Efficacy and Safety of Efpeglenatide Versus Dulaglutide in Patients With Type 2 Diabetes Mellitus Inadequately Controlled With Metformin: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy and Safety of Efpeglenatide Versus Dulaglutide in Patients With Type 2 Diabetes Mellitus Inadequately Controlled With Metformin (NCT03684642), a randomized phase 3 study. The primary outcome is Change From Baseline to Week 56 in HbA1c.

2. Clinical Question

In practical statistical terms, the primary comparison asks how the adjusted average change in HbA1c differed between each efpeglenatide regimen and dulaglutide. It is not a comparison of individual responses, and an average difference does not describe the full range of outcomes among participants. The results are specific to this treatment setting, follow-up period, and analysis population.

3. Trial Design

This was a phase 3, randomized, parallel-group, active-controlled study with no masking. The official title describes it as a 56-week, multicenter study. The registry lists enrollment of 908 and the status as terminated. The reason for termination is not reported in the registry record, so the status alone does not establish why the study ended or whether termination was related to efficacy, safety, recruitment, or another circumstance.

There were three treatment groups: efpeglenatide 4 mg, efpeglenatide 6 mg, and dulaglutide 1.5 mg. Each was administered once weekly by subcutaneous injection with metformin as background therapy. The regimens included dose escalation before reaching their listed doses. The record describes planned treatment through Week 56 and states that the primary endpoint covered baseline to Week 56.

Randomization helps form groups for comparative assessment, while the parallel design means participants were assigned to one of the regimens rather than crossing between treatments. The study was open-label, meaning masking was not used. General statistical context: lack of masking can matter for outcomes influenced by participant or investigator expectations, although laboratory-based HbA1c is less dependent on subjective rating than many clinical outcomes. The record does not report further details about safeguards against bias.

The primary outcome analysis used the modified intent-to-treat population. Its registry definition includes participants who completed study treatment; participants who discontinued treatment and completed or discontinued the study before early termination; and participants who discontinued treatment before early termination and discontinued the study. The group-specific numbers reported for outcome analyses are not the same thing as the trial’s enrollment figure. The record does not explain the relationship between all enrollment and analysis counts in detail.

The trial start date is 2018-09-26 and the primary completion date is 2020-10-13. The dates and status describe the registry record; they do not, by themselves, indicate how the results should be interpreted. The original sample-size justification and the trial’s prespecified statistical plan 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: A total of 908 participants were randomized in 1:1:1 ratio to either efpeglenatide 4 milligram (mg), efpeglenatide 6 mg, or dulaglutide 1.5 mg treatment arms, stratified by screening glycated hemoglobin (HbA1c) values (less than \[\<\]8%, greater than or equal to \[\>=\]8 percent \[%\]) and by body mass index (BMI) (\<30 kg/m\^2 and \>=30 kg/m\^2) on Day 1.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Analysis was performed on modified intent-to-treat (mITT) population which included participants who completed study treatment; or who discontinued study treatment and completed/discontinued study before early termination; or who discontinued treatment before early termination and discontinued study due to early termination; or who discontinued treatment due to early termination within 30 days of target Week 56 visit.

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: Analysis was performed on intent-to-treat (ITT) population, which included all randomized participants, analyzed according to the treatment group allocated by randomization.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline to Week 56 in HbA1c, measured over Baseline to Week 56.

The primary endpoint was change from baseline to Week 56 in HbA1c, reported as an adjusted least-squares mean in percentage of HbA1c. HbA1c reflects average blood glucose over a period of time and is widely used as a measure of glycemic control. In general, an endpoint expressed as change focuses on how the measured value moved from the starting point, rather than only describing the value at follow-up.

The registry reports adjusted least-squares means and standard errors from an analysis of covariance, or ANCOVA, model. Such adjustment can account for modeled covariates and can produce a comparison that is not simply the raw difference between group averages. The exact covariates used in this model are not reported in the registry record. Missing values for the primary analysis were imputed using a baseline-observation-carried-forward-like multiple-imputation method.

For a non-inferiority question, the relevant issue is whether the observed difference and its uncertainty are compatible with a prespecified acceptable loss relative to the active comparator. General statistical context: the margin defines that acceptable loss and must be specified before results are interpreted. The non-inferiority margin and the registry’s full decision rule are not reported in the registry record. Therefore the reported estimates and intervals can be described, but they do not by themselves establish whether the prespecified non-inferiority criterion was met.

7. Results Posted in the Registry

• Efpeglenatide 6 mg: mean -1.17 versus -1.12 on control; difference -0.05 (95% CI -0.22 to 0.12), standardized effect d = -0.05; z-test p = 0.556.
• Dulaglutide 1.5 mg: mean -1.09 versus -1.12 on control; difference 0.03 (95% CI -0.14 to 0.20), standardized effect d = 0.03; z-test p = 0.724.

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 From Baseline to Week 56 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:
  • Efpeglenatide 6 mg: mean -3.04 versus -2.87 on control; difference -0.17 (95% CI -1.99 to 1.65), standardized effect d = -0.02; z-test p = 0.854.
  • Dulaglutide 1.5 mg: mean -2.81 versus -2.87 on control; difference 0.06 (95% CI -1.74 to 1.86), standardized effect d = 0.01; z-test p = 0.948.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Efpeglenatide 6 mg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Dulaglutide 1.5 mg and the control. The interval gives the range of effects that remain compatible with the data.

Number of Participants With HbA1c < 7.0 %

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Efpeglenatide 6 mg: 61.8% (157/254) versus 60.3% (155/257) on control; risk difference 1.5 percentage points (95% CI -7.0 to 10.0), risk ratio 1.02 (0.89–1.18); unadjusted two-proportion z-test p = 0.728.
  • Dulaglutide 1.5 mg: 60.0% (150/250) versus 60.3% (155/257) on control; risk difference -0.3 percentage points (95% CI -8.8 to 8.2), risk ratio 0.99 (0.86–1.15); unadjusted two-proportion z-test p = 0.943.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Efpeglenatide 6 mg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Dulaglutide 1.5 mg and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline to Week 56 in Fasting Plasma Glucose (FPG)

• 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:
  • Efpeglenatide 6 mg: mean -1.57 versus -1.81 on control; difference 0.24 (95% CI -0.18 to 0.66), standardized effect d = 0.10; z-test p = 0.258.
  • Dulaglutide 1.5 mg: mean -1.71 versus -1.81 on control; difference 0.10 (95% CI -0.32 to 0.52), standardized effect d = 0.04; z-test p = 0.637.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Efpeglenatide 6 mg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Dulaglutide 1.5 mg and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes broaden the picture beyond average HbA1c change. They include body-weight change, the number of participants below the stated HbA1c threshold at Week 56, change in fasting plasma glucose, and hypoglycemia outcomes measured by participants with events and events per participant-year. Taken together, these outcomes cover related but distinct aspects of glycemic control, weight, and safety. They should not be treated as interchangeable measures or as direct substitutes for the primary endpoint.

For the secondary outcomes with numerical results posted, the group summaries and unadjusted comparisons generally do not show a clear separation between the study groups. As with the primary recalculations, these comparisons are descriptive and unadjusted; the registry does not report sponsor analyses for these secondary outcomes. The posted secondary findings do not supply a separate basis for establishing the primary non-inferiority objective, which depends on the primary endpoint, the formal analysis, and the non-inferiority margin.

The registry lists hypoglycemia outcomes, including documented symptomatic episodes below the stated threshold and severe episodes, but their numerical results are not reported in the registry record. The record therefore does not support a comparison of those event outcomes across groups from the posted secondary results.

General statistical context: when a study examines multiple secondary endpoints and comparisons, the chance of at least one apparently unusual result can increase if each is assessed without accounting for multiplicity. The registry does not report a multiplicity adjustment or a hierarchy for interpreting the secondary endpoints. Unadjusted secondary results therefore carry less confirmatory weight than a prespecified primary analysis, and any individual finding should be interpreted cautiously rather than isolated from the broader set of outcomes.

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

• Change From Baseline to Week 56 in Body Weight
• Number of Participants With HbA1c < 7.0 %
• Change From Baseline to Week 56 in Fasting Plasma Glucose (FPG)
• Number of Participants With At Least One Hypoglycemic Events (Documented Symptomatic Hypoglycemia <3.0 mmol/L [<54 mg/dL], Severe Hypoglycemia)
• Number of Hypoglycemic Events (Documented Symptomatic Hypoglycemia <3.0 mmol/L [<54 mg/dL] and Severe Hypoglycemia) Per Participant-Year

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is a continuous measure, and the registry reports adjusted least-squares means with standard errors. The analysis compares mean change in HbA1c from baseline to Week 56 among the treatment groups. General statistical context: a mean condenses a distribution to a central value, so it is useful for comparing average changes but does not show the distribution’s shape, the number of participants who improved, or how responses differed within a group.

What does an adjusted least-squares mean represent?

An ordinary arithmetic mean is calculated directly from observed values. A least-squares mean is an estimate from a fitted statistical model, often adjusted for specified covariates. ANCOVA combines outcome information with baseline or other modeled variables to estimate group means under the model. Here, the registry states that ANCOVA was used to account for missing data, but does not list the covariates or give enough detail to reproduce the sponsor’s model. The adjusted values should therefore be distinguished from the summary-statistic recalculations, which use the posted means and standard errors rather than the full model.

The three adjusted HbA1c changes are -1.12, -1.17, and -1.09 percentage of HbA1c for efpeglenatide 4 mg, efpeglenatide 6 mg, and dulaglutide 1.5 mg, respectively. Each has a standard error of 0.06. General statistical context: a standard error describes the uncertainty of an estimated mean under the analysis, not the variation of individual participant measurements. It should not be interpreted as a standard deviation or as a range containing most individual outcomes.

How should the non-inferiority comparisons be read?

Non-inferiority asks whether a new treatment is not worse than a comparator by more than a prespecified margin. It is not the same as testing whether the treatments differ at all. The sponsor reports differences and 95% confidence intervals for both comparisons: -0.03 with an interval from -0.20 to 0.14 for efpeglenatide 4 mg versus dulaglutide, and -0.08 with an interval from -0.25 to 0.09 for efpeglenatide 6 mg versus dulaglutide. The record does not report the non-inferiority margin. Without it, the reported intervals cannot be checked against the boundary that would define non-inferiority.

General statistical context: a confidence interval describes a range of effect values compatible with the data and the assumptions of the analysis at the stated confidence level. It gives information about both the estimate and its precision. An interval that includes zero indicates that the data are compatible with no difference as well as with differences in either direction, but zero alone is not the relevant threshold for a non-inferiority decision. That decision depends on the prespecified margin and the analysis plan.

What do the superiority p-values test?

The sponsor reports ANCOVA superiority p-values of 0.706 for efpeglenatide 4 mg versus dulaglutide and 0.343 for efpeglenatide 6 mg versus dulaglutide. General statistical context: a p-value describes how unusual data at least as extreme as those observed would be under a specified null model, assuming the model and analysis conditions. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of an effect. A lack of statistical evidence for superiority is not proof that two treatments are equivalent or that a non-inferiority criterion has been met.

The registry reports no p-value for the sponsor’s non-inferiority comparisons. The margin, confidence-interval procedure, and other details needed to reconstruct the full non-inferiority testing approach are not reported in the registry record. The superiority analyses and non-inferiority analyses should be kept distinct because they address different hypotheses.

What do the recalculations add?

Unadjusted recalculations use the posted group summaries to estimate simple pairwise differences and confidence intervals or p-values. For the primary outcome, the recalculated difference for efpeglenatide 6 mg relative to efpeglenatide 4 mg is -0.05, with an unadjusted p-value of 0.556. The recalculated difference for dulaglutide relative to efpeglenatide 4 mg is 0.03, with an unadjusted p-value of 0.724. These are not adjusted sponsor estimates and should not replace the sponsor’s reported model-based analyses. They are a transparent, approximate comparison of the summary statistics.

For the recalculations, the posted standard errors are used to approximate standard deviations. This makes the calculations approximate, particularly because adjusted least-squares means and their standard errors do not necessarily behave like raw sample means and standard deviations. The recalculated comparisons do not incorporate the sponsor’s model, missing-data procedure, or any unreported analysis details. The registry record does not identify a discrepancy flag between these calculations and the sponsor’s reported analysis.

Why can over-dispersion matter?

General statistical context: for count outcomes, a simple model may assume that variability is tied closely to the average count. If real counts vary more than that assumption allows, there is over-dispersion. Ignoring it can make standard errors too small and intervals too narrow. The registry lists a secondary hypoglycemia outcome as number of events per participant-year, but does not post numerical event-rate results or describe a count model for it. Over-dispersion cannot be evaluated from the event-rate results because those results are not reported in the registry record.

The primary endpoint is not a count outcome, so count over-dispersion is not the central issue for its mean comparison. More generally, appropriate uncertainty estimates depend on the outcome type and the fitted model. The details needed to evaluate model diagnostics, covariate choices, and sensitivity analyses are not reported in the registry record.

11. Confidence Intervals and Precision

The sponsor’s primary non-inferiority estimates have 95% confidence intervals. For efpeglenatide 4 mg versus dulaglutide, the reported difference is -0.03, with a 95% confidence interval from -0.20 to 0.14. For efpeglenatide 6 mg versus dulaglutide, the difference is -0.08, with a 95% confidence interval from -0.25 to 0.09. The intervals include zero and extend to values on both sides of zero, so the posted estimates do not pinpoint a single direction of difference.

General statistical context: interval width communicates precision. A narrower interval indicates that the estimate is more tightly constrained under the analysis assumptions; a wider interval leaves more uncertainty about the size and direction of the effect. A confidence interval is not a guarantee that the true effect lies inside the particular interval. It is a procedure with a stated long-run coverage property when its assumptions hold.

For this trial, interval interpretation is limited by missing design details. The non-inferiority margin is not reported in the registry record, so the intervals cannot be judged against the boundary needed for that objective. The registry also does not provide sponsor confidence intervals for the superiority analyses. The recalculated intervals are unadjusted approximations from posted summary statistics and should not be treated as substitutes for model-based intervals.

The displayed group standard errors for the primary adjusted means are 0.06. A standard error is uncertainty around a mean estimate, not a measure of participant-to-participant variability. The registry does not report the primary group standard deviations. Precision in a group mean also does not alone establish the precision of a between-group effect; the comparison interval is the more direct summary for that question.

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 these continuous and binary outcomes, not total event counts. The planning illustration uses the observed effect as its starting point, two-sided significance level 0.05, equal allocation between the two arms, and a normal approximation. It estimates the participants per arm needed to detect the stated fraction of the observed effect. These assumptions describe the planning illustration, not the trial’s original design; the original design assumptions are not reported in the registry record.

The 80% power column describes a hypothetical design that would have an 80% probability of detecting the specified effect under the stated assumptions if that effect were true. The 90% power column describes the corresponding hypothetical design with 90% probability. These figures are planning illustrations, not observed power. Observed power is not a useful substitute for reading the estimate and its uncertainty after the trial has been conducted.

The rows for 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect illustrate how the assumed target effect changes the planning requirement. Smaller effects require more participants per arm to detect under the same assumptions. For example, the primary comparison planning figures for 100% of the observed efpeglenatide 6 mg versus efpeglenatide 4 mg effect are 5,776 participants per arm for 80% power and 7,733 for 90% power. At 50% of that observed effect, the corresponding figures are 23,103 and 30,929 participants per arm.

These are not recommendations for a future trial and should not be interpreted as evidence about the trial’s actual power. They are sensitive to the assumed effect and the simplifying assumptions used in the illustration. The registry’s original sample-size justification and its own power 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 = -0.050; pooled standard deviation σ = 0.959.
2. n = 2 × (1.96 + 0.842)² × 0.959² / -0.050² = 5,775.7, rounded up to 5,776 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 <- -0.0499999999999998                      # observed difference in means
sigma <- 0.959078634477895                     # 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(3684642)
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 = 5,776 (80% power), 7,733 (90% power)
 75% of observed effect: n per arm = 10,268 (80% power), 13,746 (90% power)
 50% of observed effect: n per arm = 23,103 (80% power), 30,929 (90% power)
power.t.test: 5776.7 
Simulated power at n = 5,776 per arm: 0.800 (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 outcome analyses are model-based ANCOVA results for the modified intent-to-treat population, with least-squares mean differences reported for the non-inferiority comparisons. Those analyses are the registry’s reported primary analyses. They should be distinguished from unadjusted recalculations, which use posted summary statistics and approximate standard deviations rather than the full sponsor model.

For the primary endpoint, the sponsor reports a difference of -0.03 and a 95% confidence interval from -0.20 to 0.14 for efpeglenatide 4 mg versus dulaglutide. For efpeglenatide 6 mg versus dulaglutide, the reported difference is -0.08 and the 95% confidence interval is from -0.25 to 0.09. The sponsor’s superiority analyses report p-values of 0.706 and 0.343, respectively. The recalculated primary comparisons instead produce unadjusted differences of -0.05 for efpeglenatide 6 mg versus efpeglenatide 4 mg and 0.03 for dulaglutide versus efpeglenatide 4 mg.

Differences between sponsor estimates and recalculations can arise because adjustment, missing-data handling, and model specification affect estimates and uncertainty. The recalculations do not reproduce the ANCOVA. The registry does not report the non-inferiority margin or enough model detail to independently reconstruct the sponsor’s full analysis. No discrepancy between sponsor results and recalculations is indicated in the registry record.

15. Multiplicity, Missing Data and Analysis Populations

The registry reports several outcome measures and multiple between-group analyses. It does not report whether the sponsor adjusted for multiple comparisons, used a testing hierarchy, or designated particular secondary analyses as confirmatory. Without that information, the interpretation of multiple p-values is uncertain. General statistical context: assessing many hypotheses can increase the chance of at least one result appearing statistically notable by chance, so a prespecified multiplicity strategy helps define how evidence should be judged across outcomes.

The primary outcome description states that adjusted least-squares means and standard errors came from ANCOVA, and that missing values were imputed with a baseline-observation-carried-forward-like multiple-imputation method. The registry does not give the imputation model, assumptions about missingness, number or generation of imputed datasets, or sensitivity analyses. The method description is therefore informative but not sufficient to assess how robust the primary estimate is to alternative assumptions about missing data.

The analysis population is described as modified intent-to-treat and includes participants meeting specified treatment and study completion or discontinuation conditions. This differs from an uncomplicated analysis of every randomized participant, and the registry’s definition should be considered when interpreting the results. The record does not report detailed participant flow or explain all reasons for missing outcome data in the material available for these results.

The registry lists the trial as terminated, but the reason for termination is not reported in the registry record. It also does not report interim analyses, stopping rules, the protocol’s original sample-size justification, or the full statistical analysis plan. These omissions limit assessment of whether analyses and decision rules followed a prespecified plan. They do not establish that any particular analysis was changed; that information is simply 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: acute myocardial infarction (0.7%), atrioventricular block complete (0.7%), cholelithiasis (0.7%), atrial fibrillation (0.7%), pulmonary embolism (0.7%).

The posted serious-adverse-event counts are 20 among 313 participants at risk in the efpeglenatide 4 mg group, 23 among 292 at risk in the efpeglenatide 6 mg group, and 20 among 302 at risk in the dulaglutide 1.5 mg group. The registry also reports deaths of 0, 0, and 1, respectively. These figures describe the recorded groups and should be read as counts, not as adjusted comparisons or estimates of causality.

The registry lists serious adverse-event terms including acute myocardial infarction, atrioventricular block complete, cholelithiasis, atrial fibrillation, and pulmonary embolism, each reported at 0.7%, and coronary artery disease at 0.6%. The listed term percentages are not a substitute for the full safety record. The registry does not provide a comparative statistical analysis of serious adverse events here, and the counts alone do not establish whether observed events were caused by a study treatment.

Other adverse-event counts are also recorded: 202 among 313 at risk for efpeglenatide 4 mg, 180 among 292 at risk for efpeglenatide 6 mg, and 178 among 302 at risk for dulaglutide. The definitions and event-level detail needed to interpret this category further are not reported in the registry record. Safety figures should be considered alongside the duration and population of the study, without inferring rare-event risks from these counts alone.

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

The primary results describe adjusted average HbA1c changes through Week 56 in the registry-defined modified intent-to-treat population. The three group estimates are similar in magnitude, and the sponsor’s reported between-group confidence intervals extend across zero. This means the posted estimates do not establish a clear direction of difference on their own. It does not show that the treatments are identical, nor does it answer the formal non-inferiority question without the margin and decision rule.

The sponsor’s non-inferiority comparisons and superiority p-values address different statistical questions. A non-significant superiority result does not establish non-inferiority, equivalence, or equal effects. General statistical context: evidence for non-inferiority requires an interval or test assessed against a prespecified margin. Because that margin is not reported in the registry record, the primary objective cannot be independently evaluated from the registry entries alone.

The secondary measures can add context about weight, glycemic control, and hypoglycemia, but they do not replace the primary analysis. Some numerical secondary results are posted, while numerical hypoglycemia findings are not reported in the registry record. The secondary comparisons are unadjusted in the recalculations, and the registry does not report a multiplicity strategy. Their evidential role is therefore more limited than that of a complete prespecified analysis.

The safety counts describe events recorded in the treatment groups. They do not prove causation or establish that one regimen is safer than another. The registry’s terminated status should not be taken as evidence of a particular outcome, since the reason for termination is not reported in the registry record.

Overall, these results support a careful descriptive reading of average outcomes and their uncertainty, rather than a categorical claim that one treatment is better, equivalent, non-inferior, or safer. The registry record gives useful result estimates but omits several elements needed to assess the full planned statistical decision process.

18. Why This Trial Matters Statistically

This trial illustrates why the type of statistical question matters. The primary objective was non-inferiority, while the registry also reports superiority analyses. Those hypotheses require different interpretations: a test that fails to demonstrate superiority cannot be substituted for a formal non-inferiority assessment. The reported confidence intervals are therefore most informative when considered with the non-inferiority margin, which is not reported in the registry record.

It also illustrates the difference between a model-based result and an unadjusted recalculation. The sponsor’s adjusted estimates account for the ANCOVA model and missing-data procedure described in the record. Simple recalculations from the posted summary statistics provide a useful check on the scale and direction of the comparisons, but they cannot reproduce the sponsor’s model. Understanding which estimate answers which question is essential to sound interpretation.

The registry reports group means, standard errors, comparisons, and selected safety counts. Those summaries are useful, but they cannot convey all individual outcome variation or the full analysis process. Important details—including the non-inferiority margin, covariates, multiplicity plan, sample-size justification, and reason for termination—are not reported in the registry record. These omissions constrain what can be concluded from the posted results.

Finally, the trial demonstrates why numerical similarity should not be equated with proof of sameness. Confidence intervals express uncertainty, and statistical conclusions depend on the prespecified threshold relevant to the research question. Reading estimates, intervals, analysis populations, missing-data methods, and outcome multiplicity together gives a more defensible account than relying on a single p-value or a visual impression of the group averages.

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 does not report the non-inferiority margin, which is central to judging the stated primary objective. It also does not provide the full statistical analysis plan, including the covariates in ANCOVA, detailed confidence-interval procedures, or a complete description of the non-inferiority decision rule. The sponsor’s superiority results are reported as p-values without corresponding effect estimates or confidence intervals in the registry analysis details.

Missing-data handling is described as a baseline-observation-carried-forward-like multiple-imputation method, but the imputation model and sensitivity analyses are not reported in the registry record. The primary analysis uses a modified intent-to-treat population with a specific completion and discontinuation definition. That population may not represent all enrolled participants, and complete participant-flow details are not reported with these results.

The recalculated comparisons are unadjusted approximations based on posted summary statistics, including standard errors used to approximate standard deviations. They are not a replication of the sponsor’s ANCOVA. Secondary endpoint results do not include numerical hypoglycemia findings, and no multiplicity strategy is reported. Thus the secondary evidence is incomplete for a broad comparison across all listed outcomes.

The open-label design means masking was not used. The registry lists the study as terminated, but the reason for termination and any implications for follow-up or analysis are not reported in the registry record. The original sample-size justification, interim analyses, and detailed trial design assumptions are also not reported. These specific omissions limit independent assessment of the planned analysis and the context for interpreting the results.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Efficacy and Safety of Efpeglenatide Versus Dulaglutide in Patients With Type 2 Diabetes Mellitus Inadequately Controlled With Metformin — NCT03684642.
• Registry record last updated 2021-11-01; status last verified 2021-10.
• Results first posted 2021-11-01.
• 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

The registry describes a phase 3, randomized, open-label, parallel-group study of efpeglenatide versus dulaglutide added to metformin in participants with type 2 diabetes inadequately controlled with metformin. The primary endpoint was adjusted change in HbA1c from baseline to Week 56 in a modified intent-to-treat population, with missing values handled using a baseline-observation-carried-forward-like multiple-imputation method.

The posted group means show HbA1c decreases in all three groups. Sponsor-reported non-inferiority estimates include 95% confidence intervals, but the non-inferiority margin is not reported in the registry record. Recalculated comparisons are unadjusted and approximate, and should be distinguished from the sponsor’s ANCOVA analyses. Some secondary results and safety counts are available, while numerical hypoglycemia results are not reported in the registry record. The study status is terminated; the reason is not reported in the registry record.