Safety and Efficacy of AVP-923 in the Treatment of Central Neuropathic Pain in Multiple Sclerosis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safety and Efficacy of AVP-923 in the Treatment of Central Neuropathic Pain in Multiple Sclerosis (NCT01324232), a randomized phase 2 study. The primary outcome is Association Between the Dextromethorphan Plasma Concentration and the Change From Baseline Pain Rating Scale (PRS) Score to the Average Pain Rating Scale Score During Days 57 Through 84.

2. Clinical Question

The trial also examined changes in pain ratings and several other measures, including fatigue, disability, multiple sclerosis impact, sleep, neuropsychological screening, depression, and cognitive performance. The central statistical distinction is between the primary concentration–pain association and comparisons of outcomes across randomized treatment groups. The former asks whether participants with different measured plasma concentrations also tended to have different pain-score changes. The latter asks how outcome summaries compare between assigned groups. These questions use different information and should not be treated as interchangeable.

The registry record lists 209 participants enrolled and identifies the study as completed. It does not report a clinical threshold for an important change in the pain rating scale, so the numerical changes cannot be classified as clinically important from the registry information alone. Nor does the primary result by itself establish whether any treatment group differed from placebo.

3. Trial Design

This was a phase 2, randomized, quadruple-masked, parallel-group, placebo-controlled, multicenter dose-finding study. The official title describes three dose levels of AVP-923 and placebo. The intervention groups are named AVP-923-20, AVP-923-30, and AVP-923-45; the registry record does not specify the units or administration schedule behind those labels. The total enrollment reported is 209.

The registry gives a study start date of 2011-09-08 and a primary completion date of 2013-09-26, and lists the overall status as completed. The listed objectives were to evaluate safety, tolerability, and efficacy of three AVP-923 doses for central neuropathic pain in participants with multiple sclerosis. Randomization and masking are relevant design protections: random assignment supports comparison of groups, while masking can reduce the influence of treatment knowledge on participants’ reports and trial conduct. The record does not describe how randomization was generated or implemented, or which specific parties were included in the quadruple masking.

The primary outcome combines pain ratings with measured dextromethorphan plasma concentrations. Pain ratings were collected on a scale from 0 to 10, with 0 meaning no pain and 10 meaning the worst possible pain. The primary description defines baseline from recent pre-baseline ratings and the post-baseline value from ratings during Days 57 through 84, subject to stated rules when ratings were unavailable in those windows. Plasma concentrations were measured on Days 22 and 50.

Secondary outcomes include pain, fatigue, disability, multiple sclerosis impact, sleep quality, neuropsychological screening, depression, and cognitive performance. The registry record does not report the planned sample-size justification, the intended treatment duration beyond the outcome timeframes, or a complete statistical analysis plan. Those omissions limit what can be concluded about the design choices beyond the information explicitly recorded.

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: After screening procedures, participants underwent a 1-week washout period for all analgesic medications, with the exception of ibuprofen in doses that did not exceed 800 milligrams per day (mg/day).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Modified Intention-To-Treat (MITT) Population: all participants in the ITT Population (randomized participants) who received at least one dose of study drug, provided a Baseline PRS score, and had at least one post-Baseline PRS assessment. Analysis was based on the randomized treatment assigned (regardless of actual treatment received).

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 Association Between the Dextromethorphan Plasma Concentration and the Change From Baseline Pain Rating Scale (PRS) Score to the Average Pain Rating Scale Score During Days 57 Through 84, measured over Baseline; Days 57 through 84 (PRS score); Days 22 and 50 (dextromethorphan plasma concentrations).

The primary endpoint links dextromethorphan plasma concentration with change in the Pain Rating Scale (PRS). Participants rated pain over the prior 12 hours on a 0-to-10 scale, where 0 indicates no pain and 10 indicates the worst possible pain. The endpoint’s change score is defined as the post-baseline PRS average minus the baseline average. Under that definition, a negative change means the later average rating was lower than the baseline average.

Baseline PRS was defined as the average of ratings in the last 7 days before the baseline visit. If fewer than 4 ratings were available in that period, the registry says the average of up to 7 of the most recent pre-baseline ratings was used. The post-baseline score was the average for Days 57 through 84. If no PRS ratings were available during those days, the average of the last 7 available post-baseline ratings was used.

The concentration measure was based on the average logarithms of dextromethorphan plasma concentrations on Days 22 and 50. The registry describes calculating a Pearson correlation between concentration and change in PRS across the reporting groups. In general, an exposure–outcome association can help assess whether measured drug exposure tracks symptom change. It is not the same as a randomized treatment-versus-placebo effect, and an association alone does not establish that concentration caused a change in pain. The concentration summary values themselves are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The trial’s main statistical questions involve distinct estimands: an association between measured concentration and pain change, and comparisons of outcome changes across randomized groups. Keeping those questions separate is essential. A correlation describes how two measured quantities vary together. A treatment comparison estimates a contrast between groups assigned to different interventions. The primary reported Pearson correlation of 0.002 addresses the first question, not a dose-versus-placebo contrast.

What does a Pearson correlation describe?

In general, Pearson correlation summarizes the direction and strength of a linear relationship between two variables. A value near zero means the analysis found little linear co-variation in the data. It does not rule out a curved or otherwise non-linear relationship, and it can be affected by unusual observations or by the range of values represented. The registry reports a two-sided test with p = 0.983, but no confidence interval for the correlation. Without an interval, the range of associations compatible with the data is not directly available in the record. The plasma concentration summaries are also not reported, limiting independent assessment of the observed exposure range.

How should the PRS summaries be read?

The primary PRS arm summaries are medians and full ranges. A median identifies the middle of the observed distribution; the full range gives its lowest and highest reported values. These summaries are useful when the data are skewed or include extreme observations, but a range is not a confidence interval and does not measure uncertainty in the median in the same way. The reported values describe change from baseline, calculated as post-baseline minus baseline. Negative values therefore correspond to lower post-baseline ratings under the stated scoring rule. The registry does not report a confidence interval for each arm’s median.

Some secondary summaries are least-squares means with standard errors. In general, a least-squares mean is a model-based adjusted estimate, while a standard error describes the estimate’s sampling uncertainty under the fitted model. It should not be read as the spread of individual participant outcomes. The registry reports ANCOVA for several secondary analyses. ANCOVA is a regression approach that can compare groups while accounting for covariates included in the model. The covariates and full model specification are not reported in the registry record, so the adjusted estimates cannot be fully reconstructed from the posted summaries.

What is the role of the recalculated comparisons?

The displayed recalculations are simple, unadjusted comparisons reconstructed from posted summary statistics, not replacements for the sponsor’s model-based analyses. For example, the recalculated unadjusted mean difference for AVP-923-20 versus placebo in the secondary pain-change outcome is -0.03, with p = 0.948. The recalculated comparison for AVP-923-30 versus placebo is -0.37, with p = 0.42. The corresponding sponsor-reported adjusted analysis for AVP-923-20 versus placebo gives an adjusted mean difference of -0.04, a 95% confidence interval from -0.94 to 0.87, and p = 0.938. These figures are related but not identical estimands: the recalculation is unadjusted and uses approximated variability, whereas the sponsor analysis is adjusted.

In general, a standard error can be used to approximate a standard deviation when combined with a sample size, but such a reconstruction depends on assumptions about the reported estimate and analysis. The recalculated comparisons therefore serve as an analytic cross-check, not as a more authoritative result. The sponsor’s reported analysis is the primary source for its adjusted inference. The registry record does not give all details needed to independently reproduce the ANCOVA.

What do the planning illustrations mean?

The sample-size illustrations use the observed effect and then consider planning for 100%, 75%, and 50% of that effect. They assume a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. These are planning illustrations, not the trial’s documented design calculation, and not observed power. They describe the number of participants per arm that would be needed under those assumptions to target the stated power for a two-arm comparison. The registry record does not report the trial’s own sample-size justification or show that the study was designed around these recalculated effects.

In general, a smaller target effect requires more information to distinguish it from random variation, so planning requirements increase as the assumed effect is reduced. The illustrations depend heavily on the observed effect, which may itself be unstable. They should not be taken as evidence that the completed trial had a particular probability of detecting an effect. They also do not account for all possible design features, such as the model covariates, multiplicity, or missing-data assumptions, unless explicitly included in the stated assumptions.

Why can over-dispersion matter?

In general, over-dispersion means outcome variability exceeds what a simpler statistical model expects. If a model underestimates variability, standard errors may be too small and uncertainty may appear narrower than warranted. Conversely, accounting for extra variability can widen uncertainty. The PRS results include full ranges, but the registry does not report a distributional model or an over-dispersion assessment for the primary outcome. No claim about over-dispersion in this trial can therefore be made from the registry record.

Finally, p-values and confidence intervals address sampling uncertainty under specified models; they do not correct automatically for multiple testing, bias from missing outcomes, or limitations in how outcomes were measured. Those considerations are especially relevant when a study reports several outcome domains and multiple comparisons. The statistical plan for handling those issues is not reported in the registry record.

10. Analyses Reported by the Sponsor

The sponsor-reported primary analysis is the Pearson correlation of 0.002, with a two-sided p-value of 0.983. That is the result corresponding to the stated primary question about concentration and change in PRS. The registry does not report a confidence interval for that correlation. The arm-level medians and ranges provide descriptive context, but they do not constitute a sponsor-reported adjusted comparison of each dose with placebo for the primary association endpoint.

For the secondary pain-change outcome, the sponsor reports an overall ANCOVA p-value of 0.887. For AVP-923-20 versus placebo, the reported adjusted mean difference is -0.04, with a 95% confidence interval from -0.94 to 0.87 and p = 0.938. The simple unadjusted recalculation gives an estimate of -0.03 and p = 0.948. The close numerical estimates are reassuring as a basic check, but the recalculation is not the same analysis: it does not reproduce the sponsor’s adjustment and uses approximate variability.

For fatigue, the sponsor reports an overall ANCOVA p-value of 0.973; the recalculated unadjusted comparison for AVP-923-20 versus placebo has an estimate of 1.32 and p = 0.585. For multiple sclerosis impact, the overall ANCOVA p-value is 0.478, while the sponsor reports an adjusted AVP-923-20 versus placebo mean difference of 0.49 with a 95% confidence interval from -6.76 to 7.74. The recalculated estimate is 0.5. These comparisons are checks from summary data, not substitutes for the sponsor’s analysis. There is no reported discrepancy flagged for these results.

11. Multiplicity, Missing Data and Analysis Populations

The registry describes a modified intention-to-treat population for the primary outcome: randomized participants who received at least one dose, had a baseline PRS score, and had at least one post-baseline PRS assessment. The analysis description says it was based on randomized treatment assignment, although the registry text is truncated after that statement. The listed population therefore excludes randomized participants who did not meet one or more of those outcome and treatment conditions. The record does not report the number excluded for each reason or a corresponding full intention-to-treat sensitivity analysis.

The primary outcome description specifies rules for constructing PRS averages when measurements were sparse. Baseline uses the average from the last 7 pre-baseline days; if fewer than 4 scores were present, it uses up to 7 of the most recent available pre-baseline scores. The post-baseline value uses Days 57 through 84; if there are no scores in that window, it uses the average of the last 7 available post-baseline scores. These rules describe how the endpoint was derived. They do not, by themselves, explain how participants with no usable post-baseline observations were handled in the analysis population.

The registry record does not report a complete missing-data strategy, sensitivity analyses for missing outcomes, or assumptions about why data might be missing. It also does not report the planned control of multiplicity across the primary and secondary outcomes, or across treatment-group comparisons. With multiple registered outcomes and analyses, that omission makes it difficult to identify which secondary tests, if any, were intended to be confirmatory. The full statistical analysis plan, including interim analyses and prespecified model details, is not reported in the registry record.

In general, excluding participants without usable outcome data can affect comparability if missingness is related to treatment or outcome. Sensitivity analyses assess whether conclusions change under alternative plausible assumptions. Without such analyses in the registry record, the robustness of the posted estimates to missing data cannot be evaluated. These are limits on interpretation, not evidence that the reported analyses were necessarily biased.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 3% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: duodenitis (2.0%), pneumonia (1.9%), tonsillitis (1.9%), vertigo (1.9%), diplopia (1.9%).

The registry reports serious adverse events among participants at risk in each group: 1 in placebo, 1 in AVP-923-20, 1 in AVP-923-30, and 2 in AVP-923-45. The corresponding numbers at risk were 49, 54, 53, and 53. The registry reports no deaths in any group. These counts describe events recorded during the study; they do not establish that an event was caused by the assigned treatment.

The record also lists serious-event terms and percentages: duodenitis at 2.0%, and pneumonia, tonsillitis, vertigo, diplopia, and cellulitis at 1.9% each. The listing does not identify which treatment group each term occurred in, and the displayed percentages should not be interpreted as comparative treatment rates without that context. The registry does not provide event timing, duration, severity detail beyond the serious-event classification, or an analysis comparing safety between groups.

In general, adverse-event counts in a trial describe what was observed in its participants; they do not by themselves quantify risk in other populations or establish causation. The numbers should be read alongside the study’s size, follow-up, and reporting completeness. The registry’s summary is not a substitute for a full safety dataset.

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

The posted primary result addresses whether measured dextromethorphan concentration tracked change in pain rating across the reporting groups. The Pearson correlation of 0.002, with p = 0.983, shows that the analysis did not detect a linear association in these data. It does not show that the true correlation is exactly zero, and it does not establish that AVP-923 has no treatment effect. A concentration association and a randomized treatment comparison answer different questions.

The arm-level median changes describe the observed PRS change distributions. Because change is calculated as post-baseline minus baseline, negative values represent lower average pain ratings after baseline. But the medians and full ranges do not provide a direct adjusted estimate of each treatment’s effect relative to placebo. They also do not establish how participants experienced the changes or whether those changes were meaningful to them. A clinical-importance threshold is not reported in the registry record.

The secondary analyses add information about other outcomes, but the registry’s reported overall tests do not provide strong evidence of group differences. That should not be taken as proof of equivalence. Demonstrating equivalence would require a defined margin and an analysis designed to assess whether plausible effects fall within that margin; neither is reported in the registry record. Wide uncertainty can leave both benefit and no meaningful difference compatible with the data.

The safety summaries report serious adverse-event counts and no deaths in the listed groups, but they cannot establish whether events were caused by treatment or compare risks precisely. Likewise, missing details about multiplicity, missing-data sensitivity analyses, and the full analysis plan constrain the strength of conclusions. The findings should therefore be read as the results recorded for this phase 2 dose-finding study, not as a definitive answer to every clinical or safety question.

14. Why This Trial Matters Statistically

This study illustrates why the statistical question must be identified before interpreting a result. Its primary endpoint is a concentration–outcome association, while many readers may initially expect a direct comparison of each dose with placebo. The reported correlation is not a treatment effect. Keeping those estimands separate prevents a near-zero association from being misread as proof that randomized treatment groups were the same.

The trial also shows how a single outcome can be represented in several ways. The primary PRS summaries are medians and ranges, while secondary pain results include adjusted means and model-based comparisons. Those summaries serve different purposes and cannot be directly substituted for one another. The adjusted analysis uses a model; a simple recalculation from summary statistics can check broad numerical consistency but cannot recover the full model or its assumptions.

Multiple outcomes are another important statistical feature. Pain, fatigue, disability, and other measures address different aspects of the study question, but each additional analysis increases the need for a prespecified interpretation strategy. Without a reported multiplicity plan, isolated secondary findings would warrant particular caution. The record’s incomplete reporting of some outcomes and analysis details further limits assessment of the full evidence pattern.

Finally, planning illustrations based on observed effects are not evidence of the original design’s power or justification. They are conditional calculations that depend on assumptions and on effect estimates from the observed data. Distinguishing planning from observed results is important when using a completed study to inform future research.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

Several limitations are specific to the information recorded for this trial. The primary endpoint is a correlation between plasma concentration and PRS change, but the concentration summaries are not reported in the registry record. That prevents assessment of the exposure distribution and makes it harder to evaluate whether the measured concentrations covered a useful range. A confidence interval for the primary correlation is also absent, leaving its precision unclear.

The primary outcome description supplies rules for deriving the PRS averages, but it does not report the complete missing-data strategy or sensitivity analyses. The modified intention-to-treat definition requires at least one post-baseline PRS assessment, so participants without such an assessment are not part of that stated population. The registry does not report how many participants were excluded under this condition or whether conclusions were tested in another population.

The study is identified as phase 2 and dose-finding. The registry does not report the sample-size justification, the complete statistical analysis plan, the covariates in the ANCOVA models, or the multiplicity strategy. These missing details limit assessment of how the analyses were selected and how much weight to give secondary comparisons. The trial’s masked and randomized design is reported, but operational details about randomization and masking are not.

Although the registry lists additional secondary outcomes, numerical results for all listed measures are not reported in the record. Safety information is also presented as summary counts and selected serious-event terms without group attribution for each listed term. These reporting limits do not establish that an analysis or event was absent; they define what can be evaluated from this record.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Safety and Efficacy of AVP-923 in the Treatment of Central Neuropathic Pain in Multiple Sclerosis — NCT01324232.
• Registry record last updated 2021-11-22; status last verified 2021-10.
• Results first posted 2021-11-22.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

The registry describes a completed phase 2, randomized, quadruple-masked, parallel, placebo-controlled, multicenter dose-finding study of AVP-923 for central neuropathic pain in participants with multiple sclerosis. It reports enrollment of 209, with three AVP-923 groups and a placebo group.

The primary outcome assesses the association between dextromethorphan plasma concentration and change in PRS. The reported Pearson correlation is 0.002, with p = 0.983. The record also reports secondary outcome summaries and serious adverse-event counts, but does not include numerical results for every listed secondary measure. Key details, including the primary correlation’s confidence interval, the complete analysis plan, and a multiplicity strategy, are not reported in the registry record.