Dose Response Study of GSK2330672 for the Treatment of Pruritus in Participants With Primary Biliary Cholangitis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Dose Response Study of GSK2330672 for the Treatment of Pruritus in Participants With Primary Biliary Cholangitis (NCT02966834), a randomized phase 2 study. The primary outcome is Mean Change From Baseline at Week 16 in the Mean Worst Daily Itch Score.

2. Clinical Question

The central statistical question is about the size and precision of the difference in change, not simply whether scores moved from baseline within a group. A decrease in the score represents a decrease in reported itch severity on the scale. The placebo comparison helps put the changes in the context of the randomized control group, while the confidence intervals indicate the uncertainty around the reported estimates.

3. Trial Design

This completed Phase 2 study, NCT02966834, is titled Dose Response Study of GSK2330672 for the Treatment of Pruritus in Participants With Primary Biliary Cholangitis. Its official title describes a randomized, double-blind, multi-dose, placebo-controlled study. It used a parallel-group design and included placebo and multiple GSK2330672 dose regimens. The registry identifies GlaxoSmithKline as the sponsor and reports enrollment of 147.

The intervention descriptions list GSK2330672 20 mg once daily, 90 mg once daily, 180 mg once daily, 40 mg twice daily, and 90 mg twice daily. Each active regimen included matching placebo tablets to preserve blinding. The brief study summary describes a prospectively defined adaptive design using interim data to inform and potentially optimize doses, and says an additional dose regimen could be added. The registry record does not report how that adaptive process was implemented or whether a regimen was added.

The listed primary endpoint was measured at baseline and Week 16. The study summary describes participation as including up to 45 days of screening and 24 weeks of study, including follow-up. The registry lists a start date of 2017-01-11, primary completion on 2020-04-15, and status as completed.

For the primary outcome, the analysis population was defined as randomized participants who received at least one dose, had a baseline assessment, and had at least one on-treatment assessment. The registry says only participants with data available at the specified data points were analyzed. The number contributing to the primary outcome differs by arm; those analyzed counts are not the same as the registry’s overall enrollment. Detailed allocation procedures, dose assignment probabilities, and the statistical rules for any adaptive decisions are not reported in the registry record.

4. Participant Flow and Accounting

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

Before assignment: A total of 147 adult participants were randomized in this study.

Main Study Period (Up to 12 Weeks)

Reasons for not completing this period:

Final Study Period (Up to 4 Weeks)

Follow-up (Up to 4 Weeks)

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-to-Treat (ITT) Population comprised of all randomized participants who received at least one dose of study treatment, had a Baseline and at least one on-treatment assessment. Only those participants with data available at the specified data points were analyzed.

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: Baseline characteristics is presented for Safety Population which consisted of all randomized participants who took at least one dose of randomized study treatment.

Only the first 6 of 7 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Mean Change From Baseline at Week 16 in the Mean Worst Daily Itch Score, measured over Baseline and Week 16.

The primary endpoint was mean change from baseline at Week 16 in Mean Worst Daily Itch Score. Participants rated itching on a numerical rating scale from 0, no itching, to 10, the worst imaginable itching. For each day, the Worst Daily Itch Score was the highest, or most severe, rating recorded that day. The mean score for the endpoint was calculated from the worst daily scores during the 7 days before the Week 16 visit.

Baseline was the average of scores during the 7 days before the Week 4 visit, identified as Visit 3. Change from baseline was calculated as the post-baseline value minus the baseline value. Consequently, a negative change represents a lower score at the endpoint than at baseline. The endpoint captures a participant-reported symptom over several days rather than relying on a single rating at the visit.

The registry reports least-squares mean changes and 95% confidence intervals. It describes an analysis of covariance (ANCOVA) including treatment group and centered baseline Mean Worst Daily Itch Score. Adjusting for baseline can account for initial differences in scores when estimating group means. The registry record does not report additional model details beyond that description. The endpoint is useful for describing change in the symptom the study set out to examine, but these estimates alone do not establish how participants experienced other aspects of the condition or how the intervention affected outcomes not included in this endpoint.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: it is a score measured on a numerical scale. The registry reports least-squares mean change, an estimate from the stated analysis of covariance, rather than only a raw average change. The ANCOVA included treatment group and centered baseline Mean Worst Daily Itch Score. In general, an analysis of covariance uses a baseline measure to adjust comparisons for initial outcome values. Centering is a way of expressing the baseline variable relative to a reference value; the registry does not give further details about the centering convention or other model terms.

What do the group estimates describe?

Each least-squares mean change describes the estimated average change for one randomized group under the reported model. The placebo estimate was -1.73, and estimates for the five active regimens ranged from -2.19 to -2.86. These are changes from baseline, not direct comparisons between an active regimen and placebo. Because the outcome is scored so that lower values indicate less severe itching, a negative change represents a decrease in the measured score. The registry reports a 95% confidence interval for each group estimate to convey uncertainty around that estimated mean.

What does a confidence interval communicate?

In general, a confidence interval describes the precision of an estimate under the statistical procedure used to construct it. A narrower interval indicates greater precision than a wider interval, while neither width nor location alone establishes clinical importance. For a difference between groups, an interval that includes zero is compatible with no difference under the interval procedure; it is also compatible with nonzero differences in the directions represented by its endpoints. The primary comparison intervals all include zero, and several extend to values on both sides of it. The registry record does not report a p-value for these primary comparisons.

Why distinguish a change from a comparison?

A within-group change asks how the estimated score at follow-up differs from that group’s baseline. A randomized comparison asks how the estimated change in one group differs from the change in the control group. These are not interchangeable. The posted placebo change is also negative, so a negative change in an active group does not by itself indicate an advantage over placebo. The reported net mean differences are the relevant posted contrasts for the active-versus-placebo question.

What does adjustment for baseline do?

In general, baseline adjustment can improve the comparison by accounting for initial outcome values, particularly when baseline scores are associated with follow-up scores. The registry states that treatment group and centered baseline score were included in the ANCOVA. It does not report further details such as how missing measurements were handled in the model, the covariance assumptions, or whether other covariates were included. Those details should not be inferred from the use of ANCOVA alone.

How should the planning illustrations be read?

The planning illustrations are based on observed effects and assumptions stated in the registry: a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and effects set at stated fractions of the observed effect. They are planning illustrations, not the trial’s actual design justification and not observed power. In general, planning estimates depend on assumptions about the effect to detect and variability in the outcome. Smaller effects to detect require more participants for a chosen power target than larger effects, all else being equal. The registry does not report the trial’s own sample-size justification or the assumptions used to select its enrollment.

The primary analysis population also affects interpretation. It included randomized participants who received at least one dose, had baseline data, and had at least one on-treatment assessment; only those with data at the specified points were analyzed. This definition is more limited than a claim that every randomized participant contributed to every outcome estimate. The registry record does not report additional missing-data procedures, sensitivity analyses, or model diagnostics, so their influence cannot be evaluated from the recorded information.

10. Analyses Reported by the Sponsor

The registry reports sponsor analyses as net mean differences between placebo and each GSK2330672 regimen, with 95% confidence intervals. For 20 mg once daily the difference was -0.47, with an interval from -1.75 to 0.82; for 90 mg once daily it was -0.88, from -2.07 to 0.31; for 180 mg once daily it was -0.88, from -2.03 to 0.28; for 40 mg twice daily it was -1.13, from -2.29 to 0.03; and for 90 mg twice daily it was -0.53, from -1.71 to 0.65. The registry does not report p-values for these contrasts.

The sponsor’s posted estimates are the primary source for interpreting the study’s own analysis. Any simple recalculation from arm-level summary statistics would not replace the sponsor’s analysis, which may use participant-level data and the stated ANCOVA. The posted primary analysis description identifies treatment group and centered baseline score as ANCOVA terms, while the registry record does not provide enough detail to reproduce that model from the arm summaries alone.

All five reported intervals include zero. That means the posted comparisons retain uncertainty about the direction and size of the difference; it should not be converted into a claim that the regimens were equivalent or that there was no effect. The registry does not report a prespecified equivalence or non-inferiority margin. The arm-level least-squares mean changes and the net comparisons are complementary summaries, but the comparison estimates are the relevant quantities for reading the active-versus-placebo question.

11. Multiplicity, Missing Data and Analysis Populations

The registry brief summary describes a prospectively defined adaptive design that would use interim data to inform and potentially optimize the dose regimens under investigation. It also says that an additional regimen could be added. This is important context: adaptive dose selection and multiple comparisons can affect how uncertainty should be interpreted. However, the registry record does not report the adaptation rules, the timing or results of interim analyses, whether any dose was added, or how the primary analysis accounted for adaptation.

The record lists five active regimens in the outcome results, each compared with placebo. It does not report a multiplicity adjustment for these primary comparisons, a testing hierarchy, or whether the confidence intervals were adjusted for the number of comparisons. Accordingly, the reported 95% confidence intervals should be read as posted, without assuming family-wise error control or another multiplicity procedure. The record also does not report a prespecified threshold for declaring a dose-response pattern.

Missing-data handling is only partly described. For the primary endpoint, the analysis population comprised randomized participants who received at least one dose, had baseline data, and had at least one on-treatment assessment. Only participants with data at the specified points were analyzed. The registry record does not report whether missing outcome values were imputed, how withdrawals or intermittent missing assessments were handled, or whether sensitivity analyses tested the effect of missing data. It also does not report the reasons for missing observations by group.

The outcome-specific numbers analyzed vary among arms. This means that the estimates are based on the available observations contributing to that outcome, rather than necessarily on every participant enrolled or randomized. The registry does not report an intention-to-treat analysis that included every randomized participant regardless of treatment received or data availability; the stated ITT definition itself requires treatment exposure and outcome measurements.

The planning illustrations use the observed effect and stated assumptions, including a two-sided significance level of 0.05, equal allocation, and a normal approximation. They are not the trial’s sample-size justification and are not observed power. The trial’s original sample-size rationale, any interim decision thresholds, and its complete prespecified statistical plan are not reported in the registry record.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: lower respiratory tract infection (6.2%), lumbar spinal stenosis (4.3%), constipation (3.7%).

The registry reports safety counts separately for the main study period, final study period, and follow-up. During the main study period, serious adverse events were reported for 1 participant in the GSK2330672 90 mg once-daily group; the serious-event count was 0 for placebo and the other listed regimens. During follow-up, the serious adverse-event count was 1 in the 20 mg once-daily group and 1 in the 180 mg once-daily group, and 0 in the other listed groups. The final study period lists 0 serious adverse events for each group. The registry lists 0 deaths in every group and period shown.

Other adverse events were reported in multiple groups and periods. The registry identifies lower respiratory tract infection at 6.2%, lumbar spinal stenosis at 4.3%, and constipation at 3.7% among the top serious adverse event terms. These term percentages should be read as posted; the registry information used does not provide the event-level context needed to interpret their timing, recurrence, or relationship to treatment.

Safety counts describe events recorded in this study, not causal attribution. The registry record does not establish that an event was caused by a particular regimen. Group sizes and exposure periods differ, and the recorded counts should not be combined across periods to create a new total. Sparse serious-event counts also provide limited information about uncommon harms. These safety data are descriptive and should be considered alongside the study’s design and the limits of the registry reporting.

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

The primary results describe estimated change in a participant-reported itch score from baseline to Week 16. Every listed group had a negative estimated change, including placebo. The active-versus-placebo net estimates were also negative, but every reported 95% confidence interval included zero. These findings show the size and uncertainty of the posted estimates; they do not establish a definitive difference between any regimen and placebo.

An interval spanning zero does not prove that the groups were identical. It indicates that the data and analysis remain compatible with no difference as well as with differences in either direction across the interval. The estimates also do not answer whether any possible difference is important to participants, because the registry record does not provide a threshold for clinical importance. Nor does it report an equivalence margin that would support a conclusion of equivalence.

The results apply to the outcome definition and analysis population used: worst daily itch ratings averaged over the 7 days before the Week 16 visit, compared with the baseline period before the Week 4 visit, among participants meeting the stated analysis-population criteria and having the relevant measurements. They do not by themselves establish effects on other symptoms, laboratory measures, longer-term outcomes, or populations not represented in the trial.

The registry’s adaptive-design description and the range of studied doses make the study relevant to evaluating dose-response questions. But the adaptation details and multiplicity strategy are not reported in the registry record, limiting interpretation of the set of dose comparisons as a confirmatory dose-selection analysis. Secondary outcomes offer additional endpoints, but many numeric results are not reported and unadjusted analyses require caution.

In short, the registry supports a precise description of the reported estimates and their uncertainty, not a broad treatment recommendation. The posted results should be read as evidence from this trial, with the stated analysis population, endpoint, and reporting limitations in view.

14. Why This Trial Matters Statistically

This trial is statistically informative because it compares placebo with several dose regimens while measuring a symptom on a continuous numerical scale. A dose-ranging design can help describe whether estimated outcomes vary across regimens, while randomization and blinding are design features intended to support interpretable comparisons. The registry’s adaptive-design description adds another statistical consideration: dose optimization based on interim data can affect how estimates and uncertainty should be evaluated.

The primary endpoint is based on repeated daily itch ratings summarized over a week, rather than a single measurement. That definition can capture symptom experience across several days, but it also makes the analysis dependent on participants having data in the specified measurement windows. The registry’s analysis-population criteria and its statement that only available measurements were analyzed are therefore central to understanding which observations contribute to the estimates.

The results also illustrate why within-group change is not the same as a randomized treatment contrast. Placebo participants had a negative mean change, as did participants in each active regimen. The net comparisons address the question of difference from placebo, and their confidence intervals show substantial uncertainty around the estimates. Reading the arm changes alone could obscure this distinction.

Finally, multiple doses, multiple outcomes, and a potentially adaptive design raise questions about multiplicity, prespecification, and missing data. The registry gives some information about the intended adaptive approach and the ANCOVA, but does not report the full statistical plan, adaptation rules, or missing-data methods. These gaps matter because they limit independent assessment of how the collection of estimates should be interpreted. The record remains useful for examining estimation, precision, endpoint definition, and the distinction between posted analyses and unadjusted secondary comparisons.

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 primary comparison intervals are broad and include zero for every listed regimen, leaving uncertainty about the magnitude and direction of differences from placebo. The analyzed counts vary across arms, and the registry does not explain the reasons for missing outcome data or report sensitivity analyses. The primary analysis population also excludes randomized participants who did not receive at least one dose or lacked the specified baseline and on-treatment assessments, according to the registry definition.

The study summary describes a prospective adaptive design using interim data, with the possibility of adding a dose regimen. The registry record does not report the interim results, adaptation rules, or how such decisions were incorporated into estimation and inference. It also does not report a multiplicity strategy for the several dose comparisons or the numerous secondary outcomes. This limits assessment of the strength of the entire set of findings, beyond the uncertainty shown in each posted interval.

Although the primary ANCOVA terms are described, the registry does not report full model specifications, missing-data procedures, sensitivity analyses, or diagnostics. It does not report a clinical-importance threshold for the itch score, so the practical importance of a particular possible difference cannot be determined from the statistical estimates alone.

Secondary outcome reporting is incomplete in the information available: many named outcomes do not have numeric results reported in the registry record. The detailed secondary comparisons are unadjusted recalculations from posted summary statistics and should not be confused with sponsor analyses or confirmatory tests. Safety data are also descriptive; serious-event counts are small, and the registry figures do not establish causal relationships. These limitations do not erase the observed estimates, but they constrain what can be concluded from them.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Dose Response Study of GSK2330672 for the Treatment of Pruritus in Participants With Primary Biliary Cholangitis — NCT02966834.
• Registry record last updated 2021-05-04; status last verified 2021-04.
• Results first posted 2021-05-04.
• 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

NCT02966834 is a completed Phase 2 randomized, double-blind, parallel-group, placebo-controlled study sponsored by GlaxoSmithKline. The registry reports enrollment of 147 participants and describes evaluation of GSK2330672 dose regimens for pruritus in primary biliary cholangitis.

The primary outcome was mean change from baseline at Week 16 in Mean Worst Daily Itch Score, rated on a scale from 0 to 10. The registry reports least-squares mean changes and 95% confidence intervals, with ANCOVA including treatment group and centered baseline score. The five reported active-versus-placebo net mean difference intervals all include zero.

The registry describes an adaptive design, but does not report its detailed interim rules or the full statistical plan. Missing-data methods and a multiplicity strategy are also not reported in the registry record. Secondary outcomes and safety counts add context, but their reporting and inferential limitations should be kept in view.