BI 10773 (Empagliflozin) Cardiovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients (EMPA-REG OUTCOME).: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for BI 10773 (Empagliflozin) Cardiovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients (EMPA-REG OUTCOME). (NCT01131676), a randomized phase 3 study. The primary outcome is Time to the First Occurrence of Any of the Following Adjudicated Components of the Primary Composite Endpoint (3-point MACE): CV Death (Including Fatal Stroke and Fatal MI), Non-fatal MI (Excluding Silent MI), and Non-fatal Stroke..

2. Clinical Question

3. Trial Design

EMPA-REG OUTCOME was a phase III, multicentre, international, randomized, double-blind, parallel-group study. The registry describes the study as a cardiovascular safety study of BI 10773, the development name for empagliflozin, compared with usual care in people with type 2 diabetes and increased cardiovascular risk. The outcome groups identify placebo as the comparator. The study was completed, and the registry lists enrollment as 7064.

The treatment arms were empagliflozin 10 mg, empagliflozin 25 mg, and placebo. Each empagliflozin dose was administered orally once daily; placebo tablets matched BI 10773. The primary endpoint results list 2333 participants in the placebo group, 2345 in the 10 mg group, and 2342 in the 25 mg group. A pooled empagliflozin group is also reported for the primary endpoint.

The primary outcome was assessed from randomisation to each participant’s end of observation, up to 4.6 years. The endpoint was time to the first occurrence of an adjudicated component of 3-point MACE. The registry reports additional outcomes, including a broader cardiovascular composite, silent myocardial infarction, and hospitalised heart failure. The exact protocol, allocation concealment procedures, treatment adherence, follow-up completeness, and original statistical analysis plan are not reported in the registry record.

Randomization and masking are design features intended to reduce bias in treatment comparisons. In general, random assignment helps balance prognostic factors across groups on average, while masking can limit differences in care or outcome assessment related to knowledge of treatment. Those design features do not remove uncertainty from the estimates, and the registry record does not provide enough detail to assess every aspect of trial conduct.

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: Randomisation stratified by:BMI at randomisation(\<30/≥30 kg/m2),HbA1c at screening (\<8.5%/ ≥8.5%);geographical region(North America including Australia and New Zealand,Latin America,Europe,Africa,Asia);renal function at screening (normal:eGFR ≥90 mL/min, mild impairment:60 mL/min ≤ eGFR ≤89 mL/min, moderate impairment:30 mL/min ≤ eGFR≤59 mL/min).

Discontinuation From Treatment

Reasons for not completing this period:

Discontinuation From Study (Treated Set)

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: TS

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: Treated set (TS): Consists of all randomised patients who received at least 1 dose of study medication. Patients were assigned to treatment groups according to the randomisation scheme.

6. Primary Endpoint

The registry lists the primary outcome as Time to the First Occurrence of Any of the Following Adjudicated Components of the Primary Composite Endpoint (3-point MACE): CV Death (Including Fatal Stroke and Fatal MI), Non-fatal MI (Excluding Silent MI), and Non-fatal Stroke., measured over From randomisation to individual end of observation, up to 4.6 years.

The primary endpoint is time to the first adjudicated occurrence of cardiovascular death, non-fatal myocardial infarction, or non-fatal stroke. Cardiovascular death includes fatal stroke and fatal myocardial infarction. Non-fatal myocardial infarction excludes silent myocardial infarction. This combination is commonly called 3-point MACE in the registry description.

A time-to-first-event composite brings several clinically serious outcomes into one endpoint. It allows an analysis to consider both whether a participant experiences an event and when the first event occurs during observation. Using a first event also means that later events in a participant do not add further events to this endpoint. General: a composite can increase the number of observed outcomes compared with studying a single component, but it can also combine events that differ in frequency and clinical meaning. The registry does not report component-specific primary endpoint results in the record.

The registry gives the observation timeframe as from randomisation to individual end of observation, up to 4.6 years, and reports the percentage of participants with an event. It also reports a Cox proportional hazards analysis for the pooled empagliflozin group versus placebo. Those are different summaries: a percentage describes the proportion with an event over the recorded observation, while a hazard ratio summarizes the relative event rate over time under the model. They should not be treated as interchangeable measures.

Adjudication indicates that events were assessed according to specified criteria, but the adjudication procedures and their operational details are not reported in the registry record. The endpoint’s interpretation therefore rests on the stated event definitions and the analyses actually reported.

7. Results Posted in the Registry

• Empagliflozin 10 mg: 10.4% (244/2345) versus 12.1% (282/2333) on control; risk difference -1.7 percentage points (95% CI -3.5 to 0.1), risk ratio 0.86 (0.73–1.01); unadjusted two-proportion z-test p = 0.069.
• Empagliflozin 25 mg: 10.5% (246/2342) versus 12.1% (282/2333) on control; risk difference -1.6 percentage points (95% CI -3.4 to 0.2), risk ratio 0.87 (0.74–1.02); unadjusted two-proportion z-test p = 0.087.

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.

Percentage of Participants With the Composite of All Events Adjudicated (4-point MACE): CV Death (Including Fatal Stroke and Fatal MI), Non-fatal MI (Excluding Silent MI), Non-fatal Stroke and Hospitalization for Unstable Angina Pectoris

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Empagliflozin 10 mg: 12.8% (300/2345) versus 14.3% (334/2333) on control; risk difference -1.5 percentage points (95% CI -3.5 to 0.4), risk ratio 0.89 (0.77–1.03); unadjusted two-proportion z-test p = 0.128.
  • Empagliflozin 25 mg: 12.8% (300/2342) versus 14.3% (334/2333) on control; risk difference -1.5 percentage points (95% CI -3.5 to 0.5), risk ratio 0.89 (0.77–1.03); unadjusted two-proportion z-test p = 0.132.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Empagliflozin 10 mg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Empagliflozin 25 mg and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants With Silent MI

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Empagliflozin 10 mg: 1.6% (19/1174) versus 1.2% (15/1211) on control; risk difference 0.4 percentage points (95% CI -0.6 to 1.3), risk ratio 1.31 (0.67–2.56); unadjusted two-proportion z-test p = 0.434.
  • Empagliflozin 25 mg: 1.6% (19/1204) versus 1.2% (15/1211) on control; risk difference 0.3 percentage points (95% CI -0.6 to 1.3), risk ratio 1.27 (0.65–2.50); unadjusted two-proportion z-test p = 0.479.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Empagliflozin 10 mg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Empagliflozin 25 mg and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants With Heart Failure Requiring Hospitalisation (Adjudicated)

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Empagliflozin 10 mg: 2.6% (61/2345) versus 4.1% (96/2333) on control; risk difference -1.5 percentage points (95% CI -2.5 to -0.5), risk ratio 0.63 (0.46–0.87); unadjusted two-proportion z-test p = 0.004.
  • Empagliflozin 25 mg: 2.8% (66/2342) versus 4.1% (96/2333) on control; risk difference -1.3 percentage points (95% CI -2.3 to -0.2), risk ratio 0.68 (0.50–0.93); unadjusted two-proportion z-test p = 0.015.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Empagliflozin 10 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 proportion with the event 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 secondary outcomes broaden the description beyond the primary 3-point MACE endpoint. The registry lists a 4-point cardiovascular composite that adds hospitalisation for unstable angina, silent myocardial infarction, hospitalised heart failure, and albuminuria-related outcomes. The record also describes a composite microvascular outcome. These endpoints capture different events and are not interchangeable with the primary outcome.

Across the secondary results with detailed figures, the pattern is not uniform: the broader cardiovascular composite and hospitalised heart failure have lower posted percentages in the empagliflozin groups than in placebo, while silent myocardial infarction has higher posted percentages. This variation cautions against treating the secondary results as one consistent measure of a single treatment effect. The associated estimates and intervals should be assessed separately, with attention to how many outcomes were examined.

Secondary results can provide context for the primary finding, but they do not independently confirm it simply because several estimates point in a similar direction. The primary endpoint was designated as primary in the registry; the secondary endpoints answer additional questions. The registry record does not report a complete multiplicity adjustment strategy. Without knowing whether and how the family of secondary tests was controlled, unadjusted secondary p-values carry less weight: repeated testing increases the chance of apparently notable findings arising by chance.

The record lists six secondary outcomes as posted, but the full protocol and a complete analysis plan are not reported in the registry record. Interpretation should therefore distinguish the prespecified role of the primary endpoint from exploratory or otherwise secondary comparisons, without assuming that every secondary result was protected against multiplicity.

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

• Percentage of Participants With the Composite of All Events Adjudicated (4-point MACE): CV Death (Including Fatal Stroke and Fatal MI), Non-fatal MI (Excluding Silent MI), Non-fatal Stroke and Hospitalization for Unstable Angina Pectoris
• Percentage of Participants With Silent MI
• Percentage of Participants With Heart Failure Requiring Hospitalisation (Adjudicated)
• Percentage of Participants With New Onset Albuminuria
• Percentage of Participants With New Onset Macroalbuminuria
• Percentage of Participants With the Composite Microvascular Outcome

9. Statistical Interpretation

10. Statistical Methods Explained

The primary reported analysis uses a Cox proportional hazards model. General: a Cox model analyzes the timing of events while accounting for participants’ differing follow-up times, and its hazard ratio compares event rates between groups over the observation period. A hazard ratio of 1 corresponds to equal modeled hazards; values below 1 indicate a lower modeled hazard in the numerator group, and values above 1 indicate a higher modeled hazard. The model relies on assumptions, including proportional hazards. Whether those assumptions were assessed, and how, is not reported in the registry record.

The reported primary hazard ratio is 0.86 for all empagliflozin versus placebo. Its 95.02% confidence interval is 0.74 to 0.99. The interval expresses uncertainty in the estimated hazard ratio under the model and analysis. It is not a range containing a known fixed percentage of individual outcomes. Because this is the pooled comparison, it should not be read as a separate estimate for either dose.

The posted percentages are event proportions for each group. They offer an intuitive description of how often the endpoint was recorded over the available observation, but they do not use the timing information in the same way as a Cox model. Participants could have different observation durations, and the registry specifies follow-up up to 4.6 years. The exact handling of censoring and missing outcome data is not reported in the registry record.

The recalculated comparisons use unadjusted risk differences for each dose versus placebo. General: a risk difference is the event proportion in one group minus the event proportion in the comparison group; zero represents equal proportions. The recalculated confidence intervals describe uncertainty around those differences. For empagliflozin 10 mg, the reported recalculated difference is -0.017, with an interval from -0.035 to 0.001. For empagliflozin 25 mg, it is -0.016, with an interval from -0.034 to 0.002. These are simple calculations from posted summary statistics, not adjusted estimates from the trial’s time-to-event model.

The recalculations also report relative risks and odds ratios. General: a relative risk compares the event probabilities between groups, while an odds ratio compares event odds. Both express relative contrasts, unlike the risk difference, which expresses an absolute contrast. The registry’s listed recalculated relative-risk intervals for the 10 mg and 25 mg comparisons are 0.732 to 1.012 and 0.740 to 1.021, respectively. The corresponding odds-ratio intervals are 0.704 to 1.013 and 0.712 to 1.023. These recalculated measures do not replace the sponsor’s analysis and are not adjusted for other variables.

Unadjusted p-values are also listed for the dose-specific recalculations: 0.069 for 10 mg and 0.087 for 25 mg. General: a p-value measures how incompatible the observed data are with a specified null model, conditional on the analysis assumptions. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. Interpretation depends on the analysis being tested, and these unadjusted values should not be confused with the sponsor-reported p-value for the pooled Cox analysis.

The discrepancy flag matters because the recalculated comparisons and the sponsor’s analysis do not form a single interchangeable result. The sponsor’s own analysis takes precedence for the trial’s primary inference. The registry reports a pooled hazard ratio and separate dose-level percentages, while the recalculations compare each dose separately using unadjusted summary-statistic methods. Differences in estimand, model, and analysis can lead to different estimates and uncertainty.

General: confidence intervals and p-values are related ways of expressing sampling uncertainty, but neither supplies information about every source of bias. Randomization, masking, event adjudication, outcome definitions, follow-up, and the handling of missing data all matter. The registry record describes randomized, double-blind, parallel-group conduct and adjudicated endpoints, but details of the full statistical plan, model diagnostics, missing-data methods, and analysis populations beyond the stated total study population are not reported in the registry record.

11. Confidence Intervals and Precision

Precision is reflected in the width of an estimate’s confidence interval: narrower intervals indicate less sampling uncertainty under the method used, while wider intervals indicate greater uncertainty. An interval does not guarantee that the true value lies within it; it describes the behavior of an interval procedure over repeated samples under its assumptions.

For the registry-reported pooled primary analysis, the hazard ratio is 0.86 and its 95.02% confidence interval runs from 0.74 to 0.99. The interval is entirely below 1, though its upper limit is close to 1. This indicates that the reported estimate is not infinitely precise and leaves a range of plausible relative hazards under the model. It does not state how much absolute event probability differs for an individual participant.

The dose-specific recalculated risk-difference intervals are -0.035 to 0.001 for 10 mg and -0.034 to 0.002 for 25 mg. Each interval includes zero. In general, an interval spanning zero for a difference is compatible with no difference as well as with values on either side of zero under the analysis used. These recalculated intervals are unadjusted and do not supersede the pooled sponsor-reported time-to-event result.

Precision also varies across outcomes. For instance, the registry reports fewer silent myocardial infarctions than primary endpoint events, and the recalculated intervals for that secondary outcome are relatively broad. General: estimates based on less information tend to be less precise, especially for uncommon outcomes. The registry record does not provide a formal precision target, the original sample-size justification, or the assumptions used for the trial’s design, so those cannot be assessed from the reported results.

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 display is a planning illustration based on the observed effect, not the trial’s original design calculation. For binary outcomes, the figures are participants per arm. For time-to-event outcomes, the figures are total events. The displayed primary and secondary comparisons discussed in the recalculation material are binary event outcomes, so their planning figures are per arm.

The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. The original design assumptions and the trial’s own sample-size justification are not reported in the registry record.

The 80% power column gives the planned number of participants per arm under those assumptions for an 80% chance of detecting the specified effect if that effect is the true one. The 90% power column gives the corresponding planning number for a 90% chance. These are design probabilities under assumed conditions, not guarantees about a particular study. Smaller effects require more participants to detect with the same power, so planning for 75% or 50% of the observed effect gives larger requirements than planning for 100% of it.

For example, for the 10 mg primary comparison, planning for 100% of the observed effect gives 5,536 participants per arm for 80% power and 7,410 participants per arm for 90% power. Planning for 50% of the observed effect gives 22,864 participants per arm for 80% power and 30,607 participants per arm for 90% power. These are illustrations based on recalculated effects, not the observed power of the completed trial.

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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.121; treatment proportion p₁ = 0.104; average p̄ = 0.112.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.4468 = 0.8757.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.4466 = 0.3759.
4. n = (0.8757 + 0.3759)² / (-0.0168)² = 5,535.0, rounded up to 5,536 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
p0 <- 0.12087441063009                          # control proportion
p1 <- 0.104051172707889                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(1131676)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 5,536 (80% power), 7,410 (90% power)
 75% of observed effect: n per arm = 10,002 (80% power), 13,389 (90% power)
 50% of observed effect: n per arm = 22,864 (80% power), 30,607 (90% power)
power.prop.test: 5535 
Simulated power at n = 5,536 per arm: 0.800 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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-reported primary analysis compares all empagliflozin with placebo using a Cox proportional hazards model. The registry lists a hazard ratio of 0.86, a 95.02% confidence interval from 0.74 to 0.99, and a superiority-analysis p-value of 0.038. It also records a separate non-inferiority-or-equivalence analysis with the same estimate and interval and a p-value of < 0.001.

The simple recalculations are different analyses: they compare the 10 mg and 25 mg groups separately with placebo, use unadjusted risk differences, and report intervals that include zero. The registry marks a discrepancy. Consequently, the sponsor’s own analysis takes precedence for the primary trial inference; recalculated summaries should be treated as supplementary checks, not as replacements for the reported Cox analysis.

The record does not provide enough information to reconcile the discrepancy fully. It does not include the complete analysis plan, detailed model specification, or a derivation of the reported tests. The pooled sponsor estimate should not be confused with either dose-specific recalculation, and neither the estimate nor its p-value alone establishes the practical importance of the result.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when multiple hypotheses are tested, such as several secondary outcomes or more than one treatment comparison. General: as the number of tests increases, the chance of at least one apparently notable result under a collection of null hypotheses can increase. A prespecified strategy may control this risk by adjusting the testing procedure or by defining a hierarchy of hypotheses.

The registry record includes a primary endpoint, multiple secondary outcomes, two empagliflozin dose groups, and a pooled empagliflozin result. It does not report the complete statistical analysis plan or a multiplicity-control strategy for the secondary endpoints or dose-level comparisons. The recalculated p-values are explicitly unadjusted. The sponsor-reported primary analysis includes separate non-inferiority-or-equivalence and superiority-or-other entries, but the detailed procedures governing those analyses are not reported in the registry record.

Missing-data methods are also not reported in the registry record. The outcome timeframe ends at each participant’s individual end of observation, up to 4.6 years, but that description does not specify how participants with incomplete follow-up, missing endpoint information, or treatment discontinuation were handled. The registry record does not state whether sensitivity analyses were conducted for missing outcomes.

The primary population is identified as the total study population, but the detailed analysis population definition, inclusion rules for each analysis, and handling of protocol deviations are not reported in the registry record. The record does not report interim analyses or stopping rules. These omissions limit assessment of whether the reported tests were fully prespecified and how robust the conclusions are to alternative analysis choices. They do not establish that these procedures were absent; they mean the relevant details are not reported in 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: angina unstable (3.7%), cardiac failure (2.4%), pneumonia (2.3%), myocardial infarction (2.1%), acute myocardial infarction (2.0%).

The registry reports serious adverse events for each treatment group. Serious adverse events affected 988 participants in the placebo group, 876 in the empagliflozin 10 mg group, and 913 in the empagliflozin 25 mg group. Other adverse events affected 1690, 1556, and 1551 participants, respectively. The registry lists each group’s number at risk alongside these counts.

The most frequently listed serious adverse event terms include unstable angina at 3.7%, cardiac failure at 2.4%, pneumonia at 2.3%, myocardial infarction at 2.1%, acute myocardial infarction at 2.0%, and coronary artery disease at 2.0%. These terms describe reported events and should not be added together: a participant may have more than one event term, and the record does not establish that the categories are mutually exclusive.

These counts and percentages are descriptive safety summaries. The registry does not provide a comparative statistical analysis for the serious adverse event figures, details of exposure-adjusted rates, or a full account of event severity and causality assessment in the record. They should not be interpreted as a complete measure of comparative safety or as evidence of a cause-and-effect relationship.

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

The posted results describe a randomized, double-blind comparison of empagliflozin and placebo in a population with type 2 diabetes and increased cardiovascular risk. For the primary composite endpoint, the registry reports similar percentages for the two dose groups and the pooled empagliflozin group, all below the placebo percentage. The sponsor-reported pooled Cox analysis gives a hazard ratio of 0.86, with a 95.02% confidence interval from 0.74 to 0.99 and a superiority-analysis p-value of 0.038.

These findings are estimates from a particular endpoint, population, follow-up period, and analysis. The registry record does not establish that the relative hazard applies to other populations, other outcomes, or other durations of follow-up. A composite endpoint also combines cardiovascular death, non-fatal myocardial infarction, and non-fatal stroke; a pooled estimate cannot show whether each component contributed similarly. Component-specific primary results are not reported in the registry record.

The dose-specific unadjusted recalculations have intervals that include zero, while the sponsor-reported pooled time-to-event interval is below 1. These results are not contradictory versions of the same exact estimate: they concern different treatment contrasts and use different methods. The record flags a discrepancy, and the sponsor’s own analysis takes precedence for the primary inference.

The result does not establish the size of benefit or harm for an individual, does not describe every possible safety outcome, and does not resolve uncertainty about missing-data handling or multiplicity. Confidence intervals express statistical uncertainty under their analysis assumptions; they do not capture all limitations in trial conduct or generalizability. The registry findings are a statistical account of the posted results, not medical advice or a recommendation about treatment.

18. Why This Trial Matters Statistically

This trial is statistically informative because it used randomization, double masking, and a parallel-group design to compare empagliflozin doses with placebo in a population selected for type 2 diabetes and increased cardiovascular risk. Its primary endpoint combines adjudicated cardiovascular death, non-fatal myocardial infarction, and non-fatal stroke, and its follow-up extended from randomisation to individual end of observation, up to 4.6 years.

The record illustrates why the analysis method matters. Posted event percentages are straightforward descriptive summaries, whereas the sponsor-reported Cox analysis incorporates event timing and individual follow-up. The simple recalculated risk differences answer dose-specific questions and do not substitute for the pooled time-to-event analysis. The discrepancy flag makes this distinction especially important: the sponsor’s own analysis takes precedence for the primary inference.

The trial also illustrates the role of precision. The sponsor-reported hazard ratio is accompanied by a confidence interval, while recalculated intervals for dose-specific risk differences include zero. Reading estimates alongside their intervals gives more information than relying on a single p-value or a percentage alone.

Finally, the collection of secondary outcomes demonstrates why multiplicity and endpoint hierarchy matter. Additional endpoints can describe different aspects of cardiovascular and microvascular outcomes, but they should not be treated as repeated independent confirmations without considering how many tests were planned and how they were handled. The registry record does not report the full statistical plan, so the strength of conclusions beyond the primary analysis is more limited.

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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

Several trial-specific limitations affect interpretation of the registry record. The original statistical analysis plan, sample-size justification, and design assumptions are not reported in the registry record. The planning figures accompanying the recalculations are illustrations based on observed effects, not the trial’s original design calculations. The record also does not report interim analyses or stopping rules.

The primary endpoint is a composite of cardiovascular death, non-fatal myocardial infarction, and non-fatal stroke. Component-specific primary results are not reported in the registry record, so the pooled composite estimate cannot identify which components contributed most. The posted percentages and the Cox model also summarize outcomes differently, and the record does not provide all details needed to reconcile their relationship.

The registry flags a discrepancy between the sponsor’s analysis and the simple unadjusted recalculations. Those recalculations compare each dose separately with placebo, while the sponsor’s primary analysis compares pooled empagliflozin with placebo using a Cox model. The sponsor’s own analysis takes precedence, but the full analysis plan and model details are not reported in the registry record.

Multiplicity procedures, missing-data methods, and detailed analysis-population definitions are not reported in the registry record. This limits assessment of secondary findings and sensitivity to incomplete follow-up. The population is described as having type 2 diabetes and increased cardiovascular risk, but further eligibility and background-care details are not reported. The safety summaries are descriptive; exposure-adjusted comparisons and full event-assessment details are not reported in the registry record.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: BI 10773 (Empagliflozin) Cardiovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients (EMPA-REG OUTCOME). — NCT01131676.
• Registry record last updated 2016-05-16; status last verified 2016-04.
• Results first posted 2016-05-16.
• 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 EMPA-REG OUTCOME as a completed phase III, multicentre, international, randomized, double-blind, parallel-group cardiovascular safety study. It enrolled 7064 people with type 2 diabetes and increased cardiovascular risk and compared empagliflozin 10 mg, empagliflozin 25 mg, and placebo.

The primary endpoint was time to first adjudicated cardiovascular death, non-fatal myocardial infarction, or non-fatal stroke, assessed from randomisation to individual end of observation, up to 4.6 years. The sponsor-reported pooled Cox analysis gives a hazard ratio of 0.86, with a 95.02% confidence interval from 0.74 to 0.99 and a superiority-analysis p-value of 0.038. The record also flags a discrepancy with simple unadjusted recalculations; the sponsor’s own analysis takes precedence.

Secondary and safety summaries provide additional descriptive results, but multiplicity procedures, missing-data methods, and the full statistical plan are not reported in the registry record.