The CANTATA-M (CANagliflozin Treatment and Trial Analysis - Monotherapy) Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for The CANTATA-M (CANagliflozin Treatment and Trial Analysis - Monotherapy) Trial (NCT01081834), a randomized phase 3 study. The primary outcome is Change in HbA1c From Baseline to Week 26 (Main Study).

2. Clinical Question

The central statistical question concerns the estimated difference in average change, rather than whether every participant experienced a particular change. A negative treatment difference reflects a more negative estimated HbA1c change in the canagliflozin group than in the placebo group under the registry’s comparison convention. These group summaries do not describe the range of individual responses, and they do not by themselves establish what the result means for an individual patient.

3. Trial Design

CANTATA-M (NCT01081834) was a Phase 3 randomized, triple-masked, parallel-group study comparing canagliflozin with placebo as monotherapy. The official title describes a double-blind, placebo-controlled, parallel-group, multicenter study. Its stated purpose was to evaluate efficacy, safety, and tolerability in people with type 2 diabetes inadequately controlled with diet and exercise. The registry record lists 678 participants enrolled and a completed study.

The listed study start was 2010-03, and primary completion was 2011-08. The treatment groups included canagliflozin 100 mg, canagliflozin 300 mg, and placebo followed by sitagliptin. In the main study, placebo was used through Week 26; participants in that group then switched to sitagliptin through Week 52. The canagliflozin groups continued their assigned dose through Week 52 in the main study. A separate high glycemic substudy used canagliflozin treatment through Week 26.

The primary endpoint and the reported sponsor comparisons focus on the main study through Week 26. The primary analysis population was the mITT set: randomized participants who received at least one dose of study drug. Missing Week 26 values were handled using last observation carried forward, and measurements taken before glycemic rescue were used as the last observation for participants receiving rescue therapy. Further details about allocation procedures, the exact masking roles, the protocol’s original sample-size calculation, and the full statistical analysis plan are not reported in the registry record.

The post-Week-26 switch from placebo to sitagliptin matters when interpreting the longer-term descriptions. The Week 52 placebo/sitagliptin group is not a continuing placebo group. It therefore does not supply a placebo-controlled comparison through Week 52 in the way the primary Week 26 contrast does.

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: 678 patients were enrolled into the study; 587 patients in the main study and 91 patients in the high glycemic substudy. 584 patients in the main study and all 91 patients in the high glycemic substudy received at least one dose of study drug and were included in the modified intent-to-treat (mITT) analyses sets and the safety analyses sets.

Core Period: Baseline to Week 26

Reasons for not completing this period:

Extension Period: Week 26 to Week 52

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Analysis used mITT analysis set (all randomized patients who received at least 1 dose of study drug). Last-observation-carried-forward method used for missing Week 26 values. Measurements taken pre-rescue used as last observation in patients receiving glycemic rescue therapy. Table includes only patients with both baseline and post baseline values

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 Change in HbA1c From Baseline to Week 26 (Main Study), measured over Day 1 (Baseline) and Week 26.

The primary endpoint is change in HbA1c from baseline to Week 26 in the main study, expressed in percent. HbA1c reflects blood glucose over time and is commonly used as a continuous measure in diabetes trials. Measuring change allows the analysis to compare how the group average moved from baseline during the study period. The registry reports least-squares means, which are model-based adjusted estimates, and the treatment differences between each canagliflozin dose and placebo.

The analysis description identifies the mITT population and the handling of missing values: last observation carried forward for missing Week 26 measurements, with pre-rescue measurements carried forward for participants who received glycemic rescue therapy. These choices affect which observations contribute to the endpoint and how missing or post-rescue data are represented. The registry record does not report all details of the model specification or the rationale for those choices.

As a continuous endpoint, HbA1c change preserves information about the magnitude and direction of change rather than reducing each participant’s result to a yes-or-no threshold. A separate secondary outcome reports the percentage of patients whose HbA1c was below 7% at Week 26. That threshold-based measure addresses a different question: how many participants reached a specified level, rather than the average amount of change. Neither endpoint alone describes every participant’s course or resolves the broader clinical meaning of a treatment contrast.

7. Results Posted in the Registry

• Canagliflozin 100 mg: mean -0.77 versus 0.14 on control; difference -0.91 (95% CI -1.09 to -0.73), standardized effect d = -1.02; z-test p = <0.001.
• Canagliflozin 300 mg: mean -1.03 versus 0.14 on control; difference -1.17 (95% CI -1.35 to -0.99), standardized effect d = -1.31; z-test p = <0.001.

8. Secondary Efficacy Endpoints

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

Percentage of Patients With HbA1c <7% at Week 26 (Main Study)

• 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:
  • Canagliflozin 100 mg: 44.5% (85/191) versus 20.6% (39/189) on control; risk difference 23.9 percentage points (95% CI 14.8 to 33.0), risk ratio 2.16 (1.56–2.97); unadjusted two-proportion z-test p = <0.001.
  • Canagliflozin 300 mg: 62.4% (121/194) versus 20.6% (39/189) on control; risk difference 41.7 percentage points (95% CI 32.8 to 50.7), risk ratio 3.02 (2.24–4.08); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Canagliflozin 100 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 higher for Canagliflozin 300 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change in Fasting Plasma Glucose (FPG) From Baseline to Week 26 (Main Study)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Canagliflozin 100 mg: mean -27.20 versus 8.33 on control; difference -35.53 (95% CI -42.27 to -28.79), standardized effect d = -1.07; z-test p = <0.001.
  • Canagliflozin 300 mg: mean -35.00 versus 8.33 on control; difference -43.33 (95% CI -50.04 to -36.62), standardized effect d = -1.31; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Canagliflozin 100 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Canagliflozin 300 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change in 2-hour Post-prandial Glucose From Baseline to Week 26 (Main Study)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Canagliflozin 100 mg: mean -42.90 versus 5.19 on control; difference -48.09 (95% CI -59.15 to -37.03), standardized effect d = -1.02; z-test p = <0.001.
  • Canagliflozin 300 mg: mean -58.80 versus 5.19 on control; difference -63.99 (95% CI -75.02 to -52.96), standardized effect d = -1.36; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Canagliflozin 100 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Canagliflozin 300 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes cover several related measures: the percentage of participants with HbA1c below 7%, fasting plasma glucose change, post-prandial glucose change, body weight, systolic blood pressure, triglycerides, and HDL-C. The registry lists 14 secondary outcomes in total. The numerical secondary results described in the posted record include the threshold-based HbA1c outcome and changes in fasting plasma glucose and 2-hour post-prandial glucose. Results for the other listed secondary outcomes are not reported in the registry record.

The available secondary results point in a direction broadly consistent with the primary HbA1c comparison: the reported canagliflozin groups had larger changes in glucose-related measures than placebo, and a higher reported share reached the HbA1c threshold. These endpoints are related to the primary question but are not substitutes for it. The threshold outcome converts HbA1c into a category, while fasting and post-prandial glucose describe different glucose measures. They can add context, but they do not independently establish the meaning of the primary endpoint.

Secondary findings generally carry less weight when many outcomes are tested and the registry does not report how multiplicity was handled. If several hypotheses are examined, the chance of at least one apparently strong result can increase. General statistical explanation: adjustment methods or a prespecified testing hierarchy can address some multiplicity concerns, but their use and details must be reported to assess them. The registry record does not state whether such a strategy was used for the secondary outcomes.

Accordingly, the secondary results are best treated as supportive descriptions of related measures, not as a collection of separate confirmatory conclusions. The primary endpoint remains the prespecified focus identified by the registry, while the range of outcomes and the unavailable results limit a comprehensive assessment of the secondary evidence.

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

• Percentage of Patients With HbA1c <7% at Week 26 (Main Study)
• Change in Fasting Plasma Glucose (FPG) From Baseline to Week 26 (Main Study)
• Change in 2-hour Post-prandial Glucose From Baseline to Week 26 (Main Study)
• Percent Change in Body Weight From Baseline to Week 26 (Main Study)
• Change in Systolic Blood Pressure (SBP) From Baseline to Week 26 (Main Study)
• Percent Change in Triglycerides From Baseline to Week 26 (Main Study)
• Percent Change in High-density Lipoprotein Cholesterol (HDL-C) From Baseline to Week 26 (Main Study)
• Percentage of Patients With HbA1c <7% at Week 26 (High Glycemic Substudy)

9. Statistical Interpretation

10. Statistical Methods Explained

What does a least-squares mean represent?

General statistical explanation: a least-squares mean is an estimate from a statistical model, rather than simply the arithmetic average of the observed measurements. Model-based estimates can account for variables included in the model and present group estimates on a common adjusted basis. The registry identifies the primary analysis as ANCOVA and reports least-squares mean changes and their contrasts. The precise covariates and full model specification are not reported in the registry record, so the adjustment should not be described in more detail.

The main result is a contrast: the estimated change in one canagliflozin group minus the estimated change in placebo. A negative contrast means the estimated change is more negative in the canagliflozin group under that subtraction order. This direction is important because reversing the order would reverse the sign without changing the underlying comparison.

What does ANCOVA contribute?

General statistical explanation: analysis of covariance (ANCOVA) compares outcome values across groups while incorporating one or more covariates into a model. In a clinical trial, baseline measurement is often relevant to a change outcome, but the exact variables used are a protocol and analysis-plan matter. The registry identifies ANCOVA for the sponsor-reported primary comparisons, but the full list of model terms and assumptions is not reported in the registry record. The adjusted estimates from that model are therefore the appropriate primary figures to quote from the posted results.

How should the recalculations be distinguished?

The approximate recalculations are simple, unadjusted comparisons derived from posted summary statistics. They are not a reproduction of ANCOVA. For the primary endpoint, those recalculated mean differences are -0.91 for canagliflozin 100 mg versus placebo and -1.17 for canagliflozin 300 mg versus placebo. The corresponding sponsor-reported ANCOVA estimates are -0.91 and -1.16. Since the recalculation uses summary information and does not reproduce the adjusted model, it is best read as a separate check of the direction and approximate scale, not as a replacement for the registry’s analysis.

The spread used in the continuous-outcome recalculations is approximate: it is inferred from reported standard errors and sample sizes. That approximation is useful for a simple comparison but does not recover participant-level data, the original model, or any adjustments used in the sponsor analysis. In particular, the recalculated p-values are unadjusted and should not be mistaken for the ANCOVA results.

What do p-values say?

General statistical explanation: a p-value describes how compatible the observed data are with a specified null model, under the assumptions of the test. A small p-value is not the probability that the null hypothesis is true, nor does it quantify the size or clinical importance of an effect. The sponsor reports p <0.001 for both primary comparisons. The registry record does not report a detailed multiplicity strategy for these comparisons, so interpretation should not imply that all possible testing adjustments have been accounted for.

Why does over-dispersion matter?

General statistical explanation: over-dispersion occurs when the variability in data is greater than expected under a model’s assumptions. If a model underestimates variability, its standard errors can be too small and its confidence intervals too narrow. The primary endpoint is analyzed as a continuous measure using ANCOVA, and the registry record does not identify a count model or report an over-dispersion assessment. Therefore, over-dispersion is not a specific reported issue for this primary analysis; the general concept is relevant when choosing or checking models for other kinds of data, especially counts.

How are missing and rescue measurements handled?

The registry states that last observation carried forward was used for missing Week 26 values. For participants receiving glycemic rescue therapy, measurements taken before rescue were used as the last observation. General statistical explanation: carrying an earlier observation forward fills a later missing value with a prior measurement, making an assumption about the unobserved value. This approach does not directly represent uncertainty about the missing measurement, and its consequences depend on why data are missing and how outcomes change over time. The registry does not report enough detail to evaluate those assumptions or compare this approach with alternative missing-data analyses.

What is the analysis population?

The primary analysis used the mITT analysis set, defined as all randomized participants who received at least one dose of study drug. This is narrower than all randomized participants if any randomized participant did not receive a dose. The registry does not provide the full participant flow or a separate sensitivity analysis in the registry record described for the primary endpoint. Readers should therefore understand the estimate as applying to the stated analysis set and the specified endpoint-handling rules.

11. Confidence Intervals and Precision

The sponsor-reported 95% confidence interval for the canagliflozin 100 mg comparison ranges from -1.088 to -0.729, around a least-squares mean difference of -0.91. For canagliflozin 300 mg, the interval ranges from -1.342 to -0.985, around a difference of -1.16. These intervals describe uncertainty in the estimated group contrasts under the reported analysis; they are not ranges containing a specified proportion of individual participants’ outcomes.

General statistical explanation: a confidence interval gives a range of estimates compatible with the data and the statistical procedure at the stated confidence level. Its width reflects both the observed variability and the amount of information available, along with features of the analysis. An interval that excludes zero indicates that zero is not among the values in that interval. The registry reports p <0.001 for both comparisons as well, but the interval is more informative about the range of contrast values compatible with the model than a thresholded p-value alone.

The intervals are estimates for group mean differences, not prediction intervals for a future participant. They also do not incorporate uncertainty about unreported design choices, model specification, missing-data assumptions, or the broader set of outcomes tested. The registry record does not report the full statistical analysis plan or a multiplicity adjustment for the primary comparisons, so those aspects cannot be assessed from the posted summary.

The endpoints are reported with different degrees of rounding. The sponsor’s adjusted results should be interpreted using their reported precision; the approximate summary-statistic recalculations should not be treated as more precise merely because a recalculation can produce additional digits. Precision in the displayed number is not the same as certainty about the result.

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 expressed as participants per arm for continuous and binary outcomes in these comparisons. They are not total event counts: the trial outcomes summarized in this analysis are not time-to-event calculations. Read each row as a hypothetical planning scenario based on detecting a stated fraction of the observed effect, not as the actual enrollment target or the number enrolled in CANTATA-M.

The listed 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 trial’s own original design assumptions and sample-size justification are not reported in the registry record. The illustrations are therefore not evidence that the original study was designed using these values.

The 80% power column gives the estimated participants per arm for a planning scenario with an 80% chance of detecting the specified effect if the assumptions hold and that effect is the true one. The 90% power column gives the corresponding number per arm for a higher chance of detecting that same specified effect. These are prospective planning quantities, not a measure of what the completed trial achieved.

Rows that plan for a smaller fraction of the observed effect require more participants per arm than rows based on the full observed effect. This is expected: smaller differences are harder to distinguish from random variation when other assumptions remain unchanged. The figures are sensitive to the estimated effect used as their starting point, so planning directly from a trial’s observed effect can yield optimistic illustrations. This is not observed power.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -0.910; pooled standard deviation σ = 0.896.
2. n = 2 × (1.96 + 0.842)² × 0.896² / -0.910² = 15.2, rounded up to 16 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -0.91                      # observed difference in means
sigma <- 0.895975643919719                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(1081834)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 16 (80% power), 21 (90% power)
 75% of observed effect: n per arm = 28 (80% power), 37 (90% power)
 50% of observed effect: n per arm = 61 (80% power), 82 (90% power)
power.t.test: 16.2 
Simulated power at n = 16 per arm: 0.796 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The registry’s sponsor-reported primary comparisons use ANCOVA and report least-squares mean differences with 95% confidence intervals and p-values. Those are the primary estimates to use when describing the trial’s posted analysis. The simple recalculations use unadjusted summary statistics and an approximate spread inferred from reported standard errors and sample sizes; they do not duplicate the sponsor’s adjusted analysis.

For canagliflozin 100 mg, the recalculated mean difference is -0.91, matching the sponsor-reported estimate of -0.91 at the displayed precision. For canagliflozin 300 mg, the recalculated estimate is -1.17, while the sponsor-reported ANCOVA estimate is -1.16. The recalculated p-values are unadjusted, whereas the registry reports p <0.001 for each sponsor comparison. These approaches answer closely related questions but are not interchangeable.

The registry record does not identify a discrepancy flag for the primary result. The sponsor’s ANCOVA estimates remain the appropriate reference because they are the trial’s reported analysis; the recalculations offer a transparent, approximate comparison from posted summaries, rather than an alternative trial result.

15. Multiplicity, Missing Data and Analysis Populations

The registry reports a primary endpoint and multiple secondary outcomes, with 16 outcome measures and 16 analyses listed overall. When multiple outcomes or contrasts are tested, multiplicity becomes relevant: even if every null hypothesis were true, testing many hypotheses can increase the chance of observing at least one small p-value by chance. General statistical explanation: a prespecified hierarchy, family-wise error adjustment, or other plan can control some of this risk, but the method and the family of tests must be known to interpret adjusted claims.

The registry record reports p-values for the primary comparisons and for some secondary comparisons, but does not report whether multiplicity adjustment was used or how the primary and secondary tests were ordered. The original statistical analysis plan and protocol-level multiplicity strategy are not reported in the registry record. Therefore, the posted p-values should be read as reported for their specific analyses, without assuming a correction that is not described.

Missing data were addressed using last observation carried forward for missing Week 26 values. Measurements before glycemic rescue were used as the last observation for participants receiving rescue therapy. This is an explicit analysis choice, but the registry summary does not provide the amount or pattern of missingness, the reasons for missing observations, or sensitivity analyses using alternative assumptions. General statistical explanation: missing-data methods rely on assumptions about the unobserved outcomes, and different assumptions can affect an estimate.

The primary analysis population is the mITT set: randomized participants who received at least one dose of study drug. The registry description does not give a complete participant-flow account for the primary endpoint in this material, nor does it report a comparison of the primary conclusion across alternative analysis populations. The result should therefore be understood in the stated analysis population and under the stated imputation and rescue rules.

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: hepatic enzyme increased (2.3%), urticaria (1.0%), intestinal obstruction (0.5%), abscess limb (0.5%), cellulitis (0.5%).

The registry reports serious adverse events for main-study groups over Baseline to Week 26 and Baseline to Week 52, as well as for canagliflozin groups in the high glycemic substudy through Week 26. For the main study through Week 26, serious adverse events affected 4 of 192 participants in the placebo group, 8 of 195 in the canagliflozin 100 mg group, and 2 of 197 in the canagliflozin 300 mg group. Through Week 52, the reported counts were 11 of 192 for placebo/sitagliptin, 11 of 195 for canagliflozin 100 mg, and 5 of 197 for canagliflozin 300 mg.

The exposure periods and treatment descriptions differ across these groups. In particular, the placebo/sitagliptin group switched treatment after Week 26, so its Week 52 count is not a 52-week placebo-only comparison. The high glycemic substudy counts are also from a distinct subgroup and should not be combined with the main-study groups. The registry record does not report death counts for these groups.

Serious adverse event counts are a limited safety summary. They do not describe all adverse events, duration, severity distribution, or causal attribution. The record lists hepatic enzyme increased at 2.3%, urticaria at 1.0%, and intestinal obstruction, abscess limb, cellulitis, and ankle fracture at 0.5% each among the listed serious terms. These figures should be read as registry-reported safety information, not as an independent comparison or medical recommendation.

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

The posted primary analysis estimates that average HbA1c change from baseline to Week 26 differed between each canagliflozin dose group and placebo in the main study. Both sponsor-reported ANCOVA contrasts were negative, their reported 95% confidence intervals did not include zero, and each p-value was <0.001. These statements describe the direction, uncertainty, and statistical evidence of the reported group comparisons.

They do not show that every participant had the same outcome, that the group averages predict a particular individual’s result, or that the observed changes necessarily have a specific clinical importance. The registry record does not provide a clinical threshold for judging the size of the primary difference. Statistical evidence about an average contrast is distinct from a judgment about patient-level value, tolerability, or suitability.

The primary inference is limited to the main study’s Week 26 comparison with placebo, in the defined mITT population and using the stated missing-data and rescue-measurement procedures. It does not establish a placebo-controlled treatment difference through Week 52 because the placebo group switched to sitagliptin after Week 26. Longer-term data in the canagliflozin groups and placebo/sitagliptin group have a different comparison structure.

Secondary glucose-related outcomes are directionally consistent with the primary result in the posted numerical summaries, but the registry does not report a multiplicity strategy. Their evidential role is therefore more limited. Nor do the posted summaries alone resolve the effects of missing data, model assumptions, or variation among participants.

The appropriate conclusion is statistical and bounded: the registry reports a difference in mean HbA1c change between each canagliflozin group and placebo at Week 26 under its stated analysis. Further claims about clinical benefit, risks for a particular person, or treatment choice go beyond what these trial summaries establish.

18. Why This Trial Matters Statistically

This trial illustrates how a randomized comparison can be summarized using both adjusted model estimates and simpler calculations from posted group summaries. The ANCOVA contrasts are the sponsor-reported primary results, while the approximate unadjusted recalculations offer a distinct check on direction and scale. Keeping those analyses separate is important: similar estimates do not make the methods equivalent, and recalculation does not reconstruct the original participant-level analysis.

The study also shows why trial timing and comparator changes matter. The primary question is a Week 26 comparison with placebo. Because the placebo group switched to sitagliptin afterward, the later period cannot be interpreted as continued placebo control. A clear account of the treatment sequence prevents an inaccurate reading of follow-up results.

Endpoint handling is another important statistical feature. The analysis used an mITT population, last observation carried forward for missing Week 26 values, and pre-rescue measurements for participants receiving glycemic rescue. These rules help define the estimate, but their assumptions and sensitivity to alternatives cannot be fully examined from the registry summary alone.

Finally, the distinction between statistical evidence and clinical meaning remains central. Confidence intervals show uncertainty around group contrasts, while the registry does not supply a threshold for determining the clinical importance of the estimated HbA1c differences. The trial therefore provides a useful example of how design, analysis choices, follow-up structure, and reporting completeness shape the conclusions that can responsibly be drawn from trial results.

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.

The registry summary does not report the original sample-size justification, the full statistical analysis plan, or the detailed covariates and model specification for ANCOVA. This limits independent assessment of the analysis decisions and how closely they followed a prespecified plan. Allocation procedures beyond the randomized designation, the exact roles covered by triple masking, and details of site-level implementation are also not reported in the registry record.

The primary analysis used last observation carried forward for missing Week 26 values and pre-rescue observations for participants receiving glycemic rescue. The amount and pattern of missing data, reasons for missingness, and sensitivity analyses under alternative assumptions are not reported in the registry summary. These omissions make it difficult to judge how dependent the estimates are on the stated handling rules.

Multiplicity handling is not described despite the listing of multiple outcomes and analyses. The numerical secondary results available in the record cover only some of the listed outcomes; results for other listed secondary measures are not reported in the registry record. This limits the assessment of the overall outcome pattern and the evidential weight of secondary findings.

The primary endpoint is a group-average change, so it does not convey the distribution of individual responses. The registry summary also does not provide a threshold for the clinical importance of the estimated difference. Safety counts are reported for different follow-up periods and treatment sequences; the placebo/sitagliptin group changes treatment after Week 26. Death counts are not reported in the registry record.

Finally, the primary contrast is limited to Week 26 and the stated mITT population. The posted result should not be generalized to an uninterrupted 52-week placebo comparison or to groups beyond the study population described in the record.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: The CANTATA-M (CANagliflozin Treatment and Trial Analysis - Monotherapy) Trial — NCT01081834.
• Registry record last updated 2017-02-23; status last verified 2013-06.
• Results first posted 2013-06-03.
• 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

CANTATA-M (NCT01081834) was a completed Phase 3 randomized, triple-masked, parallel-group study of canagliflozin monotherapy in people with type 2 diabetes inadequately controlled with diet and exercise. The registry lists 678 participants. The primary endpoint was change in HbA1c from baseline to Week 26 in the main study.

The sponsor’s ANCOVA analysis reported a least-squares mean difference versus placebo of -0.91 for canagliflozin 100 mg and -1.16 for canagliflozin 300 mg, with p <0.001 for each comparison. The reported 95% confidence intervals were -1.088 to -0.729 and -1.342 to -0.985, respectively. The primary analysis used the mITT set and last observation carried forward for missing Week 26 values, with pre-rescue measurements used for participants receiving rescue therapy.

The primary placebo comparison ends at Week 26; the placebo group then switched to sitagliptin. The original sample-size justification, detailed model specification, missing-data sensitivity analyses, and multiplicity strategy are not reported in the registry record.