data support; the intervals show that the posted comparison remains imprecise.

Another statistical issue is denominator transparency. Total enrollment is 58, while the endpoint and safety summaries use different denominators. Without participant flow and analysis-population definitions, readers cannot establish why the numbers differ or how missing outcome information was handled. Clear reporting of these details is important for understanding which participants the estimates describe.

Finally, the registry names several intended analytic methods but does not post their results or the original analysis plan. This distinction between an analysis description and an analysis result is essential when evaluating trial evidence. The trial is statistically informative as an example of why outcome definitions, denominators, precision, and complete reporting all matter alongside a p-value.

17. 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 principal limitation of the posted primary result is that it is a count, although the endpoint is described as EFS estimated over time. No 3-year Kaplan–Meier estimate, log-rank result, or Cox model result is posted. The event definition, event dates, censoring rules, and completeness of follow-up are not reported in the registry record. As a result, the posted counts cannot fully address the stated time-to-event question.

The outcome denominators are 22 for RABVD and 26 for ABVD, while overall enrollment is 58. Safety denominators are 26 and 32. Participant flow and reasons for these differences are not reported. The registry also does not state the analysis population or how missing outcome data were handled. These omissions limit assessment of selection and missing-data effects.

The unadjusted confidence intervals are broad, and no adjusted comparison is posted. Although the endpoint description mentions Cox regression with multiple covariates, the selected covariates, model results, and model assumptions are not reported. The registry record also does not state whether the statistical plan included interim analyses, how multiplicity was handled, or how the original sample size was justified.

The adverse-event summary is limited: the listed serious-event terms are not assigned to a specific arm, and the calculation basis for the percentages is not reported. The record does not provide further event-level detail. The arm treatment description is also truncated, leaving parts of the regimen schedule unreported. These limitations concern the detail available in the registry record and do not by themselves indicate that analyses or information were absent from other study documents.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Combination Chemotherapy With or Without Rituximab in Treating Participants With Stage III-IV Classic Hodgkin Lymphoma — NCT00654732.
• Registry record last updated 2020-02-28; status last verified 2020-02.
• Results first posted 2020-02-28.
• 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.

21. Record Summary

NCT00654732 is a completed randomized phase II parallel-group trial sponsored by M.D. Anderson Cancer Center. It enrolled 58 participants with advanced-stage classical Hodgkin lymphoma and poor-risk features, comparing rituximab plus ABVD with ABVD alone. The trial was open-label, with no masking.

The primary outcome is labeled Event-free Survival (EFS) Rate, with a timeframe from the start of study treatment up to 3 years. The registry describes Kaplan–Meier estimation, a log-rank test, and Cox regression, but posts counts of participants with an event: 17 of 22 in the RABVD arm and 20 of 26 in the ABVD arm. The unadjusted risk difference is -0.004, with a 95% confidence interval from -0.242 to 0.235 and p = 0.977.

The posted count-based comparison is imprecise and is not a reported 3-year time-to-event estimate. Participant flow, analysis populations, missing-data methods, and results of the stated time-to-event analyses are not reported in the registry record.