Study to Compare Oxymorphone Extended-Release (Opana ER) Versus Oxycodone Controlled-Release (Oxycontin): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study to Compare Oxymorphone Extended-Release (Opana ER) Versus Oxycodone Controlled-Release (Oxycontin) (NCT00955110), a randomized phase 4 study. The primary outcome is High VAS - Emax (mm).

2. Clinical Question

3. Trial Design

This was a Phase 4 randomized crossover study, with quadruple masking listed in the registry record. Its official title describes an exploratory, single-dose, double-blind, randomized, placebo-controlled crossover study in healthy non-dependent recreational opioid users. The study status is completed. The record lists start and primary-completion dates as 2009-06 and 2009-09.

The treatment conditions with posted primary outcome summaries are placebo, oxymorphone ER 15 mg, oxymorphone ER 30 mg, oxycodone CR 30 mg, and oxycodone CR 60 mg. The primary outcome population is defined as participants who received all 5 treatments and had no major protocol deviations or other circumstances that would exclude them from analysis. The pharmacokinetic and pharmacodynamic analyses used that Per Protocol population, and the record says missing values were not imputed.

The registry reports enrollment of 78. Separately, the primary outcome summaries list 35 for each treatment condition. Because the design is crossover, the treatment-condition summaries should not be read as counts of separate parallel groups or added to infer a number of unique participants. The registry record does not report the allocation sequence, treatment order, washout interval, or details about how masking was maintained. Hydromorphone is named in the intervention list, but its place in the primary outcome comparisons is not reported in the registry record.

The record reports one primary outcome measure and six sponsor analyses. It does not report the original sample-size justification, a detailed prespecified statistical analysis plan, or whether interim analyses were planned or conducted. The study is described as exploratory, but that label does not supply those missing design details.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: A double-blind crossover qualification phase (hydromorphone 8 mg vs. placebo) was followed by a washout period and then randomization to a 5-period, 10-sequence crossover treatment phase. Approximately 40 qualified subjects were to be enrolled in the treatment phase in order to ensure that at least 30 subjects completed all 5 periods of the study.

Qualification Phase

Reasons for not completing this period:

Treatment Phase

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Per Protocol population: Subjects who received all 5 treatments and who had no major protocol deviations or other circumstances that would exclude them from the analysis. The pharmacokinetic and pharmacodynamic analyses were performed using the Per Protocol population. No imputation of missing values was performed.

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 High VAS - Emax (mm), measured over High VAS was administered at pre-dose, 0.5, 1, 1.5, 2, 3, 4, 5, 6, 8, 12, and 24 hours post-dose..

The primary endpoint, High VAS - Emax (mm), measures a participant’s response to the prompt I am feeling high. Participants used a laptop to mark a position on a horizontal line anchored by Definitely not and Definitely so. The scale ran from 0 to 100, with higher positions representing a stronger endorsement of the statement. The registry record describes the outcome as a mean of a continuous measure and reports a standard deviation for each treatment condition.

High VAS was administered before dosing and at 0.5, 1, 1.5, 2, 3, 4, 5, 6, 8, 12, and 24 hours after dosing. Emax is part of the endpoint name, but the registry record does not report the exact rule used to derive the Emax value from those repeated assessments. The results therefore describe the posted Emax summaries without assuming a more specific derivation.

A self-rated scale is used to quantify a subjective experience in a consistent numerical form, allowing the study to compare average responses across treatment conditions. It is not an objective measure of every effect of a treatment, and a score on this scale alone does not establish broader clinical consequences. The outcome is also specific to the stated population, single-dose crossover setting, and subjective question.

7. Results Posted in the Registry

• Oxymorphone ER 15 mg: mean 30.60 versus 23.10 on control; difference 7.50 (95% CI -8.79 to 23.79), standardized effect d = 0.22; z-test p = 0.367.
• Oxymorphone ER 30 mg: mean 62.50 versus 23.10 on control; difference 39.40 (95% CI 23.82 to 54.98), standardized effect d = 1.18; z-test p = <0.001.
• Oxycodone CR 30 mg: mean 81.40 versus 23.10 on control; difference 58.30 (95% CI 44.86 to 71.74), standardized effect d = 2.03; z-test p = <0.001.
• Oxycodone CR 60 mg: mean 93.00 versus 23.10 on control; difference 69.90 (95% CI 58.07 to 81.73), standardized effect d = 2.77; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

General: A continuous outcome is measured along a numerical scale, so summaries can describe its average and variability. The registry identifies the endpoint as a mean and gives a standard deviation for each treatment condition. The mean locates the center of the posted scores; the standard deviation describes how spread out those scores were around the mean. A mean does not show the full distribution, and a standard deviation is not a confidence interval for the mean.

What does a mean difference represent?

General: A mean difference expresses the distance between two group or treatment-condition means on the outcome’s original scale. In the simple recalculations, the reference is placebo and the estimate is the active-condition mean minus the placebo mean. Thus a positive estimate indicates a higher posted mean for the active condition in that comparison. The recalculations are unadjusted and use the posted summary statistics; they are not the sponsor’s ANCOVA results.

For example, the unadjusted estimated difference for oxymorphone ER 15 mg versus placebo is 7.5. Its interval extends from -8.786 to 23.79, and the unadjusted p-value is 0.367. The estimates for the other active conditions are 39.4, 58.3, and 69.9, each with p < 0.001. These figures describe the recalculated comparisons, not a substitute for the trial’s reported analysis.

Why does the crossover design matter?

General: In a crossover design, the same participant can contribute observations under more than one treatment condition. Statistical comparisons may therefore need to account for the within-participant relationship between observations. That relationship can affect the estimated uncertainty. The simple recalculations use posted summary statistics and do not use a reported within-participant covariance. The registry record does not give the details needed to reconstruct a paired analysis from those summaries. This is an important reason not to treat the recalculations as equivalent to the sponsor’s model-based results.

The sponsor reports ANCOVA and least-square-mean differences. General: ANCOVA is a modeling approach that estimates group or treatment means while accounting for variables included in the model. A least-square mean is a model-based estimate rather than simply the raw arithmetic mean. The registry record does not report the ANCOVA covariates, model specification, or other implementation details, so those features should not be inferred. The sponsor’s reported values are the trial’s own analyses and take precedence over the recalculations.

How should p-values be read?

General: A p-value describes how compatible the observed result is with a specified null model, under the assumptions of the analysis. It is not the probability that the null hypothesis is true, nor does it measure the size or importance of an effect. The recalculated p-value for oxymorphone ER 15 mg versus placebo is 0.367; the other recalculated comparisons have p < 0.001. The sponsor’s adjusted comparisons report p = 0.395 for placebo versus oxymorphone ER 15 mg and p < 0.001 for the other listed comparisons. No multiplicity adjustment is described in the registry record.

The method for the sponsor analyses is identified as ANCOVA, but further details are not reported in the registry record. The record does not report whether the tests were one-sided or two-sided, how repeated measurements were handled in the model, or whether the analysis accounted for treatment sequence or period. The recalculation assumptions used for planning illustrations are stated separately and should not be mistaken for the sponsor’s analysis assumptions.

What does the endpoint’s scale allow?

General: A numerical scale can make subjective responses comparable in a structured way, but it does not make the underlying experience objective. The 0-to-100 anchors define the response range. The posted means and differences remain measures of participants’ reported feeling high; they are not direct measures of other outcomes. The registry record does not report a threshold for interpreting a particular score change as important.

Results also depend on the analysis population. The record defines the Per Protocol population by completion of all 5 treatments and the absence of major protocol deviations or other exclusionary circumstances. No imputation of missing values was used. The record does not report how many otherwise eligible observations were missing at each assessment or how missingness might affect the result.

10. Confidence Intervals and Precision

A confidence interval gives a range of values compatible with the estimate under the model and its assumptions. General: a narrower interval indicates greater statistical precision, while a wider interval indicates more uncertainty about the estimate. An interval should be read with its estimate and analysis method; it does not guarantee that the true value lies inside it.

In the sponsor’s ANCOVA results, placebo versus oxymorphone ER 15 mg had an estimated least-square-mean difference of 5.5, with a 95% confidence interval from -7.3 to 18.3. This interval includes zero, and the reported p-value is 0.395. Placebo comparisons for oxymorphone ER 30 mg, oxycodone CR 30 mg, and oxycodone CR 60 mg had estimates of 38.0, 56.0, and 66.8, respectively; their 95% confidence intervals were 25.2 to 50.8, 43.1 to 68.9, and 53.9 to 79.7. Each had p < 0.001.

The sponsor’s active-treatment comparisons were -28.8 for oxymorphone ER 30 mg versus oxycodone CR 60 mg, with a 95% confidence interval from -41.6 to -16.0, and -50.5 for oxymorphone ER 15 mg versus oxycodone CR 30 mg, with a 95% confidence interval from -63.4 to -37.5. Both had p < 0.001. The sign of each estimate follows the order of groups in the reported comparison.

The unadjusted recalculations also include intervals, but they are not the sponsor’s model-based intervals. They are based on posted summary statistics and do not use a reported within-participant covariance for the crossover design. Differences between the recalculated and sponsor estimates or intervals are therefore not evidence that one can be substituted for the other. The registry record does not report a clinically meaningful threshold for the High VAS scale, so precision and statistical evidence do not alone determine practical importance.

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 illustrations express required participants per arm for 80% power and participants per arm for 90% power. They are planning illustrations based on the observed effect, not the original study’s design calculation. The 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. These illustrations are not observed power.

For each comparison, the rows represent planning for 100%, 75%, and 50% of the observed effect. The columns indicate the planning target: 80% power or 90% power. In plain terms, power is the chance that a planned study would detect an effect of the specified size under the assumptions if that effect were present. A higher target power generally requires more participants for the same assumed effect. The registry record does not report the original sample-size justification or the assumptions used when designing the trial.

The figures are per arm, not a total across both arms and not total events. Because this study used a crossover design, the simplified equal-allocation illustration should not be read as a reconstruction of the actual design or analysis. Smaller effects require more participants than larger effects under the stated planning framework. The values are derived from the observed comparisons and are consequently sensitive to how large those observed effects were. They should be interpreted as hypothetical planning references, rather than as evidence about the study’s achieved power or as a recommendation for a future study.

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, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 7.500; pooled standard deviation σ = 34.760.
2. n = 2 × (1.96 + 0.842)² × 34.760² / 7.500² = 337.2, rounded up to 338 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 <- 7.5                      # observed difference in means
sigma <- 34.7598403333502                     # 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(955110)
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 = 338 (80% power), 452 (90% power)
 75% of observed effect: n per arm = 600 (80% power), 803 (90% power)
 50% of observed effect: n per arm = 1,349 (80% power), 1,806 (90% power)
power.t.test: 338.1 
Simulated power at n = 338 per arm: 0.795 (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.

13. Analyses Reported by the Sponsor

The registry reports six sponsor analyses using ANCOVA and least-square-mean differences. These are the trial’s own adjusted analyses. The simple recalculations are unadjusted comparisons reconstructed from the posted means and standard deviations. Because the study is a crossover trial and the recalculations do not use a reported within-participant covariance, the two approaches need not yield the same estimates or uncertainty.

The discrepancy indicator is true. Accordingly, the sponsor’s own analysis takes precedence over the recalculated comparisons when describing the trial’s reported findings. For example, the recalculated placebo comparison for oxymorphone ER 15 mg is 7.5 with p = 0.367, whereas the sponsor reports an ANCOVA least-square-mean difference of 5.5 with p = 0.395. For placebo versus oxymorphone ER 30 mg, the recalculated estimate is 39.4 and the sponsor estimate is 38.0. The recalculated estimates for placebo versus oxycodone CR 30 mg and 60 mg are 58.3 and 69.9; the sponsor estimates are 56.0 and 66.8.

These differences reflect distinct calculation approaches and should not be presented as interchangeable versions of one result. The registry record does not provide enough model detail to independently reproduce the sponsor’s ANCOVA. The sponsor estimates, confidence intervals, and p-values should therefore be treated as the reported trial analyses, while the recalculations serve only as unadjusted checks based on posted summaries.

14. Multiplicity, Missing Data and Analysis Populations

The registry reports six sponsor analyses for the primary outcome, including placebo comparisons and comparisons between active-treatment conditions. It does not report a multiplicity adjustment or a prespecified hierarchy for interpreting these tests. When several comparisons are examined, the chance of at least one apparently unusual result can increase if each test is interpreted without accounting for the full set. General: multiplicity procedures can control or describe that issue, but the appropriate method depends on the trial’s design and analysis plan. No such method is reported in the registry record, so the listed p-values should be read as reported rather than assumed to be adjusted.

The record identifies the primary endpoint and says that the pharmacokinetic and pharmacodynamic analyses used the Per Protocol population. That population included participants who received all 5 treatments and had no major protocol deviations or other circumstances that would exclude them from analysis. The record also states that missing values were not imputed. It does not report the number or pattern of missing endpoint observations, the handling of missing measurements in each sponsor model, or whether another analysis population was used for sensitivity checks.

The registry record does not report a detailed statistical analysis plan, the ANCOVA covariates, or the handling of treatment period, sequence, or repeated assessments. It also does not report whether interim analyses were planned or conducted. Those omissions limit independent assessment of how the models were specified and how the set of comparisons was chosen. They do not change the fact that the record identifies the reported sponsor analyses, but they constrain interpretation of the p-values and the ability to reproduce the results.

15. 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 5% reporting threshold.

Most frequent non-serious adverse event terms

For each listed treatment-phase group, the registry record reports zero participants affected by serious adverse events: placebo, 0 of 38; oxymorphone ER 15 mg, 0 of 37; oxymorphone ER 30 mg, 0 of 38; oxycodone CR 30 mg, 0 of 40; and oxycodone CR 60 mg, 0 of 40. These are the group-specific at-risk and affected counts in the record. The serious adverse event terms list is empty.

Zero affected in the posted counts means the record lists no serious adverse events for those treatment-phase groups. It does not establish that a treatment has no risk, and it should not be generalized beyond the study. The groups have different listed at-risk counts, and the study’s crossover design means treatment-phase group counts should not be treated as separate parallel-group samples or summed to infer unique participants. The registry record does not report death counts; those values are not reported in the registry record.

The record separately lists other adverse events, but serious adverse events are distinct from that broader category. A serious-event count of zero does not mean there were no other adverse events. These safety counts are descriptive and do not, on their own, establish comparative safety.

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

The results describe how participants in this study rated feeling high after the listed single-dose treatment conditions. The posted mean scores were 23.1 for placebo, 30.6 for oxymorphone ER 15 mg, 62.5 for oxymorphone ER 30 mg, 81.4 for oxycodone CR 30 mg, and 93.0 for oxycodone CR 60 mg. The sponsor’s ANCOVA comparisons show estimates and uncertainty for selected pairs of conditions; the recalculated comparisons are unadjusted and should not replace those sponsor analyses.

The endpoint is subjective, and it is specific to a particular wording, scale, population, and study setting. The study was described as involving healthy non-dependent recreational opioid users and used a crossover design with single-dose conditions. The record does not establish how these findings apply to people with other characteristics or to other dosing situations. It also does not report outcomes that would establish effects on broader clinical endpoints.

A low p-value does not itself establish that a difference is clinically important, and an interval that excludes zero does not supply a threshold for meaningful change. The registry record does not report such a threshold. Conversely, a result with p = 0.395 in the sponsor’s placebo versus oxymorphone ER 15 mg analysis does not prove that the conditions are equivalent; the interval extends from -7.3 to 18.3 and is compatible with a range of differences under the analysis.

The findings should therefore be read as comparisons on the posted High VAS - Emax measure, not as a complete assessment of treatment effects. The record’s stated exclusions, lack of imputation, absent details about the statistical model, and absence of a reported multiplicity procedure are relevant to how much can be inferred. These results do not constitute medical advice or a recommendation about treatment.

17. Why This Trial Matters Statistically

This study illustrates how a clearly defined subjective endpoint can be represented numerically and compared across randomized treatment conditions. The High VAS scale gives an explicit prompt and anchored response range, while the Emax summary condenses repeated assessments into the reported endpoint. At the same time, the record does not explain the exact derivation of Emax, showing why endpoint labels and administration schedules do not always provide a complete analysis definition.

The crossover design makes the distinction between raw treatment summaries and model-based comparisons especially important. The posted means and standard deviations support simple unadjusted recalculations, but those summaries do not include the within-participant relationship needed to reproduce a crossover analysis. The sponsor’s ANCOVA results are therefore central to describing the trial’s reported comparisons, while the recalculations are useful only as a separate, limited check.

The record also demonstrates why statistical significance, precision, and practical meaning should be kept distinct. Some sponsor intervals exclude zero and have p < 0.001, while the placebo versus oxymorphone ER 15 mg interval includes zero. None of these results supplies a threshold for a meaningful change on the scale. The endpoint’s numerical nature does not resolve that interpretive question.

Finally, the study’s findings are bounded by its defined population, single-dose setting, and reported outcome. Missing details about the original sample-size rationale, multiplicity handling, model specification, and interim analyses affect reproducibility and the scope of interpretation. Recognizing those limits is part of reading a clinical trial result statistically rather than treating a list of estimates as a complete account of evidence.

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 study is described as exploratory and single-dose, and the population is healthy non-dependent recreational opioid users. That setting limits how far the findings can be generalized. The primary outcome is a self-rated feeling, rather than a broader clinical outcome, and the registry record does not report a threshold that would define a meaningful score difference.

The record identifies a crossover design but does not report treatment order, sequence allocation, washout details, or how period and sequence effects were addressed. It also does not report the exact method for deriving Emax from the scheduled assessments. The sponsor analysis is identified as ANCOVA, but its model specification and covariates are not reported. These omissions prevent an independent reconstruction of the adjusted results from the registry record.

The primary outcome analysis used a Per Protocol population, defined by receipt of all 5 treatments and absence of major protocol deviations or other exclusionary circumstances. Missing values were not imputed. The record does not report the pattern or extent of missing endpoint observations, or sensitivity analyses that might show how missingness or the analysis population affected findings.

Although six sponsor analyses are listed, multiplicity handling is not reported. The original sample-size justification, interim analysis plans, and a detailed prespecified statistical analysis plan are also not reported in the registry record. The listed enrollment and treatment-condition summary counts should not be combined to infer unique participants, particularly given the crossover design. Finally, the serious adverse event counts are descriptive and do not establish comparative safety.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Study to Compare Oxymorphone Extended-Release (Opana ER) Versus Oxycodone Controlled-Release (Oxycontin) — NCT00955110.
• Registry record last updated 2017-10-05; status last verified 2017-09.
• Results first posted 2011-07-14.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

22. Record Summary

The completed Phase 4 study was randomized, quadruple-masked, and crossover. Its official title describes a single-dose comparison of extended-release oxymorphone and controlled-release oxycodone in healthy non-dependent recreational opioid users. Registry enrollment was 78, and the primary outcome summaries report 35 for each listed treatment condition.

The primary outcome was High VAS - Emax (mm), a continuous mean score based on the statement I am feeling high. The sponsor reported six ANCOVA comparisons. The record states that the primary analyses used a Per Protocol population and that missing values were not imputed.

The sponsor’s analyses take precedence over unadjusted recalculations from the posted summary statistics because the recalculations do not reproduce the sponsor’s adjusted crossover analysis. Details about multiplicity, model specification, the original sample-size rationale, and interim analyses are not reported in the registry record. Serious adverse events were listed as affecting zero participants in each treatment-phase group; death counts are not reported.