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 reported in the ClinicalTrials.gov record.
1. Trial at a Glance
CANTATA was a completed phase 3 randomized, parallel-group, triple-masked trial in type 2 diabetes. The registry reports an enrollment of 344 and three arms, with 16 statistical analyses posted, including two primary-endpoint analyses.
| Feature | CANTATA |
|---|---|
| Trial name | CANTATA-MP Trial (CANagliflozin Treatment and Trial Analysis - Metformin and Pioglitazone) |
| Phase | Phase 3 |
| Status | Completed |
| Therapeutic area | Endocrinology |
| Condition | Diabetes Mellitus, Type 2 |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Triple |
| Primary purpose | Treatment |
| Enrollment | 344 |
| Lead sponsor | Janssen Research & Development, LLC |
| Start | 2010-06 |
| Primary completion | 2011-11 |
| ClinicalTrials.gov | NCT01106690 |
2. Clinical Question
The central statistical question was whether each canagliflozin dose group differed from the placebo/sitagliptin comparator in the change in HbA1c from baseline to Week 26. The registered primary endpoint was analyzed as a superiority comparison.
Population
The registry identifies the condition as Diabetes Mellitus, Type 2. The ClinicalTrials.gov record does not provide a detailed baseline demographic table or additional eligibility characteristics.
Intervention
The trial data identify canagliflozin as a study intervention, with statistical comparisons reported for canagliflozin 100 mg and canagliflozin 300 mg.
Comparator
The primary statistical analyses compare each canagliflozin group with Placebo/Sitagliptin.
Primary question
Does canagliflozin produce a different change in HbA1c from baseline to Week 26 compared with placebo/sitagliptin?
3. Trial Design
Placebo/Sitagliptin
- Placebo and sitagliptin are identified in the trial intervention data.
- This group serves as the comparator in the posted primary and secondary analyses.
Canagliflozin
- Canagliflozin 100 mg
- Canagliflozin 300 mg
- Each dose is separately compared with placebo/sitagliptin in the posted analyses.
The ClinicalTrials.gov record also list metformin and pioglitazone among the interventions. The available trial data do not provide a full treatment-regimen description sufficient to reconstruct every component of each arm, so this page does not infer additional regimen details.
4. Endpoints
Primary Endpoint
| Endpoint | Definition / time frame | Posted analysis |
|---|---|---|
| Change in HbA1c From Baseline to Week 26 | Day 1 (Baseline) and Week 26. The registry describes the table as showing the least-squares mean change in HbA1c from Baseline to Week 26 for each treatment group, with treatment differences defined as each canagliflozin group minus placebo. | ANCOVA; least-squares mean difference |
Secondary Endpoints With Posted Statistical Analyses
| Endpoint | Time frame | Method | Effect measure |
|---|---|---|---|
| Percentage of Patients With HbA1c <7% at Week 26 | Week 26 | Logistic regression | Odds ratio |
| Change in Fasting Plasma Glucose (FPG) From Baseline to Week 26 | Day 1 (Baseline) and Week 26 | ANCOVA | Least-squares mean difference |
| Change in Homeostasis Model Assessment (HOMA2-%B) From Baseline to Week 26 | Day 1 (Baseline) and Week 26 | ANCOVA | Least-squares mean difference |
| Percent Change in Body Weight From Baseline to Week 26 | Day 1 (Baseline) and Week 26 | ANCOVA | Least-squares mean difference |
| Change in Systolic Blood Pressure (SBP) From Baseline to Week 26 | Day 1 (Baseline) and Week 26 | ANCOVA | Least-squares mean difference |
| Percent Change in Triglycerides From Baseline to Week 26 | Day 1 (Baseline) and Week 26 | ANCOVA | Least-squares mean difference |
| Percent Change in High-density Lipoprotein Cholesterol (HDL-C) From Baseline to Week 26 | Day 1 (Baseline) and Week 26 | ANCOVA | Least-squares mean difference |
5. Analysis Population and Missing Data
The posted analyses used the mITT analysis set, defined in the ClinicalTrials.gov record as all randomized patients who received at least 1 dose of study drug.
All randomized patients who received at least 1 dose of study drug.
The registry states that LOCF was used for missing Week 26 values.
The analysis population is important because the estimate is not described as coming from every randomized patient regardless of treatment exposure. Instead, the registry wording explicitly defines the mITT set around randomization plus receipt of at least 1 dose.
The use of last-observation-carried-forward also means that a missing Week 26 measurement was not simply ignored in these analyses. An earlier observed value was carried forward according to the stated method. That creates a specific modeling assumption about what the unobserved Week 26 value should be represented by; it should therefore be considered part of the interpretation of the reported estimates.
6. Statistical Methodology
ANCOVA for continuous change endpoints
The primary endpoint and most posted secondary continuous endpoints were analyzed using analysis of covariance (ANCOVA). The registry reports the effect measure as a least-squares mean difference.
For a baseline-adjusted clinical endpoint, ANCOVA estimates treatment-group differences while accounting for baseline information included in the model. The ClinicalTrials.gov record identifies the method as ANCOVA but do not provide the complete model specification or list every covariate.
The distinction between an ordinary arithmetic mean difference and a least-squares mean difference matters. A least-squares mean is a model-based adjusted mean. Consequently, the reported estimate is not necessarily obtained by simply subtracting the raw observed group means.
Logistic regression for HbA1c target attainment
The secondary endpoint "Percentage of Patients With HbA1c <7% at Week 26" was analyzed using logistic regression, with an odds ratio as the effect measure.
An odds ratio above 1 indicates higher odds of meeting the HbA1c <7% criterion in the canagliflozin group under the fitted model. It is not itself a risk ratio or a percentage-point difference in response rates.
Superiority framework
The primary and secondary analyses are identified as superiority analyses. The question is therefore whether the estimated treatment difference is compatible with no difference and whether the observed data provide statistical evidence for a difference under the specified testing framework.
Confidence intervals
The primary analyses report two-sided 95% confidence intervals. A confidence interval describes statistical uncertainty around the estimated treatment difference under the model and sampling framework. It does not describe the range of individual patient responses.
P-values
The registry reports p-values alongside the effect estimates and confidence intervals. A p-value quantifies the compatibility of the observed data with a specified null hypothesis under the statistical model; it does not measure the magnitude or clinical importance of the treatment effect.
7. Primary Results: HbA1c Change at Week 26
The primary endpoint was "Change in HbA1c From Baseline to Week 26," assessed from Day 1 (Baseline) to Week 26. Both posted primary analyses used the mITT analysis set, with LOCF for missing Week 26 values, and both used ANCOVA.
Canagliflozin 100 mg vs Placebo/Sitagliptin
Least-squares mean difference in HbA1c change
95% CI: -0.811 to -0.437 · P < 0.001
Effect measure: least-squares mean difference; two-sided 95% CI; superiority analysis.
The estimate of -0.62 means that the modeled least-squares mean change in HbA1c was 0.62 percentage points lower in the canagliflozin 100 mg group than in the placebo/sitagliptin group, based on the reported comparison.
The estimate does not mean that every patient experienced a 0.62 percentage-point reduction, nor does it describe the absolute HbA1c value at Week 26. It is a between-group difference in modeled change from baseline.
The 95% CI from -0.811 to -0.437 describes the uncertainty around that estimated difference. Because the entire interval is below 0, the interval is consistent with a negative treatment difference under the reported model.
The P < 0.001 value is evidence against the corresponding null hypothesis under the reported statistical test. It does not indicate that the treatment effect is "greater" because the p-value is smaller, and it does not measure the clinical magnitude of the HbA1c difference.
Interpretation should also retain the stated analysis population and missing-data method: this result is based on the mITT set with LOCF for missing Week 26 values. The ClinicalTrials.gov record does not provide enough detail to independently assess the full ANCOVA covariate specification.
Canagliflozin 300 mg vs Placebo/Sitagliptin
Least-squares mean difference in HbA1c change
95% CI: -0.951 to -0.575 · P < 0.001
Effect measure: least-squares mean difference; two-sided 95% CI; superiority analysis.
The estimate of -0.76 means that the modeled least-squares mean change in HbA1c was 0.76 percentage points lower in the canagliflozin 300 mg group than in the placebo/sitagliptin group, using the reported analysis.
Again, this is a between-group mean difference in change, not a statement that each individual patient experienced a 0.76 percentage-point reduction. It also does not directly provide the proportion of patients who achieved a particular HbA1c target.
The 95% CI from -0.951 to -0.575 quantifies uncertainty around the estimated least-squares mean difference. The entire interval lies below 0, so the reported interval is consistent with a negative difference throughout its stated range.
The P < 0.001 result addresses statistical evidence against the null hypothesis; it should not be interpreted as a measure of effect size. The magnitude of the treatment difference is communicated by the estimate and its confidence interval.
As with the 100 mg comparison, the interpretation is conditioned on the mITT analysis set and LOCF handling of missing Week 26 observations.
| Primary comparison | Estimate | 95% CI | P-value | Method |
|---|---|---|---|---|
| Canagliflozin 100 mg vs placebo/sitagliptin | -0.62 | -0.811 to -0.437 | <0.001 | ANCOVA |
| Canagliflozin 300 mg vs placebo/sitagliptin | -0.76 | -0.951 to -0.575 | <0.001 | ANCOVA |
8. Secondary Results: HbA1c <7% at Week 26
The registry reports a binary secondary endpoint measuring the percentage of patients with HbA1c <7% at Week 26. Logistic regression was used, with odds ratios as the effect measure.
Canagliflozin 100 mg vs Placebo/Sitagliptin
Odds ratio for HbA1c <7%
95% CI: 1.26 to 4.57 · P = 0.007
An odds ratio of 2.40 means that the estimated odds of having HbA1c <7% at Week 26 were 2.40 times the corresponding odds in the placebo/sitagliptin group, under the reported logistic regression analysis.
Canagliflozin 300 mg vs Placebo/Sitagliptin
Odds ratio for HbA1c <7%
95% CI: 2.73 to 10.60 · P < 0.001
An odds ratio of 5.38 means that the estimated odds of having HbA1c <7% at Week 26 were 5.38 times the corresponding odds in the placebo/sitagliptin group, under the reported logistic regression analysis.
| Primary comparison | Odds ratio | 95% CI | P-value | Method |
|---|---|---|---|---|
| Canagliflozin 100 mg vs placebo/sitagliptin | 2.40 | 1.26 to 4.57 | 0.007 | Logistic regression |
| Canagliflozin 300 mg vs placebo/sitagliptin | 5.38 | 2.73 to 10.60 | <0.001 | Logistic regression |
9. Secondary Results: Fasting Plasma Glucose
Canagliflozin 100 mg
Least-squares mean difference in FPG change
95% CI: -36.96 to -21.78 · P < 0.001
Canagliflozin 300 mg
Least-squares mean difference in FPG change
95% CI: -43.30 to -28.11 · P < 0.001
Both FPG comparisons were analyzed with ANCOVA in the mITT analysis set using LOCF for missing Week 26 values. The estimates are between-group differences in change from baseline, not absolute Week 26 FPG values.
10. Secondary Results: HOMA2-%B
Canagliflozin 100 mg
Least-squares mean difference in HOMA2-%B change
95% CI: 9.315 to 19.236 · P < 0.001
Canagliflozin 300 mg
Least-squares mean difference in HOMA2-%B change
95% CI: 12.293 to 22.166 · P < 0.001
The positive estimates indicate higher modeled changes in HOMA2-%B relative to placebo/sitagliptin in the reported comparisons. The registry supplies the endpoint as HOMA2-%B and reports ANCOVA with a least-squares mean difference; the ClinicalTrials.gov record does not provide additional clinical interpretation of this measure.
11. Secondary Results: Body Weight
Canagliflozin 100 mg
Least-squares mean difference in percent change in body weight
95% CI: -3.6% to -1.8% · P < 0.001
Canagliflozin 300 mg
Least-squares mean difference in percent change in body weight
95% CI: -4.6% to -2.8% · P < 0.001
These estimates represent modeled between-group differences in percent change from baseline to Week 26. They do not represent the percentage of patients losing weight, nor do they establish that the same magnitude occurred in every participant.
12. Secondary Results: Systolic Blood Pressure
Canagliflozin 100 mg
Least-squares mean difference in SBP change
95% CI: -6.879 to -1.251 · P = 0.005
Canagliflozin 300 mg
Least-squares mean difference in SBP change
95% CI: -6.281 to -0.643 · P = 0.016
The registry reports both SBP analyses as ANCOVA comparisons of change from baseline to Week 26. The confidence intervals quantify uncertainty around the modeled treatment differences; the p-values address statistical evidence against the corresponding null hypotheses.
13. Secondary Results: Triglycerides
Canagliflozin 100 mg
Least-squares mean difference in percent change
95% CI: -12.1% to -0.9% · P = 0.034
Canagliflozin 300 mg
Least-squares mean difference in percent change
95% CI: -28.1% to -5.8% · P = 0.003
14. Secondary Results: HDL-C
Canagliflozin 100 mg
Least-squares mean difference in percent change
95% CI: 1.2 to 8.5 · P = 0.010
Canagliflozin 300 mg
Least-squares mean difference in percent change
95% CI: 2.8 to 10.2 · P < 0.001
The positive estimates indicate higher modeled percent changes in HDL-C relative to placebo/sitagliptin in the reported comparisons. As with the other ANCOVA results, the estimates are adjusted between-group differences rather than raw within-person changes.
15. Complete Statistical Results Table
The table below brings the posted analyses together in one place. Every estimate and interval is reproduced from the ClinicalTrials.gov record without rounding or recalculation.
| Endpoint | Comparison | Method | Estimate | 95% CI | P-value |
|---|---|---|---|---|---|
| Change in HbA1c | Canagliflozin 100 mg vs placebo/sitagliptin | ANCOVA | -0.62 | -0.811 to -0.437 | <0.001 |
| Change in HbA1c | Canagliflozin 300 mg vs placebo/sitagliptin | ANCOVA | -0.76 | -0.951 to -0.575 | <0.001 |
| HbA1c <7% | Canagliflozin 100 mg vs placebo/sitagliptin | Logistic regression | OR 2.40 | 1.26 to 4.57 | 0.007 |
| HbA1c <7% | Canagliflozin 300 mg vs placebo/sitagliptin | Logistic regression | OR 5.38 | 2.73 to 10.60 | <0.001 |
| Change in FPG | Canagliflozin 100 mg vs placebo/sitagliptin | ANCOVA | -29.4 | -36.96 to -21.78 | <0.001 |
| Change in FPG | Canagliflozin 300 mg vs placebo/sitagliptin | ANCOVA | -35.7 | -43.30 to -28.11 | <0.001 |
| Change in HOMA2-%B | Canagliflozin 100 mg vs placebo/sitagliptin | ANCOVA | 14.28 | 9.315 to 19.236 | <0.001 |
| Change in HOMA2-%B | Canagliflozin 300 mg vs placebo/sitagliptin | ANCOVA | 17.23 | 12.293 to 22.166 | <0.001 |
| Percent change in body weight | Canagliflozin 100 mg vs placebo/sitagliptin | ANCOVA | -2.7 | -3.6 to -1.8 | <0.001 |
| Percent change in body weight | Canagliflozin 300 mg vs placebo/sitagliptin | ANCOVA | -3.7 | -4.6 to -2.8 | <0.001 |
| Change in SBP | Canagliflozin 100 mg vs placebo/sitagliptin | ANCOVA | -4.07 | -6.879 to -1.251 | 0.005 |
| Change in SBP | Canagliflozin 300 mg vs placebo/sitagliptin | ANCOVA | -3.46 | -6.281 to -0.643 | 0.016 |
| Percent change in triglycerides | Canagliflozin 100 mg vs placebo/sitagliptin | ANCOVA | -12.1 | -12.1 to -0.9 | 0.034 |
| Percent change in triglycerides | Canagliflozin 300 mg vs placebo/sitagliptin | ANCOVA | -16.9 | -28.1 to -5.8 | 0.003 |
| Percent change in HDL-C | Canagliflozin 100 mg vs placebo/sitagliptin | ANCOVA | 4.8 | 1.2 to 8.5 | 0.010 |
| Percent change in HDL-C | Canagliflozin 300 mg vs placebo/sitagliptin | ANCOVA | 6.5 | 2.8 to 10.2 | <0.001 |
16. Statistical Methods Explained
Why was ANCOVA used?
ANCOVA is a natural framework for continuous follow-up outcomes and change-from-baseline analyses because it can estimate adjusted treatment-group differences while accounting for baseline information included in the model. In CANTATA, the registry explicitly identifies ANCOVA as the method for the primary HbA1c endpoint and most secondary continuous endpoints.
What does a least-squares mean difference of -0.62 mean?
It means the modeled change in HbA1c from baseline to Week 26 was estimated to be 0.62 percentage points lower for canagliflozin 100 mg than for placebo/sitagliptin. The "least-squares" qualifier indicates that the estimate comes from the statistical model rather than necessarily being a simple subtraction of raw sample means.
What does an odds ratio of 2.40 mean?
For the Week 26 HbA1c <7% endpoint, an OR of 2.40 means that the estimated odds of meeting the endpoint were 2.40 times the odds in the placebo/sitagliptin group. It does not mean that the probability was 2.40 times as large, because odds and probability are different quantities.
Why report a confidence interval as well as a p-value?
The p-value addresses evidence against a null hypothesis, whereas the confidence interval communicates the estimated effect and its statistical precision. In CANTATA, the intervals also show the range of treatment differences compatible with the reported model and data under the stated confidence framework.
Why does LOCF matter?
Missing Week 26 measurements were handled using last-observation-carried-forward. That means missing observations were represented using the last available observation rather than being discarded from the analysis. The resulting estimate therefore depends partly on the assumption implicit in that imputation rule.
What does the mITT population mean here?
The registry definition is all randomized patients who received at least 1 dose of study drug. Thus, the posted estimates are not described as an analysis of every randomized patient regardless of whether study treatment was received.
Can the p-value tell us whether the HbA1c difference is clinically important?
No. A p-value is a measure of statistical evidence under a specified null hypothesis. Clinical importance requires consideration of the magnitude of the effect, its uncertainty, the endpoint itself, and the broader clinical context. The p-value alone cannot make that determination.
17. Multiple Comparisons and Interpretation of the Secondary Analyses
The ClinicalTrials.gov record identifies two primary analyses and a larger set of secondary statistical analyses. The primary endpoint was evaluated separately for canagliflozin 100 mg and 300 mg versus placebo/sitagliptin, while additional endpoints were also tested.
Primary endpoint structure
Two primary analyses were posted for the same registered HbA1c endpoint: one comparing canagliflozin 100 mg with placebo/sitagliptin and one comparing canagliflozin 300 mg with placebo/sitagliptin.
Secondary endpoint structure
The registry reports additional analyses covering HbA1c target attainment, FPG, HOMA2-%B, body weight, SBP, triglycerides, and HDL-C.
Because multiple treatment comparisons and multiple endpoints are reported, the collection of p-values should not automatically be interpreted as though every one were an independent confirmatory hypothesis test. The ClinicalTrials.gov record identifies the hypothesis type as superiority but do not provide a complete multiplicity-adjustment strategy. This page therefore reports the nominal values exactly as registered without assigning them a stronger confirmatory interpretation than the source supports.
18. Safety Results
The ClinicalTrials.gov record reports serious adverse events by treatment group for two different time frames. Because the denominators and time frames differ, the figures should be read exactly as reported rather than combined into a single safety rate.
| Group | Time frame | Serious adverse events | Affected / at risk |
|---|---|---|---|
| Placebo/Sitagliptin | Baseline to Week 26 | Serious adverse events | 5 / 115 |
| Canagliflozin 100 mg | Baseline to Week 2 | Serious adverse events | 3 / 113 |
| Canagliflozin 300 mg | Baseline to Week 2 | Serious adverse events | 4 / 114 |
| Placebo/Sitagliptin | Baseline to Week 52 | Serious adverse events | 6 / 115 |
| Canagliflozin 100 mg | Baseline to Week 5 | Serious adverse events | 8 / 113 |
| Canagliflozin 300 mg | Baseline to Week 5 | Serious adverse events | 7 / 114 |
19. What the Effect Measures Do — and Do Not — Mean
A least-squares mean difference such as -0.62 is a model-based estimate of the difference between treatment groups in the analyzed outcome. For CANTATA's HbA1c endpoint, it represents a difference in change from baseline to Week 26.
It does not mean that every individual patient had the same change, and it does not provide the distribution of individual treatment responses.
An odds ratio such as 2.40 compares odds, not probabilities. The relationship between odds and probability is nonlinear, so an odds ratio should not be restated as a percentage increase in the probability of achieving HbA1c <7% without additional information.
A 95% confidence interval such as -0.811 to -0.437 describes statistical uncertainty around the estimated treatment difference under the specified model and sampling framework. It is not a range containing 95% of individual patient responses.
A p-value such as <0.001 is evidence concerning a null hypothesis under the statistical test. It is not an effect-size metric, does not quantify clinical importance, and should not be used as a substitute for the estimate and confidence interval.
20. Why the Primary Results Are Statistically Informative
The primary endpoint provides a useful example of how several statistical quantities should be interpreted together rather than separately.
| Component | What it contributes |
|---|---|
| Randomization | Provides the design basis for comparing the randomized treatment groups. |
| mITT analysis set | Defines which randomized participants were included in the posted efficacy analyses. |
| ANCOVA | Provides the reported model for the continuous change endpoint. |
| Least-squares mean difference | Expresses the modeled between-group difference in change. |
| 95% confidence interval | Shows uncertainty and precision around the estimated difference. |
| P-value | Addresses statistical evidence against the relevant null hypothesis. |
| LOCF | Defines how missing Week 26 values were handled in the posted analysis. |
| Superiority hypothesis | Frames the analysis as a test for a treatment difference rather than non-inferiority. |
The most informative reading therefore starts with the effect estimate, examines the confidence interval, considers the analysis population and missing-data method, and only then interprets the p-value. This sequence helps prevent a common error in clinical-trial reporting: treating statistical significance as if it were synonymous with effect magnitude.
21. Important Limitations and Interpretation Issues
- Incomplete model specification: the ClinicalTrials.gov record identifies ANCOVA and logistic regression but do not provide the complete statistical model specification or all covariates used.
- LOCF assumptions: missing Week 26 observations were handled with last-observation-carried-forward. This is an imputation assumption and can affect the resulting estimates.
- mITT population: the posted efficacy analyses used all randomized patients who received at least 1 dose of study drug rather than a purely all-randomized analysis population.
- Multiple analyses: the registry contains two primary analyses and multiple secondary analyses. The ClinicalTrials.gov record does not provide a complete multiplicity-adjustment strategy.
- Registry endpoint-type inconsistencies: some outcomes are labeled "Binary" in the endpoint-type field even though the associated analysis is ANCOVA with a least-squares mean difference. The page reports the statistical method and effect measure as reported in the registry.
- Safety time frames differ: the serious-adverse-event counts posted on ClinicalTrials.gov for different groups use different follow-up windows and should not be interpreted as a single harmonized safety comparison.
- No baseline table reported: the trial data do not provide detailed baseline demographic or clinical characteristics, so no baseline balance analysis is presented.
- No subgroup analyses reported: the data do not contain subgroup estimates or interaction tests, so treatment-effect heterogeneity cannot be assessed from the ClinicalTrials.gov record.
- No time-to-event analysis: the statistical analyses posted on ClinicalTrials.gov are ANCOVA and logistic regression. Survival-analysis methods such as Kaplan-Meier estimation and Cox proportional-hazards modeling are therefore not part of this analysis.
22. Why This Trial Matters Statistically
CANTATA is useful as a teaching case because its registry results illustrate several core ideas in clinical-trial biostatistics without requiring a time-to-event endpoint.
| Concept | How it appears in CANTATA |
|---|---|
| Randomization | The study used randomized allocation in a parallel design. |
| Blinding | The registry describes the trial as triple masked. |
| ANCOVA | Used for the primary HbA1c change endpoint and multiple continuous secondary endpoints. |
| Least-squares mean difference | Used as the reported effect measure for ANCOVA analyses. |
| Logistic regression | Used for the binary HbA1c <7% endpoint. |
| Odds ratio | Expresses the treatment comparison for the HbA1c <7% endpoint. |
| Confidence intervals | Quantify statistical uncertainty around the reported estimates. |
| P-values | Provide hypothesis-test evidence but do not measure effect magnitude. |
| Missing-data handling | LOCF was used for missing Week 26 values. |
| Analysis population | The posted analyses used an mITT set consisting of randomized patients who received at least 1 dose. |
| Multiple endpoints | Primary and secondary outcomes generated a broad set of statistical comparisons. |
23. Related Tutorials
Learn more about the methods used in this trial:
24. Related Calculators
25. Sources
- ClinicalTrials.gov: CANTATA-MP Trial (NCT01106690).
- PubMed: PMID 28327140.
- PubMed: PMID 28241822.
- PubMed: PMID 28197834.
- PubMed: PMID 27977934.
- PubMed: PMID 27600862.
Continue through the Clinical Biostats knowledge graph
Connect the statistical methods used in CANTATA with deeper tutorials and calculators for ANCOVA, logistic regression, confidence intervals, odds ratios, p-values, and randomized-trial design.
26. Record Summary
CANTATA provides a useful example of a randomized phase 3 trial in which the principal statistical analysis is based on ANCOVA rather than time-to-event methods. The registered primary endpoint was change in HbA1c from baseline to Week 26, and the posted analyses estimated least-squares mean differences for canagliflozin 100 mg and 300 mg versus placebo/sitagliptin. Additional analyses used logistic regression and odds ratios for the HbA1c <7% endpoint, while ANCOVA was used for several metabolic and cardiovascular measures.
The most informative statistical reading combines the effect estimate, its confidence interval, the p-value, the mITT analysis population, and the LOCF missing-data method. The secondary results add information across several physiological endpoints, but their interpretation should remain distinct from the primary endpoint because the ClinicalTrials.gov record does not provide a complete multiplicity strategy.