A Study of LY3556050 in Adult Participants With Diabetic Peripheral Neuropathic Pain: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of LY3556050 in Adult Participants With Diabetic Peripheral Neuropathic Pain (NCT06074562), a randomized phase 2 study. The primary outcome is Mean Change From Baseline in Average Pain Intensity (API) as Measured by Weekly Average of Numeric Rating Scale (NRS) - Bayesian Model Averaging (BMA) Dose-response Model.

2. Clinical Question

The clinical question is therefore about the Week 12 pattern of average pain scores across the recorded mazisotine regimens relative to placebo, within the trial’s defined outcome population. It is not simply a comparison of raw endpoint scores: the analysis uses change from baseline and a model that incorporates repeated measurements and other stated covariates. The posted primary contrasts are posterior mean differences with 95% credible intervals. Those intervals are important because they show the range of values compatible with the fitted model and the available data, rather than presenting a point estimate alone.

The registry record also describes the study as lasting approximately 24 weeks across 3 study periods, while the assigned treatment period in the arm descriptions is 12 weeks. The record does not report enough detail to characterize the other study periods. This page focuses on the posted Week 12 outcome summaries and does not infer what happened during unreported periods.

3. Trial Design

This was a phase 2, randomized, double-blind, placebo-controlled, parallel-group dose-ranging study. It enrolled 404 participants and is listed as completed. The registry record gives a start date of 2023-10-05 and a primary completion date of 2025-06-11. The condition is diabetic peripheral neuropathy, and the study title specifies adult participants with diabetic peripheral neuropathic pain.

Participants were assigned to placebo or one of the mazisotine regimens recorded as 50 mg, 200 mg, or 400 mg. The placebo and 50 mg arms used twice-daily dosing for 12 weeks. The 200 mg and 400 mg regimens used weekly titration, with the stated target maintenance doses reached by Week 3 and Week 5, respectively, and maintained up to Week 12. These arm descriptions explain the intended treatment schedules; the record does not provide individual adherence or exposure data.

Blinding and random assignment are features intended to reduce systematic differences in treatment assignment and outcome assessment. They do not by themselves eliminate uncertainty, missing outcome data, or differences that can arise by chance. The registry record does not report the randomization sequence method, allocation concealment details, or how blinding was maintained in practice.

The primary outcome is change from baseline in average pain intensity measured by weekly average of a numeric rating scale, analyzed at Baseline and Week 12. The primary description identifies a mixed-model repeated-measures analysis followed by a Bayesian model averaging dose-response model. The primary analysis population is restricted to participants who were randomized, received at least one dose, and had evaluable data for the specific outcome; results are stated to be reported by treatment regimen. The registry record does not report a full flow of participants through the study or the reasons participants may lack evaluable outcome data.

In addition to the primary outcome, the record lists secondary measures covering worst pain intensity, pain interference, sleep, physical function, and patient global impressions. The posted tables should be read outcome by outcome because the evaluable sample can vary across measures. The tables and plots contain the detailed arm summaries; this discussion explains how to interpret them without reproducing them.

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: All randomized participants who received at least one dose of study drug and had evaluable data for this specific outcome. As pre-specified in the statistical analysis plan (SAP), outcomes were analyzed and reported by treatment regimen.

5. Baseline Characteristics

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

Population: All randomized participants.

6. Primary Endpoint

The registry lists the primary outcome as Mean Change From Baseline in Average Pain Intensity (API) as Measured by Weekly Average of Numeric Rating Scale (NRS) - Bayesian Model Averaging (BMA) Dose-response Model, measured over Baseline, Week 12.

The primary endpoint measures change from baseline in average pain intensity at Week 12. Participants rated their average pain over the past 24 hours on a numeric rating scale from 0 to 10, where 0 means no pain and 10 means pain as bad as you can imagine. A higher score indicates a worse outcome, meaning greater pain severity. The outcome is expressed as a mean change, so it describes the average shift in scores over the specified period rather than a count of participants reaching a particular threshold.

A weekly average is used to summarize repeated ratings, which can make the endpoint less dependent on a single day’s report. The registry record describes a mixed-model repeated-measures analysis that includes visit and specified baseline and treatment-context factors. Repeated-measures modeling uses information across measured visits while accounting for the structure of observations within participants; the record does not detail every modeling decision or the treatment of all missing observations.

The dose-response framing is relevant because the study compares several active regimens with placebo rather than evaluating only one active dose. The reported primary results are posterior mean differences from a Bayesian model averaging dose-response model, accompanied by 95% credible intervals. The model’s estimates summarize the fitted dose-response analysis; they are not the same as a set of raw arm means. The endpoint is useful for assessing the study’s stated pain-intensity question, but it does not alone describe other aspects of pain, function, sleep, safety, or the overall experience of participants.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary result combines repeated-measures modeling with Bayesian model averaging for a dose-response analysis. These are distinct steps. The mixed-model repeated-measures model produces adjusted dose-response values, and the Bayesian model averaging approach is then used to estimate the primary dose-response contrasts. The registry record identifies this structure but does not report all implementation details, such as the candidate dose-response models or prior distributions.

What does the repeated-measures model do?

Repeated observations from the same participant are related rather than independent. A mixed-model repeated-measures analysis is a general approach for handling this structure while estimating mean patterns over visits. The registry description says the model included fixed, centered, categorical effects for region, baseline average pain score, concurrent use of allowed concomitant medications, and visit. The allowed-medication variable was based on the interactive web response system. Centering a covariate is a general modeling practice that expresses it relative to a reference level or mean; the record does not state the exact centering convention used.

Adjustment can improve comparability of estimated outcomes when the specified factors are relevant to outcome variation. It does not mean that every possible source of variation has been controlled, and it does not remove the need to assess the analysis population and missing data. The record does not report the full covariance structure, degrees-of-freedom method, or detailed model diagnostics.

What does Bayesian model averaging mean?

In general, Bayesian model averaging combines results across candidate models, weighting them according to their support under the data and the modeling framework. It can reflect uncertainty about which dose-response shape best describes the data, rather than relying on a single assumed shape. The registry record names the method and reports posterior means and 95% credible intervals. It does not report the candidate models, prior choices, or model weights, so those specifics cannot be assessed from the posted description.

A credible interval is a Bayesian interval summary: under the fitted model and its assumptions, it describes a range of parameter values with the stated posterior credibility. This differs conceptually from a confidence interval, which is defined through the long-run behavior of a procedure across repeated samples. The primary result description calls the intervals credible intervals, and that is the appropriate label when explaining the reported posterior summaries. Although the outcome metadata also uses confidence-interval wording, the narrative analysis description identifies posterior 95% credible intervals.

How should the secondary recalculations be read?

The selected secondary tables report least-squares means and standard errors by arm. A least-squares mean is a model-adjusted mean, while a standard error describes uncertainty in an estimated mean; it is not the same quantity as the spread of individual participant measurements. The recalculated contrasts in the tables are simple, unadjusted comparisons reconstructed from the posted summaries. The calculations use an approximate pooled standard deviation where indicated, so they are not identical to the sponsor’s fitted models.

These recalculated mean differences and intervals are useful for showing the scale and uncertainty of selected contrasts in a consistent way. Their p-values are unadjusted. They should not be confused with multiplicity-adjusted tests, formal sponsor analyses, or the Bayesian primary dose-response results. The registry record does not list sponsor-reported analysis entries for these secondary outcomes in the displayed summaries.

What do the planning illustrations show?

The sample-size section uses the observed effect as a planning reference and then considers smaller fractions of that effect. Its stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The displayed quantities are participants per arm for 80% power and participants per arm for 90% power. In general, a planning calculation asks how many participants might be needed to detect a specified effect under stated assumptions. It does not measure the power actually achieved by a completed study.

Because the reference effect comes from the observed data, these values can be unstable when an estimate is small or imprecise. They also depend on assumptions that may differ from the trial’s original design. The registry record does not report the original sample-size justification, so the displayed planning illustrations should not be attributed to the sponsor’s design rationale. They are an aid to understanding how effect size and variability shape planning, not evidence that the trial was or was not adequately sized.

What is not specified?

The registry record does not report detailed handling of missing outcome data, sensitivity analyses, interim analyses, or the complete statistical analysis plan. It does state that the outcome population was specified in the analysis plan and that outcomes were analyzed by treatment regimen. That statement does not establish which missing-data assumptions or multiplicity procedures were used. Readers should distinguish what is explicitly described from what would require the full protocol or analysis plan.

10. Analyses Reported by the Sponsor

The primary contrasts shown in the posted results are the sponsor’s Bayesian model averaging dose-response estimates. They are reported as posterior mean differences with 95% credible intervals, and no primary p-values are listed. These model-based results are the relevant sponsor analysis for the primary endpoint; they should not be replaced by a simple recalculation from arm-level means, which are not reported for that endpoint in the available result entries.

The secondary comparison tables include recalculated, unadjusted contrasts derived from the posted least-squares means and standard errors. Those calculations are descriptive checks based on summary statistics and approximate variability, not a reproduction of the sponsor’s full analysis models. Differences can arise because the recalculations do not use all participant-level information or the complete model specification.

No discrepancy is flagged between a sponsor result and a recalculated result in the record. Even so, the distinction in methods remains important: the sponsor’s primary analysis uses Bayesian model averaging, while the secondary recalculations are simple unadjusted summaries. The record does not contain sponsor-reported analysis entries for the selected secondary contrasts shown in the recalculation tables. Any interpretation should identify which source and method underlie the estimate rather than treating all numbers as interchangeable.

11. Multiplicity, Missing Data and Analysis Populations

The registry record identifies one primary outcome and lists multiple secondary outcomes, but a full multiplicity strategy is not reported in the registry record. That leaves uncertainty about whether and how the sponsor controlled the overall chance of false-positive findings across secondary endpoints, dose comparisons, or other analyses. The primary endpoint’s Bayesian dose-response model may address dose-response model uncertainty in its own framework, but that is not the same as a reported family-wise multiplicity procedure across all study outcomes.

Missing-data handling is also not fully described. The primary analysis population includes randomized participants who received at least one dose and had evaluable data for the specific outcome. This is an outcome-specific restriction; the posted participant counts can differ across outcomes. The record does not report the number or reasons for missing measurements, the assumptions made about missingness, or sensitivity analyses testing alternative assumptions. A repeated-measures model can use available observations under its modeling assumptions, but the registry description alone does not establish whether those assumptions are suitable for the observed missingness pattern.

The record states that outcome analyses were reported by treatment regimen and that the population was specified in the statistical analysis plan. It does not give enough detail to assess all analysis-population decisions, protocol deviations, or any estimand strategy. The enrollment count should not be treated as the denominator for every reported outcome: the posted outcome populations are smaller and differ by measure.

Interim analyses and any associated stopping or decision rules are not reported in the registry record. The original sample-size justification is also not reported in the registry record. The planning illustrations on this page are separate calculations based on observed effects and stated assumptions; they do not recover the original design rationale. These gaps limit how fully readers can assess the analysis plan and the role of unreported decisions.

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: ovarian cancer stage iv (1.6%), gastroenteritis (1.3%), pneumonia (1.2%), diabetic ketoacidosis (1.2%), basal cell carcinoma (1.2%).

The posted safety summary reports serious adverse events among 4 of 162 participants in the placebo group, 1 of 79 in the 50 mg group, 3 of 81 in the 200 mg group, and 3 of 80 in the 400 mg group. It also reports other adverse events affecting 18, 17, 23, and 13 participants in those respective groups. The record lists one death in the 200 mg group and none in the other groups.

These are counts of participants, not a measure of causality or severity beyond the registry’s serious-event classification. The figures alone do not show whether an event was judged related to study treatment, when it occurred, whether a participant had more than one event, or the duration of observation for each individual. The record includes a list of serious-event terms and percentages, but those entries do not establish a comparative risk pattern without further context.

Safety comparisons should be considered in light of the group sizes and the study’s duration and design. The registry record’s summary is not a substitute for detailed event narratives or a full safety analysis. No causal conclusion should be drawn from the counts alone.

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

The primary result is a set of model-based estimates of the difference in average pain change for the listed mazisotine regimens relative to placebo at Week 12. The posterior means are close to zero, and each reported 95% credible interval includes zero. The results therefore convey substantial uncertainty about the direction and size of the comparisons, rather than a precise separation from placebo in this analysis.

That interpretation is limited to the endpoint, time point, population, and model described in the registry record. It is not a statement about every pain-related outcome, the whole study period, or every participant. The primary analysis population required receipt of at least one dose and evaluable data for the outcome, and the registry does not provide a complete account of missing measurements or sensitivity analyses.

The intervals are not proof that the treatments are equivalent, nor do they show that meaningful differences are impossible. Establishing equivalence or non-inferiority requires a prespecified margin and an analysis designed for that question; such a margin is not reported in the registry record. Conversely, a point estimate alone should not be read as evidence of a definite benefit or harm. Precision and the modeling assumptions matter.

Secondary results may offer context about other aspects of pain and functioning, but unadjusted calculations across multiple outcomes have greater risk of chance findings. The planning illustrations do not show observed power and do not determine whether the study’s original design was adequate. Safety counts describe events recorded in the registry, but do not establish whether treatment caused them. Taken together, the posted results support a careful description of estimates and uncertainty, not a clinical recommendation or a broad conclusion beyond the analyses reported.

14. Why This Trial Matters Statistically

This trial is statistically informative because it uses a dose-ranging design and a model intended to estimate a dose-response pattern, rather than relying solely on a single active-versus-placebo contrast. Dose-response analyses can be useful when the scientific question concerns how estimates change across several regimens. Bayesian model averaging can, in general, account for uncertainty across candidate model forms, although the registry record does not disclose the candidate models or prior choices needed to assess that feature in detail.

The posted result also illustrates why effect estimates should be read with their intervals. The central estimates alone are small in magnitude, but the intervals show the remaining uncertainty around each comparison. This is especially relevant when several doses are being evaluated: seeing a set of point estimates without considering their precision can suggest a pattern more definite than the data support.

The study provides results for pain intensity alongside outcomes addressing interference, sleep, function, and patient impressions. These domains can complement one another, while also creating a multiplicity challenge. Separating the primary analysis from secondary exploratory calculations helps readers avoid treating a selected secondary contrast as if it were the trial’s main confirmatory result.

Finally, the registry record makes clear that the primary analysis population is outcome-specific and that the treatment schedules differ by regimen. These details matter when interpreting both estimates and safety summaries. The recalculated tables and planning illustrations further demonstrate how summary-data calculations can help describe uncertainty, while remaining distinct from the sponsor’s full model and from the original trial design. The value of the record is in making these distinctions visible, not in turning uncertain estimates into a definitive clinical conclusion.

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.

The registry record does not report arm-level primary endpoint means or their dispersion in the result entries, so readers cannot reconstruct the primary contrasts from the primary arm summaries. The reported primary estimates are Bayesian model averaging results, but the registry description does not give the candidate dose-response models, prior distributions, or model weights. This limits independent assessment of the analysis choices.

For selected secondary outcomes, the recalculated comparisons rely on posted least-squares means and standard errors, with approximate variability used in the calculations. They are unadjusted summaries and are not a reproduction of the sponsor’s full model. The secondary p-values do not account for the number of outcomes and comparisons. Detailed sponsor analyses for the selected secondary contrasts are not reported in the registry record.

Outcome-specific evaluable populations vary, and the record does not provide a complete account of missing observations, reasons for missingness, or sensitivity analyses. The primary analysis population excludes participants without evaluable outcome data by its stated definition. Without further details, readers cannot determine how robust the estimates are to alternative assumptions about missing data.

The registry record does not report the original sample-size justification, detailed randomization or allocation procedures, a complete multiplicity plan, interim analyses, or all protocol and analysis-plan details. The study is described as lasting approximately 24 weeks across 3 periods, but the available treatment descriptions focus on the 12-week dosing period. The presented safety counts do not include event-level context or attribution. These limitations do not invalidate the posted results, but they constrain how much can be inferred from the registry summaries alone.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study of LY3556050 in Adult Participants With Diabetic Peripheral Neuropathic Pain — NCT06074562.
• Registry record last updated 2026-07-30; status last verified 2026-07.
• Results first posted 2026-07-30.
• 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

This completed phase 2 study enrolled 404 adults with diabetic peripheral neuropathic pain. It was randomized, double-blind, placebo-controlled, and parallel-group, comparing placebo with mazisotine regimens recorded as 50 mg, 200 mg, and 400 mg. The primary outcome was change from baseline to Week 12 in average pain intensity, measured by a weekly average of the numeric rating scale.

The primary analyses used a mixed-model repeated-measures model followed by Bayesian model averaging for dose response. The reported posterior mean differences are close to zero, and the 95% credible intervals include zero for each listed comparison. Selected secondary results and safety counts are available in the posted tables. Secondary recalculations are unadjusted, and the registry record does not report complete details for multiplicity, missing-data sensitivity analyses, or the original sample-size justification. The findings should be read as estimates with uncertainty, within the limits of the recorded analyses.