A Phase 2b Diabetic Kidney Disease Study: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Phase 2b Diabetic Kidney Disease Study (NCT04170543), a randomized phase 2 study. The primary outcome is Percent Change From Baseline to Day 169 (Week 24) in UACR - Per Protocol Population.

2. Clinical Question

3. Trial Design

This was a randomized, quadruple-masked, parallel-group Phase 2b study of MEDI3506 in subjects with diabetic kidney disease. The registry lists an enrollment of 609 and identifies AstraZeneca as the sponsor. The study status is completed; the start date is 2019-11-18 and the primary completion date is 2023-05-16.

The five arm descriptions identify four MEDI3506 dose groups and a volume-matched placebo group. Each description specifies Dapagliflozin from Day 85 to Day 168. The registry lists the MEDI3506 dose arms as 30 mg, 60 mg, 120 mg, and 300 mg, in addition to placebo. The primary endpoint was measured through Day 169, described as Week 24.

Randomization and masking describe design features, not a guarantee that all potential sources of bias were eliminated. The registry does not report the randomization method, concealment procedures, or further operational details of masking in the record. The record identifies the trial as parallel-group but does not report a sample-size justification or detailed allocation ratio.

The reported primary endpoint analysis used the Per Protocol Population, rather than all randomized participants. The arm counts for that analysis are 103 for placebo, 81 for MEDI3506 30 mg, 78 for 60 mg, 72 for 120 mg, and 131 for 300 mg. These are analysis counts, not the enrollment total or the safety denominators.

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: The following data were excluded from this analysis: * Data from patients enrolled in Phase 2a of the study: This phase was discontinued due to the COVID-19 pandemic. * Data from patients enrolled at a site that was discovered to not follow Good Clinical Practice (GCP) guidelines. * Data from patients not treated.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Per Protocol Population: All randomized participants who received any study drug and did not violate any relevant important protocol deviations affecting the primary efficacy endpoint.

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 Percent Change From Baseline to Day 169 (Week 24) in UACR - Per Protocol Population, measured over From baseline to Day 169.

The primary endpoint was percent change from baseline to Day 169 (Week 24) in UACR, the urinary albumin-to-creatinine ratio, in the Per Protocol Population. UACR combines urinary albumin and creatinine measurements into a ratio. In this trial, measurements were collected as triplicates at baseline and planned visits; each triplicate was averaged using a geometric mean. Baseline was defined as the geometric mean of UACR measurements before the first study-treatment dose.

Percent change expresses the endpoint as a change relative to each participant’s baseline rather than as an absolute UACR value. The registry analysis modeled the logarithm of the post-baseline-to-baseline UACR ratio, then reported least-squares mean percent changes and differences in percent change. These model-based summaries account for specified treatment, visit, baseline, region, and SGLT2i terms.

UACR is used in kidney-disease research as a measure of urinary albumin relative to creatinine. A change in this measure can characterize a treatment-group pattern over the study period, but it is not by itself a complete account of kidney health or of clinical outcomes. The registry’s primary endpoint concerns the measured ratio at the stated time point; it does not establish what an individual participant’s longer-term course would be.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary analysis used a mixed model with repeated measures, or MMRM. In general, an MMRM analyzes repeated observations from participants while accounting for the fact that observations from the same participant are related. This approach can use measurements across protocol-specified visits rather than reducing each participant’s information to a single observation. The registry says the model included UACR values from baseline up to Day 169.

The response was the logarithm of the ratio of post-baseline UACR to baseline UACR. In general, logarithms are useful for ratio outcomes because they express multiplicative changes on an additive model scale. The reported summaries are least-squares means of percent change and differences in percent change. Least-squares means are model-adjusted estimates, rather than simple arithmetic averages of the observed values. The registry does not describe the conversion from the model scale to percent change in further detail.

The listed fixed effects were treatment, visit, treatment-by-visit interaction, SGLT2i, region (Japan or ROW), baseline log UACR, and baseline log UACR-by-visit interaction. In general, fixed effects represent model terms used to estimate systematic differences associated with the included factors. The treatment-by-visit interaction allows modeled treatment patterns to vary by visit. The baseline-by-visit interaction allows the association between baseline log UACR and the modeled response to vary across visits. The record does not give further information about the coding of these terms or the covariance structure used for repeated observations.

The registry reports 90% confidence intervals for the primary endpoint estimates and between-group differences. In general, a confidence interval communicates estimation uncertainty: narrower intervals indicate greater precision under the model assumptions, while wider intervals indicate less precision. A confidence interval that includes zero for a difference is compatible with both a difference in either direction and no difference at the interval’s confidence level. It does not establish that all values in the interval are equally likely, nor does it prove equivalence.

For the primary analysis, the Per Protocol Population comprised randomized participants who received any study drug and did not violate relevant important protocol deviations affecting the primary efficacy endpoint. This population definition focuses the analysis on participants meeting the specified protocol conditions, but it is not the same as an analysis of all randomized participants. The registry also lists secondary outcomes using both Per Protocol and Full Analysis populations. The record does not provide further population definitions for those secondary analyses.

Missing-data handling is described for certain intercurrent events. If a participant was lost to follow-up, discontinued treatment due to an adverse event, or used prohibited medication, UACR data were treated as missing on or after the event, with no imputation. In general, no imputation means missing observations were not filled in with substituted values. This does not, by itself, explain how the model handles all missing observations or establish that missingness is harmless. MMRM estimates rely on assumptions about the observed data and missingness; the record does not report sensitivity analyses that test those assumptions.

The registry reports four sponsor analyses, each comparing one MEDI3506 group with placebo using MMRM. The estimated differences, confidence intervals, and p-values are adjusted-model results, not unadjusted arithmetic comparisons reconstructed from the arm summaries. No discrepancy is identified in the record. The statistical plan, including any prespecified multiplicity procedure, interim analyses, or sample-size justification, is not reported in the registry record.

General statistical context: over-dispersion occurs when variability in the data is greater than a chosen model assumes. Accounting for over-dispersion can improve uncertainty estimates in settings where it applies. The endpoint is analyzed with an MMRM, and the registry does not report an over-dispersion assessment or a separate over-dispersion adjustment. There is therefore no basis in the record to claim that such an issue was present or that an adjustment was used.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary comparisons as MMRM least-squares mean differences in percent change, with 90% confidence intervals and p-values. The estimates are -2.85 for 30 mg versus placebo, -19.52 for 60 mg versus placebo, -17.47 for 120 mg versus placebo, and -7.22 for 300 mg versus placebo. Their respective p-values are 0.834, 0.117, 0.177, and 0.538.

These are model-based comparisons. They should not be confused with subtracting one arm’s displayed least-squares mean from another: the sponsor’s analysis uses the stated repeated-measures model and its covariate adjustments. The record identifies no discrepancy between the posted sponsor results and any recalculation. No recalculated comparison estimates are reported in the record, so there is no separate unadjusted result to interpret against the sponsor’s analysis.

The confidence intervals for all four reported differences include zero. The registry results therefore show uncertain estimates rather than a clear separation from zero at the reported interval level. Whether the statistical plan included multiplicity adjustment is not reported in the registry record, so the p-values should be interpreted without assuming a particular adjustment.

11. Multiplicity, Missing Data and Analysis Populations

The record lists 11 outcome measures, 20 analyses, and 9 posted secondary outcomes. The secondary outcomes include UACR changes and reductions at Day 85 and Day 169, immunogenicity, and asymptomatic positive COVID-19 tests. Multiple outcomes, visits, populations, and dose comparisons create opportunities for multiple statistical testing. In general, when many hypotheses are tested, the chance of at least one apparently notable result can increase unless the analysis plan addresses multiplicity.

The record does not report whether the four primary dose-versus-placebo comparisons were covered by a multiplicity procedure, whether a hierarchy of hypotheses was prespecified, or whether the reported p-values were adjusted. It also does not report an interim-analysis plan. The primary comparisons should consequently be interpreted as posted, without assuming family-wise error control or another correction. The absence of reported details is not evidence that no procedure was used; the procedure is not reported in the registry record.

The primary analysis population was the Per Protocol Population: randomized participants who received any study drug and had no relevant important protocol deviations affecting the primary efficacy endpoint. This differs from the Full Analysis Population named for some secondary outcomes. The registry does not provide the full definition of that population in the record. The primary analysis counts were 103 for placebo, 81 for 30 mg, 78 for 60 mg, 72 for 120 mg, and 131 for 300 mg.

For loss to follow-up, treatment discontinuation due to an adverse event, or prohibited medication use, UACR data were treated as missing on or after the event and no imputation was performed. In general, missingness can affect an estimate if the chance of having observed data is related to the outcome. The MMRM’s use of available repeated measurements does not remove the need to consider assumptions about missing observations. Sensitivity analyses and further missing-data assumptions 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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: atrial fibrillation (2.2%), hyperkalaemia (2.1%), acute kidney injury (1.5%), covid-19 (1.1%), septic shock (1.1%).

The safety results use different at-risk counts from the primary efficacy analysis. The serious adverse event counts and denominators were: placebo, 15 of 133; MEDI3506 30 mg, 14 of 95; 60 mg, 5 of 89; 120 mg, 12 of 93; and 300 mg, 11 of 148. These are observed counts in the safety groups and should not be substituted for the Per Protocol Population counts.

The registry lists serious adverse event terms including atrial fibrillation (2.2%), hyperkalaemia (2.1%), acute kidney injury (1.5%), Covid-19 (1.1%), septic shock (1.1%), and cardio-respiratory arrest (1.1%). It also records one death in each of the five groups. The record does not establish from these figures whether an event was caused by study treatment.

Counts and listed percentages describe events recorded during the study; they do not by themselves establish comparative risk or explain event severity, timing, or relatedness. The registry record does not report those details in the safety information. Safety comparisons also require caution when group sizes and follow-up differ or when event counts are limited. These figures are descriptive and should not be read as a complete account of safety.

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

The posted results address a specific question: the model-estimated difference in percent change in UACR from baseline to Day 169 between each MEDI3506 dose group and placebo in the Per Protocol Population. They show negative least-squares mean percent changes in every group, including placebo, and provide a separate model-adjusted comparison for each dose. All four comparison confidence intervals include zero, and the reported p-values are 0.834, 0.117, 0.177, and 0.538.

These findings do not demonstrate that the groups are equivalent. A nonsignificant test is not an equivalence test, and the registry does not report an equivalence margin. Nor do the estimates establish that there is no possible treatment-group difference: the confidence intervals include a range of values, reflecting uncertainty. The registry also does not report a threshold for deciding whether a difference in UACR change would be clinically important.

The results do not answer every question about kidney disease, longer-term outcomes, or treatment beyond the endpoint and follow-up specified. UACR change is a measured endpoint, not a complete measure of disease course. The record’s findings are from a Per Protocol analysis, and missing-data handling and the treatment of multiplicity require consideration when interpreting the comparisons.

The safety counts are likewise descriptive. They identify serious adverse events and deaths recorded in each safety group, but do not establish causality or provide a complete comparative safety assessment. Taken together, the registry results characterize endpoint estimates, their statistical uncertainty, and recorded safety events. They should not be interpreted as medical advice or as a definitive conclusion about treatment benefit or harm.

14. Why This Trial Matters Statistically

This trial illustrates how a clinical endpoint and its statistical analysis jointly shape interpretation. UACR was measured repeatedly, with triplicates averaged using geometric means, and analyzed on a log-ratio scale using an MMRM. The reported least-squares means and differences are therefore adjusted model estimates, not straightforward summaries of raw observations.

It also illustrates the importance of keeping within-group changes separate from between-group comparisons. Each arm’s estimated percent change describes its model-based change from baseline, while the treatment comparison asks how that estimate differs from placebo. Because the placebo group also had a negative estimated change, interpreting an active group’s estimate alone would not answer the comparative question.

The confidence intervals make precision central to interpretation. The four dose-versus-placebo intervals include zero, and the registry does not report a multiplicity procedure or a clinical-importance threshold. Statistical significance, clinical importance, and certainty are distinct considerations; the posted estimates do not settle all three.

Finally, the primary analysis population, missing-data rules, and safety denominators differ in purpose. Understanding which participants contribute to each result is essential when reading a randomized trial. This record makes the endpoint model and some missing-data rules explicit, while leaving the statistical plan and other design details unreported.

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 results are reported for the Per Protocol Population rather than all randomized participants. This population excludes participants with relevant important protocol deviations affecting the primary efficacy endpoint, and the registry does not report the number or nature of those deviations. The record also does not provide a full definition of the Full Analysis Population referenced for secondary outcomes.

The primary analysis uses a specified repeated-measures model, but the record does not report the covariance structure, model diagnostics, or sensitivity analyses. UACR observations were treated as missing on or after certain intercurrent events, with no imputation. Further assumptions about missing data and analyses testing their effect are not reported in the registry record.

The registry reports four dose-versus-placebo comparisons but does not report whether multiplicity was addressed. It also does not report a sample-size justification, interim-analysis details, or a prespecified clinical-importance threshold. The sample-size planning illustration is based on the observed effect and stated planning assumptions; such illustrations are planning calculations, not observed power or evidence about the trial’s original design.

The primary endpoint is measured through Day 169, and the listed analysis concerns change in UACR rather than a broader set of long-term clinical outcomes. The safety results provide counts and selected serious adverse event terms, but the record does not report further detail on event relatedness, timing, or the complete clinical context. These limits constrain how far the reported statistical findings can be interpreted.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Phase 2b Diabetic Kidney Disease Study — NCT04170543.
• Registry record last updated 2024-07-10; status last verified 2024-07.
• Results first posted 2024-07-10.
• 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

NCT04170543 is a completed randomized, quadruple-masked, parallel-group Phase 2b study of MEDI3506 in subjects with diabetic kidney disease, sponsored by AstraZeneca. The registry lists enrollment of 609 and five groups: placebo and MEDI3506 at 30 mg, 60 mg, 120 mg, or 300 mg, with Dapagliflozin specified from Day 85 to Day 168 in each arm description.

The primary endpoint was percent change in UACR from baseline to Day 169 in the Per Protocol Population. The sponsor reported four MMRM comparisons against placebo, each with a 90% confidence interval that includes zero. The registry also reports serious adverse event counts and one death in each group. Details on multiplicity, sample-size justification, and interim analyses are not reported in the registry record.