A Study to Test Whether Different Doses of BI 690517 Alone or in Combination With Empagliflozin Improve Kidney Function in People With Chronic Kidney Disease: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Test Whether Different Doses of BI 690517 Alone or in Combination With Empagliflozin Improve Kidney Function in People With Chronic Kidney Disease (NCT05182840), a randomized phase 2 study. The primary outcome is Change From Treatment Period Baseline in Log Transformed Urine Albumin Creatinine Ratio (UACR) Measured in First Morning Void (FMV) Urine After 14 Weeks - All Patients.

2. Clinical Question

UACR is the urine albumin-creatinine ratio. The trial measured it to assess changes in urinary albumin relative to creatinine, an indicator used in kidney research. The primary outcome focuses on the mean change in this continuous measure, rather than a clinical event or a direct measure of kidney function. The stated question concerns the average change over the treatment period; it does not by itself establish what any change means for an individual participant or longer-term health outcomes.

The trial also assessed whether participants had a UACR decrease of at least 30% in secondary analyses. These responder analyses address a different question from the primary mean-change analysis: they classify participants according to whether they crossed a specified threshold. The registry record presents these secondary outcomes under several approaches to missing measurements.

3. Trial Design

This was a phase 2, randomized, double-blind, placebo-controlled, parallel dose-group trial. The registry record describes masking as quadruple. Enrollment was 714, and the trial status is completed. The recorded start date is 2022-01-11 and the primary completion date is 2023-06-19. The sponsor was Boehringer Ingelheim.

The study included a run-in period followed by a treatment period. During the run-in period, participants received empagliflozin 10 mg or placebo matching empagliflozin. The brief summary describes this period as lasting 2 months. In the treatment period, participants received BI 690517 at 3 mg, 10 mg, or 20 mg, or placebo matching BI 690517, while continuing their assigned background medication. This creates dose groups across both background-medication assignments. The registry record describes the treatment period as 14 weeks.

The brief summary says participants were in the study for about 6 months, with about 12 visits. Where possible, about 4 visits could take place at home, and trial staff could also make contact by phone or video call. Participants collected urine samples at home for analysis. These details describe the planned participant experience; they do not establish how many visits each participant completed.

Randomization and placebo control support comparisons between assigned groups, while masking is intended to reduce the influence of expectations on care and assessment. The parallel-group design compares groups followed over the same treatment period rather than having each participant receive multiple BI 690517 doses. Further details about randomization procedures, allocation concealment, and any prespecified sample-size justification 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: 2 patients of the arm "Run-in period: Placebo to empagliflozin 10 mg" who did not complete the Run-in period (RP) were mistakenly randomized to the treatment period (TP) but were not treated during the TP with BI 690517 or placebo to BI 690517.

Only the first 6 of 10 groups are shown.

Run-in Period

Reasons for not completing this period:

Completed RP and Randomized to TP

Reasons for not completing this period:

Patients Starting the Treatment Period

Treatment Period

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set (FAS): This patient set included all randomised patients to one of 3 doses of BI 690517 (3mg QD, 10 mg QD or 20 mg QD) or placebo matching BI 690517 who had at least one baseline measurement of UACR at Week -2, -1, or 0 (Week 6, 7 or 8 of the Run-in Period) and at least one post-baseline measurement when patients were still on treatment with BI 690517 or placebo matching BI 690517.

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: Randomised Set (RS): This patient set included all entered and randomised patients based on the treatment groups they were randomised to at the randomisation prior to the treatment period (i.e. at second randomisation (R2)), regardless of being treated by BI 690517 or not.

Only the first 6 of 9 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Change From Treatment Period Baseline in Log Transformed Urine Albumin Creatinine Ratio (UACR) Measured in First Morning Void (FMV) Urine After 14 Weeks - All Patients, measured over The MMRM model is a longitudinal analysis and it incorporated UACR measurements from baseline (Week 6,7 or 8 of the Run-in period) and Week 6, Week 10 and Week 12-14 of Treatment period..

The primary endpoint is change from treatment-period baseline in log-transformed UACR measured in first morning void urine after 14 weeks. UACR relates the amount of albumin in urine to creatinine in the same sample. Albumin in urine is used as a marker in kidney research, and the ratio helps interpret urinary albumin while accounting for urine concentration. First morning void sampling standardizes the type of urine sample used for measurement.

The endpoint is continuous and is analyzed on a logarithmic scale. A change below zero represents a decrease in log-transformed UACR from baseline; a value above zero represents an increase on that scale. A mean summarizes the group’s average change, while its confidence interval communicates uncertainty around that estimate. The endpoint is not a direct count of kidney failure events, nor is it a measure of change in kidney filtration reported in the posted results.

The registry record reports adjusted means and confidence intervals from a mixed model for repeated measures. It states that UACR measurements at baseline and during treatment were incorporated, including measurements at Week 6, Week 10, and Week 12-14 of the treatment period. The baseline measurements were from Week 6,7 or 8 of the run-in period. Using repeated measurements allows the model to use the longitudinal pattern rather than relying only on a single recorded value at the endpoint visit.

The outcome is useful for comparing short-term changes in a kidney-related biomarker between randomized groups. Its interpretation remains bounded by the measurement and follow-up period: a change in UACR is not, on its own, evidence of a change in longer-term clinical outcomes.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary analysis uses a mixed model for repeated measures, abbreviated MMRM. In general, this type of model estimates average outcomes over time while accounting for the fact that repeated measurements from the same participant are related. Treating those observations as if they came from unrelated people would generally misrepresent their dependence. The model can use the recorded pattern across visits to estimate change at the endpoint rather than using only one visit.

The registry record says the MMRM was estimated using restricted maximum likelihood. In general, restricted maximum likelihood is a method for estimating model variance components while accounting for the estimation of fixed effects. The record specifies fixed effects for treatment at each visit, baseline at each visit, baseline, visit, treatment, background medication, and randomization stratum as main effects, as well as random effects of patient. The statistical interpretation depends on this model structure and the data available for fitting it.

The primary outcome is log-transformed. In general, a logarithmic transformation changes the scale on which differences are modeled and can be useful when the original measure has a skewed distribution or when relative rather than additive patterns are relevant. On the log scale, a mean difference is not a raw-unit difference in UACR. The reported primary estimates are therefore interpreted on the stated log scale; a transformed result should not be casually described as a percentage change unless the relevant transformation and analysis are explicitly reported.

The arm-level primary estimates are least-squares means, also called adjusted means in the registry description. In general, adjusted means are model-based estimates that account for factors included in the statistical model. They are not necessarily identical to simple arithmetic averages of the observed values. The reported contrast is a mean difference between a BI 690517 dose group and placebo matching BI 690517. The 95% confidence interval expresses the uncertainty around that model-based contrast under the analysis assumptions.

The registry also reports MCPMod Emax, linear, and exponential model fits. In general, MCPMod is a framework for evaluating dose-response patterns using candidate models. The Emax model allows a pattern that approaches a plateau; a linear model represents a straight-line pattern across dose; an exponential model represents another possible shape. The reported fit p-values indicate evidence against the corresponding model-fit null tests as specified by the sponsor, but the complete model-selection procedure and details needed to interpret those tests are not reported in the registry record. The fit results do not replace the pairwise MMRM estimates.

Secondary responder outcomes classify whether UACR decreased by at least 30%. For binary outcomes, a risk difference describes the difference in the proportions meeting the definition between groups. Odds ratios compare the odds of meeting the definition. These measures answer related but distinct questions, and their numerical scales are not interchangeable. The posted results include logistic regression odds ratios and also permit simple unadjusted re-calculations from the reported counts. Such recalculations do not reproduce adjusted sponsor models and should be interpreted accordingly.

Missing-data methods matter because a participant may lack a measurement at the target visit. The registry lists multiple imputation, missing as non-responder, last observation on treatment carried forward, and complete-case analysis for secondary responder outcomes. In general, multiple imputation estimates missing values using a statistical model and carries uncertainty from imputation into the analysis. Missing as non-responder counts participants without an observed response as not meeting the responder criterion. Last observation carried forward uses an earlier on-treatment observation in place of the missing target measurement. Complete-case analysis uses participants with the necessary recorded measurements. Each method has assumptions, and results can vary across approaches.

Sample-size figures based on observed effects are planning illustrations, not observed power and not evidence that the trial was designed with those sample sizes. The stated illustration assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is a stated fraction of the observed effect. In general, smaller target effects require larger samples for a given power, while these calculations are sensitive to assumptions about the event rate, effect, and analysis. The trial’s own sample-size justification is not reported in the registry record.

10. Analyses Reported by the Sponsor

The registry’s sponsor-reported primary analyses use MMRM and give mean differences of -0.168, -0.486, and -0.421 for 3 mg, 10 mg, and 20 mg BI 690517 versus placebo matching BI 690517. The associated 95% confidence intervals and p-values are reported with the primary outcome. Those results are the sponsor’s own model-based analyses and are the appropriate source for describing the trial’s prespecified comparisons as recorded.

Simple recalculations from posted summary statistics are unadjusted and do not reproduce the MMRM. They can serve as a check on the direction or scale of a comparison when the necessary summary data are available, but they do not account for the repeated-measures model or its covariates. In this case, the registry record does not flag a discrepancy between the sponsor analysis and the recalculated results. The sponsor’s analysis therefore remains the primary reference for interpreting the reported comparison.

The registry additionally reports dose-response model fits using MCPMod. These address dose-response patterns rather than serving as recalculations of the pairwise MMRM contrasts. The record does not fully describe the model-selection steps, multiplicity handling, or a hierarchy among the statistical tests. Accordingly, the reported p-values should be read as the tests attached to their specified analyses, not as a single overall probability statement about all outcomes and contrasts.

11. Multiplicity, Missing Data and Analysis Populations

The registry record identifies a primary outcome and reports secondary outcomes, but does not report a multiplicity-adjustment procedure or a testing hierarchy covering the primary dose comparisons, dose-response models, and secondary analyses. Multiplicity matters because testing several contrasts or outcomes increases the chance that at least one result will appear unusual under a null hypothesis. Without a reported adjustment plan, individual secondary p-values should not be treated as if they arose from a single isolated test.

The primary outcome model incorporates measurements from baseline and multiple treatment visits. A repeated-measures model can make use of incomplete longitudinal records under assumptions about the relationship between missingness and observed data. The registry description does not state the missing-data assumption used for the primary MMRM, how departures from that assumption were evaluated, or whether sensitivity analyses were conducted for the primary endpoint. Those details are not reported in the registry record.

For secondary UACR responder outcomes, the registry explicitly lists multiple imputation, missing as non-responder, last observation on treatment carried forward, and complete-case analysis. In general, these strategies answer the response question under different assumptions. Missing as non-responder is conservative in the sense that missing participants do not count as responders, but it may not reflect the actual unobserved response. Carrying a prior measurement forward assumes that the earlier value is an appropriate stand-in for the missing target measurement. Complete-case estimates can be affected if participants with missing measurements differ from those retained. Multiple imputation relies on its imputation model and assumptions about missingness. The posted analyses provide alternative views but do not remove uncertainty caused by missing data.

The primary population description identifies a Full Analysis Set and says participants had to be randomized to a BI 690517 dose or matching placebo and have baseline UACR measurements. The registry text available for the additional eligibility criterion is incomplete; the remaining details are not reported in the registry record. The full analysis-population definitions, the number with missing primary endpoint data, and the reasons for missingness are not reported in the registry record for this account. The sample-size justification and any interim-analysis plan are also 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: adrenal insufficiency (2.8%), acute kidney injury (2.8%), cardiac failure congestive (1.4%), subcapsular renal haematoma (1.4%), sudden death (1.4%).

The registry reports serious adverse events and other adverse events by study group, including separate groups for the empagliflozin run-in and the treatment period. Those groupings should be kept distinct: run-in-period figures are not the same population or treatment exposure as the later dose-group figures. The counts are presented for participants at risk and participants affected, with deaths listed separately. They are descriptive counts, not adjusted comparisons between randomized groups.

In the treatment-period groups, serious adverse events are reported for participants assigned to each combination of background medication and BI 690517 dose or matching placebo. The registry’s serious-event terms include Adrenal insufficiency at 2.8%, Acute kidney injury at 2.8%, Cardiac failure congestive at 1.4%, Subcapsular renal haematoma at 1.4%, Sudden death at 1.4%, and Dengue fever at 1.4%. These listed percentages are registry figures for reported terms; the record excerpt does not supply enough context to treat them as comparative risks across all groups.

Death counts are reported by group, including deaths in both run-in groups and several treatment groups. The figures alone do not establish whether an event was caused by a study treatment. Interpretation would require details such as event timing, adjudication, exposure, and clinical context, which are not reported in the registry record. The safety counts should therefore be read descriptively rather than as proof of a treatment-related effect or absence of risk.

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

The posted primary results show modeled decreases in log-transformed UACR in all four primary groups, with the placebo-matching group’s confidence interval including zero. The MMRM contrasts estimate greater decreases relative to placebo matching BI 690517 for each BI 690517 dose. The 10 mg and 20 mg estimates are more negative than the 3 mg estimate, while the 10 mg estimate is numerically more negative than the 20 mg estimate. These are group-level findings over the stated treatment period, with uncertainty described by the reported confidence intervals.

These results do not establish that the same change occurred for every participant. A group mean can conceal variation in individual measurements, and its confidence interval is not a prediction interval for individual outcomes. A reported p-value also does not measure clinical importance. The registry record does not provide a threshold for interpreting the size of a change in log-UACR as clinically important.

The primary endpoint is a urine biomarker, not a direct measure of kidney filtration, kidney failure, survival, or quality of life. The trial’s treatment period and endpoint are limited to the reported short-term assessment. The results therefore do not demonstrate that any observed biomarker change persists or translates into longer-term clinical benefit. They also do not establish the comparative effects of BI 690517 doses against one another, since the listed primary contrasts compare each dose with placebo matching BI 690517.

The study includes people with and without type 2 diabetes and uses either empagliflozin or matching placebo as background medication. The primary result is reported for all patients, with background medication included in the model. It should not be assumed to apply identically to each subgroup; subgroup-specific primary estimates are not reported in the registry record. Generalization to people outside the study population or to different treatment durations is likewise not established by these results.

Safety findings are recorded as counts and listed terms. They do not, without fuller comparative context, determine whether an event was caused by study treatment. Taken together, the record supports a statistical description of short-term UACR changes and reported events, not medical advice or a conclusion about treatment for an individual.

14. Why This Trial Matters Statistically

This trial illustrates why endpoint scale and analysis method matter in clinical research. The primary outcome is log-transformed UACR, and the sponsor reports adjusted means from a repeated-measures model. Reading those estimates as raw UACR units, or as direct clinical-event effects, would change their meaning. The model-based contrast and its confidence interval are the relevant quantities for the stated primary comparison.

It also illustrates the difference between a mean-change outcome and a responder outcome. A mean summarizes the distribution on a continuous scale; a responder analysis applies a threshold and counts participants who meet it. These approaches can tell related but non-identical stories, especially when missing data are handled in different ways. The registry’s multiple responder analyses make clear that conclusions depend partly on the assumptions used for incomplete observations.

The dose groups and background-medication assignments make interpretation more involved than a single two-arm comparison. The primary analysis accounts for background medication, while the reported pairwise contrasts compare BI 690517 doses against matching placebo. The absence of a reported direct active-dose comparison means the numerical ordering of dose estimates should not be treated as proof that one active dose differs from another.

Finally, the record reports primary and secondary tests as well as dose-response model fits, but does not report a multiplicity plan. This is important for calibrating the strength of secondary findings. Clear separation between primary inference, secondary exploration, model fit, and safety description helps keep the conclusions aligned with what each analysis can support.

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 endpoint is a biomarker measured over a 14-week treatment period. The registry record summarized for this analysis does not report longer-term clinical kidney outcomes, so it cannot show whether short-term UACR changes predict outcomes for participants in this trial. Nor does the primary outcome directly measure filtration or kidney function, despite the trial title referring to kidney function.

The main reported primary estimates are for all patients, and the model includes background medication and randomization stratum. The record does not give subgroup-specific primary results for people with and without type 2 diabetes or for the two background-medication assignments. It also does not report direct statistical comparisons among the active BI 690517 doses. The dose estimates can be described side by side, but their numerical ordering does not establish a dose-to-dose difference.

The primary analysis relies on a repeated-measures model. The registry description identifies model terms and visits, but does not report the primary model’s missing-data assumption, sensitivity analyses for that assumption, or details of model diagnostics. The analysis-population definition is also incomplete in the available record description. These omissions limit assessment of how robust the primary inference is to incomplete follow-up and analysis choices.

The registry lists multiple secondary analyses and dose-response model fits but does not report a multiplicity procedure or testing hierarchy. Nominal secondary p-values consequently warrant caution. The record also does not report the trial’s sample-size justification, interim-analysis plan, or full protocol-level statistical plan. These are important details for evaluating which hypotheses were prespecified and how the study was sized.

Safety results are presented as descriptive counts by group, with serious-event terms and deaths. The available figures do not provide full information on event timing, exposure-adjusted rates, causality assessments, or comparative uncertainty. They therefore cannot alone establish differences in treatment-related risk.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study to Test Whether Different Doses of BI 690517 Alone or in Combination With Empagliflozin Improve Kidney Function in People With Chronic Kidney Disease — NCT05182840.
• Registry record last updated 2024-10-22; status last verified 2024-10.
• Results first posted 2024-08-29.
• 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

NCT05182840 is a completed phase 2 randomized, double-blind, placebo-controlled, parallel dose-group trial sponsored by Boehringer Ingelheim. It enrolled 714 participants with chronic kidney disease, including people with and without type 2 diabetes. Participants received BI 690517 at 3 mg, 10 mg, or 20 mg, or matching placebo, with empagliflozin or matching placebo as background medication.

The primary endpoint was change from treatment-period baseline in log-transformed first morning void UACR after 14 weeks. The registry reports adjusted mean changes and MMRM comparisons against placebo matching BI 690517. The dose-group contrasts were negative, with the 10 mg estimate numerically more negative than the 20 mg estimate. The record also includes responder analyses using different approaches to missing data and descriptive safety counts.

The results describe short-term group-level biomarker changes. Multiplicity handling, the primary missing-data assumption, the trial’s sample-size justification, and several details of the statistical plan are not reported in the registry record.