Dapagliflozin Evaluation in Patients With Inadequately Controlled Type 1 Diabetes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Dapagliflozin Evaluation in Patients With Inadequately Controlled Type 1 Diabetes (NCT02268214), a randomized phase 3 study. The primary outcome is Adjusted Mean Change in HbA1c From Baseline at Week 24.

2. Clinical Question

3. Trial Design

NCT02268214 is a completed Phase 3, randomized, double-blind, placebo-controlled, parallel-group study sponsored by AstraZeneca. The registry lists an enrollment of 833. The study started on 2014-11-11, and primary completion was recorded as 2017-01-04. The registry describes fourfold masking, but does not report further details about how masking was implemented.

Participants were assigned to one of three groups: dapagliflozin 5 mg plus insulin, dapagliflozin 10 mg plus insulin, or placebo plus insulin. The assigned tablet was taken once daily for 52 weeks. This three-group structure allows each dose group to be compared with a shared placebo group. The registry does not report the method used to generate the randomization sequence, the allocation concealment procedures, or the intended allocation ratio.

The primary endpoint was adjusted mean change in HbA1c from baseline at Week 24, analyzed using a repeated measures model. The registry identifies the primary analysis population as randomized participants who took at least one dose of double-blind study medication during the short-term double-blind period. It also states that the first 55 randomized participants were excluded from the full analysis dataset because of a randomization system error. The effect of that exclusion on the trial’s conduct or results is not reported in the registry record.

The registry lists additional outcomes involving insulin dose, body weight, continuous glucose monitoring measures, and participants meeting a specified HbA1c reduction criterion without severe hypoglycemia events. The original sample-size justification, planned analysis details beyond those posted, and any interim analyses are not reported in the registry record.

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: 833 participants were randomized to a treatment group. Of the 771 participants not randomized to a treatment group: 585 No longer met study criteria, 125 withdrew consent, 26 were lost to follow-up, and 35 did not continue for other reasons

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All randomized subjects who took at least one dose of double-blind study medication during the short-term double-blind period. The first 55 randomized subjects will be excluded from the full analysis dataset due to the presence of a randomization system error.

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.

Population: Safety Analysis Set: It consists of all subjects who received at least one dose of double-blind study medication during the short-term double-blind treatment period.

6. Primary Endpoint

The registry lists the primary outcome as Adjusted Mean Change in HbA1c From Baseline at Week 24, measured over From Baseline to Week 24.

The primary endpoint measures the change in HbA1c from the start of the study to Week 24. HbA1c is a continuous measure of average blood glucose over time, expressed as a percentage of hemoglobin. A change from baseline describes how the measure moved during follow-up; a between-group contrast compares that change in one randomized group with the change in another.

The registry reports an adjusted mean change, estimated using a repeated measures model. Adjustment means that the reported estimate comes from a statistical model rather than simply subtracting two unadjusted group averages. The registry description specifies repeated measurements, but does not report all model details, such as the included covariates or the covariance structure.

A Week 24 endpoint provides a defined follow-up time at which the randomized groups can be compared. It is one measure of glycemic change and does not encompass every outcome relevant to the study. The endpoint should be read alongside the other posted outcomes and the safety information, rather than as a complete account of the trial. The registry record does not report a clinically meaningful threshold for interpreting the size of the HbA1c difference.

7. Results Posted in the Registry

• Dapagliflozin 5 mg + Insulin: mean -0.45 versus -0.03 on control; difference -0.42 (95% CI -0.57 to -0.27), standardized effect d = -0.49; z-test p = <0.001.
• Dapagliflozin 10 mg + Insulin: mean -0.47 versus -0.03 on control; difference -0.44 (95% CI -0.59 to -0.29), standardized effect d = -0.51; 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.

Adjusted Mean Percent Change in Total Daily Insulin Dose From Baseline at Week 24

• 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:
  • Dapagliflozin 5 mg + Insulin: mean -7.74 versus 1.16 on control; difference -8.90 (95% CI -13.27 to -4.53), standardized effect d = -0.35; z-test p = <0.001.
  • Dapagliflozin 10 mg + Insulin: mean -12.16 versus 1.16 on control; difference -13.32 (95% CI -17.62 to -9.02), standardized effect d = -0.54; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Dapagliflozin 5 mg + Insulin 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 Dapagliflozin 10 mg + Insulin than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Adjusted Mean Percent Change in Body Weight From Baseline at Week 24

• 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:
  • Dapagliflozin 5 mg + Insulin: mean -3.00 versus 0.05 on control; difference -3.05 (95% CI -3.71 to -2.39), standardized effect d = -0.80; z-test p = <0.001.
  • Dapagliflozin 10 mg + Insulin: mean -3.67 versus 0.05 on control; difference -3.72 (95% CI -4.37 to -3.07), standardized effect d = -0.98; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Dapagliflozin 5 mg + Insulin 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 Dapagliflozin 10 mg + Insulin than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Adjusted Mean Change in 24-hour Mean Continuous Glucose Monitoring Glucose From Baseline at Week 24

• 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:
  • Dapagliflozin 5 mg + Insulin: mean -10.28 versus 5.06 on control; difference -15.34 (95% CI -20.63 to -10.05), standardized effect d = -0.52; z-test p = <0.001.
  • Dapagliflozin 10 mg + Insulin: mean -12.97 versus 5.06 on control; difference -18.03 (95% CI -23.37 to -12.69), standardized effect d = -0.61; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Dapagliflozin 5 mg + Insulin 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 Dapagliflozin 10 mg + Insulin than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes cover several related aspects of glycemic control and treatment experience: daily insulin dose, body weight, continuous glucose monitoring measures, and a combined HbA1c reduction and severe hypoglycemia criterion. The posted continuous outcomes generally show changes in the direction of lower measured values in the dapagliflozin groups relative to placebo, while the reported categorical outcome asks how many participants met a defined response criterion without severe hypoglycemia events. Taken together, the direction of the continuous results is broadly consistent with the primary HbA1c contrast, but these measures are distinct outcomes and should not be treated as interchangeable confirmation of one another.

Each secondary measure has its own scale, analysis population, and potential sources of uncertainty. Continuous glucose monitoring outcomes, for example, describe glucose readings over a specified period, whereas insulin dose and body weight describe different features of treatment and participants. Their estimates may support a broader description of the posted results, but they do not change the definition of the primary endpoint.

The registry lists multiple outcomes and analyses. When many outcomes or comparisons are examined, the chance of finding nominally small p-values by chance can increase if no multiplicity procedure is applied. The registry record does not report whether secondary analyses were adjusted for multiplicity or which outcomes were designated as confirmatory. For that reason, unadjusted secondary results generally carry less inferential weight than a prespecified primary analysis. They are useful as supporting or exploratory evidence, but should be interpreted cautiously and in context rather than as independent proof.

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

• Adjusted Mean Percent Change in Total Daily Insulin Dose From Baseline at Week 24
• Adjusted Mean Percent Change in Body Weight From Baseline at Week 24
• Adjusted Mean Change in 24-hour Mean Continuous Glucose Monitoring Glucose From Baseline at Week 24
• Adjusted Mean Change in 24-hour Continuous Glucose Monitoring MAGE From Baseline at Week 24
• Adjusted Mean Change in Percent 24-hour Continuous Glucose Monitoring Glucose > 70 and <= 180 (mg/dL) From Baseline at Week 24
• Subjects With HbA1c Reduction From Baseline to Week 24 (LOCF) >= 0.5% and Without Severe Hypoglycemia Events

9. Statistical Interpretation

10. Statistical Methods Explained

The primary measure is continuous: HbA1c can take a range of values rather than falling into a simple yes-or-no category. The registry reports adjusted mean changes and between-group contrasts. A mean summarizes the group’s average value, while adjustment uses a model to account for the analysis structure or other modeled terms. The registry names a repeated measures model but does not report the full specification. In general, repeated-measures methods use information from measurements collected across follow-up and account for their relationship within a participant. Exactly which visits and model terms contributed to the posted estimate are not reported in the registry record.

What does the group contrast mean?

A contrast expresses how the outcome estimate in one group differs from the estimate in another. For the primary endpoint, a negative contrast means the estimated change in HbA1c was more negative in the dapagliflozin group than in the placebo group. That is a description of the direction of the contrast, not a statement about clinical importance. The reported comparison for 5 mg is a mean difference. The sponsor’s 10 mg comparison is reported as a median difference, while the simple recalculation is a mean difference. Means and medians describe different features of a distribution, so the two kinds of estimate should be distinguished.

What does the standard error describe?

The registry lists standard errors alongside adjusted means. A standard error describes the estimated sampling uncertainty of an estimate: it reflects how much estimates might vary across repeated studies under the model and its assumptions. It is not the same as the spread of individual participants’ measurements. The posted standard errors are used in the recalculated comparisons, but those calculations do not reproduce the sponsor’s full repeated measures model.

How should p-values be read?

The sponsor reports p < 0.001 for both primary comparisons. The recalculated unadjusted comparisons report p <0.001. A p-value is calculated under a null hypothesis and a statistical model; smaller values indicate that the observed data would be less compatible with that null model if its assumptions hold. A p-value does not measure the probability that the null hypothesis is true, the probability that the result will replicate, or the clinical importance of the estimate. The registry does not report whether or how the primary p-values were adjusted for multiple testing.

Why distinguish the sponsor analysis from a recalculation?

The sponsor’s primary results use a repeated measures model and report confidence intervals and p-values. The recalculated contrasts are simple, unadjusted comparisons made from the posted group estimates and standard errors. Because the recalculation does not have the full participant-level data or the model specification, it cannot incorporate all features of the sponsor’s analysis. The recalculated estimates are useful as a transparent check on the direction and approximate scale of the posted summary statistics, but should not be presented as a replication of the sponsor’s model.

What does the analysis population imply?

The stated primary population consists of randomized participants who took at least one dose of double-blind study medication during the short-term double-blind period. This definition excludes randomized participants who did not meet that dosing condition. The registry also says that the first 55 randomized participants were excluded from the full analysis dataset because of a randomization system error. The record does not explain whether sensitivity analyses examined alternative populations, how participants with missing endpoint measurements were handled, or how the exclusion affected the estimates. These omissions limit independent assessment of the robustness of the reported analysis.

What is not specified?

The registry record does not report the original sample-size justification, detailed model terms, missing-data assumptions, multiplicity procedures, or interim-analysis plan. General statistical practice can describe why these choices matter, but it cannot establish which choices were made in this trial. The posted results should therefore be interpreted according to the methods that are explicitly identified, without assuming additional protocol details.

11. Confidence Intervals and Precision

A confidence interval gives a range of values compatible with the data and the analysis model at the stated confidence level. It expresses uncertainty around an estimate; it is not a range containing a fixed proportion of individual participants’ outcomes. In general, narrower intervals indicate greater statistical precision, while wider intervals indicate more uncertainty. A confidence interval that does not include the null value is often associated with a small p-value under a corresponding test, but the interval also shows the range of effect sizes that remain compatible with the data.

For the sponsor’s 5 mg comparison, the reported mean difference is -0.42, with a 95% confidence interval from -0.56 to -0.28. For the 10 mg comparison, the sponsor reports a median difference of -0.45, with a 95% confidence interval from -0.58 to -0.31. Both intervals lie on the negative side of zero. Their widths convey that the estimates are not exact, even though the p-values are small. The 10 mg interval applies to a median difference, not a mean difference.

The recalculated intervals are based on unadjusted contrasts from posted summary statistics rather than the sponsor’s repeated measures model. They should therefore be interpreted as approximate, model-independent checks, not as the trial’s primary confidence intervals. The sponsor’s analysis is the appropriate reference for the model-based estimates described in the registry. The registry does not report a prespecified threshold for a clinically important difference, so statistical precision alone cannot determine whether an estimate is important in practice.

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 planning figures are participants per arm for 80% power or participants per arm for 90% power. They are not total events. The estimates use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. For each illustration, the effect to detect is set as a stated fraction of the observed effect. These are planning assumptions for an illustration; the trial’s own original design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives the estimated number of participants per arm for a design with an 80% chance of detecting the specified effect under the assumptions used. The 90% column gives the corresponding planning number for a 90% chance. These are design probabilities under an assumed effect and model, not guarantees about what a future study will find. For example, the primary 5 mg comparison’s illustration at 100% of the observed effect gives 66 participants per arm for 80% power and 89 participants per arm for 90% power. At 50% of the observed effect, the corresponding figures are 264 and 354 participants per arm.

Smaller effects are harder to distinguish from random variation, so detecting a smaller fraction of the observed effect generally requires more participants. A higher target power also requires a larger planned sample under the same assumptions. These planning numbers are derived from the observed results and should not be mistaken for evidence that the completed trial had a particular probability of producing a significant result. 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 = -0.420; pooled standard deviation σ = 0.861.
2. n = 2 × (1.96 + 0.842)² × 0.861² / -0.420² = 65.9, rounded up to 66 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 <- -0.42                      # observed difference in means
sigma <- 0.860804620603684                     # 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(2268214)
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 = 66 (80% power), 89 (90% power)
 75% of observed effect: n per arm = 118 (80% power), 157 (90% power)
 50% of observed effect: n per arm = 264 (80% power), 354 (90% power)
power.t.test: 66.9 
Simulated power at n = 66 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 sponsor’s primary results use a repeated measures model and report a mean difference of -0.42 for dapagliflozin 5 mg versus placebo, with a 95% confidence interval from -0.56 to -0.28 and p < 0.001. For dapagliflozin 10 mg versus placebo, the sponsor reports a median difference of -0.45, with a 95% confidence interval from -0.58 to -0.31 and p < 0.001.

The simple recalculation based on posted summary statistics gives a mean difference of -0.42 for 5 mg and -0.44 for 10 mg, with unadjusted p-values of <0.001. The 5 mg point estimates are the same at the displayed precision. The 10 mg estimates are close but are not the same measure: the sponsor’s result is a median difference, whereas the recalculation is a mean difference. The model-based and unadjusted analyses also use different methods, so differences in estimates or intervals are not necessarily evidence of an error.

The registry indicates no discrepancy between the sponsor’s reported analyses and the posted results. The sponsor’s model-based analysis remains the primary reported analysis; recalculations are a supplementary check using summary statistics, not a substitute for the full model or its participant-level data.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study tests more than one outcome, dose, time point, or comparison. Each test has its own chance of a false-positive result under its null hypothesis. Conducting many tests without an adjustment can increase the overall chance that at least one result appears statistically significant by chance. Common general approaches include controlling a family-wise error rate or a false discovery rate, but the choice depends on the objectives and analysis plan.

The registry lists a primary endpoint and multiple secondary outcomes, and reports 14 analyses. It does not report the full testing hierarchy, whether a multiplicity adjustment was used, or which secondary comparisons were considered confirmatory. Thus, the posted p-values should be read as reported for their individual analyses; the record does not establish whether they control an overall error rate across the set of outcomes and comparisons.

The registry identifies a repeated measures model for the primary endpoint, but does not give its full specification. The primary analysis population is described as randomized participants who took at least one dose of double-blind medication during the short-term double-blind period. The first 55 randomized participants were excluded from the full analysis dataset because of a randomization system error. Details about the timing and pattern of missing observations, imputation or likelihood assumptions, sensitivity analyses, and alternative analysis populations are not reported in the registry record.

Missing data can affect an estimate when the probability of an observation being absent is related to the outcome or to participant characteristics. Statistical methods often rely on assumptions about such missingness. Because the registry does not state the methods or assumptions used, their impact cannot be evaluated from the posted record alone. The same caution applies to any unreported interim analyses or further analysis-plan decisions: their existence and handling are not reported in the registry record.

16. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

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: diabetic ketoacidosis (3.6%), hypoglycaemia (1.4%), ketoacidosis (1.4%), ankle fracture (0.8%), vitreous haemorrhage (0.7%).

The posted safety counts describe participants at risk and participants affected by serious adverse events or other adverse events in each group. For serious adverse events, the reported affected counts are 37 of 277 in the 5 mg group, 40 of 296 in the 10 mg group, and 30 of 260 in the placebo group. The deaths recorded are 0, 0, and 1, respectively. These are descriptive counts; they do not by themselves establish whether groups differ statistically or whether an event was caused by study treatment.

The registry lists diabetic ketoacidosis at 3.6%, hypoglycaemia at 1.4%, ketoacidosis at 1.4%, ankle fracture at 0.8%, vitreous haemorrhage at 0.7%, and ketosis at 0.7% among the top serious-event terms. The record does not provide enough context in this summary to assign those percentages to a particular arm or to infer comparative risks from them. Event counts should be interpreted with the number at risk, follow-up, event definitions, and clinical context in mind. The registry figures alone do not establish a causal relationship or provide a complete assessment of safety.

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

The posted primary results show lower adjusted mean HbA1c changes from baseline in both dapagliflozin groups than in the placebo plus insulin group at Week 24. The sponsor’s reported comparisons have negative estimates and confidence intervals on the negative side of zero, with p < 0.001 for each comparison. The registry’s secondary results cover additional measures, but their interpretation depends on their own analysis populations, methods, and the unreported approach to multiplicity.

These findings describe group-level statistical results for the participants included in the posted analyses. They do not establish what an individual participant’s outcome would be, nor do they answer every question about longer-term outcomes. A statistically clear between-group estimate is not by itself a judgment about clinical importance. The registry does not report a threshold for a clinically meaningful HbA1c difference, and no such threshold can be inferred from the p-value alone.

The trial included three groups and the reported primary comparisons are each against placebo. The registry results do not establish a direct statistical comparison between the 5 mg and 10 mg doses. The estimates being numerically close does not demonstrate that the doses are equivalent, nor does it establish that one dose is better than the other.

Interpretation is also constrained by details that are not reported: the original sample-size justification, complete model specification, missing-data procedures, multiplicity handling, and further sensitivity analyses. The stated exclusion of the first 55 randomized participants due to a randomization system error is relevant when considering the analysis population. The results should therefore be understood as the reported estimates and uncertainty for this trial, rather than as a complete account of every methodological or clinical question.

18. Why This Trial Matters Statistically

This trial is statistically informative because it reports a randomized, placebo-controlled comparison with a defined continuous primary endpoint and estimates for two active-dose groups. Its parallel-group design supplies separate comparisons against the same placebo group, while the Week 24 endpoint gives a common point for assessing change in HbA1c. The sponsor also reports confidence intervals and p-values, allowing readers to consider both the estimated contrasts and their statistical uncertainty.

The record illustrates why an estimate and its method should be read together. The sponsor’s repeated measures model provides the primary reported analysis, while recalculations from group summary statistics provide simpler unadjusted contrasts. These methods address the same broad group comparison but are not interchangeable. The difference between a mean and median estimate for the 10 mg comparison further underscores the importance of checking which measure is being reported.

The trial also highlights interpretive issues common to clinical research: a stated analysis population, an exclusion due to a randomization system error, multiple secondary outcomes, and incomplete reporting of missing-data and multiplicity methods. These details affect how much confidence readers can place in the broader set of results and how far conclusions can be generalized. The posted findings are therefore most useful when the effect estimates, uncertainty, analysis methods, and record-level omissions are considered together.

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 limitations are specific to what the registry record reports. The primary analysis excluded the first 55 randomized participants from the full analysis dataset because of a randomization system error. Although the record gives this reason, it does not describe the error in detail, its consequences, or sensitivity analyses assessing the effect of the exclusion.

The registry names a repeated measures model but does not report the full model specification. It also does not report detailed missing-data methods, assumptions, or the amount and pattern of missing endpoint information. The original sample-size justification, any interim analyses, and the complete statistical analysis plan are not reported in the registry record. Multiplicity procedures for the numerous secondary outcomes and comparisons are also not reported.

The primary endpoint group summaries and recalculated contrasts are based on posted aggregate statistics. They cannot reproduce a participant-level model or independently verify every analysis choice. In addition, the sponsor’s 10 mg primary result is a median difference, whereas the recalculation is a mean difference; the two estimates represent different summaries. The registry does not report a threshold for clinical importance or a direct dose-versus-dose primary comparison.

Safety counts are descriptive, and the posted summary does not establish causality or a formal between-group safety comparison. The serious-event term percentages are not accompanied by arm-specific denominators. These limitations do not negate the reported estimates, but they restrict how much can be concluded from the registry record alone.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Dapagliflozin Evaluation in Patients With Inadequately Controlled Type 1 Diabetes — NCT02268214.
• Registry record last updated 2018-09-13; status last verified 2018-08.
• Results first posted 2018-03-29.
• 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

NCT02268214 is a completed Phase 3 randomized, double-blind, placebo-controlled, parallel-group trial sponsored by AstraZeneca. The registry lists 833 participants and compares dapagliflozin 5 mg plus insulin and dapagliflozin 10 mg plus insulin with placebo plus insulin. The primary endpoint is adjusted mean change in HbA1c from baseline at Week 24, analyzed using a repeated measures model.

The sponsor reports a negative primary comparison for each dose versus placebo, with 95% confidence intervals and p < 0.001. The primary analysis population excludes the first 55 randomized participants because of a randomization system error. The registry also posts secondary outcome and safety results, but does not report several details needed for a full independent assessment, including the original sample-size justification, missing-data methods, and multiplicity procedures.