Efficacy and Safety Study of Canagliflozin (TA-7284) in Combination With Insulin in Patients With Type 2 Diabetes Mellitus: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy and Safety Study of Canagliflozin (TA-7284) in Combination With Insulin in Patients With Type 2 Diabetes Mellitus (NCT02220920), a randomized phase 4 study. The primary outcome is Change in HbA1c From Baseline.

2. Clinical Question

This is a question about a measured laboratory outcome over the study period, not a comparison of insulin with no insulin. The registry describes a randomized, parallel-group study and identifies placebo as the comparator. The primary endpoint is change in HbA1c, a continuous measure, and its reported comparison accounts for baseline through an ANCOVA analysis. A negative difference indicates that the estimated change was more negative in the canagliflozin-plus-insulin group than in the placebo-plus-insulin group. The numerical estimate and its uncertainty are interpreted in the context of the stated time point and analysis population.

The record also identifies fasting plasma glucose, body-weight change, blood pressure, and adverse-event measures among secondary outcomes. These can add context to the primary comparison, but they are separate outcomes and do not substitute for the primary endpoint. Registry details about participant eligibility beyond the type 2 diabetes and insulin context, and about the criteria used to select participants, are not reported in the registry record.

3. Trial Design

The registry identifies NCT02220920 as a completed Phase 4 study sponsored by Tanabe Pharma Corporation. It describes randomized allocation, a parallel-group model, and multicenter conduct. The study title identifies the design as double-blind, while the registry masking field says triple. The record does not explain which participant, investigator, or other roles were masked, so the masking detail should not be inferred beyond those descriptions.

The comparison was canagliflozin (TA-7284) added to insulin versus placebo added to insulin. The registry states that the study aimed to assess efficacy and safety over 16 weeks. The primary endpoint timeframe is baseline and Week 16. The overall enrollment is recorded as 146. The primary endpoint results are reported for 76 participants in the canagliflozin-plus-insulin group and 70 in the placebo-plus-insulin group; the analysis population is specified as the full analysis set with last observation carried forward. These endpoint analysis counts should not be treated as a different statement of enrollment.

The registry records a study start of 2014-08 and a primary completion of 2015-07, and lists the status as completed. The trial is described as placebo-controlled and parallel-group, so the primary contrast is between two concurrently assigned groups. Randomization supports a between-group comparison, although the registry record does not report details of the randomization sequence or concealment procedures.

The record gives no detailed arm descriptions beyond the intervention names and their combination with insulin. Dose schedules, insulin-management rules, adherence procedures, eligibility criteria, or protocol deviations are not reported in the registry record. Original design assumptions, including the original sample-size justification and any interim-analysis plan, are also not reported in the registry record. Those omissions limit what can be concluded about operational details, but they do not alter the interpretation of the posted endpoint estimates.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full analysis set, last observation carried forward

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, measured over baseline and Week 16.

The primary endpoint is change in HbA1c from baseline to Week 16, measured in percent. HbA1c reflects average blood glucose over a period of time and is commonly used in diabetes trials as a continuous outcome for comparing changes between groups. In this study, the endpoint is a change score rather than a single Week 16 measurement, so it expresses how the measured value shifted from baseline during the stated study period.

The registry reports least-squares means for the full analysis set, using last observation carried forward. The reported group estimates are -0.97 for canagliflozin plus insulin and 0.13 for placebo plus insulin. The sponsor's ANCOVA comparison is a least-squares mean difference of -1.10, with a 95% confidence interval from -1.33 to -0.87 and p < 0.001. The negative contrast indicates a more negative estimated change in the canagliflozin-plus-insulin group than in the placebo-plus-insulin group.

As a general statistical point, a continuous endpoint preserves information about the magnitude of change rather than reducing participants to categories. Interpretation still depends on the measurement, the analysis population, missing-data handling, and uncertainty around the group contrast. The registry record does not report a threshold for a clinically important HbA1c change, so statistical evidence should not be equated with a judgment about clinical importance.

7. Results Posted in the Registry

• Canagliflozin (TA-7284) ＋Insulin: mean -0.97 versus 0.13 on control; difference -1.10 (95% CI -1.32 to -0.88), standardized effect d = -1.61; 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

• 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 (TA-7284) ＋Insulin: mean -34.10 versus -1.40 on control; difference -32.70 (95% CI -46.28 to -19.12), standardized effect d = -0.78; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Canagliflozin (TA-7284) ＋Insulin than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percent Change in Body Weight

• 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 (TA-7284) ＋Insulin: mean -2.13 versus 0.24 on control; difference -2.37 (95% CI -3.08 to -1.66), standardized effect d = -1.09; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Canagliflozin (TA-7284) ＋Insulin than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes with detailed numerical comparisons reinforce a consistent direction in the posted continuous measures: the group receiving canagliflozin with insulin has more negative estimated changes than the placebo-plus-insulin group for fasting plasma glucose and percent change in body weight. These comparisons are separate from the primary HbA1c endpoint and use their own units and scales. They provide additional context about the pattern of measured outcomes during the study period, but they should not be combined into a single overall effect.

The registry lists change in blood pressure and the percentage of participants with adverse events and with hypoglycemia or decreased blood glucose among the secondary outcomes. Detailed numerical group results for those measures are not reported in the registry record. Their inclusion in the outcome list does not establish whether groups differed on them, nor does it show the direction or size of any difference.

The primary endpoint was designated as the main efficacy comparison, whereas secondary outcomes ordinarily address additional questions. As a general statistical point, testing several outcomes without an adjustment can increase the possibility of finding at least one apparently notable result by chance. The posted secondary p-values are unadjusted, and the registry does not report a multiplicity plan. Consequently, the secondary findings carry less confirmatory weight than the primary comparison and are best interpreted as supplementary evidence, not as independent proof of multiple separate conclusions.

The coherent direction across the primary and two numerically reported secondary measures is a descriptive pattern. It does not resolve questions about clinical importance, missing data, or the reliability of analyses whose full specifications are not reported. Results for blood pressure and the listed hypoglycemia-related outcomes cannot be inferred from the results for HbA1c, fasting plasma glucose, or body weight.

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

• Change in Fasting Plasma Glucose
• Percent Change in Body Weight
• Change in Blood Pressure
• Percentage of Participants With "Adverse Events" and "Hypoglycemia and Blood Glucose Decreased"

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is continuous: each participant contributes a change in HbA1c, and the group comparison concerns the average estimated change. The registry reports least-squares means, which are model-based group estimates, rather than only raw arithmetic averages. It identifies ANCOVA as the method for the sponsor's primary comparison and reports a least-squares mean difference. The record does not give the full model specification, including all covariates or the exact parameterization, so those details are not reported in the registry record.

What does ANCOVA contribute?

As a general statistical explanation, analysis of covariance compares groups while accounting for one or more covariates. For a baseline-to-follow-up outcome, baseline measurement is often relevant to the estimated follow-up contrast or change comparison. The registry identifies ANCOVA and reports least-squares mean differences, but does not provide enough detail to establish every variable included in that model. The reported estimate should therefore be read as the sponsor's adjusted model-based contrast, not as an independently reconstructed raw-data analysis.

A least-squares mean is an estimate from the fitted model, evaluated in a way that accounts for the model's covariates. It can differ from an unadjusted mean computed directly from individual observations. In this record, no individual-level observations are available in the posted summary, so the displayed group estimates and contrast are interpreted as reported summary statistics. The separately labelled recalculation uses the posted group summaries and is not a replacement for the sponsor's ANCOVA.

What does the standard error describe?

The primary group estimates are reported with standard errors of 0.08 in each group. As a general explanation, a standard error describes the estimated sampling uncertainty of an estimate, under the assumptions used to calculate it. It is not the spread of individual participant values. The registry labels the dispersion as standard error, and the primary endpoint record gives no confidence limits for the separate group estimates. The sponsor does report a confidence interval for the contrast, which is more directly useful for judging the precision of the between-group estimate.

The posted recalculation is an unadjusted re-analysis based on the summary estimates and their standard errors. Its approximate standard deviations are inferred from those summaries, as indicated by the record. Such a recalculation cannot reproduce the sponsor's ANCOVA exactly because the model adjustment and individual-level data are not available. It is useful as a consistency check on the direction and approximate magnitude of the contrast, not as an alternative primary analysis. The record marks no discrepancy between the recalculation and the sponsor's analysis.

Why can over-dispersion matter?

As a general concept, over-dispersion means that observed variability is greater than a statistical model expects. When it is present and not accounted for, uncertainty estimates can be too small, making intervals narrower and tests more likely to signal evidence than they should. This is most directly a concern for outcomes modeled as counts or proportions, where the variance assumptions can be restrictive. The primary HbA1c outcome is continuous, and the registry's posted summary does not report an over-dispersion assessment. It would therefore be inappropriate to claim that over-dispersion was present or that a particular correction was applied.

For the secondary continuous outcomes, the registry also reports least-squares means and standard errors, with ANCOVA results for the stated comparisons. The same broad interpretation applies: model-based contrasts and their intervals quantify group differences, but complete assessment of the analysis requires model specifications and diagnostics. Those details are not reported in the registry record. The record likewise does not describe whether all secondary outcomes were prespecified in a statistical analysis plan or how their analyses were selected.

How should the p-values be read?

As a general explanation, a p-value measures how compatible the observed data are with a specified null model, assuming the model and analysis conditions hold. It is not the probability that the null hypothesis is true, nor does it measure the size or practical importance of an effect. The sponsor reports p < 0.001 for the primary comparison. That value is evidence against a zero-difference null under the reported ANCOVA analysis, but it does not address all possible sources of bias, generalizability, or clinical importance.

The recalculated p-value is also reported as p < 0.001. Its unadjusted nature matters: it is based on the posted summary statistics rather than the sponsor's full model, and it does not account for multiple outcomes. When many outcome tests are considered, unadjusted p-values should not be interpreted as if only one comparison had been planned. The registry record does not report whether a formal multiplicity strategy was specified.

What does the analysis population mean?

The primary population is named as the full analysis set, and the registry specifies last observation carried forward for missing endpoint values. In general, a full analysis set is intended to retain participants according to a stated analysis principle following randomization, but the precise definition can vary. The registry does not provide the detailed population rules, the number of missing measurements, or a breakdown of exclusions. Last observation carried forward uses a participant's last observed value to stand in for a missing later value. As a general matter, this assumes that the carried-forward value is an adequate representation of the unobserved value, an assumption that may not hold in all circumstances.

The posted results thus support interpretation of the stated model-based comparison, but not a detailed audit of all analysis decisions. The statistical analysis plan, model diagnostics, covariate details, handling of protocol deviations, and sensitivity analyses are not reported in the registry record. Those omissions are important when assessing robustness, particularly because the primary endpoint uses an imputation approach rather than only observed Week 16 measurements.

11. Confidence Intervals and Precision

The sponsor-reported primary contrast is -1.10, with a 95% confidence interval from -1.33 to -0.87. The interval gives a range of values compatible with the data under the ANCOVA model and its assumptions. Because the full interval lies below zero, the result is consistent with a negative between-group contrast rather than a null contrast. Its width also shows that the estimate is not a single exact value: values across the interval represent uncertainty about the magnitude of the contrast.

As a general statistical explanation, a confidence interval reflects both the observed data and the method used to construct it. A 95% confidence interval is not a statement that there is a 95% probability that this particular fixed effect lies inside the interval. Rather, the method is designed to cover the true value at the stated rate over repeated uses under its assumptions. Here, those assumptions include the sponsor's ANCOVA model and the analysis population and missing-data method identified in the record.

The posted recalculation uses the group summary estimates and standard errors and is labelled approximate. It gives a contrast close to the sponsor's reported -1.10 estimate, and the registry marks no discrepancy. The recalculation should not be treated as having the same model basis as the sponsor's analysis. The sponsor's confidence interval is the primary interval to use when describing the registry's own adjusted result.

Precision is not the same as importance. The confidence interval quantifies uncertainty in the statistical contrast, but the registry does not report a threshold that would define a clinically important difference. The interval also does not account for unreported issues such as sensitivity to the missing-data approach or multiplicity across outcomes. Its interpretation is therefore strongest for the stated primary analysis, population, and timeframe.

12. Planning a Similar Trial: Sample Size

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

The sample-size figures are planning illustrations based on the observed effect, not the trial's original design calculation. For these continuous outcomes, the figures are participants per arm, not total events. The stated 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 80% power column describes the estimated number of participants per arm for a future study to have an 80% chance of detecting the chosen effect under those assumptions, if that effect and the other assumptions are correct. The 90% power column uses a higher target chance of detection and therefore generally calls for more participants per arm. For example, planning for 50% of the observed primary effect corresponds to 25 participants per arm for 80% power and 33 participants per arm for 90% power in the posted illustration.

Smaller effects are harder to distinguish from random variation, so planning to detect a smaller fraction of the observed effect requires more participants per arm. The figures are conditional on the planning assumptions and the observed effect used as their reference; a future study's actual variability, analysis, and effect could differ. They are not observed power, and they do not establish that the original trial was designed using these assumptions. The original trial's own sample-size justification and planning assumptions are not reported in the registry record.

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 = -1.100; pooled standard deviation σ = 0.684.
2. n = 2 × (1.96 + 0.842)² × 0.684² / -1.100² = 6.1, rounded up to 7 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 <- -1.1                      # observed difference in means
sigma <- 0.684105255059483                     # 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(2220920)
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 = 7 (80% power), 9 (90% power)
 75% of observed effect: n per arm = 11 (80% power), 15 (90% power)
 50% of observed effect: n per arm = 25 (80% power), 33 (90% power)
power.t.test: 7.2 
Simulated power at n = 7 per arm: 0.792 (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 reports the primary comparison using ANCOVA as a least-squares mean difference of -1.10, with a 95% confidence interval from -1.33 to -0.87 and p < 0.001. This is the registry's own adjusted analysis and should anchor interpretation of the primary result.

The recalculation uses the posted group estimates and standard errors to form an unadjusted comparison. It is an approximate check based on summary statistics, not a reproduction of the sponsor's ANCOVA using individual-level data. Its estimated contrast is -1.1, and its interval and p-value are broadly consistent with the sponsor's reported result. The registry marks no discrepancy between the two analyses.

The distinction matters because model adjustment can affect both the estimate and its uncertainty, and the summary-level recalculation cannot account for covariates or analysis details that are not included in the posted summaries. The recalculation can help readers see whether the group summaries point in the same direction as the sponsor's contrast, but it should not be treated as a substitute for the reported analysis. The sponsor's adjusted estimate is the appropriate figure to quote as the trial's own primary result.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies one primary endpoint, change in HbA1c from baseline to Week 16, and lists multiple secondary outcomes. The primary result is reported with an ANCOVA estimate, confidence interval, and p-value. The record does not report a multiplicity-adjustment strategy or a hierarchy for interpreting the secondary outcomes. Without such information, the secondary p-values should be read as unadjusted comparisons rather than as a set of tests with a controlled overall error rate. This is one reason to give the designated primary comparison more weight than the secondary findings.

The primary analysis population is the full analysis set, with last observation carried forward. The record does not report how many measurements were missing, why they were missing, or how many values were carried forward. It also does not report sensitivity analyses using alternative missing-data assumptions. As a general statistical concern, carrying a last observed value forward can misrepresent a later outcome if the outcome would have changed after that observation; the direction and size of any resulting bias cannot be established from the registry summary.

The registry does not report the detailed definition of the full analysis set, the handling of protocol deviations, or the exact model covariates. Nor does it report an interim-analysis plan, stopping rules, or the original sample-size justification. These items are not evidence that particular procedures were or were not used; they are simply not reported in the registry record.

These gaps affect how fully the reported precision and robustness can be assessed. They do not erase the posted result, but they limit independent evaluation of the assumptions behind the analysis and of whether the conclusions would remain similar under other reasonable analysis choices. The reported estimate should therefore be described with its stated population, timeframe, model, and missing-data approach rather than generalized beyond them.

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: cataract (1.4%), retinal detachment (1.3%), vitreous haemorrhage (1.3%), alcoholic liver disease (1.3%).

The registry reports serious adverse-event figures for the groups using at-risk counts of 75 for canagliflozin plus insulin and 71 for placebo plus insulin. The numbers reported as affected by serious adverse events are 3 and 1, respectively. The figures for other adverse events are 39 and 34, respectively, using the same at-risk counts for those group entries. These are descriptive counts as recorded in the registry, not adjusted comparisons.

The record lists cataract at 1.4%, retinal detachment at 1.3%, vitreous haemorrhage at 1.3%, and alcoholic liver disease at 1.3% among its top serious-event terms. The registry's death counts are not reported in the registry record. It does not provide enough information in these figures alone to assess event timing, duration, severity beyond the serious-event classification, relatedness to study treatment, or whether a participant had more than one event.

As a general interpretive point, raw counts do not establish that an intervention caused an event or that one group has a reliably different safety profile. The study's 16-week period and the registry's reported summaries are the context for these observations; they cannot by themselves resolve uncommon or longer-term risks. No medical conclusion should be drawn from these counts alone.

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

The primary result means that, in the registry's full analysis set and under its stated last-observation-carried-forward approach, the estimated change in HbA1c differed between the randomized groups over baseline to Week 16. The sponsor's ANCOVA estimate was -1.10, with a 95% confidence interval from -1.33 to -0.87 and p < 0.001. This describes a statistical contrast between canagliflozin added to insulin and placebo added to insulin in this study.

It does not establish what would happen in people who differ from the studied population, in other treatment settings, or over periods beyond the recorded study timeframe. The registry provides no clinical-importance threshold, so the statistical result alone does not determine whether the difference is clinically meaningful. Nor does a small p-value measure the probability that the result will be reproduced or that any particular participant will experience a particular change.

The primary analysis is model-based and uses last observation carried forward for the stated analysis population. The record does not report missing-data counts, sensitivity analyses, or full ANCOVA specifications. The secondary comparisons add descriptive context but are unadjusted for multiplicity, and detailed numerical results for some listed secondary outcomes are not reported. Safety counts are likewise descriptive and do not prove causation or settle comparative safety.

Randomization and a placebo comparator support a structured between-group comparison, but a registry summary is not a complete account of every design and analysis detail. The result should be understood as evidence about the reported endpoint, groups, and timeframe, with uncertainty described by the confidence interval and limitations described by the registry record. It is not a personalized prediction or a treatment recommendation.

18. Why This Trial Matters Statistically

Statistically, this trial is useful because it presents a randomized, placebo-controlled comparison in which both groups received insulin and the added study treatment differed. That structure makes the primary contrast specific: it concerns canagliflozin as an add-on versus placebo as an add-on, rather than an insulin-versus-no-insulin comparison. The parallel-group design and reported ANCOVA estimate give readers a direct between-group result for the principal continuous outcome.

The registry reports both group estimates and a sponsor-adjusted contrast with a confidence interval. That supports interpretation of direction and precision rather than relying on a p-value alone. A summary-level recalculation provides a consistency check, while its approximate and unadjusted basis illustrates why a sponsor's model-based result and a simple re-analysis should not be treated as interchangeable.

The record also demonstrates why statistical evidence needs context. The primary analysis uses last observation carried forward, while details about missingness, model covariates, and sensitivity analyses are not reported. Multiple secondary outcomes are listed, but no multiplicity strategy is reported. Those details affect the strength and scope of conclusions even when the primary p-value is small.

Finally, the posted sample-size illustrations show how planning depends on the effect chosen to detect and on stated assumptions. They are not evidence about the original design or about the probability that the completed study detected an effect. Separating the trial's observed estimates, uncertainty, planning illustrations, and unreported design information helps prevent statistical significance from being mistaken for a complete assessment of importance or generalizability.

19. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

Several trial-specific limitations arise from the registry entries available. The primary endpoint uses the full analysis set with last observation carried forward, but the number and pattern of missing Week 16 measurements are not reported in the registry record. No alternative missing-data analyses are reported, so the sensitivity of the primary estimate to that imputation choice cannot be assessed from the posted results.

The sponsor's ANCOVA result is reported, but the full model specification, including its covariates and diagnostics, is not reported in the registry record. Randomization details and concealment procedures are also not reported. The record describes the title as double-blind and the masking field as triple, without explaining the masked roles. These omissions limit an independent review of the design and analysis procedures.

The registry lists several secondary outcomes, but detailed numeric results are not posted for blood pressure or the adverse-event and hypoglycemia-related outcome measures. It does not report a multiplicity strategy for the secondary analyses. Therefore, those outcomes cannot be fully evaluated from the record, and the reported unadjusted secondary p-values carry less confirmatory weight.

The study is described as lasting 16 weeks, which confines interpretation to the reported period. The record does not report a clinical-importance threshold for the primary HbA1c difference or a sample-size justification for the original trial. The safety summary gives counts but does not provide enough context to establish causality or evaluate longer-term events. These limitations concern what can be established from the registry record; they do not imply that unreported procedures or analyses were absent.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Efficacy and Safety Study of Canagliflozin (TA-7284) in Combination With Insulin in Patients With Type 2 Diabetes Mellitus — NCT02220920.
• Registry record last updated 2026-01-08; status last verified 2025-12.
• Results first posted 2016-12-15.
• 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

NCT02220920 is a completed, randomized, parallel-group, placebo-controlled Phase 4 study of canagliflozin (TA-7284) added to insulin versus placebo added to insulin in participants with type 2 diabetes mellitus. The registry records enrollment as 146 and describes the planned study period as 16 weeks. The primary endpoint is change in HbA1c from baseline to Week 16 in the full analysis set with last observation carried forward.

The sponsor's ANCOVA estimate for the primary between-group difference is -1.10, with a 95% confidence interval from -1.33 to -0.87 and p < 0.001. Numerically reported secondary comparisons include fasting plasma glucose and percent change in body weight; detailed numerical results for other listed secondary measures are not reported in the registry record. Serious and other adverse-event counts are descriptive, and death counts are not reported in the registry record.

Interpretation is limited by the unreported details of the original design assumptions, model specifications, missing-data pattern, sensitivity analyses, and multiplicity approach. The results describe the stated comparison, analysis population, and timeframe.