Efficacy and Safety of Empagliflozin (BI 10773) Versus Placebo and Sitagliptin Over 24 Weeks in Patients With Type 2 Diabetes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy and Safety of Empagliflozin (BI 10773) Versus Placebo and Sitagliptin Over 24 Weeks in Patients With Type 2 Diabetes (NCT01177813), a randomized phase 3 study. The primary outcome is Change From Baseline in Glycosylated Haemoglobin (HbA1c) After 24 Weeks.

2. Clinical Question

The trial was designed to investigate efficacy, safety and tolerability over 24 weeks in patients whose glycaemic control remained insufficient with diet and exercise. The primary endpoint was change in glycosylated haemoglobin, or HbA1c, from baseline to the end of the treatment period. HbA1c is a continuous measure, so the reported mean changes describe the average change within each group; between-group mean differences describe how those average changes compare.

The randomized comparisons are the main basis for interpreting the primary outcome. The open-label group has a different treatment and analysis context: it was descriptive, and its results should not be treated as a randomized comparison of the same kind. The registry record does not report a distinct primary question for the open-label group beyond estimating efficacy and safety in patients with very poor glycaemic control.

3. Trial Design

This was a phase 3, randomized, double-blind, parallel-group study sponsored by Boehringer Ingelheim. The registry record lists total enrollment of 986 and describes a 24-week treatment period. Participants were assigned to study groups, and the randomized treatment groups were masked. The parallel-group design means participants remained in their assigned groups rather than crossing from one treatment to another during the comparison.

The listed randomized groups included placebo, empagliflozin 10 mg, empagliflozin 25 mg and sitagliptin 100 mg. The outcome data also include an empagliflozin 25 mg open-label group. The registry describes this group separately and states that its statistics are descriptive. It should therefore be distinguished from the blinded randomized groups when reading between-group estimates.

The primary outcome was change from baseline in HbA1c after 24 weeks, with the timeframe specified as baseline and day 169. Baseline was defined as the last observation before the start of randomized treatment, or before open-label treatment for the open-label group. The primary population is described as the FAS and open-label set, with last observation carried forward used for both sets.

The registry record lists body weight and systolic and diastolic blood pressure as secondary outcomes. It reports 4 outcome measures and 8 analyses. The original sample-size justification, planned interim analyses, full allocation details, and detailed masking procedures are not reported in the registry record. The posted planning illustrations are not the trial’s original design calculations.

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: FAS and open-label set, last observation carried forward (LOCF) was used as the imputation rule for both sets

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: Full analysis set (FAS) including all randomised and treated patients who had a baseline HbA1c value (non-open label groups) and Open-label set including all patients entered in the empagliflozin 25 mg open-label treatment arm.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Glycosylated Haemoglobin (HbA1c) After 24 Weeks, measured over Baseline and day 169.

The primary endpoint measures the change in glycosylated haemoglobin (HbA1c) from baseline after 24 weeks. HbA1c is a continuous measure reported in percent of HbA1c. The registry defines baseline as the last observation before randomized treatment starts, or before open-label treatment for participants in the open-label group. The stated timeframe is baseline and day 169.

As a general statistical explanation, a continuous endpoint preserves information about the magnitude of change rather than classifying participants only into categories. The mean change summarizes the average change in a group. Comparing mean changes across groups can therefore describe the average separation between assigned treatments on this measure.

HbA1c is used in this trial as the measure of glycaemic control, matching the registry’s stated aim of studying treatment in patients with insufficient glycaemic control. The endpoint alone does not describe every aspect of diabetes or establish the importance of a given change for an individual patient. Nor does an average group change show how much participants’ individual responses varied.

The record distinguishes unadjusted measured values from adjusted statistical analyses. It reports adjusted analyses for the randomized comparisons and descriptive statistics for the open-label group. Last observation carried forward was used as the imputation rule for the FAS and open-label set.

7. Results Posted in the Registry

• Empagliflozin10 mg: mean -0.66 versus 0.06 on control; difference -0.72 (95% CI -0.87 to -0.57), standardized effect d = -0.87; z-test p = <0.001.
• Empagliflozin 25 mg: mean -0.77 versus 0.06 on control; difference -0.83 (95% CI -0.98 to -0.68), standardized effect d = -1.00; z-test p = <0.001.
• Sitagliptin 100 mg: mean -0.65 versus 0.06 on control; difference -0.71 (95% CI -0.85 to -0.57), standardized effect d = -0.95; z-test p = <0.001.
• Empagliflozin 25 mg OL: mean -3.10 versus 0.06 on control; difference -3.16 (95% CI -3.60 to -2.72), standardized effect d = -2.52; z-test p = <0.001.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Change From Baseline to Week 24 in Body Weight

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Empagliflozin10 mg: mean -2.26 versus -0.33 on control; difference -1.93 (95% CI -2.40 to -1.46), standardized effect d = -0.75; z-test p = <0.001.
  • Empagliflozin 25 mg: mean -2.48 versus -0.33 on control; difference -2.15 (95% CI -2.61 to -1.69), standardized effect d = -0.86; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Empagliflozin10 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Empagliflozin 25 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The posted secondary outcomes cover change in body weight and change in systolic and diastolic blood pressure. Across these outcomes, the pattern is not identical for every measure or treatment comparison. The body-weight summaries show decreases in the empagliflozin groups, while the sitagliptin group’s posted mean change is in the opposite direction from those decreases. The blood-pressure endpoint includes both systolic and diastolic measures, so it adds a different aspect of the clinical profile than either HbA1c or body weight.

These outcomes are related to the primary result in that they describe changes over the same treatment period and provide additional context about the groups. They do not replace the primary HbA1c comparison, and they do not independently establish the overall effect of treatment. Secondary endpoints can be useful for describing a broader set of outcomes, but their interpretation depends on the analysis plan, measurement quality, missing-data handling and how many comparisons were made.

The recalculated secondary comparisons are unadjusted summaries. The registry also reports sponsor analyses for some comparisons, but the record does not provide a complete multiplicity procedure for interpreting all secondary tests together. When many endpoints or comparisons are examined, the chance of at least one apparently notable result can increase. As a general statistical explanation, adjustment for multiplicity or a clearly prespecified hierarchy can help control how such results are interpreted. The relevant plan is not reported in the registry record.

Accordingly, the secondary results should be read as supportive descriptions rather than as independent confirmation of the primary outcome. The aggregate record does not establish whether these endpoints were designated as confirmatory, how their testing order was set, or whether all relevant comparisons were included in a multiplicity adjustment. Their pattern adds context, but it does not resolve those uncertainties.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Change From Baseline to Week 24 in Body Weight
• Change From Baseline to Week 24 in Systolic and Diastolic Blood Pressure (SBP and DBP)

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous, and the reported summaries include group means and standard errors. The registry describes measured values as unadjusted, while the statistical analyses are adjusted. Keeping those two kinds of quantities separate is important: an unadjusted summary describes the observed group values, whereas an adjusted analysis estimates a comparison after accounting for factors in the model. The registry names ANCOVA as the sponsor’s method for the primary comparisons.

What does ANCOVA contribute?

As a general statistical explanation, analysis of covariance (ANCOVA) compares group outcomes while adjusting for covariates included in its model. In a clinical trial, adjustment can account for baseline values or other prespecified variables, if they are part of the model. The registry record identifies ANCOVA but does not report the complete model specification, the covariates used, or the detailed estimation procedure. Those aspects cannot be inferred from the method name alone.

The sponsor reports adjusted mean differences for empagliflozin 10 mg and 25 mg versus placebo. The primary comparisons are reported with 97.5% confidence intervals and p-values. The simple recalculations, by contrast, compare the posted unadjusted group means and use the summary statistics. They are useful as a check on the direction and approximate size of the unadjusted contrast, but they do not reproduce the adjusted ANCOVA model. The registry record reports no discrepancy flag for these results.

How should a mean difference be read?

A mean difference expresses the difference between the average outcome in two groups, using the order of groups specified for the comparison. Here, the recalculated comparisons use placebo as the reference and report negative values when the treatment-group mean change is lower than the placebo-group mean change. For example, the recalculated mean difference for empagliflozin 10 mg versus placebo is -0.72. The sponsor’s adjusted estimate for that comparison is -0.74. These are distinct estimates from different analyses, not competing versions of the same calculation.

As a general statistical explanation, a confidence interval describes a range of values compatible with the data under the analysis and its assumptions. An interval entirely on one side of zero is consistent with a nonzero mean difference under that model. The interval’s width also conveys uncertainty: a narrower interval indicates greater precision than a wider interval, assuming comparable methods and scale. A p-value summarizes how incompatible the observed data are with a specified null hypothesis under the model; it is not the probability that the null hypothesis is true or a measure of clinical importance.

What do standard errors and standardized effects describe?

The reported arm summaries give standard errors for the mean changes. As a general statistical explanation, a standard error describes the sampling uncertainty of an estimated mean, rather than the spread of individual participant outcomes. The recalculation inputs also include pooled standard deviations and standardized effect estimates. Standardization expresses a mean difference relative to a measure of variability, which can help describe the difference on a scale independent of the original units. These standardized estimates are derived from the posted summary statistics and are not the sponsor’s adjusted analysis.

Standard errors, standard deviations and confidence intervals have different meanings. A standard deviation describes participant-level variability; a standard error describes uncertainty in an estimated mean. The registry provides mean changes and standard errors for the arms, while the recalculation uses the reported summaries to derive comparison quantities. The registry record does not report individual participant outcome distributions, so the aggregate summaries cannot show the full shape or range of individual responses.

What does the open-label result represent?

The open-label arm’s primary statistics are descriptive according to the registry. Its mean change and its recalculated contrast with placebo should not be interpreted as evidence from the same randomized, blinded comparison as the other groups. The open-label group is listed with 87 participants, while the randomized placebo group is listed with 228 for the primary outcome. The design and analysis context differ, so a large numerical contrast does not on its own establish a causal treatment effect.

The registry record also does not report the full ANCOVA specification, a detailed missing-data sensitivity analysis, or a full multiplicity plan. These omissions limit independent assessment of model choices and robustness. The appropriate reading is therefore to treat the sponsor’s adjusted analyses as the reported primary inferential results, use unadjusted recalculations as descriptive checks, and interpret all estimates alongside their uncertainty and analysis context.

11. Confidence Intervals and Precision

The sponsor’s adjusted primary comparisons are reported with 97.5% confidence intervals. For empagliflozin 10 mg versus placebo, the adjusted mean difference is -0.74, with an interval from -0.90 to -0.57. For empagliflozin 25 mg versus placebo, it is -0.85, with an interval from -1.01 to -0.69. Both intervals lie below zero, indicating that the reported adjusted estimates are separated from zero under the sponsor’s analysis. The intervals communicate uncertainty about the size of the mean difference; they do not identify a single exact effect.

The recalculated unadjusted comparisons have intervals from -0.873 to -0.567 for empagliflozin 10 mg, from -0.983 to -0.677 for empagliflozin 25 mg, and from -0.849 to -0.571 for sitagliptin. These are based on the posted unadjusted summaries and should not be confused with the sponsor’s adjusted confidence intervals. The recalculated intervals are reported with more decimal places in the underlying statistics; their precision should not be mistaken for greater certainty.

As a general statistical explanation, confidence intervals reflect both the estimated effect and its uncertainty under a particular analysis. Their width depends on the variability and amount of information available, as well as the model and confidence level. A confidence interval does not show the range of outcomes for individual participants, and it does not establish clinical importance. The registry record does not report a clinically important threshold for HbA1c change. Precision is therefore best read as statistical uncertainty about average group differences, not as a judgment about the value of those differences for each person.

12. 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 assumptions. For these continuous outcomes, the figures are participants per arm, not total events. The stated illustration assumes a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect equal to the stated fraction of the observed effect. The trial’s own sample-size justification is not reported in the registry record.

The 80% power column gives the estimated participants per arm for a planning scenario with an 80% chance of detecting the specified effect under those assumptions, if that effect and the assumptions were correct. The 90% power column gives the corresponding participants per arm for a 90% chance. These are design calculations for hypothetical future studies, not descriptions of the probability that this completed study found a result.

The rows planning for 100%, 75% and 50% of the observed effect illustrate how the assumed effect size changes the required sample size. A smaller effect is harder to distinguish from random variation, so detecting it generally requires more participants. The listed counts are rounded planning outputs, not observed enrollment targets or evidence about the trial’s actual power. This is not observed power.

Interpret each comparison’s planning figures in the context of its own observed estimate and variability. The open-label comparison has a different analysis context from the randomized comparisons, so its planning illustration should not be treated as an equivalent randomized design recommendation. The original protocol assumptions, including any allowance for attrition or other design considerations, are not reported in the registry record.

13. 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 = -0.720; pooled standard deviation σ = 0.829.
2. n = 2 × (1.96 + 0.842)² × 0.829² / -0.720² = 20.8, rounded up to 21 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 <- -0.72                      # observed difference in means
sigma <- 0.828944710661292                     # 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(1177813)
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 = 21 (80% power), 28 (90% power)
 75% of observed effect: n per arm = 37 (80% power), 50 (90% power)
 50% of observed effect: n per arm = 84 (80% power), 112 (90% power)
power.t.test: 21.8 
Simulated power at n = 21 per arm: 0.782 (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.

14. Analyses Reported by the Sponsor

The sponsor’s own primary analyses are adjusted ANCOVA comparisons against placebo. For empagliflozin 10 mg, the reported adjusted mean difference is -0.74, with a 97.5% confidence interval from -0.90 to -0.57 and p < 0.001. For empagliflozin 25 mg, the adjusted mean difference is -0.85, with a 97.5% confidence interval from -1.01 to -0.69 and p < 0.001.

The recalculated results use the posted unadjusted arm means and standard errors. They yield mean differences of -0.72 for empagliflozin 10 mg and -0.83 for empagliflozin 25 mg, with recalculated p-values of <0.001 for both. The recalculated estimates are close to, but not identical to, the sponsor’s adjusted estimates. The registry explicitly distinguishes unadjusted measured values from adjusted statistical analyses, so the two sets of estimates should not be treated as the same calculation.

The registry does not flag a discrepancy between the posted summaries and the sponsor’s analyses. The sponsor’s adjusted results are the trial’s reported inferential analyses; the simple recalculations are checks based on aggregate unadjusted summaries. The registry record does not provide the complete ANCOVA model details needed to reproduce the sponsor’s estimates independently.

15. Multiplicity, Missing Data and Analysis Populations

The registry record reports a primary outcome and secondary outcomes, along with multiple analyses. It gives sponsor-reported ANCOVA results for the randomized primary comparisons against placebo and reports secondary analyses for body weight. However, the multiplicity strategy is not reported in the registry record. It is therefore not possible to determine from this record whether the sponsor controlled the overall type I error across primary comparisons, secondary endpoints, doses, or other analyses, or whether a prespecified testing hierarchy was used.

As a general statistical explanation, testing multiple hypotheses creates more opportunities for a result to appear statistically notable by chance. Procedures such as a prespecified hierarchy or multiplicity adjustment can address this issue, but the record does not identify such a procedure. The p-values should therefore be interpreted in the context of the reported comparisons rather than as a complete account of evidence across all outcomes.

The primary population is described as the FAS and open-label set. The registry record does not expand the FAS acronym or define the exact inclusion rules for that analysis population. It states that last observation carried forward was used as the imputation rule for both the FAS and open-label set. As a general statistical explanation, LOCF substitutes the last observed value when a later measurement is missing. This method can make assumptions about what would have happened after the last observation; the registry record does not report sensitivity analyses that assess the effect of alternative missing-data assumptions.

The record does not report the amount or pattern of missing primary outcome data, reasons for missingness, or the exact number of participants contributing to each adjusted analysis. It also does not report a detailed analysis plan or the covariates in the ANCOVA model. These omissions limit assessment of the robustness of the adjusted findings and the extent to which missing data or modeling decisions could influence them. They do not invalidate the posted results, but they restrict what can be independently verified from the registry record.

16. 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: diabetic neuropathy (1.1%), angina unstable (1.1%), soft tissue injury (1.1%), lung neoplasm malignant (0.4%), metastases to central nervous system (0.4%).

The serious adverse event figures give, for each group, the number of participants at risk and the number affected by one or more serious adverse events. The registry lists 6 affected among 229 at risk in placebo, 8 among 224 in the empagliflozin 10 mg group, 5 among 223 in the empagliflozin 25 mg group, 6 among 223 in the sitagliptin 100 mg group, and 3 among 87 in the open-label empagliflozin 25 mg group. These are counts as recorded, not a statistical comparison of treatment risks.

The registry lists diabetic neuropathy, unstable angina and soft tissue injury at 1.1% each among the leading serious-event terms. Lung neoplasm malignant, metastases to central nervous system and goitre are each listed at 0.4%. Those terms describe reported events; the record does not establish that study treatment caused them. Death counts are not reported in the registry record.

As a general statistical explanation, serious adverse event counts are one part of safety reporting. Differences in counts alone do not establish that one group is safer or riskier, particularly without an analysis plan, detailed event timing, participant-level information and adequate follow-up context. The open-label group should also be considered separately from the randomized blinded groups.

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

The posted results describe average changes in HbA1c over the recorded 24-week treatment period. The sponsor’s adjusted comparisons against placebo are negative, with reported confidence intervals below zero and p-values below < 0.001 for both empagliflozin doses. The unadjusted recalculations show a similar direction. These findings concern group averages under the analyses reported in the registry; they do not predict the response of an individual participant.

The results do not, by themselves, establish the clinical importance of the observed differences. The registry record does not report a threshold for a clinically important HbA1c change, and statistical evidence against a null comparison is not the same as evidence about the value of an effect to a patient. The record also does not establish whether the observed changes persist beyond the stated treatment period or how they relate to outcomes not reported as endpoints.

The randomized, double-blind design supports interpretation of the active treatment comparisons as randomized group comparisons, but the available registry information does not include every detail needed to assess the analysis independently. The sponsor’s estimates are adjusted, while the recalculations use unadjusted summaries; the open-label group is descriptive and is not equivalent to the randomized comparisons. The use of LOCF and the absence of a reported multiplicity plan are additional considerations.

Safety counts should also be interpreted cautiously. The serious adverse event figures summarize participants affected in each group, but they do not establish causation or provide a complete account of all possible safety outcomes. Taken together, the registry record supports a description of the reported results and their statistical uncertainty, not individual treatment recommendations or conclusions beyond the measured outcomes and timeframe.

18. Why This Trial Matters Statistically

This trial is statistically useful because it combines a randomized, double-blind, parallel-group comparison with a continuous primary outcome and multiple active treatment groups. The posted summaries allow readers to distinguish average within-group change from between-group differences, and the sponsor’s ANCOVA results provide adjusted comparisons against placebo. The registry also makes clear that measured values and adjusted analyses are not the same quantities.

The recalculations illustrate how aggregate summaries can be used to check the approximate unadjusted contrast and its uncertainty, while also showing why they should not replace the sponsor’s model-based analysis. The differences between the adjusted and unadjusted estimates are modest in the posted results, but the methods answer different questions and use different information. Reading both together helps clarify what can and cannot be reproduced from summary statistics alone.

The record also demonstrates why design context matters. The open-label group is descriptive, not a parallel randomized comparison of the same inferential type. Secondary outcomes add information beyond HbA1c, but interpretation depends on multiplicity handling, which is not reported. The sample-size illustrations can help readers understand how assumed effect size and desired power affect planning, while remaining distinct from the original trial design and from observed power.

Finally, the available record makes several limits visible: the full ANCOVA model, original sample-size justification, multiplicity strategy and detailed missing-data sensitivity analyses are not reported. Those gaps are relevant to evidence appraisal, without erasing the information that is available about the trial’s design, outcomes and posted estimates.

19. 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.

Several limitations are specific to the registry information available for this trial. The sponsor’s primary comparisons are identified as ANCOVA, but the covariates and full model specification are not reported in the registry record. This prevents complete reconstruction of the adjusted estimates from the record alone. The sponsor’s confidence intervals and p-values can be described as reported, but their precise model inputs cannot be independently checked using the posted aggregate summaries.

The primary outcome description distinguishes unadjusted measured values from adjusted analyses, and the recalculations are based on the unadjusted summaries. Therefore, recalculated estimates should not be treated as reproductions of the sponsor’s adjusted analyses. The registry does not report a discrepancy, but the analysis approaches differ.

The record states that LOCF was used for the FAS and open-label set, but does not report the amount or pattern of missing data, reasons for missingness, or sensitivity analyses using alternative assumptions. The FAS is not fully defined in the record. The multiplicity strategy across the primary and secondary analyses is also not reported.

The open-label group is described as descriptive and differs from the randomized blinded groups. Its result cannot be interpreted as if it were a randomized comparison. The record also does not report the trial’s original sample-size justification, detailed analysis plan, or interim analysis information. Serious adverse event counts are available, but death counts are not reported in the registry record. These limits constrain interpretation of precision, robustness and safety without changing the values that the record reports.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Efficacy and Safety of Empagliflozin (BI 10773) Versus Placebo and Sitagliptin Over 24 Weeks in Patients With Type 2 Diabetes — NCT01177813.
• 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.

23. Record Summary

The registry describes a completed phase 3, randomized, double-blind, parallel-group study of empagliflozin, placebo and sitagliptin over 24 weeks in patients with type 2 diabetes. Total enrollment was 986. The primary outcome was change from baseline in HbA1c, with the primary analysis population described as the FAS and open-label set and LOCF used for both.

The sponsor reported adjusted ANCOVA comparisons of both empagliflozin doses against placebo, with 97.5% confidence intervals and p-values below < 0.001. Unadjusted recalculations from the posted group summaries give similar-direction estimates but are not the same analyses. Body weight and blood pressure were listed as secondary outcomes. The registry reports serious adverse event counts by group, while death counts are not reported in the registry record.

The original sample-size justification, full model specification, multiplicity plan and missing-data sensitivity analyses are not reported in the registry record. The open-label statistics are descriptive.