Efficacy and Safety of Empagliflozin (BI 10773) in Patients With Type 2 Diabetes and Renal Impairment: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy and Safety of Empagliflozin (BI 10773) in Patients With Type 2 Diabetes and Renal Impairment (NCT01164501), a randomized phase 3 study. The primary outcome is HbA1c Change From Baseline in Patients With Mild or Moderate Renal Impairment.

2. Clinical Question

3. Trial Design

This was a Phase III, randomized, double-blind, placebo-controlled, parallel-group study sponsored by Boehringer Ingelheim. The registry lists 741 participants enrolled and the study status as completed. The recorded start was 2010-07, and primary completion was 2012-07. The official study title describes once-daily BI 10773 at 10 mg and 25 mg, given as an add-on to pre-existing antidiabetic therapy over 52 weeks.

The stated study population had type 2 diabetes, renal impairment and insufficient glycaemic control. The primary endpoint is narrower: it is the HbA1c change in participants described as having mild or moderate renal impairment, measured from baseline to 24 weeks. The posted primary comparison is between the 25 mg group and placebo. Although the overall study included a 10 mg group, no 10 mg result is included in the posted primary comparison.

Randomization is intended to distribute known and unknown participant characteristics across treatment groups, while double-blinding is intended to reduce the influence of treatment knowledge on participant care and outcome assessment. These are general design principles; the record does not report the randomization procedure, how allocation was concealed, or which parties were blinded in practice. It identifies matching placebo tablets and a parallel-group model, but the detailed visit schedule and other operational features are not reported in the registry record.

The registry’s primary analysis population included randomized participants who received at least one dose and had a baseline HbA1c measurement. Values after antidiabetic rescue therapy were set to missing, and last observation carried forward was used for missing-value imputation. Further details of the original design, including the original sample-size justification and prespecified statistical plan, 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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full analysis set (FAS) which included all randomised patients, treated with at least one dose of trial medication, who had a baseline HbA1c value. Values after start of antidiabetic rescue therapy were set to missing and last observation carried forward (LOCF) was used for imputation of missing values.

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 are presented for treated patients only.

6. Primary Endpoint

The registry lists the primary outcome as HbA1c Change From Baseline in Patients With Mild or Moderate Renal Impairment, measured over Baseline and 24 weeks.

The primary endpoint is change in HbA1c from baseline to 24 weeks among participants with mild or moderate renal impairment. HbA1c is a continuous measure reported as a percentage of HbA1c. A change score describes how the measurement at follow-up differs from the participant’s baseline measurement; a negative change indicates a decrease from baseline. The posted group summaries are adjusted means, and the sponsor’s analysis reports the mean difference between the 25 mg group and placebo.

HbA1c is used in diabetes trials as a measure of glycaemic control over time. In statistical terms, a continuous endpoint preserves more information about the magnitude of individual measurements than a simple classification into categories. Comparing mean changes can therefore describe both the direction and average size of the measured difference between groups. This general explanation does not establish how the endpoint relates to outcomes beyond the measurements reported in this trial.

The endpoint is defined for the 24-week assessment, even though the official study title describes treatment over 52 weeks. The registry record does not explain why this endpoint time point was selected or provide a separate primary result at 52 weeks. The primary result should accordingly be interpreted as a comparison of mean HbA1c change at the specified 24-week time point, not as a result for every outcome or duration considered in the study.

7. Results Posted in the Registry

• Empa 25mg: mean -0.46 versus 0.05 on control; difference -0.51 (95% CI -0.62 to -0.40), standardized effect d = -0.76; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary comparison uses a continuous outcome: change in HbA1c from baseline to 24 weeks. For a continuous measure, a group mean gives the average observed value in that group, while a mean difference expresses the contrast between two group means. A negative mean difference in this setting indicates a lower mean change for Empa 25mg than for placebo. The reported contrast is -0.51 percentage of HbA1c.

What does ANCOVA contribute?

The sponsor reports an analysis of covariance, or ANCOVA, for the comparison. In general, ANCOVA compares group outcomes while adjusting for selected covariates. Adjustment may improve precision or account for baseline differences when the model is appropriately specified. The registry describes the group means as adjusted but does not list the covariates, their handling, or the full model specification. Those details are not reported in the registry record, so the exact adjustment cannot be reconstructed from the entry.

The sponsor’s analysis is described as a mean difference in final values. The primary endpoint description, by contrast, specifies change from baseline. Both are part of the posted analysis information, but the registry does not give enough model detail to explain fully how the ANCOVA formulation relates to the adjusted group summaries. The reported treatment contrast and interval remain the principal inferential results available in the record.

What does the confidence interval express?

In general, a confidence interval is a range of values compatible with the observed data under a particular statistical procedure and its assumptions. It conveys uncertainty about the estimated mean difference; it does not show the range of individual participants’ outcomes. The sponsor reports a 95% confidence interval from -0.62 to -0.39. Because this interval lies below zero, it indicates that the values in the reported interval all point toward a lower mean change in the 25 mg group than in the placebo group. The interval is an inferential summary, not a guarantee that the true value lies within the endpoints.

How should the p-values be read?

The sponsor reports p < 0.001 for its ANCOVA analysis. The simple unadjusted recalculation reports p < 0.001. In general, a p-value quantifies how unusual results at least as discrepant as those observed would be if a specified null hypothesis and statistical model were correct. It is not the probability that the null hypothesis is true, and it does not measure clinical importance. The two p-value statements use different thresholds and analyses; they should not be read as identical numerical results.

The unadjusted recalculation uses posted summary statistics and is not the sponsor’s ANCOVA model. Recalculations of this kind can help readers understand the approximate contrast implied by summary values, but they cannot recover adjustments or other features absent from those summaries. The registry marks no discrepancy between the posted analysis and the recalculation. The sponsor’s analysis is therefore the reported trial analysis, while the recalculation is a supplementary, unadjusted check.

What does the analysis population mean?

The record describes the full analysis set as all randomized participants treated with at least one dose of trial medication who had a baseline HbA1c value. This definition excludes people who did not meet one or more of those conditions. It is not identical to all enrolled participants, and the primary group sizes are specific to the posted endpoint analysis. The registry does not report a separate intention-to-treat analysis or a per-protocol analysis for this endpoint.

For values after the start of antidiabetic rescue therapy, the record says the values were set to missing. It also says last observation carried forward, or LOCF, was used to impute missing values. In general, LOCF substitutes a participant’s last observed value for a later missing measurement. This approach can be simple to implement, but it assumes that the last observed value is an adequate stand-in for the missing outcome. If outcomes would have changed after that last measurement, the substitution may not reflect the unobserved value. The registry does not report how many values were missing, how many participants received rescue therapy, or sensitivity analyses using alternative missing-data methods.

What assumptions remain unclear?

Statistical methods rely on assumptions about the data and the analysis model. For a mean comparison, relevant considerations can include the distribution of outcomes, the relationship between covariates and outcome in an adjusted model, and the handling of missing observations. These are general considerations, not claims about this trial. The registry does not report diagnostics, model covariates, a detailed prespecified analysis plan, or interim analyses. It also does not report a multiplicity adjustment for testing multiple outcomes, doses or subgroups.

Four outcome measures and four analyses are listed, but no secondary results are posted. Without the full testing plan, it is not possible to determine whether the primary comparison was interpreted within a larger family of formal hypotheses or whether any multiplicity procedure was planned. The primary result can be described on its own terms, while the broader inferential context remains unavailable in the registry record.

10. Confidence Intervals and Precision

The sponsor’s estimate of the adjusted mean difference is -0.51 percentage of HbA1c. Its 95% confidence interval extends from -0.62 to -0.39. The interval describes uncertainty around the estimated group contrast under the sponsor’s analysis; it does not describe the spread of individual participant outcomes. Because both endpoints are negative, the reported interval is consistent with a lower average change in the Empa 25mg group than in the placebo group.

In general, a confidence interval combines an estimate with information about its precision. A narrower interval indicates less statistical uncertainty about the estimated contrast than a wider interval, given the model and assumptions used. The interval is not a prediction interval for an individual participant, nor does it include every value that might occur in a replication. Its interpretation depends on the statistical procedure being appropriate for the data and analysis.

The registry also gives an unadjusted recalculation based on posted summary statistics. Its values are not identical in form to the sponsor’s adjusted ANCOVA result. The recalculation is a straightforward summary-based check, while the sponsor’s model is the trial analysis. The record marks no discrepancy between them. The sponsor’s interval is the clearest reported measure of precision for the primary analysis, and its endpoints are quoted at the precision used in the registry.

Although the interval is on one side of zero, that fact does not establish clinical importance. The registry does not report a threshold for a clinically meaningful difference, nor does it provide additional context that would determine whether the estimated magnitude matters for outcomes beyond this HbA1c endpoint. Precision around a statistical estimate and importance for decisions are related but distinct questions.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size figures are planning illustrations based on the observed effect, not the trial’s original design calculation. The original design assumptions and sample-size justification are not reported in the registry record. These planning entries give participants per arm, not total participants or total events, because this is a continuous-outcome comparison.

The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. Each planning scenario uses a stated fraction of the observed effect as the effect to detect. The entries for 100% of the observed effect are 28 participants per arm for 80% power and 37 participants per arm for 90% power. For 75% of the observed effect, the corresponding figures are 49 and 66 participants per arm. For 50% of the observed effect, they are 110 and 147 participants per arm.

In general, the 80% power column means the planned sample size is calculated to give an 80% chance of detecting the specified effect under the stated assumptions if that effect is present. The 90% column uses a higher target chance of detection for the same assumed effect, and therefore calls for a larger sample in each listed scenario. Smaller effects are harder to distinguish from random variation and require more participants under these assumptions.

These figures are not observed power. They do not describe the probability that the completed trial found a result, and they are not evidence that the original trial was designed using these assumptions. They are illustrations conditional on the stated planning model and the observed endpoint effect.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -0.510; pooled standard deviation σ = 0.673.
2. n = 2 × (1.96 + 0.842)² × 0.673² / -0.510² = 27.3, rounded up to 28 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -0.51                      # observed difference in means
sigma <- 0.672908369511678                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(1164501)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 28 (80% power), 37 (90% power)
 75% of observed effect: n per arm = 49 (80% power), 66 (90% power)
 50% of observed effect: n per arm = 110 (80% power), 147 (90% power)
power.t.test: 28.3 
Simulated power at n = 28 per arm: 0.795 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Analyses Reported by the Sponsor

The sponsor’s posted analysis is an ANCOVA comparison of Empa 25mg with placebo. It reports a mean difference of -0.51 percentage of HbA1c, a 95% confidence interval from -0.62 to -0.39, and p < 0.001. This is the trial’s reported inferential result for the primary endpoint.

A separate recalculation from the posted group summaries is unadjusted. It reports the same mean difference at the displayed precision and p < 0.001. Because this recalculation does not reproduce the sponsor’s ANCOVA adjustment, it is a check on the summary figures rather than an alternative reconstruction of the trial model. The registry marks no discrepancy between the two analyses.

The p-value thresholds are reported differently: the sponsor’s result is below < 0.001, while the unadjusted recalculation is below 0.001. These statements should be attributed to their respective analyses rather than treated as the same test statistic. The sponsor’s analysis remains the direct report of the trial’s analysis; the recalculation offers a simpler summary-based comparison. The registry does not provide the full ANCOVA specification, so a reader cannot reproduce all details of the sponsor’s result from the entry alone.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists four outcome measures and four analyses, but no secondary outcome results are posted. The primary result is for HbA1c change at 24 weeks in participants with mild or moderate renal impairment, comparing Empa 25mg with placebo. The record does not describe whether other outcomes, doses or subgroups were included in a formal family of hypothesis tests, or whether an adjustment for multiple comparisons was planned. A multiplicity plan is not reported in the registry record.

In general, testing several hypotheses can increase the chance of obtaining at least one apparently unusual result by chance. Statistical procedures can account for that risk, but the appropriate approach depends on which tests were planned and how they relate to one another. Since the registry does not report the complete testing strategy, it is not possible to characterize the primary p-value within a documented multiplicity framework. This does not change what the posted primary estimate says; it limits interpretation of that result in relation to unreported tests.

The analysis population is described as randomized participants who received at least one dose and had a baseline HbA1c value. Values after the start of antidiabetic rescue therapy were set to missing, and LOCF was used for missing values. The record does not give counts for missing measurements or rescue therapy, and it does not report sensitivity analyses for these choices. In general, missing-data methods require assumptions about outcomes that were not observed; LOCF assumes the last observed value can stand in for a later missing value. The extent to which that assumption matters cannot be assessed from the posted information.

The registry does not report the detailed statistical analysis plan, original sample-size justification, covariates in the ANCOVA, or interim analyses. It also does not provide an alternative analysis population or a full account of deviations from any planned analysis. These absences make it difficult to assess the complete decision process behind the reported result, while leaving the stated estimate, interval and analysis population available for interpretation.

15. Safety Results

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

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: acute coronary syndrome (1.0%), oesophageal varices haemorrhage (1.0%), chest pain (1.0%), cholecystitis acute (1.0%), osteoarthritis (1.0%).

The safety figures report participants at risk and participants affected by serious or other adverse events. In the placebo group, 319 were at risk for each reported event category; 44 were affected by serious adverse events and 207 by other adverse events. In the Empa 10mg group, 98 were at risk; 6 were affected by serious adverse events and 64 by other adverse events. In the Empa 25mg group, 321 were at risk; 40 were affected by serious adverse events and 187 by other adverse events.

These are counts, not comparative risk estimates. No percentages need to be inferred from them to read the registry entry, and the figures do not by themselves establish whether differences between groups are meaningful. The adverse-event terms listed as serious include Acute coronary syndrome, Oesophageal varices haemorrhage, Chest pain, Cholecystitis acute, Osteoarthritis and Benign prostatic hyperplasia, each reported as 1.0%. The record does not identify the group associated with each listed term in this summary.

Deaths are not reported in the registry record. The available figures should be read as the posted adverse-event counts, not as a complete account of every safety outcome or as evidence about causality. The registry does not provide the detailed definitions, follow-up context, or comparative statistical analysis needed for a fuller assessment.

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

The posted primary result is an adjusted mean difference in HbA1c change of -0.51 percentage of HbA1c for Empa 25mg compared with placebo at 24 weeks. The 95% confidence interval runs from -0.62 to -0.39, and the sponsor reports p < 0.001. This describes the average contrast observed in the analyzed population and the uncertainty reported around that contrast.

The result does not establish that every participant’s HbA1c changed in the same direction or by a similar amount. A group mean can conceal variation among individuals. Nor does the result establish effects on outcomes other than HbA1c, on participants outside the stated mild or moderate renal impairment group, or at time points other than the reported 24-week assessment. No secondary results are posted in the registry record.

The study title describes treatment over 52 weeks, but the primary endpoint result in the record is for baseline to 24 weeks. The 52-week description should not be taken as evidence that this specific result measures a 52-week effect. The primary comparison also addresses Empa 25mg and placebo; the record’s primary results do not provide an estimate for the 10 mg group.

Statistical evidence against a zero difference is distinct from clinical importance. The registry does not state a threshold for a clinically meaningful HbA1c difference or provide a broader outcome-based assessment. The posted estimate and interval therefore support a statistical description of this endpoint, but they cannot alone answer how important the measured contrast is in a wider clinical context.

The analysis used an ANCOVA method and adjusted means, but the covariates and full model specification are not reported. The handling of rescue therapy and missing values is described as setting post-rescue values to missing and using LOCF for missing values; the number of affected participants and sensitivity analyses are not reported. These features constrain interpretation of how robust the estimate is to analysis choices. Nothing in the posted primary result, on its own, establishes the overall balance of efficacy and safety.

17. Why This Trial Matters Statistically

This trial illustrates why an endpoint, its analysis population and its assessment time must be read together. The study title describes a 52-week add-on treatment study, while the primary posted endpoint is HbA1c change at 24 weeks in participants with mild or moderate renal impairment. Treating those descriptions as interchangeable would blur the specific question answered by the reported estimate.

The record also distinguishes the sponsor’s adjusted ANCOVA result from a simple unadjusted recalculation based on summary statistics. Both report a mean difference of -0.51 at the displayed precision, but they are not the same analysis. The sponsor reports a 95% confidence interval from -0.62 to -0.39 and p < 0.001; the recalculation reports p < 0.001. This is a useful example of how a summary-based check can be informative while remaining separate from the model used for the trial’s reported analysis.

Another statistical consideration is that the analysis set is defined by randomization, treatment received and baseline measurement, with post-rescue values set to missing and LOCF used for imputation. These decisions influence which measurements contribute and how unavailable values are represented. The record does not report the missing-data counts or alternative analyses, so readers cannot judge their impact from the entry alone.

Finally, the registry offers planning illustrations based on the observed effect, not the trial’s original design assumptions. Keeping those calculations separate from the completed trial prevents a planning exercise from being mistaken for evidence about the study’s design or observed power. Together, these distinctions help keep the statistical reading aligned with what the registry actually reports.

18. 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 result concerns a specific subgroup described as having mild or moderate renal impairment, a specific dose comparison of Empa 25mg with placebo, and a 24-week HbA1c change endpoint. It should not be generalized from the registry entry to more severe renal impairment, the 10 mg dose, other outcomes, or the full 52-week period described in the official title. Separate results for these questions are not reported in the registry record.

The primary analysis population required randomization, at least one dose of trial medication and a baseline HbA1c value. The posted group sizes are 282 for placebo and 284 for Empa 25mg, while overall enrollment was 741. The entry does not provide a participant flow account explaining the relationship among enrollment, treatment groups and the endpoint analysis population.

Missingness and rescue therapy are relevant to interpretation. The record says post-rescue values were set to missing and LOCF was used for missing values, but it does not report the number of affected participants, the amount of missing data, or sensitivity analyses. The full ANCOVA specification and covariates are also not reported, limiting independent assessment of the adjusted estimate.

The registry does not report the original sample-size rationale, a full multiplicity strategy, interim analyses, or secondary results. The safety summary gives event counts but does not report deaths, and the serious adverse-event terms are not assigned to individual groups in the listed summary. These reporting limits do not negate the posted primary estimate, but they restrict broader conclusions about the trial’s design, robustness and overall outcomes.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Efficacy and Safety of Empagliflozin (BI 10773) in Patients With Type 2 Diabetes and Renal Impairment — NCT01164501.
• Registry record last updated 2014-06-16; status last verified 2014-05.
• Results first posted 2014-06-16.
• 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.

22. Record Summary

NCT01164501 is a completed Phase III, randomized, double-blind, placebo-controlled, parallel-group study sponsored by Boehringer Ingelheim. The registry lists 741 participants enrolled. It describes once-daily BI 10773, also named empagliflozin, at 10 mg and 25 mg as add-on therapy to existing antidiabetic treatment, with a 52-week study duration in the official title.

The primary result is HbA1c change from baseline to 24 weeks among participants with mild or moderate renal impairment. For Empa 25mg versus placebo, the sponsor reports an adjusted mean difference of -0.51 percentage of HbA1c, with a 95% confidence interval from -0.62 to -0.39 and p < 0.001. The entry describes the analysis population and LOCF handling of missing values, but does not report the full model specification, original sample-size justification, multiplicity strategy or secondary results.