Gastrointestinal Sensorimotor Dysfunctions in Diabetes Mellitus: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Gastrointestinal Sensorimotor Dysfunctions in Diabetes Mellitus (NCT02170870), a randomized phase 2 study. The primary outcome is Mean Intestinal Chemosensitivity to Lipids Perfusion.

2. Clinical Question

3. Trial Design

The registry describes a randomized, triple-masked, parallel phase 2 study sponsored by Mayo Clinic. Enrollment was 104. The study began in 2014-06, reached primary completion on 2017-11-17, and is listed as completed. The record does not describe the specific people or roles included in the masking designation, nor does it give further details about the randomization procedure or allocation concealment.

The study included healthy controls, participants with diabetes mellitus, and participants with functional dyspepsia. Within each population, the listed arms received exendin 9-39 or placebo. Exendin 9-39 was given intravenously as a 1,200 pmol/kg bolus followed by an infusion at 300 pmol/kg/min. Placebo was normal saline prepared to match the appearance of exendin 9-39. The arm descriptions also specify lipid infusion of 66.7 mL Microlipid, at 0.5 gm/mL diluted in water to 222 ml.

The primary outcome was assessed during the duodenal lipid infusion procedure on Day 1, approximately 2 hours after infusion. For the functional dyspepsia population, enteral lipid infusion was not performed in 2 placebo subjects and 3 exendin 9-39 subjects because a nasoduodenal tube could not be placed. The primary endpoint description identifies these participants as excluded from that endpoint’s assessment. The registry record does not report a detailed analysis population definition or further handling of missing endpoint data.

A secondary outcome is listed for gastric emptying rate in participants with diabetes mellitus or non-ulcer dyspepsia compared with placebo. The posted results include two analyses, both described as ANOVA. The registry record does not report the original statistical analysis plan, the trial’s original sample-size assumptions, or the precise roles of all study populations in the randomization.

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: Functional Dyspepsia arm - enteral lipid infusion was not performed because a nasoduodenal tube could not be placed in 2 functional dyspepsia placebo subjects and 3 functional dyspepsia exendin 9-39 subjects.

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.

Only the first 6 of 7 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Mean Intestinal Chemosensitivity to Lipids Perfusion, measured over Day 1, approximately 2 hours after infusion.

The primary endpoint, mean intestinal chemosensitivity to lipid perfusion, is a continuous symptom score. During duodenal lipid infusion, participants rated nausea, fullness, bloating, abdominal pain, belching, and burning at 15 minute intervals. Each symptom used a Visual Analogue Scale marked from 0 (minimum value) to 4 (maximum value), with categories absent (0), light (1), moderate (2), severe (3), and intolerable (4). The endpoint score averaged ratings for nausea, fullness, bloating, and abdominal pain over the 2 hour infusion. It did not average all six symptoms listed in the description.

The endpoint is intended to quantify symptoms during nutrient exposure as an indicator of intestinal chemosensitivity. A mean makes repeated symptom ratings easier to summarize and compare across groups, while retaining the direction of the scale: higher scores mean a worse outcome. The reported arm values are means, with standard deviations as the dispersion measure. Because the endpoint depends on completing the enteral infusion procedure, participants whose nasoduodenal tube could not be placed did not contribute to the functional dyspepsia primary-endpoint summaries.

The registry record does not report a separate endpoint result for each symptom, nor does it report the distribution of individual participants’ scores. Thus, the mean describes the group’s average symptom score, while the standard deviation describes variation among participants; neither alone shows how many individuals experienced a particular symptom severity.

7. Results Posted in the Registry

• Healthy Controls Placebo: mean 0.36 versus 0.22 on control; difference 0.14 (95% CI 0.06 to 0.22), standardized effect d = 1.47; z-test p = <0.001.
• Diabetics Exendin 9-39: mean 0.55 versus 0.22 on control; difference 0.33 (95% CI 0.27 to 0.39), standardized effect d = 3.60; z-test p = <0.001.
• Diabetics Placebo: mean 0.65 versus 0.22 on control; difference 0.43 (95% CI 0.35 to 0.51), standardized effect d = 3.15; z-test p = <0.001.
• Functional Dyspepsia Exendin 9-39: mean 1.40 versus 0.22 on control; difference 1.18 (95% CI 0.75 to 1.61), standardized effect d = 1.73; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean and standard deviation describe?

General statistical explanation: A mean is an average of the measured scores in a group. A standard deviation describes how dispersed individual observations are around that average. The registry reports the endpoint as a mean with standard deviation. For example, the healthy-control exendin 9-39 arm had a mean of 0.22 and a standard deviation of 0.05; the functional dyspepsia exendin 9-39 arm had a mean of 1.4 and a standard deviation of 0.9. These summaries describe both average symptom level and variability, but they do not reveal the full score distribution or the number of participants at each score.

What question does ANOVA address?

General statistical explanation: Analysis of variance, or ANOVA, evaluates whether the means across a set of groups are compatible with a common mean under the model assumptions. An ANOVA p-value is not, on its own, a list of which pairs differ, the size of any difference, or evidence that a particular treatment caused a difference. The registry reports ANOVA for the four healthy-control and diabetes arms, with p = 0.06, and for the four healthy-control and functional dyspepsia arms, with p < 0.001. The record does not state the exact ANOVA model, its covariates, or any follow-up comparison procedure.

What do the recalculated contrasts represent?

The recalculated results are simple unadjusted comparisons of each posted arm mean with the healthy-control exendin 9-39 mean. They report mean differences, intervals, and unadjusted p-values based on the posted summary statistics. The comparisons are not adjusted for other factors, and they are not substitutes for the registry’s ANOVA. The reference arm is a healthy-control group, so comparisons with diabetes or functional dyspepsia arms mix differences in clinical population with differences in assigned infusion. A contrast between healthy-control placebo and healthy-control exendin arms is within the same named population, but the registry record does not provide a prespecified pairwise analysis for that contrast.

The recalculated p-values are all reported as <0.001. A p-value describes how incompatible data are with a specified null model, under the assumptions of the test; it is not the probability that the null hypothesis is true. Unadjusted p-values do not account for conducting multiple comparisons. Because the registry analysis and the recalculated contrasts differ, the sponsor’s own reported analysis takes precedence when describing its conclusions.

Why does over-dispersion matter?

General statistical explanation: When observations vary substantially within a group, estimates of group means are less precise than they would be with less variation, other things being equal. This is one reason to read a mean together with its standard deviation and interval, rather than relying on a p-value alone. Variation is especially apparent in the functional dyspepsia arm summaries, where each standard deviation is 0.9, alongside means of 1.4 and 1.2. That spread means the mean does not represent every participant’s symptom score. It also underscores the value of examining the underlying distribution, which is not reported in the registry record.

Statistical models for continuous outcomes often rely on assumptions about the observations and their residual variation. The registry record does not report diagnostic checks, transformations, robust methods, or any approach to departures from model assumptions. Therefore, it is not possible to assess from the record whether the ANOVA assumptions were evaluated or whether an alternative analysis was considered.

What does a confidence interval add?

General statistical explanation: A confidence interval expresses uncertainty around an estimated effect under a statistical procedure. Its endpoints indicate a range of values compatible with the data and procedure; it does not mean that every value inside the range is equally likely, and it is not a measure of clinical importance. Intervals for the recalculated contrasts are not the same as intervals from the registry’s ANOVA, for which no interval is reported. The registry record does not state the confidence level for the recalculated intervals.

The estimated mean differences, their intervals, and the p-values are related ways to describe the recalculated comparisons, but they convey different information. A small p-value does not guarantee a precise or important effect. The intervals give a sense of statistical uncertainty in those contrasts, subject to the unadjusted calculation and its assumptions. The registry’s own ANOVA p-values should not be treated as pairwise effect estimates, because the record reports no corresponding estimated difference or interval.

The statistical plan, any interim analyses, and the original sample-size justification are not reported in the registry record. Consequently, the available results do not establish whether the posted comparisons were prespecified, whether analysis choices were changed, or how multiplicity was intended to be handled. Such missing information limits assessment of the confirmatory status of the findings.

10. Confidence Intervals and Precision

The interval estimates attached to the recalculated contrasts describe uncertainty around unadjusted mean differences relative to healthy controls receiving exendin 9-39. The registry record does not report the confidence level for these intervals, so it should not be inferred from the posted endpoints. These are recalculated intervals, not intervals reported alongside the sponsor’s ANOVA results.

The healthy-control placebo contrast is 0.14, with an interval from 0.059 to 0.222. The diabetes exendin 9-39 contrast is 0.33, with an interval from 0.275 to 0.385, and the diabetes placebo contrast is 0.43, with an interval from 0.351 to 0.509. The functional dyspepsia contrasts are 1.18, with an interval from 0.751 to 1.609, for exendin 9-39 and 0.98, with an interval from 0.563 to 1.397, for placebo. The functional dyspepsia intervals span a broader range of values than the diabetes intervals, consistent with the larger reported standard deviations in those arms. This comparison concerns the posted summaries and does not establish why variability differs.

Intervals are useful because they convey more than whether a p-value crosses a conventional threshold: they show the range of estimates compatible with the calculation. However, they do not correct for the fact that multiple comparisons were made, and the recalculated contrasts do not adjust for differences between clinical populations. The sponsor’s ANOVA results have p-values but no reported interval or effect estimate in the registry record. Precision for those analyses therefore cannot be assessed from a reported interval. The intervals should not be treated as proof of clinical importance or as a substitute for a clearly specified analysis plan.

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 original design justification. The registry record does not report the trial’s original sample-size assumptions. For these illustrations, the stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is the stated fraction of the observed effect. The figures are participants per arm, not total participants, and they are not total event counts because this is not a time-to-event calculation.

The 80% power column gives the number of participants per arm estimated to be needed for an 80% chance of detecting the specified effect under those assumptions, if that effect is truly present. The 90% power column gives the corresponding planning figure for a 90% chance. These are design probabilities under a statistical model, not probabilities that a result is correct or that a treatment will have a particular effect. The sample-size figures do not describe the power actually achieved by the completed trial: this is not observed power.

The planning rows use the full observed effect and smaller fractions of that effect. As the assumed effect gets smaller, more participants are needed for the same target power. The observed group differences may be unstable estimates of effects that would occur in a new study, particularly when variability is substantial or when comparisons span different clinical populations. A planning exercise based on an observed effect should therefore be read as an illustration, not as a recommended enrollment target. It also does not account for details that are not reported in the registry record, such as any intended adjustment for multiple comparisons or the trial’s original allocation and analysis assumptions.

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 = 0.140; pooled standard deviation σ = 0.095.
2. n = 2 × (1.96 + 0.842)² × 0.095² / 0.140² = 7.2, rounded up to 8 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 <- 0.14                      # observed difference in means
sigma <- 0.0951075945729601                     # 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(2170870)
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 = 8 (80% power), 10 (90% power)
 75% of observed effect: n per arm = 13 (80% power), 18 (90% power)
 50% of observed effect: n per arm = 29 (80% power), 39 (90% power)
power.t.test: 8.3 
Simulated power at n = 8 per arm: 0.780 (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. Analyses Reported by the Sponsor

The registry reports two sponsor analyses, each described as ANOVA. The analysis covering the healthy-control and diabetes arms has p = 0.06. The analysis covering the healthy-control and functional dyspepsia arms has p < 0.001. The record does not report an effect estimate, interval, or more detailed model description for either analysis, so the size and precision of the ANOVA results cannot be read directly from those p-values.

The recalculated results are unadjusted mean contrasts based on posted arm summaries, with healthy controls receiving exendin 9-39 as the reference. They are not a reproduction of the reported ANOVA. In particular, contrasts involving diabetes or functional dyspepsia compare different clinical populations and may also involve different assigned infusions. They do not necessarily test the same question as an ANOVA across four groups. Their unadjusted p-values should not be substituted for the reported ANOVA p-values.

Because the recalculation differs from the registry’s own analysis, the sponsor’s reported analysis takes precedence when describing the sponsor’s statistical result. The recalculated contrasts can still help illustrate the posted arm means and how they compare under a simple calculation, but they should be treated as supplementary, not as a correction to the sponsor’s analysis. The registry record does not explain the discrepancy or report enough statistical detail to resolve it.

14. Multiplicity, Missing Data and Analysis Populations

The primary outcome is a mean symptom score, and the registry lists one secondary outcome concerning gastric emptying rate. Two sponsor analyses are posted for the primary result: one covering healthy controls and participants with diabetes, and one covering healthy controls and participants with functional dyspepsia. The record does not state whether these analyses were prespecified as separate primary comparisons, whether one was considered exploratory, or how the set of tests was defined.

Multiplicity arises when several statistical tests are conducted. Repeated testing can increase the chance of obtaining at least one small p-value even when the tested hypotheses are all false. General statistical explanation: Methods to address multiplicity can adjust the testing threshold or otherwise account for the family of analyses. The registry record does not report a multiplicity adjustment, a testing hierarchy, or any planned approach. The recalculated pairwise p-values are explicitly unadjusted and should not be interpreted as multiplicity-controlled evidence.

The primary endpoint description states that enteral lipid infusion was not performed in 2 functional dyspepsia placebo subjects and 3 functional dyspepsia exendin 9-39 subjects because a nasoduodenal tube could not be placed. It is not reported in the registry record whether any other endpoint data were missing, how the excluded participants were handled in the analysis beyond their absence from the endpoint summaries, or whether missing-data sensitivity analyses were conducted. The record also does not state whether the analyses followed an intention-to-treat, per-protocol, or other analysis population definition.

These omissions matter because results can depend on which participants are included and how missing measurements are treated. The specific tube-placement issue is documented for the functional dyspepsia endpoint; its effect on estimates cannot be determined from the posted data. Demographic or clinical characteristics for included and excluded participants are not reported in the registry record, so comparability between them cannot be assessed. Nor does the record report interim analyses or details of the statistical analysis plan.

15. 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 0% reporting threshold.

Most frequent non-serious adverse event terms

The serious adverse event summaries report 0 affected participants in each of the six arms, and 0 deaths in each arm. The reported numbers at risk for serious adverse events are 13 for healthy controls receiving exendin 9-39, 11 for healthy controls receiving placebo, 20 for each diabetes arm, and 20 for each functional dyspepsia arm. These are the serious-event counts as recorded; they are not proof that serious events could not occur.

The adverse-event record also lists other adverse events separately from serious adverse events: 0 in both healthy-control arms, 2 in the diabetes exendin 9-39 arm, 1 in the diabetes placebo arm, 1 in the functional dyspepsia exendin 9-39 arm, and 0 in the functional dyspepsia placebo arm. These figures should not be confused with the serious-event counts. The record does not describe the nature, timing, severity, or attribution of those other events.

Zero reported serious events in these groups is a description of the registry record, not evidence of no risk. The registry does not provide enough detail to assess how long participants were observed for safety, how events were collected, or whether event definitions were applied consistently. The safety arm counts should also not be assumed to match the endpoint analysis counts: the endpoint description specifically notes failed tube placement for some functional dyspepsia participants.

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

The posted endpoint measures symptom ratings during intestinal lipid exposure, with higher average scores representing worse symptoms. The group means are lowest in the healthy-control arms, higher in the diabetes arms, and highest in the functional dyspepsia arms. Within the diabetes population, the mean is 0.55 with exendin 9-39 and 0.65 with placebo; within functional dyspepsia, it is 1.4 with exendin 9-39 and 1.2 with placebo. These are descriptive summaries, not sufficient evidence that exendin 9-39 caused a difference.

The sponsor’s reported analyses address groups of arms through ANOVA and give p = 0.06 for the set including diabetes and p < 0.001 for the set including functional dyspepsia. They do not provide effect estimates or confidence intervals in the registry record. The recalculated contrasts are different, unadjusted comparisons against healthy controls receiving exendin 9-39. Several of those contrasts combine differences in clinical population with differences in infusion assignment. Their results should not be interpreted as the isolated effect of exendin 9-39, and the sponsor’s analysis takes precedence when describing the sponsor’s result.

The study’s rationale concerns whether symptoms relate to sensitivity to small-intestinal nutrients and to GLP-1 signaling. The posted endpoint is relevant to that question because it records symptom scores during lipid infusion. However, the record does not establish a causal mechanism, and a symptom-score difference does not by itself demonstrate why symptoms occurred. The endpoint also averages four symptoms, so it cannot show which symptom drove a group mean or how responses varied across individuals.

The registry record does not report the original statistical plan, multiplicity strategy, missing-data methods, or confidence level for the recalculated intervals. Some functional dyspepsia participants did not complete the infusion procedure because a tube could not be placed. Those limitations constrain the strength and scope of interpretation. The results describe the posted measurements and analyses; they do not establish clinical benefit, harm, or a treatment recommendation.

17. Why This Trial Matters Statistically

This study illustrates how a clinically motivated question can be translated into a continuous symptom endpoint. Instead of relying on a single symptom or a simple yes-or-no outcome, the endpoint combines repeated ratings of nausea, fullness, bloating, and abdominal pain during intestinal lipid exposure. The definition also makes clear that the mean covers a selected set of symptoms, even though additional symptoms were recorded. That distinction matters when interpreting what a composite score represents.

The record also shows why group summaries and statistical tests should be read together. Means and standard deviations describe average symptom levels and participant variability, whereas ANOVA p-values address a group-level statistical question without giving the size of a specific contrast. The recalculated unadjusted contrasts add estimates and intervals, but they do not reproduce the reported ANOVA and do not isolate the effect of the infusion when the comparison crosses clinical populations. Statistical transparency requires keeping those analyses distinct.

Finally, the functional dyspepsia tube-placement issue demonstrates a practical connection between study procedures and the population represented in an endpoint result. A participant can be enrolled in a trial yet not contribute to a particular measurement if the procedure cannot be completed. Clear reporting of analysis populations, missing-data handling, and prespecified comparisons would help readers assess how such issues influence the interpretation. The registry record offers useful outcome and safety summaries, while leaving those methods details unreported.

18. 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 primary endpoint depends on successful placement of a nasoduodenal tube and completion of enteral lipid infusion. The registry specifically reports that this procedure was not performed in 2 functional dyspepsia placebo subjects and 3 functional dyspepsia exendin 9-39 subjects because a tube could not be placed. The record does not describe the characteristics of these participants or assess how their exclusion might affect the functional dyspepsia summaries.

The functional dyspepsia arm standard deviations are 0.9, compared with 0.05 and 0.13 in the healthy-control arms and 0.11 and 0.17 in the diabetes arms. This indicates greater observed spread in the functional dyspepsia scores, but the registry record does not provide distributions or analyses that explain that variation. Means alone may conceal differences in individual response patterns.

The contrasts recalculated from the posted summaries use healthy controls receiving exendin 9-39 as a reference. Comparisons with diabetes or functional dyspepsia arms cross clinical populations, so they cannot be interpreted as pure infusion effects. The sponsor’s ANOVA results are not accompanied by effect estimates, intervals, model details, or a stated multiplicity approach. The discrepancy between those reported analyses and the simple recalculations is not explained in the registry record.

The primary endpoint is measured during one assessment period on Day 1, approximately 2 hours after infusion. The record does not report longer-term symptom outcomes, individual symptom results, or the full distribution of scores. It also does not report the original statistical analysis plan, interim analyses, trial-specific sample-size rationale, or detailed missing-data methods. These omissions limit assessment of precision, robustness, and generalizability beyond the participants and procedure represented in the posted results.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Gastrointestinal Sensorimotor Dysfunctions in Diabetes Mellitus — NCT02170870.
• Registry record last updated 2019-04-23; status last verified 2019-03.
• Results first posted 2019-04-23.
• 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.

22. Record Summary

NCT02170870 is a completed randomized, triple-masked, parallel phase 2 study sponsored by Mayo Clinic, with enrollment of 104. It examined intestinal symptom scores during duodenal lipid infusion in healthy controls, participants with diabetes, and participants with functional dyspepsia, under exendin 9-39 or placebo assignments. The primary outcome was a mean score from four gastrointestinal symptoms, with higher scores indicating worse symptoms.

The registry reports arm means and standard deviations, plus two ANOVA p-values: 0.06 for the analysis including diabetes and < 0.001 for the analysis including functional dyspepsia. Simple recalculated contrasts are unadjusted and differ from those sponsor analyses; the sponsor’s own analysis takes precedence. The record reports 0 serious adverse events and 0 deaths across arms. It also documents failed tube placement for some functional dyspepsia participants and does not report important analysis-plan and missing-data details.