The Vascular Effects of Exenatide Versus Metformin in Patients With Pre-Diabetes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for The Vascular Effects of Exenatide Versus Metformin in Patients With Pre-Diabetes (NCT00546728), a randomized phase 4 study. The primary outcome is Change in Reactive Hyperemic Index Over the 3-month Treatment Period.

2. Clinical Question

The registry describes a randomized, parallel-group study comparing the two drug regimens. Its stated purpose included assessing vascular health during chronic treatment and during the two-hour period following a meal. The posted primary outcome, however, is specifically the change in reactive hyperemic index over the three-month treatment period. The registry record does not post an acute post-meal outcome result.

The displayed group means are 0.01 for exenatide and -0.17 for metformin. The reported mean difference for the comparison is -0.18, with an unadjusted p-value of 0.363. The confidence interval extends from a negative value to a positive value, so it includes no difference as well as differences in either direction. These results describe the registered primary measure; they do not establish what would happen for other vascular measures, other treatment durations, or other patient populations.

3. Trial Design

This is a Phase 4, randomized, parallel-group study. The registry lists an enrollment of 50 and two treatment groups: exenatide and metformin. The arm descriptions specify exenatide 10 mcg twice daily and metformin 500 mg twice daily. The posted outcome results list 25 participants in each group. The registry status is completed, with a study start of 2007-10 and primary completion in 2011-07.

Masking is listed as none. This means the registry does not describe the study as blinded; it does not, by itself, tell the reader which people involved in treatment, measurement, or analysis knew the assigned group. More detailed masking procedures are not reported in the registry record. The randomization method, concealment procedures, treatment adherence, and any changes to the registered design are also not reported in the registry record.

The primary endpoint is the change in reactive hyperemic index from baseline to three months. The registry describes it as a continuous measure reported as a mean with a standard deviation. It says all completers were included in the primary analysis. Details about how many participants completed treatment, how missing measurements were handled, or whether the analysis used any adjustments are not reported in the registry record.

The record lists one outcome measure and no posted analyses. It does not post a secondary outcome result. The stated study purpose also mentions vascular effects during a two-hour period after a meal, but the primary result available in the registry is the three-month change. The original sample-size calculation, planned statistical model, prespecified handling of multiple outcomes, interim analyses, and analysis population beyond the description of all completers are not reported in the registry record. The sponsor is listed as St. Paul Heart Clinic.

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

Analysis population for the primary outcome, as defined in the registry: All completers were included in the analysis

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 Reactive Hyperemic Index Over the 3-month Treatment Period, measured over Change from baseline to 3 months.

Reactive hyperemic index is described as a ratio of fingertip blood-flow volume after temporary occlusion to baseline blood-flow volume. The registry identifies it as a measure of endothelial function, meaning the function of the inner lining of blood vessels. It also states that higher ratio values are considered indicative of better arterial health. The endpoint is the change in this index from baseline to three months, rather than the value at a single time point.

Using a change measure focuses the comparison on how the index differs between the start of treatment and the three-month assessment. In general, a vascular measure can provide a way to quantify a physiological response, but its interpretation depends on how it was measured and on the study population and timeframe. The record does not report details of measurement procedures beyond the index description, or how any repeated measurements were combined.

The group mean change was 0.01 in the exenatide arm and -0.17 in the metformin arm. The posted mean difference is -0.18. Those summaries describe the average observed change for the participants included in the reported analysis. They do not show the individual distribution of responses, and the registry record does not report a median, participant-level data, or a separate acute post-meal result. The standard deviations describe variability around each group mean; they are not confidence intervals for the group means.

7. Results Posted in the Registry

• Metformin: mean -0.17 versus 0.01 on control; difference -0.18 (95% CI -0.57 to 0.21), standardized effect d = -0.26; z-test p = 0.363.

8. Statistical Interpretation

9. Statistical Methods Explained

The outcome is continuous and is summarized as a mean and standard deviation. The mean describes the arithmetic average change in reactive hyperemic index within a group. The standard deviation describes how dispersed individual measurements are around that group mean. It is not the uncertainty in the mean itself, and it is not a range containing every participant’s result. The registry lists the mean change as 0.01 with standard deviation 0.68 for exenatide, and -0.17 with standard deviation 0.72 for metformin.

What does the mean difference represent?

The comparison reports a mean difference of -0.18 for metformin relative to exenatide. It is the contrast between the two group means, expressed on the reactive hyperemic index change scale. A negative sign indicates a lower mean change in the metformin group in this comparison. The comparison is unadjusted: it does not account for covariates in the recalculation from the posted summary statistics. The registry record does not report an adjusted analysis or identify covariates used in any sponsor analysis.

What does the p-value say?

The unadjusted p-value is 0.363. General statistical interpretation: a p-value describes how unusual data at least as incompatible with a specified null hypothesis would be if that hypothesis and the test assumptions were true. It is not the probability that the null hypothesis is true, the probability that the result occurred by chance, or a measure of the probability that a treatment is beneficial. The registry record does not specify the test or the null hypothesis in a posted analysis plan, so the p-value should be read as the unadjusted comparison associated with the posted summary statistics rather than as evidence of a fully documented analysis procedure.

Why does variability matter?

General statistical explanation: a mean difference is estimated less precisely when individual outcomes vary substantially relative to the amount of information available. The standard deviations of 0.68 and 0.72 indicate the observed spread within the two groups. The standard error estimates uncertainty in a mean or contrast; the comparison summaries list standard errors of 0.136 and 0.144 for the two group means. These standard errors are not the same thing as the group standard deviations. Their role is to quantify uncertainty in estimated means, not participant-to-participant spread.

The record does not report evidence of over-dispersion or a count-data model. Over-dispersion is a general concern when count outcomes vary more than a model assumes; it is not the central issue for this continuous endpoint summarized by means and standard deviations. More generally, any statistical model depends on assumptions about the data and study design. The registry does not report diagnostic checks or a detailed model specification for this result.

How should the standardized estimate be read?

The standardized effect estimate is -0.257. In general, standardization expresses a contrast relative to a measure of variability, so it can describe magnitude without using the original outcome scale. It is not a separate clinical endpoint and does not remove uncertainty. Its interpretation depends on the standardization convention; the registry does not state the convention in a posted analysis description. The mean difference remains the direct result on the original index-change scale.

What can randomization and masking tell us?

The study is listed as randomized and parallel, which indicates that participants were assigned to treatment groups and followed in separate arms according to the registry’s design labels. Randomization is generally intended to support a fair comparison by reducing systematic differences in treatment assignment. The record does not describe the sequence-generation or allocation-concealment methods, so those details cannot be evaluated from the registry record. Masking is listed as none. In general, lack of masking can leave treatment knowledge available to participants or study personnel, but the record does not specify which parties knew assignments or whether outcome assessment was independently masked.

The primary analysis population is described as all completers. The record does not give the number of completers separately from the listed outcome group sizes, explain what counted as completion, or describe missing-data methods. It also does not report a prespecified statistical analysis plan, an interim-analysis plan, or an adjustment strategy for multiple comparisons. These omissions limit evaluation of how closely the posted comparison follows an original plan. They do not change the arithmetic meaning of the reported group means and unadjusted contrast.

10. Confidence Intervals and Precision

The mean difference is -0.18, and the reported confidence interval extends from a negative value to a positive value. It therefore includes zero, the value representing no difference in mean change, as well as values on either side. The interval does not support a sharply precise estimate of the direction or magnitude of the underlying difference. Its exact bounds are not reproduced because they are recorded to more than three decimal places.

General statistical explanation: a confidence interval gives a range of estimates compatible with the data under a specified method and its assumptions. It is not a guarantee that the true value lies in the interval, nor is it a range of individual participants’ outcomes. The interval’s width conveys uncertainty: broader intervals indicate less precision than narrower ones, all else being equal. The registry record does not report the interval method or a detailed statistical analysis plan.

The standard deviations, 0.68 for exenatide and 0.72 for metformin, describe variation among observed changes. They are not bounds on the uncertainty in the difference. The unadjusted p-value of 0.363 is another summary of the comparison, but it does not replace the confidence interval or quantify how large a difference could plausibly be.

The interval and point estimate apply to the primary continuous endpoint and the stated analysis of all completers. They do not quantify uncertainty for outcomes that are not posted, including the acute post-meal question mentioned in the study purpose. The record does not report a clinically important threshold for this index, so statistical precision cannot be translated into a judgment about clinical importance from the registry result alone.

11. 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 own original design assumptions. They give participants per arm, not total events. The assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. Each planning scenario uses the stated fraction of the observed effect as the effect to detect.

For planning to detect 100% of the observed effect, the illustration gives 238 participants per arm for 80% power and 319 participants per arm for 90% power. For planning for 75% of the observed effect, it gives 423 participants per arm for 80% power and 566 participants per arm for 90% power. For planning for 50% of the observed effect, it gives 951 participants per arm for 80% power and 1,273 participants per arm for 90% power.

In general, power is the probability of detecting an effect of a specified size under the planning assumptions if that effect is present. The 80% power column describes the participant count for an 80% chance of detection under those assumptions; the 90% column gives the count for a 90% chance. The illustrations show that smaller effects need more participants to detect than larger effects, when other assumptions are held fixed. These figures are not observed power and do not describe the probability that the trial’s result is correct. The original trial sample-size justification and its design assumptions are not reported in the registry record.

12. 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 = -0.180; pooled standard deviation σ = 0.700.
2. n = 2 × (1.96 + 0.842)² × 0.700² / -0.180² = 237.6, rounded up to 238 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 <- -0.18                      # observed difference in means
sigma <- 0.700285656000464                     # 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(546728)
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 = 238 (80% power), 319 (90% power)
 75% of observed effect: n per arm = 423 (80% power), 566 (90% power)
 50% of observed effect: n per arm = 951 (80% power), 1,273 (90% power)
power.t.test: 238.6 
Simulated power at n = 238 per arm: 0.795 (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.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists one outcome measure and no posted analyses. It does not list a posted secondary outcome result. The study purpose refers to both chronic vascular effects and effects during a two-hour period following a meal, while the posted primary endpoint is the three-month change in reactive hyperemic index. The record does not report a multiplicity strategy explaining how the primary outcome relates to the broader study aims or how multiple measurements, time points, or comparisons were handled.

General statistical explanation: when multiple outcomes or comparisons are examined, the chance of obtaining at least one apparently unusual result can increase. Prespecified outcome hierarchies or statistical adjustment methods can address that issue, but no such plan is reported in the registry record. There is no basis in this record to infer whether additional analyses were planned, conducted, or not posted.

The primary outcome population is described as all completers. That description does not provide the number who completed treatment, the number with an outcome measurement, or the reasons for any missing observations. The missing-data method, if any, is not reported in the registry record. The record also does not describe an intention-to-treat analysis, a per-protocol analysis, imputation, sensitivity analyses, or how participants who stopped treatment were handled.

These gaps matter because the interpretation of a treatment comparison may depend on which participants contribute data and whether missing outcomes differ systematically between groups. They also prevent an assessment of whether the posted primary analysis followed a prespecified plan. The reported group summaries can still be described as posted, but the available information does not establish how robust they are to alternative analysis populations or missing-data assumptions.

14. 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.

The registry lists 0 serious adverse events affecting participants at risk in the exenatide group and 0 in the metformin group. It also lists 0 other adverse events affecting participants at risk in each group. The at-risk counts are 25 for each group for both categories. These are the adverse-event figures recorded for the study; they do not describe the severity, timing, or specific nature of events beyond the registry’s serious-versus-other categories.

A zero recorded count means that no events were entered in that category in the registry result. It is not proof that no event occurred, and it does not quantify the risk of uncommon events with high precision. The death fields are not reported in the registry record, so no conclusion about deaths can be drawn from those fields. The record also does not provide detailed adverse-event definitions, follow-up duration for safety, or a narrative safety analysis. These summaries should be read as registry reporting, not as a complete account of safety experience.

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

The posted result describes the average change in reactive hyperemic index over three months among participants included in the completer analysis. The mean change was 0.01 for exenatide and -0.17 for metformin, with a mean difference of -0.18 for metformin relative to exenatide. The unadjusted p-value is 0.363, and the confidence interval includes zero. Taken together, the point estimate indicates a lower average change in the metformin group, but the estimate is uncertain and compatible with differences in either direction.

The result does not establish that the treatments have the same effect. A comparison that is not statistically significant is not equivalent to a demonstration of equivalence or non-inferiority; the registry does not report such a design or analysis. Nor does the result establish a clinically important difference. The record does not give a threshold for meaningful change in reactive hyperemic index.

The outcome is a physiological measure, not a clinical event outcome. The record does not post results for cardiovascular events, symptoms, or other health outcomes. It also does not post an acute post-meal result, despite that question appearing in the study purpose. The finding is limited to the registered endpoint, the three-month time frame, the described patient population, and the analysis of completers.

General interpretation requires considering the design and the quality of reporting alongside the numerical contrast. The randomized design is relevant to the comparison, while the absence of masking and unreported details about randomization, missing data, and analysis methods constrain how fully the methods can be evaluated. The registry summaries alone do not support medical advice or a conclusion about treatment choice.

16. Why This Trial Matters Statistically

This trial illustrates why a group mean difference should be read together with uncertainty and study design. The point estimate is -0.18, but the confidence interval extends across zero and the unadjusted p-value is 0.363. Focusing only on the sign of the estimate would overstate what the data establish; focusing only on the p-value would also leave out the range of differences compatible with the result.

The outcome is continuous and reported with group means and standard deviations, making it possible to distinguish average change from participant-level variability. The standard deviations of 0.68 and 0.72 are substantial in relation to the mean contrast, and the interval indicates limited precision. General statistical reasoning: variability and the amount of information both influence how precisely a difference can be estimated.

The record also provides a useful reminder that a study’s stated purpose can be broader than its posted primary result. The purpose mentions chronic and post-meal vascular effects, while the posted primary outcome concerns the three-month change. Without an acute result or a documented multiplicity plan, readers should keep those questions separate rather than treating one endpoint as answering both.

Finally, a completed randomized study may still have important reporting gaps. The record identifies the design, groups, endpoint, and group summaries, but does not report the original analysis plan, missing-data procedures, or detailed randomization methods. Recognizing both what the numerical result says and what the registry leaves unreported is central to a careful statistical reading.

17. 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 describes the primary analysis as including all completers but does not report how many participants completed treatment or explain the treatment of missing outcome measurements. This makes it difficult to assess whether the reported group summaries include everyone assigned to treatment or only those with available endpoint data. The record lists 25 participants in each outcome group, but does not explain the relationship between those counts and completion status.

Masking is listed as none, and the registry does not specify who knew treatment assignments or whether outcome assessment was masked. It also does not describe the randomization sequence or allocation concealment. The posted result is an unadjusted comparison, while the original statistical analysis plan and any adjusted analyses are not reported in the registry record.

The registry posts the three-month change in reactive hyperemic index but no acute post-meal result, even though the study purpose includes a two-hour period following a meal. It lists no secondary outcome result and does not report a multiplicity strategy. The clinical meaning threshold for the index is also not reported, so the practical importance of the observed mean difference cannot be determined from this result alone.

The adverse-event summary reports zero events in the listed categories, but does not include detailed event descriptions or a death count. Finally, the results pertain to the listed population and three-month endpoint; the registry record does not establish how well they generalize to other populations, treatment durations, or outcomes.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: The Vascular Effects of Exenatide Versus Metformin in Patients With Pre-Diabetes — NCT00546728.
• Registry record last updated 2013-11-08; status last verified 2013-09.
• Results first posted 2013-11-08.
• 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.

21. Record Summary

The registry record describes a completed Phase 4 randomized, parallel-group study of exenatide and metformin in patients with impaired glucose tolerance. The arm descriptions list exenatide 10 mcg twice daily and metformin 500 mg twice daily; masking is listed as none. Enrollment is 50, with 25 participants listed in each primary outcome group.

The primary outcome is change in reactive hyperemic index from baseline to three months. Mean change was 0.01 with standard deviation 0.68 for exenatide and -0.17 with standard deviation 0.72 for metformin. The reported mean difference is -0.18, with an unadjusted p-value of 0.363 and a confidence interval that includes zero. All completers are described as included in the analysis.

The registry lists zero serious and zero other adverse events affecting participants at risk in each group; deaths are not reported. No secondary outcome result or analysis is posted. Details of missing-data handling, the original statistical plan, and the acute post-meal outcome are not reported in the registry record.