Imatinib Treatment in Recent Onset Type 1 Diabetes Mellitus: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Imatinib Treatment in Recent Onset Type 1 Diabetes Mellitus (NCT01781975), a randomized phase 2 study. The primary outcome is Area Under the Stimulated C-peptide Curve (AUC) Mean Over the First 2 Hours of a 4 Hour Mixed Meal Tolerance Test at the 1 Year Visit.

2. Clinical Question

3. Trial Design

This was a phase II, randomized, triple-masked, parallel-group study of imatinib mesylate and placebo. The registry lists enrollment of 67 and describes the study as completed. The listed start was 2014-01, and the primary completion was 2017-05. The sponsor was the University of California, San Francisco.

The intervention arm received imatinib mesylate at 400 mg once daily; the comparator arm received placebo once daily. The posted primary-outcome summaries give 45 participants in the imatinib group and 22 in the placebo group. These are the group counts attached to those outcome summaries, not a statement about how many participants contributed data at every visit. The analysis population and the handling of participants without endpoint measurements are not reported in the registry record.

The primary endpoint was measured at Visit 9, Week 52. It used timed C-peptide measurements collected during a mixed meal tolerance test. The registry also lists secondary outcomes covering C-peptide at 24 months, HbA1c, insulin dose, severe hypoglycemic events, and adverse events. The registry lists six outcome measures overall and five posted secondary outcomes.

The registry identifies randomization, triple masking, and a parallel model, but does not describe the randomization procedure, allocation concealment, or who was masked. Details of eligibility criteria, study sites, participant flow, adherence, and any protocol deviations are not reported in the registry record. The original sample-size justification and the trial's planned statistical analysis are also not reported in the registry record. The summary therefore describes the design features documented in the record without inferring additional procedures.

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

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 Area Under the Stimulated C-peptide Curve (AUC) Mean Over the First 2 Hours of a 4 Hour Mixed Meal Tolerance Test at the 1 Year Visit, measured over Visit 9 (Week 52) at 0, 15, 30, 60, 90, 120 minutes post-dose.

The primary endpoint was the mean area under the stimulated C-peptide curve over the first two hours of a four-hour mixed meal tolerance test at the one-year visit, reported in nmol/L. C-peptide is released alongside insulin made by the body, so stimulated C-peptide provides a measure of remaining insulin-producing beta-cell function. The registry states that its AUC mean is the area under the curve divided by the 120-minute measurement interval. This expresses the accumulated response as a mean value on the measurement scale rather than as an area whose units also include time.

Measurements were scheduled at 0, 15, 30, 60, 90, and 120 minutes after the test meal. Combining repeated measurements captures the response across the test period rather than relying on a single time point. In general, a summary over time can reflect the overall level of a response during that interval, while still depending on the measurement schedule and the calculation method.

The endpoint is a continuous measure, and the registry reports group means and standard deviations. Its role in this trial was to assess beta-cell function at the primary endpoint visit. It is not, by itself, a direct measure of every clinical consequence of type 1 diabetes or a guarantee about later outcomes. Results for the primary endpoint should therefore be interpreted as evidence about this specified C-peptide measure and visit.

7. Results Posted in the Registry

• Imatinib Mesylate: mean 0.83 versus 0.78 on control; difference 0.06 (95% CI -0.12 to 0.23), standardized effect d = 0.15; z-test p = 0.508.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the endpoint summary describe?

The primary outcome is continuous: it is the mean C-peptide AUC over a defined interval, expressed in nmol/L. For a continuous endpoint, a common summary is the mean, which describes the group's average observed value, together with a measure of spread. The registry reports standard deviations, which describe variability among observations around the mean. In general, a standard deviation is not a confidence interval for the mean and does not itself express the precision of a treatment comparison.

The posted means are 0.834 for imatinib and 0.775 for placebo. The associated standard deviations are 0.446 and 0.278. The recalculated mean difference is 0.059. This difference is obtained as a simple unadjusted comparison of the posted group summaries. It does not account for possible baseline differences or other covariates, and the registry does not provide participant-level data that would allow a separate model-based re-analysis.

What is the difference between an unadjusted comparison and ANCOVA?

In general, an unadjusted comparison contrasts the observed group summaries without controlling for additional variables. ANCOVA, or analysis of covariance, compares groups using a statistical model that includes one or more covariates. When appropriate, covariate adjustment can account for measured differences and may change the estimated comparison or its uncertainty. The registry identifies the sponsor's method as ANCOVA but does not report which covariates were included, how they were measured, or how the model was specified. It is therefore not possible to reconstruct that analysis from the record.

The sponsor's reported ANCOVA estimate is 0.062, with a 90% confidence interval from 0.002 to 0.121 and p = 0.048. The separate summary-statistic re-calculation gives p = 0.508 and a confidence interval that crosses zero. The registry flags this as a discrepancy. The sponsor's own analysis takes precedence; the re-calculation is useful as a transparent, unadjusted comparison, but it is not a replacement for the sponsor's analysis.

How should a p-value be read?

In general, a p-value describes how compatible the observed data are with a specified null model, under the assumptions of the analysis. It is not the probability that the null hypothesis is true, nor does it describe the size or importance of an effect. The sponsor reports p = 0.048 for its ANCOVA superiority analysis, while the unadjusted re-calculation reports p = 0.508. The values refer to different analyses and should not be compared as if they came from the same model. The registry record does not report enough detail to explain the difference.

A superiority analysis evaluates whether the data support a difference in the direction specified by the analysis. The registry labels the reported ANCOVA result as superiority, but the underlying specification and any decision rules are not reported. The label alone does not establish what threshold, if any, was set in the original protocol.

What does a confidence interval contribute?

In general, a confidence interval gives a range of estimates compatible with the data and the statistical model at the stated confidence level. It conveys uncertainty around an estimate, rather than merely classifying a result by whether a p-value passes a threshold. The sponsor's interval is explicitly 90% and lies just above zero through 0.121. The unadjusted re-calculated interval crosses zero. Because the intervals come from different analyses, they answer related but non-identical questions under different methods.

Confidence intervals depend on model assumptions and the data used. The registry reports no ANCOVA covariates or model diagnostics, and the re-calculation is based on summary statistics. No interval should be interpreted as covering every plausible source of uncertainty, such as missing data or analysis choices.

Why can variation matter?

In general, greater variability makes a group mean harder to estimate precisely when the amount of data is held fixed. The imatinib group has a larger posted standard deviation than the placebo group. The observed spread is relevant context, but it does not alone explain the discrepancy between analyses. The re-calculated method uses the posted group sizes and standard deviations; the sponsor's ANCOVA may use information not represented by those simple summaries. Its details are not reported in the registry record.

These methods explain why the mean difference, its interval, and the p-value need to be read together, and why analysis specifications matter. The registry's sponsor analysis remains the primary reported inferential result, while the re-calculation is a separate unadjusted check based on summary data.

10. Confidence Intervals and Precision

Precision concerns how narrowly an analysis estimates the group difference. The sponsor's ANCOVA result has a 90% confidence interval from 0.002 to 0.121. Since the interval is positive but its lower bound is close to zero, the result is compatible with a small positive difference as well as larger positive values within the interval. The registry reports the sponsor estimate as 0.062. The interval's width indicates that the estimated difference is not pinned down to a single value.

The unadjusted re-calculation has a confidence interval that crosses zero. That interval therefore includes values on both sides of no difference, and the associated unadjusted p-value is 0.508. The registry does not state the confidence level for this re-calculated interval. It should not be assigned a confidence level that is not reported.

These intervals come from different analyses and should not be treated as competing versions of the same calculation. The sponsor interval belongs to ANCOVA, while the re-calculated interval uses the posted means, standard deviations, and group sizes without adjustment. The registry does not give enough detail about the ANCOVA model to establish why the precision differs.

In general, a confidence interval describes statistical uncertainty under the method's assumptions; it does not account automatically for missing data, selective reporting, or uncertainty about model choice. The registry does not report the missing-data method or analysis populations. Precision should therefore be understood as conditional on the analyses that were carried out, rather than as a complete measure of every uncertainty in the trial.

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 planning figures are participants per arm, not total event counts. The event-count convention applies to time-to-event outcomes; this primary endpoint is continuous. The planning illustration uses a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It uses the observed effect as its reference and considers planning for 100%, 75%, and 50% of that observed effect.

For planning at 100% of the observed effect, the figures are 720 participants per arm for 80% power and 964 participants per arm for 90% power. For planning at 75% of the observed effect, the corresponding figures are 1,280 and 1,714 participants per arm. For planning at 50% of the observed effect, they are 2,880 and 3,855 participants per arm.

In plain terms, the 80% power column describes the planned number per arm under which the calculation would have an 80% chance of detecting the specified effect if that effect were present and the assumptions held. The 90% power column gives the planned number per arm for a higher chance of detecting that same specified effect. These are planning illustrations, not observed power.

The smaller the effect chosen for planning, the more participants are needed to detect it under the stated assumptions. The figures are based on the observed effect and are not the trial's original design justification. The original sample-size assumptions and justification are not reported in the registry record. The calculation also does not establish that a future study would have the same variability, effect, or analysis conditions.

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 = 0.059; pooled standard deviation σ = 0.400.
2. n = 2 × (1.96 + 0.842)² × 0.400² / 0.059² = 719.8, rounded up to 720 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 <- 0.0589999999999999                      # observed difference in means
sigma <- 0.399524101516176                     # 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(1781975)
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 = 720 (80% power), 964 (90% power)
 75% of observed effect: n per arm = 1,280 (80% power), 1,714 (90% power)
 50% of observed effect: n per arm = 2,880 (80% power), 3,855 (90% power)
power.t.test: 720.8 
Simulated power at n = 720 per arm: 0.799 (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 primary analysis as ANCOVA, with an estimated difference of 0.062, a 90% confidence interval from 0.002 to 0.121, and p = 0.048. It labels the analysis as superiority. This is the trial's own reported inferential analysis of the primary outcome.

A separate recalculation from the posted group means and standard deviations gives an unadjusted mean difference of 0.059, a confidence interval that crosses zero, and p = 0.508. The recalculation is not an alternative sponsor result: it is a simple comparison reconstructed from summary statistics and does not incorporate the ANCOVA model.

The registry flags a discrepancy between these results. Under that circumstance, the sponsor's own analysis takes precedence. The two estimates and p-values should still be described separately, because they come from different methods and do not have the same interpretation. The registry does not report the ANCOVA covariates, model specification, or other details needed to reproduce the sponsor result or explain the discrepancy. The re-calculation can serve as a check on the posted group summaries, but it should not be used to override the reported analysis.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary endpoint and five posted secondary outcomes. The secondary outcomes include C-peptide AUC at 24 months, HbA1c, insulin dose, severe hypoglycemic events, and adverse events. Considering multiple endpoints or analyses raises a general statistical issue: when many hypotheses are tested, the chance of at least one apparently unusual result can increase. In general, multiplicity procedures can control or describe this issue, but they need to be prespecified and reported to interpret the family of tests.

The registry record does not report whether the trial used a multiplicity adjustment, a testing hierarchy, or another procedure for the primary and secondary outcomes. It also does not report whether additional analyses were planned or how the reported p-value relates to any broader testing strategy. The sponsor's primary p-value should therefore be read as reported, without assuming it was adjusted or unadjusted for other outcomes unless that is explicitly stated. The p-value for the summary-statistic re-calculation is described as unadjusted.

Important details about missing data and analysis populations are also not reported in the registry record. The record does not describe how missing Week 52 measurements were handled, whether any imputation was used, or how withdrawals and incomplete tests were treated. It does not specify whether the primary analysis included all randomized participants, participants who received treatment, or a subset with evaluable measurements. The group counts attached to the posted endpoint summaries do not, by themselves, establish the analysis population or account for missing measurements.

The registry identifies ANCOVA for the sponsor analysis but does not give its covariates, model assumptions, or diagnostics. It does not report sensitivity analyses assessing the effect of missing data or alternative analysis choices. These reporting gaps matter because both missingness and model specification can affect an estimate and its uncertainty. They limit independent assessment of how robust the primary result is, without establishing that any particular analysis choice was inappropriate.

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 serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: hypoglycemia (4.5%), fluid in lungs (4.5%), diabetes ketoacidosis (4.5%), suicide attempt (2.2%), folliculitis (2.2%).

The registry reports serious adverse events affecting 8 participants in the imatinib group, among 45 at risk, and 3 participants in the placebo group, among 22 at risk. It reports one death in the imatinib group and none in the placebo group. These figures describe counts recorded for the groups; the registry does not report the cause of death or whether it was judged related to study treatment.

The registry also lists serious-event terms with percentages: Hypoglycemia, 4.5%; Fluid in Lungs, 4.5%; Diabetes ketoacidosis, 4.5%; Suicide Attempt, 2.2%; Folliculitis, 2.2%; and Wheezing, 2.2%. The listed terms and percentages should be read as the registry reports them; the record does not give further detail about event timing, recurrence, severity grading, or attribution.

In general, counts of participants with at least one serious event are not necessarily counts of all serious-event episodes, since one participant may experience more than one event. The registry does not specify the event-counting convention beyond the posted fields. These descriptive safety figures do not establish a causal treatment effect. The record also reports other adverse events affecting 23 participants in the imatinib group and 12 in the placebo group. Further detail about definitions and follow-up is not reported in the registry record.

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

The central statistical result concerns stimulated C-peptide AUC at the Week 52 visit, not every feature of diabetes or long-term health. The registry's sponsor analysis reports a positive ANCOVA estimate and a 90% interval just above zero through 0.121, with p = 0.048. The unadjusted re-calculation from group summaries is less conclusive: its interval crosses zero and p = 0.508. Because the record flags a discrepancy, the sponsor's analysis takes precedence, while the recalculation remains a separate summary-statistic comparison.

The estimates describe a difference in a continuous endpoint between groups under particular analyses. They do not establish the size or durability of any effect beyond the recorded visit, and they do not establish effects on every secondary outcome. Although the registry's background summary describes an interest in possible durable remission, the primary endpoint is a C-peptide measure at one year. The information reported for that endpoint cannot answer questions about remission over an unspecified longer period.

The registry's designation of a superiority analysis and the p-value do not eliminate uncertainty. The sponsor interval's lower end is close to zero, and the analysis details needed to reproduce it are not reported. The unadjusted calculation differs, while the analysis population, missing-data approach, and multiplicity handling are also not reported in the registry record.

Safety counts are also descriptive. The serious adverse event and death figures should not be interpreted as establishing whether treatment caused an event, particularly without event-level detail or attribution. A complete assessment would require the protocol, fuller analysis documentation, and clinical context beyond the reported summaries. The posted results support careful description of the recorded outcomes, not medical advice or a universal conclusion about treatment.

17. Why This Trial Matters Statistically

This trial illustrates why statistical interpretation should keep the endpoint, analysis method, and uncertainty together. A numerical difference between group means is not the same thing as an adjusted treatment estimate, and a p-value is not a measure of effect size. Here, the sponsor's ANCOVA result and the unadjusted recalculation lead to different inferential readings. The registry flags that discrepancy, making transparent reporting of the sponsor analysis especially important.

The endpoint also demonstrates how repeated measurements can be condensed into a clinically relevant continuous summary. The stimulated C-peptide AUC mean represents response over a specified period during a mixed meal tolerance test, rather than a single timed measurement. The definition and time window matter: the primary result is tied to the first two hours of the test at the Week 52 visit, not to a general measure of all diabetes outcomes.

Precision matters alongside the point estimate. The sponsor's interval remains positive but has a lower boundary close to zero, while the unadjusted interval crosses zero. This difference emphasizes that analytic assumptions influence uncertainty and that estimates should not be presented without their method and interval.

Finally, this record shows the value of reporting the analysis population, missing-data handling, multiplicity strategy, and model specification. Those details help readers judge how robust a result is and whether it can be independently reproduced. Since several such details are not reported, interpretation should remain anchored to the reported ANCOVA result and clearly distinguish it from the unadjusted recalculation.

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 important design and analysis details unspecified. Eligibility criteria and the operational definition of recent-onset disease are not reported, nor are baseline participant characteristics. The record lists randomization and triple masking but does not describe sequence generation, allocation concealment, or which parties were masked. Site information and participant flow are not reported in the registry record.

The primary analysis is identified as ANCOVA, but its covariates and model specification are not reported. The analysis population, handling of missing endpoint measurements, and sensitivity analyses are also not reported. These omissions limit the ability to reproduce the sponsor result and to assess how incomplete follow-up might affect it. The registry flags a discrepancy between that result and the unadjusted summary-statistic recalculation, but does not explain the source of the difference.

The record does not report the trial's original sample-size justification, the planned statistical analysis, interim analyses, or a multiplicity strategy. The posted planning figures are illustrations based on the observed effect and stated assumptions, not evidence about the original design decision. The detail available for adverse events is also limited: event timing, attribution, recurrence, and the cause of the reported death are not reported.

These reporting limitations do not show that the methods were inadequate; they define what can and cannot be evaluated from the registry record. Conclusions should remain specific to the posted results and the Week 52 primary endpoint.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Imatinib Treatment in Recent Onset Type 1 Diabetes Mellitus — NCT01781975.
• Registry record last updated 2020-02-11; status last verified 2020-02.
• Results first posted 2018-08-08.
• 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.

22. Record Summary

NCT01781975 is a completed, randomized, triple-masked, parallel-group phase II study sponsored by the University of California, San Francisco. The registry lists enrollment of 67, a start of 2014-01, and primary completion in 2017-05. The comparison was imatinib mesylate, 400 mg once daily, versus once-daily placebo in recent-onset type 1 diabetes.

The primary outcome was mean stimulated C-peptide AUC over the first two hours of a mixed meal tolerance test at Week 52. The sponsor reported an ANCOVA superiority result with estimate 0.062, 90% confidence interval 0.002 to 0.121, and p = 0.048. An unadjusted re-calculation from posted summaries gives a different result; because the registry flags a discrepancy, the sponsor's own analysis takes precedence. Important details about the model, missing data, and analysis populations are not reported in the registry record.