Multicenter Trial to Evaluate the Effect of Dapagliflozin on the Incidence of Cardiovascular Events: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Multicenter Trial to Evaluate the Effect of Dapagliflozin on the Incidence of Cardiovascular Events (NCT01730534), a randomized phase 3 study. The primary outcome is Subjects Included in the Composite Endpoint of CV Death, MI or Ischemic Stroke.

2. Clinical Question

3. Trial Design

This was a phase 3, randomized, parallel-group trial with quadruple masking, described in the registry as a multicenter, double-blind, placebo-controlled study. The intervention was dapagliflozin 10 mg once daily alongside standard care. The comparator was placebo alongside standard care. The registry lists an enrollment of 17,190 and a completed study status. The recorded start date is 2013-04-25 and the primary completion date is 2018-09-11.

The study summary describes participants with type 2 diabetes who had either established cardiovascular disease or cardiovascular risk factors. The conditions listed in the record are diabetes mellitus, non-insulin-dependent, and high risk for a cardiovascular event. A full account of eligibility requirements, recruitment settings, and the number of sites is not reported in the registry record. Although the trial is identified as multicenter, the site count and site locations are not reported in the registry record.

The primary endpoint is a participant count for a composite of cardiovascular death, myocardial infarction, or ischemic stroke, with a timeframe of up to 5.2 years. The registry identifies the primary measure as safety and co-primary efficacy. It also lists a renal composite endpoint and all-cause mortality as secondary outcomes. The displayed sponsor analyses use regression and Cox methods and report hazard ratios. The trial's original sample-size justification, planned event target, detailed analysis populations, and any planned interim analyses are not reported in the registry record.

Randomization and masking describe important features of the trial design, but the registry record does not provide enough detail to assess allocation concealment, the specific people included in quadruple masking, or how masking was maintained. The arm descriptions indicate that standard care for diabetes and associated conditions accompanied both assigned treatments.

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:

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Subjects Included in the Composite Endpoint of CV Death, MI or Ischemic Stroke, measured over up to 5.2 years.

The primary endpoint counts participants included in a composite of cardiovascular death, myocardial infarction, or ischemic stroke. These are distinct cardiovascular outcomes, combined into a single endpoint for the main comparison. The registry reports participant counts by group and gives a timeframe of up to 5.2 years. It does not report the separate number of participants experiencing each component, so the contribution of each component to the composite cannot be assessed from the posted primary result.

In general, a composite endpoint can capture a range of clinically important outcomes in one trial measure. Combining outcomes can increase the number of endpoint events available for analysis compared with considering a less frequent component alone. It also changes what the overall result means: a composite estimate does not establish that each component had the same pattern or magnitude. Components can differ in frequency and clinical character, and a composite result may reflect their combined experience rather than a uniform effect across all components.

The registry describes the primary measure as safety and co-primary efficacy. The posted analysis includes a time-to-event hazard ratio as well as participant counts. A count-based summary helps readers see how many participants in each group were included in the endpoint, while the hazard ratio summarizes the comparison over follow-up using the reported Cox analysis. Component-specific estimates, the rules for confirming events, and the handling of participants whose outcome status was unavailable are not reported in the registry record.

7. Results Posted in the Registry

• Dapa 10 mg: 8.8% (756/8582) versus 9.4% (803/8578) on control; risk difference -0.6 percentage points (95% CI -1.4 to 0.3), risk ratio 0.94 (0.86–1.03); unadjusted two-proportion z-test p = 0.208.

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.

Subjects Included in the Renal Composite Endpoint: Confirmed Sustained ≥40% Decrease in eGFR to eGFR <60 ml/Min/1.73m2 and/or ESRD and/or Renal or CV Death.

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Dapa 10 mg: 4.3% (370/8582) versus 5.6% (480/8578) on control; risk difference -1.3 percentage points (95% CI -1.9 to -0.6), risk ratio 0.77 (0.68–0.88); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Dapa 10 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Subjects Included in the Endpoint of All-cause Mortality.

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Dapa 10 mg: 6.2% (529/8582) versus 6.6% (570/8578) on control; risk difference -0.5 percentage points (95% CI -1.2 to 0.3), risk ratio 0.93 (0.83–1.04); unadjusted two-proportion z-test p = 0.198.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Dapa 10 mg and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes cover a renal composite endpoint and all-cause mortality. Together, their reported results do not form a uniform pattern: the renal composite has a more pronounced statistical signal, whereas the all-cause mortality estimate is less conclusive. The primary cardiovascular composite remains the main result for the trial's central question, and secondary analyses should be interpreted in relation to that primary finding rather than as replacements for it.

The renal composite combines a sustained decrease in estimated glomerular filtration rate to a specified level, end-stage renal disease, and renal or cardiovascular death. The all-cause mortality endpoint is broader in cause than the cardiovascular-death component of the primary endpoint. Because these outcomes differ in definition and scope, their estimates do not answer the same question as the primary composite. The registry does not provide separate results for the components of the renal composite or enough detail to assess which components contributed most to its overall result.

The posted secondary analyses report hazard ratios and confidence intervals, and the renal analysis has p < 0.001. However, the registry record does not report whether secondary hypotheses were adjusted for multiple testing, whether a testing hierarchy was prespecified, or whether the analyses were intended as confirmatory. When multiple outcomes are examined, the chance of at least one apparently strong result can increase if no multiplicity procedure is used. Accordingly, unadjusted secondary p-values carry less weight than a prespecified primary analysis and should be treated as supportive or hypothesis-generating unless the analysis plan establishes otherwise.

The registry lists two secondary outcome measures and posts analyses for both. The posted data can describe the pattern across those endpoints, but cannot establish a formal hierarchy or correction that is not reported. A statistically notable secondary result does not resolve the uncertainty in the primary comparison, and a less conclusive secondary result should not be interpreted as proof of no difference. Each estimate remains subject to its own uncertainty and endpoint definition.

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

• Subjects Included in the Renal Composite Endpoint: Confirmed Sustained ≥40% Decrease in eGFR to eGFR <60 ml/Min/1.73m2 and/or ESRD and/or Renal or CV Death.
• Subjects Included in the Endpoint of All-cause Mortality.

9. Statistical Interpretation

10. Statistical Methods Explained

The registry reports a participant-count primary outcome and a sponsor analysis using regression and Cox methods, with a hazard ratio, confidence interval, and p-value. The posted count data also support an unadjusted comparison of the proportion of participants included in the endpoint. These approaches answer related but not identical questions. The count-based comparison summarizes whether a participant had an endpoint during the recorded follow-up. A Cox analysis uses time-to-event information and accounts for the timing of observed events and follow-up in estimating a hazard ratio.

What does a hazard ratio describe?

In general, a hazard is the event rate at a particular point in follow-up among participants who have not yet had the event. A hazard ratio compares the hazards in the trial groups under the fitted model. A value below 1 represents a lower estimated hazard in the dapagliflozin group than in the placebo group; a value above 1 represents a higher estimated hazard. It is not the same quantity as a risk difference or a ratio of the final participant counts. The reported primary hazard ratio is 0.93. The registry does not report diagnostic checks for the Cox model or details about how the analysis handled censoring, so those aspects cannot be evaluated from the posted record.

How should the confidence interval be read?

In general, a confidence interval describes the uncertainty around an estimate under the statistical model and analysis used. A 95% confidence interval is constructed by a method intended to cover the underlying parameter in 95% of repeated applications under the model's assumptions. It is not a statement that there is a 95% probability that a particular fixed parameter lies inside the interval. For the primary hazard ratio, the reported interval runs from 0.84 to 1.03 and includes 1, the value representing no difference in hazards. That interval therefore does not rule out no difference at the conventional two-sided 0.05 level.

What does the p-value add?

In general, a p-value measures how incompatible the observed data, or data more extreme, are with a specified null model. It does not give the probability that the null hypothesis is true, nor does it measure the size or clinical importance of an effect. The sponsor reports p = 0.172 for the primary comparison. The unadjusted recalculation from participant counts reports p = 0.208. The latter is a simple, unadjusted analysis and does not reproduce the sponsor's time-to-event modeling. The difference between these p-values is not, by itself, evidence of an inconsistency because the analyses use different summaries.

Why can the count-based and time-to-event results differ?

A participant-count comparison treats endpoint status over the recorded period as a binary outcome. A time-to-event analysis additionally uses when an endpoint occurred and the time participants remained under observation. If follow-up duration or event timing differs, a comparison based only on final counts can give a different estimate from a Cox analysis. The record gives a maximum timeframe of up to 5.2 years, but does not give the distribution of follow-up or enough detail to reconstruct the sponsor analysis. Therefore, the unadjusted recalculation is useful as a transparent check on the posted counts, but it should not be presented as the trial's primary inferential analysis.

Why does over-dispersion matter?

Over-dispersion is a general issue for outcomes where variability is greater than a statistical model expects. When it is ignored, standard errors and confidence intervals can be too narrow, making results look more precise than the data warrant. The primary outcome is recorded as participants with an event, not as a repeated event count or a continuous measure with a reported spread. The registry does not report an over-dispersion model or a separate dispersion estimate for this endpoint. It is therefore not appropriate to infer that over-dispersion was tested or handled. For the Cox analysis, the posted record identifies the method but does not provide the model diagnostics or variance details needed for a more complete assessment.

More broadly, a statistical method's label does not establish that every underlying assumption was satisfied. The record does not report the proportional-hazards assessment, covariate set, stratification details, handling of competing events, or event-adjudication procedures. Those omissions limit independent appraisal of the analysis. They do not negate the reported estimates; they define what can and cannot be evaluated from the registry entry.

11. Confidence Intervals and Precision

The sponsor's primary estimate is a hazard ratio of 0.93 with a 95% confidence interval from 0.84 to 1.03. The interval is relatively informative about the range of estimates compatible with the fitted analysis, but it includes 1, the no-difference value. It therefore does not exclude no difference in hazards at the conventional level associated with a 95% confidence interval. The reported p-value, 0.172, conveys the same broad inferential result under the sponsor's analysis: the primary result is not statistically decisive by that convention.

The interval should not be read as proving equivalence or as demonstrating that any remaining differences are unimportant. Equivalence or non-inferiority conclusions require a prespecified margin and an analysis designed for that purpose; such a margin is not reported in the registry record. Nor does the interval cover uncertainty from every possible source. It reflects sampling uncertainty under the reported model and does not resolve limitations related to endpoint definition, missing data, analysis choices, or applicability beyond the enrolled population.

The posted participant counts are 756 for dapagliflozin and 803 for placebo, with denominators of 8,582 and 8,578. The simple unadjusted risk-difference interval spans zero and its p-value is 0.208. This interval addresses an unadjusted difference in participant proportions, not the time-to-event hazard ratio. It should therefore be considered alongside, rather than in place of, the sponsor's confidence interval.

Precision also depends on the amount and distribution of event information, not only on the overall enrollment. The registry record does not report the detailed follow-up distribution, the number of participants censored, or standard errors as a separate quantity. The posted interval is the appropriate source for describing precision of the sponsor estimate, while the missing details limit further evaluation of how that precision was obtained.

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 calculation. For this binary participant-count endpoint, the figures are participants per arm, rather than total events. The planning assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect to be detected is set at the stated fraction of the observed effect. The trial's own original design assumptions and sample-size justification are not reported in the registry record.

The column for 80% power describes the number of participants per arm that the stated planning calculation estimates would be needed to have an 80% chance of detecting the selected effect under those assumptions, if that effect were the true effect. The 90% power column describes the corresponding planning target for a higher chance of detection. Neither figure is a statement about what the completed trial actually achieved. This is not observed power.

Rows that plan for 75% or 50% of the observed effect use smaller target effects than the row based on 100% of the observed effect. Smaller effects are harder to distinguish from chance variation, so the planning calculations require more participants per arm to reach the same power target. The displayed values are conditional illustrations: they depend on using the observed estimate as a planning reference, on the specified allocation and testing assumptions, and on the normal approximation. They are not guarantees that a future trial of that size would detect an effect.

The sample-size assumptions also distinguish count-based planning from time-to-event planning: time-to-event tables give total events using the Schoenfeld method, while other tables give participants per arm. Readers should identify which quantity applies before comparing values. In interpreting the primary binary endpoint, the per-arm participant counts are the relevant unit. These planning values do not change the interpretation of the observed primary result or replace the sponsor's reported analysis.

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 = 0.094; treatment proportion p₁ = 0.088; average p̄ = 0.091.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.4064 = 0.7966.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.4064 = 0.3421.
4. n = (0.7966 + 0.3421)² / (-0.0055)² = 42,548.2, rounded up to 42,549 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 <- 0.0936115644672418                          # control proportion
p1 <- 0.0880913539967374                          # 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(1730534)
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 = 42,549 (80% power), 56,960 (90% power)
 75% of observed effect: n per arm = 76,159 (80% power), 101,955 (90% power)
 50% of observed effect: n per arm = 172,520 (80% power), 230,954 (90% power)
power.prop.test: 42548.2 
Simulated power at n = 42,549 per arm: 0.797 (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 registry reports a sponsor analysis for the primary endpoint using regression and Cox methods. It gives a hazard ratio of 0.93, a 95% confidence interval from 0.84 to 1.03, and p = 0.172. That is the trial's reported time-to-event analysis and should lead the interpretation of the primary result.

The recalculation uses the posted participant counts and denominators to make a simple unadjusted risk-difference comparison. Its interval spans zero and its unadjusted p-value is 0.208. This analysis does not use the event timing or model structure reported for the sponsor analysis, so the estimates and p-values are not expected to be numerically identical. They are different summaries of the available results.

The record indicates no discrepancy between the sponsor result and the recalculation. That does not mean the methods are interchangeable: the recalculation is a transparent check based on the displayed counts, while the sponsor's analysis is the registry-reported inferential result for the primary endpoint. The sponsor estimate, its interval, and its p-value therefore take precedence. The record does not report enough modeling details to independently reproduce the Cox analysis, including the covariates, censoring approach, or model diagnostics.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a trial evaluates more than one outcome, comparison, subgroup, or analysis. In general, testing several hypotheses can increase the chance that at least one result appears statistically significant by chance. Trial protocols may address this through an ordered testing sequence, adjusted significance thresholds, or other prespecified procedures. The registry record reports a primary endpoint and two secondary outcomes, but does not report a multiplicity adjustment, a hierarchical testing strategy, or the intended confirmatory status of the secondary analyses. The posted secondary p-values should therefore be read without assuming that an adjustment was or was not applied outside what the record states.

The record also does not report how missing outcome information was handled, whether any values were imputed, or how participants with incomplete follow-up were included in each analysis. For time-to-event analyses, censoring rules matter because they define how participants with no observed event contribute information. The registry names a Cox analysis but does not describe censoring procedures, loss to follow-up, or the amount of missingness. Those details are not reported in the registry record, so their influence on the estimates cannot be assessed.

Analysis populations are another important part of interpretation. The record does not specify whether the reported sponsor results are based on an intention-to-treat population, a per-protocol population, or another defined set, nor does it report exclusions from the randomized groups for the primary analysis. Randomization is recorded, but the analysis population and the treatment of participants who did not follow assigned treatment are not reported in the registry record.

Finally, the registry does not provide the original sample-size justification, planned interim analyses, full statistical analysis plan, or details on event adjudication. These are not evidence that a particular procedure was omitted; they are unreported items that limit independent assessment. The distinction matters: the posted estimates can be described accurately while the unreported procedures remain unknown.

16. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 1% 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 (3.2%), acute myocardial infarction (3.1%), pneumonia (2.6%), cardiac failure (2.2%), angina pectoris (1.9%).

The record labels the primary measure as safety and co-primary efficacy and also posts adverse-event summaries by group. Serious adverse events affected 3,205 of 8,574 participants in the dapagliflozin group and 3,418 of 8,569 participants in the placebo group. Other adverse events affected 268 of 8,574 and 342 of 8,569 participants, respectively. The registry also reports deaths of 529 and 570 in the two groups. These are descriptive counts with the recorded denominators; they are not, by themselves, adjusted comparisons or causal estimates.

The most frequently listed serious adverse event terms include unstable angina, acute myocardial infarction, pneumonia, cardiac failure, angina pectoris, and congestive cardiac failure. The percentages attached to these terms are those reported in the registry. The record does not provide a complete account of how event terms were coded, whether events could be counted more than once per participant, or whether the term percentages use the same denominator as the group-level serious-event summaries. They should be interpreted as registry-reported event summaries rather than as a full safety analysis.

Counts do not show timing, duration, severity grading beyond the serious-event label, or the statistical uncertainty around between-group comparisons. A complete assessment would require the prespecified safety definitions and analysis details, which are not reported in the registry record.

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

The primary analysis describes the comparison between dapagliflozin added to standard care and placebo added to standard care for the cardiovascular composite endpoint. Its reported hazard ratio is 0.93, with a 95% confidence interval from 0.84 to 1.03 and p = 0.172. The estimate is compatible with a range of possible differences, and its interval includes the no-difference value. The result therefore does not provide clear statistical evidence of a difference in the primary endpoint under the reported analysis.

That interpretation is not the same as proving that the treatments have identical effects. A non-significant result can reflect an estimate near no difference, uncertainty around the estimate, or both. The confidence interval makes that uncertainty visible, but it does not establish equivalence, exclude every potentially relevant effect, or answer questions the trial did not analyze. No equivalence or non-inferiority margin is reported in the registry record.

The result also should not be translated into a claim about each component of the composite. The registry reports an aggregate endpoint for cardiovascular death, myocardial infarction, or ischemic stroke, without separate primary component estimates. Nor does the primary result answer the same question as the renal composite or all-cause mortality secondary outcomes. Those analyses have their own definitions and uncertainty, and their interpretation is limited by the absence of reported multiplicity procedures.

These findings apply to the trial as recorded: a randomized comparison in people with type 2 diabetes described as having established cardiovascular disease or cardiovascular risk factors, with treatment added to standard care and follow-up of up to 5.2 years. They do not establish what would occur in populations, settings, or follow-up periods not represented in the record. They are statistical findings, not medical advice or an individual treatment recommendation.

18. Why This Trial Matters Statistically

This trial illustrates why the statistical meaning of an endpoint and an analysis should be kept distinct. The primary endpoint is a composite participant count, while the sponsor's reported analysis is a Cox-model hazard ratio that incorporates event timing. The posted counts make the underlying group summaries easier to inspect, but a count-based recalculation does not reproduce the time-to-event analysis. Readers can see both the descriptive event totals and the reported model-based comparison without treating them as interchangeable.

It also demonstrates how uncertainty should shape interpretation. The sponsor's primary confidence interval includes the no-difference value, and the p-value is not conventionally statistically significant. A point estimate alone would conceal that uncertainty. The interval and p-value together show why a numerical direction should not be elevated into a definitive conclusion, while also avoiding the opposite mistake of claiming that the groups are proven identical.

The secondary results show a further statistical issue: findings across outcomes need to be interpreted in light of outcome selection and multiplicity. A strong nominal signal on a secondary endpoint can coexist with an inconclusive primary analysis, but its evidential weight depends on the prespecified testing plan and any adjustment for multiple testing. Those details are not reported in the registry record.

Finally, the planning illustrations clarify that sample-size calculations depend on assumptions about the effect to detect, allocation, and statistical method. They are not retrospective measures of power. Separating planning illustrations, posted counts, recalculated summaries, and sponsor analyses helps keep design, estimation, and inference from being conflated.

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 specific to this registry record affect how fully the trial can be appraised. The study summary gives a broad description of participants with type 2 diabetes and either established cardiovascular disease or cardiovascular risk factors, but detailed eligibility criteria and baseline characteristics are not reported. The degree to which the enrolled population represents particular clinical subgroups therefore cannot be determined from this record alone.

The primary composite result is not broken down into its components. Separate counts or estimates for cardiovascular death, myocardial infarction, and ischemic stroke are not reported, so it is not possible to assess whether the composite pattern was similar across these outcomes or driven by one component. The record also does not include the event-adjudication process or detailed endpoint definitions beyond the listed outcome description.

The sponsor analysis is identified as regression and Cox, but its covariates, censoring rules, proportional-hazards assessment, and analysis population are not reported. Missing data and loss to follow-up handling are also not described. These omissions restrict evaluation of the model and of how incomplete follow-up may have affected the estimate. They should not be taken as evidence that the trial did not use procedures for these issues.

The registry does not report the original sample-size justification, planned interim analyses, multiplicity strategy, or a full statistical analysis plan. This limits assessment of whether the posted analyses match a prespecified hierarchy, especially for secondary endpoints. The reported follow-up timeframe is up to 5.2 years, but its distribution is not given. Finally, safety counts are presented without full event-coding, severity, or comparative-analysis details. These reporting limits call for caution when extending the posted summaries beyond the specific outcomes and analyses recorded.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Multicenter Trial to Evaluate the Effect of Dapagliflozin on the Incidence of Cardiovascular Events — NCT01730534.
• Registry record last updated 2019-12-26; status last verified 2019-11.
• Results first posted 2019-12-26.
• 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

NCT01730534 is a completed phase 3, randomized, parallel-group, quadruple-masked, multicenter trial comparing dapagliflozin 10 mg once daily plus standard care with placebo plus standard care in people with type 2 diabetes at cardiovascular risk. The registry lists enrollment of 17,190, a start date of 2013-04-25, and a primary completion date of 2018-09-11.

The primary endpoint was the participant composite of cardiovascular death, myocardial infarction, or ischemic stroke, with follow-up of up to 5.2 years. The sponsor-reported Cox analysis gave a hazard ratio of 0.93, a 95% confidence interval from 0.84 to 1.03, and p = 0.172. The posted secondary outcomes were a renal composite and all-cause mortality. The record also includes adverse-event counts, while several details needed for full evaluation of analysis populations, missing data, multiplicity, and the original design assumptions are not reported in the registry record.