A Efficacy, Safety, and Tolerability Study of Canagliflozin in Patients With Type 2 Diabetes Mellitus With Inadequate Glycemic Control on Metformin Alone or in Combination With a Sulphonylurea: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Efficacy, Safety, and Tolerability Study of Canagliflozin in Patients With Type 2 Diabetes Mellitus With Inadequate Glycemic Control on Metformin Alone or in Combination With a Sulphonylurea (NCT01381900), a randomized phase 3 study. The primary outcome is Change in Glycosylated Hemoglobin (HbA1c) From Baseline to Week 18.

2. Clinical Question

The question concerns the effect of adding each study assignment to ongoing background therapy, rather than comparing canagliflozin with metformin or a sulphonylurea. The trial also assessed safety and tolerability, and registered secondary outcomes included fasting plasma glucose, percent change in body weight, and the proportion of participants below specified HbA1c thresholds at Week 18.

The study record describes inadequate glycemic control as the reason for considering an added study drug. It does not report a more detailed eligibility definition in the information available for these results. The results therefore describe the enrolled trial population and its specified background regimens; they should not automatically be generalized to populations or treatment settings not described in the registry record.

3. Trial Design

This was a randomized, double-blind, placebo-controlled, parallel-group phase 3 study. The registry reports enrollment of 678 participants and a completed status. The study start is recorded as 2011-08 and primary completion as 2012-11. Allocation was randomized, and masking was double. The official title describes an 18-week study.

Participants were assigned to placebo, canagliflozin 100 mg, or canagliflozin 300 mg, each once daily for 18 weeks. In every group, study treatment was added to protocol-specified doses of metformin alone or metformin together with a sulphonylurea. The placebo was described as matching. The registry lists 226 participants in the placebo group, 223 in the 100 mg group, and 227 in the 300 mg group for the primary outcome.

The primary endpoint was change in HbA1c from baseline to Week 18, analyzed as a continuous measure. The registry also lists secondary outcomes concerning fasting plasma glucose, percent change in body weight, and proportions meeting HbA1c thresholds at Week 18. The primary analysis population was the overall modified intention-to-treat set, defined as all randomized participants who received at least one dose of double-blind study drug. For missing Week 18 values, the primary outcome description states that last observation carried forward was used. The primary results included participants with both baseline and post-baseline values.

Details such as the original allocation ratio, stratification factors, country or site distribution, full eligibility criteria, and the original sample-size justification are not reported in the registry record. Nor does the record detail the planned handling of every type of protocol deviation or missing observation beyond the stated primary-outcome method.

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: The study had two background treatment strata: 331 participants in the metformin alone stratum and 347 in the metformin plus SU. In total, 678 participants were randomly allocated to the 3 treatment arms, 676 received at least 1 dose of study drug and were included in the modified intent-to-treat (mITT) analysis set and the safety analysis set.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Analysis used overall mITT analysis set (all randomized patients who received at least 1 dose of double-blind study drug). Last-observation-carried-forward method used for missing Week 18 values. Table includes only participants 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 Glycosylated Hemoglobin (HbA1c) From Baseline to Week 18, measured over Day 1 (Baseline) and Week 18.

The primary endpoint is change in glycosylated hemoglobin, or HbA1c, from baseline on Day 1 to Week 18. HbA1c is a continuous measure, and the registry expresses its change in percent. A negative change represents a decrease from baseline in the reported measure. The listed group summaries are least-squares means, which are model-based means rather than simple raw averages.

HbA1c is used in clinical trials as a measure of glycemic status over time, making change from baseline a way to compare how the measured outcome shifted during the study. In this trial, the key statistical contrast is the least-squares mean change in each canagliflozin group minus the corresponding placebo change. A negative treatment difference therefore indicates a more negative mean change in the canagliflozin group than in placebo, as recorded by the endpoint.

General statistical context: a continuous endpoint retains information about the amount of change for each participant rather than reducing the result to whether an individual did or did not cross a threshold. The endpoint’s interpretation depends on the analysis population, measurement timing, and missing-data procedure. For this primary outcome, the record specifies the overall modified intention-to-treat set, the baseline and Week 18 time points, and last observation carried forward for missing Week 18 values. It does not report a minimum clinically important difference for HbA1c, so the numerical estimates alone do not establish the clinical importance of a change.

7. Results Posted in the Registry

• Canagliflozin 100 mg: mean -0.97 versus -0.47 on control; difference -0.50 (95% CI -0.76 to -0.24), standardized effect d = -0.35; z-test p = <0.001.
• Canagliflozin 300 mg: mean -1.06 versus -0.47 on control; difference -0.59 (95% CI -0.85 to -0.33), standardized effect d = -0.41; 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.

Change in Fasting Plasma Glucose (FPG) From Baseline to Week 18

• 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 -1.44 versus -0.40 on control; difference -1.04 (95% CI -1.70 to -0.38), standardized effect d = -0.29; z-test p = 0.002.
  • Canagliflozin 300 mg: mean -1.83 versus -0.40 on control; difference -1.43 (95% CI -2.08 to -0.78), standardized effect d = -0.40; 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.

Percent Change in Body Weight From Baseline to Week 18

• 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 -2.90 versus -0.80 on control; difference -2.10 (95% CI -3.21 to -0.99), standardized effect d = -0.35; z-test p = <0.001.
  • Canagliflozin 300 mg: mean -3.10 versus -0.80 on control; difference -2.30 (95% CI -3.41 to -1.19), standardized effect d = -0.38; 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.

Percentage of Patients With Glycosylated Hemoglobin (HbA1c) <7% at Week 18

• 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: 48.9% (109/223) versus 27.9% (63/226) on control; risk difference 21.0 percentage points (95% CI 12.2 to 29.8), risk ratio 1.75 (1.37–2.25); unadjusted two-proportion z-test p = <0.001.
  • Canagliflozin 300 mg: 52.9% (120/227) versus 27.9% (63/226) on control; risk difference 25.0 percentage points (95% CI 16.2 to 33.7), risk ratio 1.90 (1.49–2.42); 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.

The registered secondary outcomes extend the primary HbA1c comparison to fasting plasma glucose, percent change in body weight, and the proportions below HbA1c thresholds at Week 18. Across the secondary results with numerical estimates, the reported contrasts follow a broadly consistent directional pattern: the continuous-outcome differences are negative, while a larger proportion of participants in each canagliflozin group reached the posted HbA1c <7% threshold than in the placebo group. These measures offer related but distinct descriptions of the study outcomes.

Fasting plasma glucose provides another continuous measure, body-weight change describes a different outcome, and an HbA1c threshold converts a continuous measure into a proportion meeting a specified cutoff. Agreement in direction can add descriptive context, but the endpoints are not interchangeable and do not each independently confirm the primary result. The registered HbA1c <6.5% outcome is not accompanied by numerical results in the registry record.

Secondary outcomes generally carry less weight for confirmatory interpretation when several outcomes or contrasts are tested and no multiplicity adjustment is reported. Testing multiple hypotheses increases the chance that at least one result appears statistically notable by chance, even when each individual test uses a conventional threshold. The registry record does not report whether the secondary tests were controlled as a family or whether they were designated confirmatory or exploratory.

For that reason, the secondary pattern is best treated as supportive description rather than as a set of separate definitive conclusions. The primary endpoint remains the prespecified focus identifiable from the posted results, but the registry does not give a detailed hierarchy among the secondary outcomes. Interpret secondary p-values and intervals in light of their multiple comparisons, distinct measurement scales, analysis methods, and the missing detail about multiplicity.

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

• Change in Fasting Plasma Glucose (FPG) From Baseline to Week 18
• Percent Change in Body Weight From Baseline to Week 18
• Percentage of Patients With Glycosylated Hemoglobin (HbA1c) <7% at Week 18
• Percentage of Patients With Glycosylated Hemoglobin (HbA1c) <6.5% at Week 18

9. Statistical Interpretation

10. Statistical Methods Explained

What is the main comparison?

The primary comparisons are each canagliflozin group against placebo. The posted arm values are least-squares mean changes in HbA1c from baseline to Week 18. The sponsor reports the difference between groups from ANCOVA, stated as the least-squares mean for the canagliflozin group minus the least-squares mean for placebo. With this ordering, a negative difference corresponds to a more negative estimated change in the canagliflozin group. The reported results are -0.51 for 100 mg and -0.59 for 300 mg, each with a 95% confidence interval and p < 0.001.

What do least-squares means represent?

General statistical explanation: a least-squares mean is a model-based estimate of a group mean. It can account for the variables included in the fitted model, unlike a simple arithmetic mean. The registry identifies ANCOVA as the method for the reported primary contrasts, but the full model specification and covariates are not reported in the registry record. The arm summaries and the reported adjusted comparisons should therefore be distinguished rather than treated as interchangeable versions of the same calculation.

The standard errors shown with the arm estimates describe the estimated sampling uncertainty of each least-squares mean. A standard error is not the same as the spread of individual participant outcomes. It quantifies uncertainty in an estimated mean under the analysis, while individual outcomes can vary more widely. The registry does not report standard deviations for the primary endpoint arm summaries.

What does the unadjusted recalculation add?

The recalculated contrasts use the posted arm means and standard errors to form simple, unadjusted comparisons. Because this approach does not reproduce the sponsor’s ANCOVA model, its estimate and confidence interval can differ from the sponsor’s reported result. The recalculated primary contrasts are -0.5 for 100 mg versus placebo and -0.59 for 300 mg versus placebo. Their confidence intervals are approximate, because the standard errors are used to reconstruct a comparison from summary information rather than from participant-level data. The recalculation is a transparent statistical cross-check, not a replacement for the sponsor’s analysis.

General statistical explanation: an adjusted analysis can differ from an unadjusted one if the model accounts for baseline or other covariates, or if the analysis population or handling of data differs. The registry identifies ANCOVA, but its model details are not reported in the registry record. It is therefore not possible to attribute the small numerical differences to a particular model feature from the posted information alone.

How should p-values be read?

A p-value describes how compatible the observed data are with a specified null model, under the assumptions of the analysis. It is not the probability that the null hypothesis is true, nor is it a measure of clinical importance. The sponsor reports p < 0.001 for both primary comparisons. That is evidence against a zero-difference null under the reported analysis, but it does not quantify how important the estimated differences are to patients or establish the result’s durability.

The recalculated p-values are unadjusted. A test is called unadjusted when its p-value has not been modified to account for testing multiple outcomes or comparisons. The registry record does not report whether a multiplicity procedure was applied to the primary family of comparisons. The sponsor’s reported primary p-values are retained as reported and should not be replaced by the approximate recalculated values.

What is the role of the analysis population?

The primary analysis used the overall modified intention-to-treat set, described as all randomized participants who received at least one dose of double-blind study drug. The endpoint description further says that the table includes only participants with both baseline and post-baseline values. Last observation carried forward was used for missing Week 18 values. These details determine which observations contribute and how some missing follow-up values are represented. The registry does not report a separate sensitivity analysis that would show how the primary estimate changes under another missing-data method.

General statistical explanation: last observation carried forward fills a missing later value with an earlier observed value. This can retain participants who lack the final measurement, but it relies on an assumption about how their outcome would have evolved after the last observation. That assumption can be unsuitable if the outcome changes over time. The method’s use is reported; its consequences for this trial cannot be assessed fully without further analyses and information about missingness.

The stated modified intention-to-treat definition differs from a population containing every randomized participant regardless of whether they received study drug. It is important to keep that definition in view when interpreting the estimate. The record does not give the number of participants excluded from the primary analysis or describe the reasons for any exclusions.

11. Confidence Intervals and Precision

The sponsor-reported 95% confidence interval for the 100 mg comparison is -0.644 to -0.367; for the 300 mg comparison it is -0.731 to -0.453. Each interval lies below zero, consistent with a negative estimated treatment difference under the stated comparison order. The width of an interval indicates how precisely the data estimate the contrast, conditional on the analysis and its assumptions. It is not a range containing every participant’s outcome.

General statistical explanation: a confidence interval gives a range of values reasonably compatible with the data under a specified procedure. Repeated use of the same procedure would produce intervals that contain the target value at the stated rate when its assumptions hold. A 95% confidence interval does not mean there is a 95% probability that this particular fixed parameter lies in the calculated interval.

The recalculated unadjusted intervals are -0.765 to -0.235 for 100 mg and -0.855 to -0.325 for 300 mg. They are wider than the sponsor-reported intervals in this record. They are approximations formed from the posted arm summaries, not the sponsor’s ANCOVA output, so the sponsor-reported intervals are the primary reference for the reported analysis. The record indicates no discrepancy requiring the sponsor’s analysis to be displaced.

Precision should not be confused with clinical importance. An estimate can be statistically precise while its practical meaning remains uncertain without a prespecified threshold for meaningful change and context about outcomes important to participants. The registry record does not report such a threshold for the primary endpoint. Nor do the intervals capture uncertainty arising from questions not represented in the statistical model, such as generalizability beyond the enrolled population or the consequences of the missing-data method.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size figures are planning illustrations based on the observed contrasts, not the trial’s original enrollment justification. For these continuous and binary outcomes, the figures are participants per arm; the guidance that uses total events applies to time-to-event outcomes, which are not the outcomes discussed in this sample-size illustration. Each planning row describes a fraction of the observed effect, such as planning for 75% of the observed effect. Smaller effects require more participants to detect under the same planning assumptions.

The stated 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 the stated fraction of the observed effect. The original trial’s design assumptions and sample-size justification are not reported in the registry record, so these illustrations should not be read as a reconstruction of the design decision that produced enrollment of 678.

The 80% power column gives the estimated participants per arm for a planning scenario designed to have an 80% chance of detecting the specified effect under the stated assumptions, if that effect and those assumptions hold. The 90% power column gives the corresponding participant count for a 90% chance. Higher planned power generally requires more participants for the same effect size. These figures are not observed power; they are prospective planning calculations based on the observed effect, not a measure of what the completed study achieved.

The sample-size curve should be read alongside the rows as an illustration of how the required participants per arm change as the target effect changes. It does not show the trial’s actual enrollment pattern or prove that a future study will attain the planned power. Since the calculations are based on the observed effect, and observed effects are uncertain, the planning figures are illustrative rather than guarantees.

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 = -0.500; pooled standard deviation σ = 1.431.
2. n = 2 × (1.96 + 0.842)² × 1.431² / -0.500² = 128.6, rounded up to 129 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 <- -0.5                      # observed difference in means
sigma <- 1.43105947048112                     # 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(1381900)
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 = 129 (80% power), 173 (90% power)
 75% of observed effect: n per arm = 229 (80% power), 307 (90% power)
 50% of observed effect: n per arm = 515 (80% power), 689 (90% power)
power.t.test: 129.6 
Simulated power at n = 129 per arm: 0.802 (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 sponsor’s primary analyses are reported as ANCOVA least-squares mean differences. For canagliflozin 100 mg versus placebo, the reported difference is -0.51, with a 95% confidence interval of -0.644 to -0.367 and p < 0.001. For 300 mg versus placebo, it is -0.59, with a 95% confidence interval of -0.731 to -0.453 and p < 0.001.

The recalculation uses the posted arm summaries to estimate simple unadjusted differences. It gives -0.5 for 100 mg and -0.59 for 300 mg, with approximate confidence intervals and unadjusted p-values. The results are broadly similar, but the 100 mg estimate and intervals are not identical. An unadjusted reconstruction from summary statistics does not reproduce an ANCOVA analysis, and the full ANCOVA model is not reported in the registry record. The reason for each numerical difference therefore cannot be determined from the record.

The registry indicates no discrepancy requiring the recalculated result to take precedence. The sponsor’s own analysis is the appropriate reference for the reported primary result; the recalculation is a supplementary check based on posted summaries. It should not be interpreted as an independent replication or as evidence that the sponsor’s analysis used the same calculation.

15. Multiplicity, Missing Data and Analysis Populations

The registry reports two primary ANCOVA comparisons, one for each canagliflozin group against placebo, and gives a p-value and 95% confidence interval for each. It also reports several secondary outcomes, some with comparisons against placebo. The statistical plan for handling multiple comparisons is not reported in the registry record. In particular, the record does not say whether the primary comparisons were adjusted as a family, whether a testing hierarchy was specified, or whether multiplicity control was applied to secondary outcomes.

General statistical explanation: when multiple hypotheses are tested, the chance of at least one false-positive result can rise if each test is interpreted in isolation. Multiplicity procedures can control that risk, but the appropriate method depends on which outcomes and comparisons were planned together. Without the plan, the reported nominal p-values should be read as reported rather than assumed to represent multiplicity-adjusted evidence.

The primary endpoint description specifies the overall modified intention-to-treat analysis set: randomized participants who received at least one dose of double-blind study drug. It also states that participants needed both baseline and post-baseline values for inclusion in the displayed primary outcome results. For missing Week 18 values, last observation carried forward was used. The record does not report a full accounting of missing primary measurements, reasons for missingness, or sensitivity analyses under alternative assumptions.

General statistical explanation: missing-data methods require assumptions about outcomes that were not observed. Last observation carried forward substitutes the most recent observed value for a later missing value; this may not reflect the participant’s unobserved trajectory. The posted method is clear, but the extent to which missingness or the imputation affected the comparison cannot be judged from the reported information alone.

The record does not report interim analyses, stopping rules, or a detailed statistical analysis plan. Those details should not be inferred from the trial design or outcome estimates.

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: tracheitis (0.4%), spinal compression fracture (0.4%), intervertebral disc protrusion (0.4%), rotator cuff syndrome (0.4%), transient ischaemic attack (0.4%).

The registry lists serious adverse events affecting 4 of 226 participants in the placebo group, 5 of 223 in the canagliflozin 100 mg group, and 4 of 227 in the canagliflozin 300 mg group. These are participant counts among those at risk in each group. The counts describe what was recorded during the study; they do not establish that an event was caused by the assigned study treatment.

The registry also lists other adverse events affecting 22 participants in the placebo group, 27 in the 100 mg group, and 27 in the 300 mg group. A list of serious-event terms includes tracheitis, spinal compression fracture, intervertebral disc protrusion, rotator cuff syndrome, transient ischaemic attack, and pneumonia, each reported as 0.4%. The record does not provide enough detail in these summaries to assess event timing, recurrence, severity beyond the serious classification, or causal attribution.

Death counts are not reported in the registry record. Event counts should be considered alongside the duration of follow-up and the number of participants at risk; the study’s treatment period was 18 weeks. These summaries do not replace a fuller safety analysis or establish comparative safety from the counts alone.

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

The result estimates how the recorded HbA1c measure changed over the study period in three randomized groups receiving study drug or matching placebo alongside metformin-based background therapy. The sponsor-reported comparisons indicate more negative mean changes in both canagliflozin groups than in placebo at Week 18, with confidence intervals below zero and p < 0.001. This is a statistical description of the primary endpoint in the stated analysis population.

It does not show how every participant’s HbA1c changed, and it does not establish the probability that a particular person would have a particular outcome. Group means summarize the distribution of measured outcomes and can conceal variation among participants. The record does not report the proportion who experienced a specified individual change in the primary endpoint.

The results do not establish the clinical importance of the estimated differences: the registry record does not report a threshold for clinically meaningful HbA1c change. Nor do these Week 18 measurements address durability beyond the reported timeframe, outcomes not measured in the posted results, or effects in populations whose characteristics are not described in the record.

The primary endpoint and secondary endpoints should be interpreted separately. Secondary findings can add context, but multiplicity procedures are not reported, and one registered secondary outcome lacks numerical results in the registry record. Similarly, the safety event counts are descriptive and do not establish causation or provide a complete account of longer-term safety.

Randomization and double masking are features of the reported design that support a controlled comparison, but they do not remove every source of uncertainty. Interpretation still depends on analysis population, missing-data handling, model specification, precision, and applicability to other settings. The posted estimates describe this trial; they are not a recommendation about treatment choices.

18. Why This Trial Matters Statistically

This trial offers a useful example of how a continuous primary endpoint, model-based group summaries, and treatment-versus-placebo contrasts fit together. The arm-level least-squares mean changes are not themselves the comparative effect; the reported ANCOVA differences directly express each canagliflozin group minus placebo. Keeping those quantities distinct helps prevent a common reading error in which a within-group change is treated as the randomized between-group contrast.

The record also makes the analysis population and missing-data approach visible. The primary outcome used a modified intention-to-treat population and last observation carried forward for missing Week 18 values. Those choices are part of what the result means, not merely technical details. The registry does not supply a full sensitivity analysis or complete model specification, so the estimate’s robustness to alternative assumptions cannot be assessed from these results alone.

Another statistical lesson is the distinction between sponsor-reported ANCOVA results and a simple recalculation from summary statistics. The recalculation is a useful check but cannot recreate the sponsor’s adjusted model. Its approximate intervals and unadjusted p-values answer a different, narrower computational question. The sponsor’s reported analysis remains the primary interpretation of the trial’s stated comparison.

Finally, the trial’s secondary outcomes illustrate why directionally consistent findings should not be treated as a collection of independent confirmations when multiplicity handling is unreported. Statistical interpretation requires attention to the outcome hierarchy, uncertainty, and missing design details as well as the point estimates.

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 record does not report the original sample-size justification, detailed model covariates, allocation ratio, stratification factors, or a full statistical analysis plan. It also does not report whether interim analyses or multiplicity procedures were planned. These omissions limit the ability to judge how closely the posted analyses follow the original protocol or how the reported tests were prioritized.

The primary endpoint used last observation carried forward for missing Week 18 values. The registry does not provide the missing-data totals, reasons for missingness, or sensitivity analyses using alternative approaches. The primary population definition also requires receipt of at least one dose and baseline plus post-baseline measurements for the displayed endpoint, so the record does not permit a complete assessment of exclusions or their effects.

The sponsor’s ANCOVA model is not fully described. Consequently, the unadjusted recalculations based on the posted summaries cannot be used to infer which covariates were included or why estimates and confidence intervals differ. These recalculations are approximate checks rather than participant-level reanalyses.

The follow-up period described for study treatment is 18 weeks. The reported outcomes therefore do not establish what happened beyond that period. The safety summary provides event counts and selected serious-event terms but does not report deaths, detailed event timing, causal assessments, or a comprehensive account of how events were evaluated. The registry record also lists an HbA1c <6.5% secondary outcome without numerical results. Together, these constraints call for a focused interpretation of the posted outcomes rather than broad conclusions beyond the record.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Efficacy, Safety, and Tolerability Study of Canagliflozin in Patients With Type 2 Diabetes Mellitus With Inadequate Glycemic Control on Metformin Alone or in Combination With a Sulphonylurea — NCT01381900.
• Registry record last updated 2014-05-09; status last verified 2014-04.
• Results first posted 2014-01-09.
• 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

NCT01381900 is a completed, randomized, double-blind, placebo-controlled, parallel-group phase 3 study of canagliflozin added to metformin alone or metformin plus a sulphonylurea in people with type 2 diabetes. Enrollment was 678, and the study treatment period was 18 weeks.

The primary outcome was change in HbA1c from baseline to Week 18. The sponsor-reported ANCOVA differences versus placebo were -0.51 for canagliflozin 100 mg and -0.59 for 300 mg; both had 95% confidence intervals below zero and p < 0.001. The primary analysis used a modified intention-to-treat population and last observation carried forward for missing Week 18 values.

Secondary outcomes included fasting plasma glucose, body-weight change, and HbA1c thresholds. Interpretation is limited by unreported details of the original statistical plan, multiplicity handling, and missing-data sensitivity analyses. The posted estimates describe the study outcomes and do not by themselves establish clinical importance or effects beyond the reported follow-up.