Return of First-phase Insulin Secretion in Type 2 Diabetes is Associated With Depletion of Pancreas Lipid: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Return of First-phase Insulin Secretion in Type 2 Diabetes is Associated With Depletion of Pancreas Lipid (NCT03430310), a randomized na study. The primary outcome is Changes in Liver Fat.

2. Clinical Question

The registry describes a randomized, parallel-group comparison without masking. Its brief summary frames the dietary question as whether a low-glycemic diet is associated with changes in ectopic lipid and beta-cell function, with particular interest in pancreas lipid and first-phase insulin secretion. It also states an interest in whether race influences the association between pancreas lipid and beta-cell function. These statements describe the investigators’ hypotheses; they are not themselves results.

The outcome with posted numerical results is liver fat, not pancreas lipid or first-phase insulin secretion. The registry lists the outcome timeframe as baseline and week 12 and describes MRI assessment of overall liver fat. The numerical comparison therefore answers a narrower question than the study’s title and broad hypothesis may suggest: how the posted liver-fat measure differed between the diet groups. Results for the stated pancreas-lipid and beta-cell questions are not reported in the registry record.

3. Trial Design

The registry identifies this as a randomized, parallel-group study with no masking. Participants were assigned to a low-glycemic diet or a control diet. The study is recorded as completed, with enrollment listed as 65. The outcome-level results list 27 participants for the low-glycemic diet group and 30 for the control group. The registry record does not explain the difference between the enrollment figure and the outcome-specific group counts.

The low-glycemic diet emphasized low-glycemic carbohydrate sources and mainly whole foods, with minimal highly processed grain products and added sugar. The control diet was described as compatible with American Diabetes Association and United States Department of Agriculture guidelines and included low-fat foods, whole grains, fruits, and vegetables. The descriptions are incomplete, so the precise dietary prescriptions, adherence measures, and degree of separation between diets are not reported in the registry record.

The recorded study start date is 2018-10-09 and the primary completion date is 2023-07-31. The registry lists the study phase as not applicable. No masking means the registry does not describe blinding of participants or other parties; details about outcome assessment masking are not reported in the registry record. Random allocation can help balance prognostic factors across groups on average, but the record does not provide the allocation procedure or concealment details.

The scientific summary emphasizes pancreas lipid, first-phase beta-cell response, and possible differences by race. The primary outcome with numerical results, however, is changes in liver fat measured by MRI at baseline and week 12. The record does not report why liver fat was selected as the posted primary outcome or how it relates to the scientific aims described in the summary. It also does not report the original sample-size justification, planned analysis population, or protocol details for handling departures from assigned diets.

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: No washout or run-in was used for this study.

Overall Study

Reasons for not completing this period:

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 Changes in Liver Fat, measured over Baseline, week 12.

The posted primary outcome is titled “Changes in Liver Fat.” It is defined in percent liver fat and assessed with magnetic resonance imaging. The timeframe is baseline and week 12. The registry’s description says the assessment concerns the overall percentage of liver fat, but it does not give further measurement or calculation details. It also does not specify whether the arm-level values displayed represent change scores, week 12 values, or another summary of the two timepoints.

Liver fat is a continuous measure: it can take a range of numerical values rather than representing a simple event count. Reporting a mean gives a summary of the group’s measured values, while a standard error describes uncertainty around that estimated mean. A comparison between group means can describe the direction and magnitude of the observed difference in the measured outcome. The registry separately reports an ANCOVA superiority analysis, but its adjusted estimate and confidence interval are not included.

MRI is the measurement method named in the registry. In general, an imaging-based measure can quantify an organ-level feature without relying only on a clinical event count. That general rationale does not establish how liver fat relates to the study’s other stated interests. The brief summary centers on pancreas lipid and first-phase insulin secretion, while the posted primary outcome concerns liver fat. Whether liver fat was intended as a proxy, a separate outcome, or part of a broader analysis is not reported in the registry record.

The distinction matters when interpreting the result: the posted numerical findings concern liver fat as recorded, and do not establish findings about pancreas lipid, beta-cell function, glucose tolerance, or race-specific associations.

7. Results Posted in the Registry

• Low Glycemic Diet: mean 5.40 versus 7.30 on control; difference -1.90 (95% CI -3.71 to -0.09), standardized effect d = -0.55; z-test p = 0.039.

8. Statistical Interpretation

9. Statistical Methods Explained

These methods describe how continuous outcomes and group comparisons are commonly interpreted. The registry identifies the sponsor’s analysis as ANCOVA and the posted outcome as a continuous measure summarized by means and standard errors. It does not report enough detail to reconstruct the sponsor’s model.

What does a mean describe?

A mean is the arithmetic average of measured values in a group. It summarizes the center of the observed distribution, but it does not show the full spread of individual measurements or whether the distribution is symmetric. Two groups can have different means even when their individual values overlap considerably. A mean difference expresses the contrast between group averages in the outcome’s measurement units. For this outcome, the unit is percent liver fat.

The posted group means are 5.4 for the low-glycemic diet group and 7.3 for the control diet group. The registry also gives standard errors of 0.7 and 0.6. A standard error reflects the estimated variability of a group mean across repeated samples under a statistical model; it is not the standard deviation of individual participant values. The record does not post group standard deviations for the outcome.

What is an unadjusted comparison?

An unadjusted comparison contrasts the groups using the posted summaries without accounting for other variables in a model. The recalculated mean difference is -1.9, and the recalculated 95% confidence interval is -3.707 to -0.093. The recalculated p-value is 0.039. These calculations use the group summaries and are not equivalent to the sponsor’s ANCOVA result.

Unadjusted analyses can be useful for describing the crude contrast between groups. They do not necessarily match a planned analysis that adjusts for baseline values or other covariates. Adjustment can change the estimated contrast and its uncertainty. The registry names ANCOVA for the sponsor’s analysis but does not report which covariates were included, how the model was specified, or whether the outcome values were baseline-adjusted changes or follow-up measurements.

What does ANCOVA generally do?

In general, analysis of covariance compares group outcomes while accounting for one or more additional variables in a regression model. In a randomized trial, a common use is to account for a baseline measure of the outcome, which can improve precision when baseline values are associated with follow-up values. ANCOVA estimates are conditional on the model and variables included. The registry reports an ANCOVA superiority test with p = 0.05 but does not provide the adjusted effect, confidence interval, covariates, model diagnostics, or missing-data approach. Those details cannot be inferred from the method name alone.

How should a p-value be read?

A p-value is a measure of how unusual data at least as incompatible with a specified null hypothesis would be under the assumptions of a particular statistical model. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The sponsor’s reported p-value is 0.05. The recalculated unadjusted p-value is 0.039. They arise from different analyses, so they should not be treated as interchangeable estimates of the same procedure.

The registry labels the sponsor’s analysis as a superiority analysis. In general, a superiority test evaluates whether the data support a difference in a specified direction or whether the groups differ, depending on the test specification. The registry does not report the exact null and alternative hypotheses, sidedness, or decision threshold used in the sponsor’s analysis. The planning illustration separately specifies a two-sided significance level of 0.05; this planning assumption does not establish the original trial analysis plan.

Why does over-dispersion matter for count outcomes?

Over-dispersion arises when count data vary more than a model assumes. It can make standard errors too small if a simple count model understates variability. It is not a central issue for the posted continuous liver-fat outcome, and the registry record does not describe a count-outcome analysis. In general, analysts should match a model to the outcome’s scale and variability rather than assume that one method suits every endpoint.

The record identifies one primary outcome and one reported analysis. It does not report the full statistical analysis plan, outcome hierarchy, covariate-selection strategy, or sensitivity analyses. The interpretation must therefore distinguish the known method label and reported results from methodological details that remain unavailable.

10. Confidence Intervals and Precision

The recalculated unadjusted mean difference is -1.9, with a 95% confidence interval from -3.707 to -0.093. The interval expresses uncertainty about the mean difference under the recalculation’s assumptions. Its endpoints are both negative, and the upper endpoint is close to zero. This conveys that the estimate is not highly precise about the size of the difference, even though the interval does not include zero in this unadjusted calculation.

A confidence interval is not a range containing a fixed proportion of individual participants’ values, nor does it mean there is a 95% probability that the true effect lies inside this particular interval. In repeated use of the same method under its assumptions, intervals constructed in this way would cover the fixed parameter at the stated rate. The interval does not account for unreported sources of uncertainty such as missing-data handling, model choice, or multiplicity.

The arm-level standard errors are 0.7 for the low-glycemic group and 0.6 for the control group. They describe uncertainty in the estimated means, not variation among participants. The recalculation’s interval is based on the posted summaries; the sponsor’s ANCOVA confidence interval is not reported in the registry record. It is therefore not possible to compare the precision of the sponsor’s adjusted estimate with that of the unadjusted calculation.

Precision and importance are separate questions. The registry does not report a threshold for a clinically important change in liver fat, and the meaning of the posted arm means in relation to baseline and week 12 is not fully clarified. The interval should be read as a statistical description of the unadjusted contrast, not as evidence about clinical relevance or the unreported pancreas and beta-cell outcomes.

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 planning illustrations show participants per arm, not total events. They use the observed effect as a reference and consider planning for 100%, 75%, and 50% of that effect. For 80% power, the listed figures are 52 participants per arm for planning at 100% of the observed effect, 93 per arm at 75%, and 208 per arm at 50%. For 90% power, the corresponding figures are 70, 124, and 279 participants per arm.

In plain terms, the 80% power column gives the estimated number of participants per arm that the illustration would require to have an 80% chance of detecting the specified effect under its assumptions. The 90% power column gives the corresponding number per arm for a higher target chance of detection. These are planning illustrations based on an observed effect, not the original trial’s documented design calculations. The trial’s own sample-size justification and original design assumptions are not reported in the registry record.

The listed assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect equal to the stated fraction of the observed effect. The figures are participants per arm because this is not a time-to-event analysis. In general, a smaller effect is harder to distinguish from random variation, so detecting a smaller effect with the same target power requires more participants.

These estimates are sensitive to the effect used as the planning target. An observed effect from a study can be an uncertain basis for planning another study, particularly when the estimate itself is imprecise. The values are not observed power and do not describe the probability that the completed trial’s result is correct or that an effect exists.

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 = -1.900; pooled standard deviation σ = 3.457.
2. n = 2 × (1.96 + 0.842)² × 3.457² / -1.900² = 52.0, rounded up to 52 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -1.9                      # observed difference in means
sigma <- 3.45669311231519                     # 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(3430310)
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 = 52 (80% power), 70 (90% power)
 75% of observed effect: n per arm = 93 (80% power), 124 (90% power)
 50% of observed effect: n per arm = 208 (80% power), 279 (90% power)
power.t.test: 52.9 
Simulated power at n = 52 per arm: 0.796 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Analyses Reported by the Sponsor

The registry reports the sponsor’s own analysis as an ANCOVA superiority test with p = 0.05. It does not provide the adjusted mean difference, confidence interval, or model details. Separately, an unadjusted recalculation from the posted arm summaries gives a mean difference of -1.9, a 95% confidence interval from -3.707 to -0.093, and p = 0.039.

The two results are not identical because the recalculation is an unadjusted contrast based on posted summary statistics, whereas the sponsor’s analysis is identified as ANCOVA. Adjustment for covariates can affect an estimate and its standard error. The record does not specify which covariates were used or provide enough information to reproduce the sponsor’s result. The registry also flags a discrepancy, so the sponsor’s own analysis takes precedence over the unadjusted recalculation.

The recalculation can help readers understand the raw contrast between the posted group means, but it should not be substituted for the sponsor’s reported analysis. The absence of an adjusted estimate and confidence interval limits assessment of the sponsor result’s magnitude and precision. The reported p-value alone does not fill that gap.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure and one reported analysis. It does not report a complete statistical analysis plan or state whether additional outcomes or comparisons were prespecified. The record lists no secondary outcomes with posted results. The brief summary discusses pancreas lipid, first-phase beta-cell response, glucose tolerance, and possible differences by race, but numerical results for those questions are not reported in the registry record.

Multiplicity refers to the increased chance of a false-positive finding when many hypotheses are tested. In general, statistical plans may address multiplicity by defining a primary outcome, ordering tests, or adjusting inference across comparisons. The available record does not state whether any such approach was planned or used. A single posted analysis cannot establish how unreported analyses were handled.

Missing-data methods are also not reported. The registry gives arm-level sample sizes for the posted outcome, but does not explain why these differ from the overall enrollment figure, whether all randomized participants had outcome measurements, or how incomplete measurements were treated. It does not report imputation, complete-case analysis, sensitivity analysis, or assumptions about missingness. These omissions matter because the analysis population and missing-data handling can affect both estimates and uncertainty.

The record does not specify whether the analysis followed an intention-to-treat approach, a per-protocol approach, or another population definition. It also does not report adherence to the assigned diets, crossovers, withdrawals, or exclusions from analysis. These details should not be inferred from randomization or from the outcome counts alone.

Overall, multiplicity control, missing-data procedures, and the analysis population are not reported in the registry record. Consequently, the reported ANCOVA p-value and the unadjusted recalculation should be interpreted with awareness that the broader analysis framework is not fully documented.

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 1% reporting threshold.

The registry reports zero serious adverse events in the low-glycemic diet group, with 32 participants listed as at risk, and zero in the control diet group, with 33 listed as at risk. It also reports zero other adverse events in each group, using the same at-risk counts. Deaths are listed as zero in both groups.

These figures describe events recorded in the registry during the study. A zero count means no events of that category were reported for the listed group; it does not prove that an event could not occur or that the diets have no risks. The registry does not provide event definitions, the observation period for adverse events, severity details, or information about collection procedures. It lists no serious-event terms.

The at-risk counts in the safety record are not the same as the outcome-level sample sizes, and the registry does not explain the difference. The figures should therefore be read as the safety counts reported, without assuming that they describe every enrolled participant or the same analysis population as the liver-fat outcome. These registry counts alone do not establish comparative safety.

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

The posted results describe a lower mean liver-fat value in the low-glycemic diet group than in the control diet group: 5.4 compared with 7.3. The unadjusted recalculation estimates a mean difference of -1.9, with a 95% confidence interval from -3.707 to -0.093 and p = 0.039. The sponsor’s reported ANCOVA analysis gives p = 0.05, but its adjusted effect estimate and confidence interval are not reported. Because the record flags a discrepancy, the sponsor’s own analysis takes precedence.

These findings do not establish the size of the sponsor’s adjusted effect, because that estimate is absent. The statistical significance labels also do not tell whether a difference is clinically meaningful. The registry does not provide a threshold for a meaningful change in liver fat. In addition, it does not clearly specify whether the arm-level values are changes from baseline or another summary of measurements taken at baseline and week 12.

The result should not be interpreted as a finding about pancreas lipid, first-phase insulin secretion, beta-cell function, glucose tolerance, or racial differences in the association between pancreas lipid and beta-cell response. Those questions appear in the study summary, but their numerical results are not reported in the registry record. Liver fat is the posted primary outcome with results, and it is not interchangeable with the other outcomes described in the study’s rationale.

The study was randomized, which supports a group comparison, but the registry does not report allocation details, adherence, missing-data methods, or the analysis population. The record also does not explain how the primary outcome relates to the broader stated hypotheses. These limits do not erase the posted numbers; they define how narrowly the result can be interpreted.

In short, the registry records a lower posted mean for the low-glycemic group and an ANCOVA p-value of 0.05. The available information supports a description of those reported results, not a broad conclusion about dietary effects on diabetes, pancreatic function, or clinical outcomes.

17. Why This Trial Matters Statistically

This trial illustrates why readers should distinguish a study’s scientific rationale from the outcome that actually has posted numerical results. The registry summary emphasizes pancreas lipid, first-phase insulin secretion, and potential differences by race. The numerical primary outcome is changes in liver fat. Treating these as equivalent would overstate what the record establishes.

It also illustrates why an unadjusted recalculation and a sponsor’s analysis can differ. The recalculation uses posted group summaries, while the sponsor identifies ANCOVA as the analysis method. Without the sponsor’s adjusted estimate, confidence interval, and model specification, the two results cannot be reconciled fully. The registry flags the discrepancy, so the sponsor’s analysis takes precedence; the recalculation remains a separate descriptive check rather than a replacement.

Precision matters as much as a p-value. The unadjusted interval runs from -3.707 to -0.093, showing substantial uncertainty about the size of the contrast. The registry does not state what magnitude would be clinically important. Statistical evidence of a difference and evidence of meaningful benefit are different questions.

The planning illustrations also show how assumptions shape sample-size requirements: smaller target effects require more participants per arm for the same power target. They are not evidence of the trial’s original planning rationale, because that rationale is not reported. Together, these issues make the study a useful example of careful reading: identify the actual endpoint, separate reported from recalculated analyses, and keep the scope of conclusions aligned with the available record.

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 registry record leaves several trial-specific questions unresolved. Although the primary outcome is titled “Changes in Liver Fat” and the timeframe is baseline and week 12, it does not clarify whether the arm-level means of 5.4 and 7.3 represent change scores, follow-up values, or another summary. The endpoint description is also brief and gives no additional measurement or calculation details beyond MRI assessment of percent liver fat.

The listed primary outcome does not match the central focus of the study summary, which discusses pancreas lipid and first-phase beta-cell response. The record does not explain whether liver fat was intended to serve as a related measure or a distinct outcome. Results for the pancreas and beta-cell questions, including any race-specific analyses, are not reported in the registry record.

The sponsor’s ANCOVA result is reported only as a superiority test with p = 0.05. Its adjusted effect estimate, confidence interval, covariates, model specification, and missing-data approach are not reported. The unadjusted recalculation therefore cannot resolve the sponsor result. The record flags a discrepancy, and the sponsor’s analysis takes precedence.

Other important details are not reported, including the original sample-size justification, allocation procedures, analysis population, adherence, withdrawals, and reasons for the difference between overall enrollment and outcome-level group counts. The diet descriptions are incomplete, and adverse-event definitions and observation procedures are not given. These gaps limit assessment of implementation, precision, and generalizability without changing the values that the registry does post.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Return of First-phase Insulin Secretion in Type 2 Diabetes is Associated With Depletion of Pancreas Lipid — NCT03430310.
• Registry record last updated 2024-10-24; status last verified 2024-10.
• Results first posted 2024-10-24.
• 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

The registry records a completed randomized, unmasked, parallel-group study of low-glycemic and control diets in adults with type 2 diabetes mellitus. Enrollment is listed as 65. The primary outcome is changes in liver fat, measured as percent liver fat by MRI at baseline and week 12.

The posted arm means are 5.4 for the low-glycemic group and 7.3 for the control group, with standard errors of 0.7 and 0.6. The sponsor reports an ANCOVA superiority analysis with p = 0.05. An unadjusted recalculation gives a mean difference of -1.9, a 95% confidence interval from -3.707 to -0.093, and p = 0.039. The record flags a discrepancy, so the sponsor’s own analysis takes precedence.

The registry does not report the adjusted effect estimate, original sample-size justification, missing-data procedures, or results for the pancreas-lipid and beta-cell questions emphasized in the study summary.