This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical results on this page are restricted to the trial data posted on ClinicalTrials.gov for RE-NOVATE.
1. Trial at a Glance
RE-NOVATE was a randomized, double-masked, parallel phase 3 trial evaluating dabigatran etexilate versus enoxaparin for prevention of venous thromboembolism following total hip arthroplasty. The registry reports a binary primary endpoint evaluated during the treatment period of 28-35 days.
| Feature | RE-NOVATE |
|---|---|
| Phase | Phase 3 |
| Condition | Venous Thromboembolism |
| Clinical setting | Prevention of venous thromboembolism following total hip arthroplasty |
| Design | Randomized, double-masked, parallel |
| Primary purpose | Prevention |
| Enrollment | 2055.0 |
| Interventions | Enoxaparin; Dabigatran etexilate |
| Primary endpoint type | Binary |
| Results posted | Yes |
| Outcome measures posted | 12 |
| Statistical analyses posted | 11 |
| Lead sponsor | Boehringer Ingelheim |
| Sponsor type | Industry |
| ClinicalTrials.gov | NCT00657150 |
2. Clinical Question
The central statistical question was whether dabigatran etexilate 220mg was non-inferior to enoxaparin for the prevention of the registered composite of total venous thromboembolic events and all-cause mortality during the treatment period.
Population
Participants enrolled in a phase 3 prevention trial following total hip arthroplasty, with venous thromboembolism as the condition of interest.
Intervention
Dabigatran etexilate, with the primary statistical comparison specifically reported as dabigatran 220mg versus enoxaparin.
Comparator
Enoxaparin.
Primary question
Is the risk difference for total venous thromboembolic event and all-cause mortality consistent with non-inferiority of dabigatran 220mg relative to enoxaparin?
3. Trial Design
Dabigatran 220mg
- Dabigatran etexilate
- Primary comparison dose: 220mg
- Evaluated against enoxaparin
Enoxaparin
- Enoxaparin
- Active comparator
- Reference group for reported risk differences
4. Endpoints
| Endpoint | Time frame | Type | Registered definition / analysis role |
|---|---|---|---|
| Number of Participants With Total Venous Thromboembolic Event and All-cause Mortality During Treatment Period | 28-35 days | Binary | Total VTE includes both proximal and distal DVT detected by routine venography, symptomatic DVT confirmed by venous duplex, ultrasound, venography or autopsy, and PE confirmed by pulmonary V-Q scintigraphy, chest x-ray, pulmonary angiography, spiral CT or autopsy. All components and deaths were centrally adjudicated according to the registry definition. |
| Number of Participants With Major Venous Thromboembolic Event and Venous Thromboembolic Event-related Mortality During Treatment Period | 28-35 days | Binary | Secondary endpoint. |
| Number of Participants With Proximal Deep Vein Thrombosis During Treatment Period | 28-35 days | Binary | Secondary endpoint. |
| Number of Participants With Total Deep Vein Thrombosis During Treatment Period | 28-35 days | Binary | Secondary endpoint. |
| Number of Participants With Bleeding Events (Defined According to Modified McMaster Criteria) During Treatment Period | 28-35 days | Binary | Secondary safety endpoint; analyses included major and clinically relevant bleeding events and any bleeding events. |
| Number of Participants With Symptomatic Deep Vein Thrombosis During Treatment Period | 28-35 days | Binary | Secondary endpoint. |
| Number of Participants With Pulmonary Embolism During Treatment Period | 28-35 days | Binary | Secondary endpoint. |
| Number of Participants Who Died During Treatment Period | 28-35 days | Binary | Secondary endpoint. |
| Number of Participants With Total Venous Thromboembolic Event (VTE) and All-cause Mortality During the Follow-up Period | 3 months | Binary | Secondary endpoint. |
5. Statistical Methodology
Binary endpoint analysis
The registry reports binary outcomes using either a normal approximation for independent binomial distributions or the Fisher exact test. The primary endpoint used the normal approximation, normalized here as a Wald / z-test approach.
A negative risk difference indicates a lower observed event probability in the dabigatran group than in the enoxaparin group. The risk difference is an absolute rather than relative treatment-effect measure.
Normal approximation / Wald approach
For the primary endpoint, the registry describes a normal approximation of the independent binomial distribution. The reported effect measure is a percentage risk difference with a 95% confidence interval. The registry-reported analysis notes specify that the reported comparison was versus enoxaparin and that the calculation was without stratification.
Fisher exact test
Several secondary binary outcomes were analyzed using Fisher exact testing. This method is especially useful when event counts are small because it does not rely on the large-sample approximation underlying a conventional Pearson chi-square test.
Analysis populations
The primary analysis population was the Full Analysis Set: randomized participants who had taken at least one dose of oral or subcutaneous trial medication, had undergone surgery, and had an evaluable negative venogram for both distal and proximal DVT in both legs. Other analyses used endpoint-specific Full Analysis Set definitions or all randomized participants who had taken at least one dose, as specified in the registry data.
Intention-to-treat principle
The registry analysis text identifies intention-to-treat analysis as a concept associated with the primary and several secondary analyses. In a randomized trial, analyzing participants according to their randomized assignment preserves the comparison created by randomization. Here, however, the registry-reported analysis population for the primary endpoint is described explicitly as the Full Analysis Set rather than simply all randomized participants.
Stratified analysis terminology
The analysis text identifies stratified analysis as an associated concept for the primary endpoint and several secondary endpoints. At the same time, the specific analysis notes for the reported risk-difference calculations state that the normal approximation of the independent binomial distribution was performed without stratification. These statements should therefore not be treated as evidence that the reported risk-difference estimate itself was a stratified estimate.
6. Primary Endpoint Result
The primary endpoint was the number of participants with total venous thromboembolic event and all-cause mortality during the 28-35 day treatment period. The registry reports a non-inferiority analysis comparing dabigatran 220mg with enoxaparin.
Risk difference: dabigatran 220mg vs enoxaparin
95% CI: -3.79% to 1.64% · P < 0.0001
Non-inferiority margin: 7.7%
| Primary endpoint | Estimate | 95% CI | P-value | Hypothesis |
|---|---|---|---|---|
| Total venous thromboembolic event and all-cause mortality during treatment period, 28-35 days | Risk difference -1.08% | -3.79% to 1.64% | < 0.0001 | Non-inferiority or equivalence |
The estimated risk difference of -1.08% means that the reported event risk was estimated to be 1.08 percentage points lower with dabigatran 220mg than with enoxaparin for the primary 28-35 day endpoint.
The estimate does not mean that dabigatran reduced relative risk by 1.08%, nor does it mean that every patient experienced a 1.08 percentage-point change in individual risk. A risk difference is an absolute group-level comparison.
The 95% confidence interval extends from -3.79% to 1.64%. It therefore describes uncertainty around the estimated absolute treatment difference. It is not a range containing the effects that individual participants must have experienced.
The reported P < 0.0001 is evidence against the null hypothesis specified for the reported non-inferiority analysis. A p-value does not measure the magnitude of the treatment effect and should not be interpreted as the probability that the treatment is effective.
For a non-inferiority design, the central question is whether the confidence interval excludes differences that would be worse than the prespecified non-inferiority margin. The registry reports a margin of 7.7%. Because the ClinicalTrials.gov record reports the margin and the 95% confidence interval, the non-inferiority conclusion is interpreted through the relationship between the interval and that margin rather than by treating the p-value as an effect-size measure.
7. Secondary Efficacy Results
Major venous thromboembolic event and VTE-related mortality
Risk difference: dabigatran 220mg vs enoxaparin
95% CI: -3.64% to -0.19% · P = 0.029
Superiority analysis
The reported risk difference of -1.91% corresponds to an estimated absolute difference of 1.91 percentage points in favor of dabigatran 220mg relative to enoxaparin for this secondary endpoint.
The 95% confidence interval, -3.64% to -0.19%, remains below zero. Under the reported superiority analysis, this is consistent with a statistically significant difference at the conventional two-sided 5% level, as reflected by P = 0.029.
The result does not establish that the treatment effect is exactly 1.91 percentage points, because the confidence interval shows statistical uncertainty around the point estimate. It also does not establish that the same effect applies to other endpoints.
The analysis is secondary rather than the registered primary endpoint. Its interpretation should therefore remain distinct from the primary non-inferiority question.
Proximal deep vein thrombosis
Risk difference: dabigatran 220mg vs enoxaparin
95% CI: -3.47% to -0.12% · P = 0.0358
Superiority analysis
The estimated risk difference of -1.79% indicates a 1.79 percentage-point lower estimated risk of proximal DVT with dabigatran 220mg than with enoxaparin for this secondary endpoint.
The 95% confidence interval ranges from -3.47% to -0.12%. Because the interval remains below zero, the reported result is consistent with the superiority hypothesis tested for this endpoint. The associated P = 0.0358 quantifies evidence against the null hypothesis under the specified analysis; it does not quantify the clinical importance or size of the effect.
This endpoint is one component of the broader thromboembolic outcome structure, so it should not be interpreted as interchangeable with the primary composite endpoint.
Total deep vein thrombosis
Risk difference: dabigatran 220mg vs enoxaparin
95% CI: -3.65% to 1.73% · P = 0.4839
Superiority analysis
The estimated risk difference was -0.96%, but the 95% confidence interval extends from -3.65% to 1.73%. Thus, the interval includes zero and is compatible with both a modest reduction and a modest increase in absolute event risk under the reported analysis.
The P = 0.4839 does not provide evidence against the superiority null hypothesis in this analysis. It also does not demonstrate that the two treatments are identical or that any clinically relevant difference has been ruled out.
8. Secondary Safety Results
The registry includes several bleeding analyses during the 28-35 day treatment period. The reported effect measures are absolute differences in percentage for specified bleeding categories.
| Safety endpoint | Effect measure | 95% CI | P-value | Hypothesis |
|---|---|---|---|---|
| Bleeding events, category major and clinically relevant bleeding events | Absolute difference 0.8% | -0.8% to 2.3% | 0.3305 | Superiority |
| Bleeding events, category any bleeding events | Absolute difference 1.4% | -1.1% to 3.9% | 0.2626 | Superiority |
| Bleeding events, category major bleeding events | Fisher exact test; no effect estimate reported | 95% CI stated in registry analysis | 0.4022 | Superiority |
The reported absolute difference of 0.8% for major and clinically relevant bleeding events has a 95% confidence interval from -0.8% to 2.3%. Because the interval includes zero, the result does not establish superiority for either treatment under this comparison.
For any bleeding events, the reported absolute difference was 1.4%, with a 95% confidence interval from -1.1% to 3.9%. Again, the interval includes zero and the reported P = 0.2626 does not establish superiority.
The Fisher exact analysis of major bleeding events has P = 0.4022, but the ClinicalTrials.gov record does not provide a corresponding effect estimate. A p-value without an effect estimate does not show how large the treatment difference was.
Serious adverse events by treatment arm
| Treatment arm | Participants with serious adverse events | At risk |
|---|---|---|
| Dabigatran 220mg | 57 | 1010 |
| Enoxaparin | 59 | 1003 |
The ClinicalTrials.gov record reports serious adverse events as 57/1010 for dabigatran 220mg and 59/1003 for enoxaparin. These counts provide an arm-level safety description; the ClinicalTrials.gov record does not provide a formal statistical comparison for these serious adverse-event counts.
9. Other Secondary Endpoint Analyses
Additional registry-posted analyses use Fisher exact testing for several binary outcomes. For these analyses, the ClinicalTrials.gov record provides the statistical test and p-value but do not provide a corresponding effect estimate and confidence-interval bounds.
| Secondary endpoint | Time frame | Method | P-value |
|---|---|---|---|
| Symptomatic deep vein thrombosis | 28-35 days | Fisher exact | 0.0612 |
| Pulmonary embolism | 28-35 days | Fisher exact | 0.6231 |
| Death during treatment period | 28-35 days | Fisher exact | 0.4977 |
| Total VTE and all-cause mortality during follow-up | 3 months | Fisher exact | 0.6870 |
10. Non-Inferiority Framework
The primary endpoint was evaluated under a non-inferiority or equivalence hypothesis, and the registry explicitly reports a 7.7% non-inferiority margin.
The exact direction of the margin must be interpreted together with the definition of the risk difference and the comparator. For RE-NOVATE, the registry defines the comparison as dabigatran 220mg versus enoxaparin and reports a risk-difference margin of 7.7%.
The primary estimate was -1.08%, with a 95% confidence interval from -3.79% to 1.64%. The entire reported interval is well within the registry-reported 7.7% non-inferiority margin in the unfavorable direction. The registry also reports P < 0.0001 for the non-inferiority analysis.
11. Statistical Methods Explained
Why was a risk difference used?
The primary endpoint is binary: each participant either experienced the registered composite event during the specified treatment period or did not. A risk difference directly compares the absolute probabilities between the two treatment groups. This makes it naturally interpretable in percentage-point terms and also allows direct comparison with a prespecified absolute non-inferiority margin.
Why was a normal approximation used?
The registry describes the primary analysis as a normal approximation of the independent binomial distribution. For sufficiently large samples, the sampling distribution of a difference between two binomial proportions can be approximated by a normal distribution, permitting construction of a confidence interval and a corresponding Wald / z-test.
Why was Fisher exact testing used for some endpoints?
Several secondary outcomes were analyzed using Fisher exact testing. This method calculates the exact probability of the observed two-group configuration under the null hypothesis and is particularly useful when event counts are small. Unlike a normal approximation, it does not depend on the same large-sample approximation.
Why is the non-inferiority margin more important than the p-value?
The non-inferiority margin defines the largest clinically acceptable loss relative to the comparator that the trial is designed to exclude. The p-value is a measure of evidence under a statistical null hypothesis; it does not define what degree of loss is acceptable. For RE-NOVATE, the registry supplies a 7.7% margin, so interpretation of the primary result requires examining the confidence interval relative to that margin.
What does a risk difference of -1.08% mean?
A risk difference of -1.08% means the estimated absolute event probability was 1.08 percentage points lower in the dabigatran 220mg group than in the enoxaparin group. It is not a relative risk, odds ratio, hazard ratio, or statement that an individual participant's probability changed by exactly 1.08 percentage points.
Why does the confidence interval matter?
The point estimate is only one representation of the treatment comparison. The 95% confidence interval communicates how precisely the analysis estimates the risk difference under the statistical model and sampling framework. For the primary endpoint, the interval from -3.79% to 1.64% gives substantially more information about uncertainty than the point estimate alone.
Why should the analysis population be stated explicitly?
Randomized trials often define more than one analysis population. Here, the primary analysis is described as a Full Analysis Set with specific requirements concerning randomization, medication exposure, surgery, and evaluable venography. Those requirements affect which participants contribute to the reported primary analysis and therefore are part of the statistical interpretation.
12. Intention-to-Treat, Analysis Sets, and Censoring
The ClinicalTrials.gov record identifies intention-to-treat analysis as an analysis concept associated with the primary and several secondary analyses. However, the primary statistical analysis record explicitly identifies the analysis population as a Full Analysis Set with additional eligibility requirements.
| Analysis concept | What the ClinicalTrials.gov record indicates |
|---|---|
| Primary analysis population | Full Analysis Set with randomization, at least one dose, surgery, and evaluable negative venogram requirements. |
| Secondary major VTE analysis | Full Analysis Set-major, with endpoint-specific requirements. |
| Proximal DVT analysis | Full Analysis Set-pDVT, with endpoint-specific evaluability requirements. |
| Total DVT analysis | Full Analysis Set-tDVT, with evaluable venography requirements. |
| Bleeding analyses | All randomized participants who had taken at least one dose of oral or subcutaneous trial medication. |
| 3-month VTE/mortality analysis | Patients with any data available during follow-up. |
Because the endpoints are binary events assessed over defined periods rather than time-to-event endpoints with reported survival curves, Kaplan-Meier estimation and Cox proportional-hazards modeling are not the statistical methods reported for these analyses. The relevant distinction is instead how participants satisfying each analysis population contributed to the binary outcome comparison.
13. Multiplicity and Multiple Secondary Endpoints
The registry contains one registered primary endpoint and multiple secondary endpoint analyses. The ClinicalTrials.gov record reports statistical tests and p-values for these analyses but do not provide a multiplicity-adjustment procedure or an alpha-allocation hierarchy.
| Endpoint role | Number represented in the ClinicalTrials.gov record | Statistical interpretation |
|---|---|---|
| Primary endpoint | 1 | Formal non-inferiority analysis with a reported 7.7% margin. |
| Secondary endpoints | Multiple | Mixture of superiority analyses using Wald / z-test and Fisher exact test. |
14. Primary Analysis in Statistical Detail
The registry reports a 95% confidence interval of -3.79% to 1.64% and a non-inferiority margin of 7.7%.
The analysis uses a difference in binomial risks rather than a ratio measure. That choice has an important consequence: the treatment effect is expressed in percentage points. A negative value favors dabigatran in the direction defined by the comparison, whereas a positive value favors enoxaparin on this absolute-risk scale.
The confidence interval crosses zero, which means the data are compatible with a small negative difference as well as a small positive difference. That fact is not inconsistent with non-inferiority. Non-inferiority does not require the confidence interval to exclude zero; it requires the interval to exclude effects beyond the prespecified unacceptable-loss margin.
This distinction is one of the most important statistical lessons from the trial. A conventional superiority analysis and a non-inferiority analysis can use the same numerical estimate while asking different questions. For RE-NOVATE, the primary question is whether dabigatran remains within the prespecified acceptable-loss boundary relative to enoxaparin.
15. Secondary Endpoint Pattern
| Endpoint | Effect estimate | 95% CI | P-value | Interpretive point |
|---|---|---|---|---|
| Major VTE and VTE-related mortality | RD -1.91% | -3.64% to -0.19% | 0.029 | Interval remains below zero. |
| Proximal DVT | RD -1.79% | -3.47% to -0.12% | 0.0358 | Interval remains below zero. |
| Total DVT | RD -0.96% | -3.65% to 1.73% | 0.4839 | Interval includes zero. |
| Major and clinically relevant bleeding | Absolute difference 0.8% | -0.8% to 2.3% | 0.3305 | Interval includes zero. |
| Any bleeding | Absolute difference 1.4% | -1.1% to 3.9% | 0.2626 | Interval includes zero. |
The pattern illustrates why endpoint-specific interpretation is important. Several secondary analyses report negative risk differences, but their confidence intervals and p-values differ. A treatment effect should not be generalized from one endpoint to another simply because the direction of the point estimates is similar.
16. Fisher Exact Analyses
Four registry-reported secondary analyses use Fisher exact testing. These analyses concern symptomatic DVT, pulmonary embolism, death during the treatment period, and total VTE plus all-cause mortality during follow-up.
Symptomatic DVT
Fisher exact P = 0.0612 for the 28-35 day treatment period.
Pulmonary embolism
Fisher exact P = 0.6231 for the 28-35 day treatment period.
Death
Fisher exact P = 0.4977 for death during the 28-35 day treatment period.
Three-month follow-up
Fisher exact P = 0.6870 for total VTE and all-cause mortality during the 3-month follow-up period.
The absence of an effect estimate in the ClinicalTrials.gov record is important. A p-value tells us about the statistical evidence under the test's null hypothesis, but without the corresponding event proportions or effect estimate it does not show the absolute magnitude or direction of the observed difference.
17. Serious Adverse Events
The ClinicalTrials.gov record provides serious adverse-event counts and denominators for the two treatment groups.
The counts are 57 affected participants among 1010 at risk for dabigatran 220mg and 59 among 1003 at risk for enoxaparin. These are descriptive arm-level data. The ClinicalTrials.gov record does not provide a formal hypothesis test, confidence interval, or risk-difference estimate for this serious-adverse-event comparison.
18. What the Risk Difference Does — and Does Not — Mean
The primary risk difference of -1.08% is an absolute difference between two group-level event risks. It means the estimated risk in the dabigatran 220mg group was 1.08 percentage points lower than the estimated risk in the enoxaparin group for the specified endpoint and treatment period.
It does not mean that each patient had a 1.08% lower individual probability of an event. It also does not mean a 1.08% relative reduction in risk.
The 95% confidence interval of -3.79% to 1.64% shows uncertainty around the estimated risk difference. The interval includes zero, so the data are compatible with no absolute difference as well as with modest differences in either direction.
For non-inferiority, however, zero is not the decisive boundary. The relevant unfavorable boundary is the prespecified 7.7% non-inferiority margin.
The primary P < 0.0001 describes statistical evidence under the non-inferiority hypothesis being tested. It does not say that the treatment effect is large, nor does it give the probability that the observed result occurred by chance.
19. Limitations
- Registry-level evidence: this analysis is restricted to the ClinicalTrials.gov record. The page does not add numerical results from publications or other external sources.
- Incomplete effect reporting for some secondary analyses: several Fisher exact analyses provide p-values but no effect estimate in the ClinicalTrials.gov record, limiting interpretation of magnitude and direction.
- Analysis-population complexity: the primary Full Analysis Set has specific requirements, while secondary endpoints use endpoint-specific populations. Results from different populations should not automatically be compared as though they came from the same participants.
- Multiplicity: the ClinicalTrials.gov record does not specify a multiplicity-adjustment strategy for the multiple secondary analyses. Nominal secondary p-values therefore require caution.
- Non-inferiority interpretation: the primary result must be interpreted against the prespecified 7.7% margin rather than by the p-value alone.
- Normal approximation: the primary risk-difference analysis uses a normal approximation to the independent binomial distribution. The adequacy of that approximation depends on the underlying event counts, which are not posted on ClinicalTrials.gov for the primary endpoint.
- Masking details: the registry classifies the study as double-masked, but the ClinicalTrials.gov record does not identify which parties were masked.
- Safety comparison: serious adverse-event counts are reported by arm, but no formal comparative estimate or confidence interval is posted on ClinicalTrials.gov for that endpoint.
- Follow-up population: the 3-month total VTE and all-cause mortality analysis uses patients with any data available during follow-up, which differs from the primary analysis population.
20. Why This Trial Matters Statistically
RE-NOVATE is a useful teaching case because it demonstrates how the same randomized trial can combine non-inferiority testing, absolute risk differences, confidence intervals, Wald / z-tests, and Fisher exact testing across a set of binary clinical outcomes.
| Concept | How it appears in RE-NOVATE |
|---|---|
| Randomization | Randomized, parallel phase 3 design with 2 treatment arms. |
| Blinding | Registry classification: double-masked. |
| Binary endpoints | Primary and secondary outcomes are reported as participant event counts over defined periods. |
| Risk difference | Primary treatment effect reported as a percentage risk difference. |
| Confidence interval | 95% intervals accompany the primary and several secondary effect estimates. |
| Wald / z-test | Used for the primary and several secondary normal-approximation analyses. |
| Fisher exact test | Used for several secondary binary outcomes. |
| Non-inferiority | Primary endpoint evaluated with a reported 7.7% non-inferiority margin. |
| Analysis populations | Full Analysis Set and endpoint-specific analysis populations are reported. |
| Intention-to-treat concept | Identified in the analysis text for the primary and several secondary analyses. |
| Safety analysis | Bleeding endpoints and serious adverse events are reported separately from thromboembolic efficacy endpoints. |
| Follow-up | A secondary total VTE and all-cause mortality endpoint was assessed at 3 months. |
21. Related Tutorials
Learn more about the methods used in this trial:
22. Related Calculators
23. Sources
- ClinicalTrials.gov: RE-NOVATE, NCT00657150.
- PubMed: PMID 26578849.
- PubMed: PMID 21225098.
Continue through Clinical Biostats
Use the trial as a practical starting point for learning non-inferiority design, binary endpoint analysis, confidence intervals, and categorical-data methods.
24. Record Summary
RE-NOVATE provides a clear example of how statistical interpretation depends on the hypothesis being tested. The primary endpoint was a binary composite assessed during 28-35 days, analyzed as a risk difference using a normal approximation and evaluated under a non-inferiority framework with a reported 7.7% margin. The estimated risk difference was -1.08%, with a 95% confidence interval from -3.79% to 1.64% and P < 0.0001.
The secondary analyses illustrate additional issues. Major VTE and VTE-related mortality had a reported risk difference of -1.91% with a 95% confidence interval of -3.64% to -0.19% and P = 0.029; proximal DVT had a risk difference of -1.79% with a 95% confidence interval of -3.47% to -0.12% and P = 0.0358; while total DVT had a risk difference of -0.96% with a 95% confidence interval of -3.65% to 1.73% and P = 0.4839. Bleeding analyses provide separate safety comparisons, and Fisher exact analyses supply p-values for several additional binary outcomes.