Study of Sitagliptin for the Treatment of Type 2 Diabetes Mellitus With Inadequate Glycemic Control on Insulin (MK-0431-260): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of Sitagliptin for the Treatment of Type 2 Diabetes Mellitus With Inadequate Glycemic Control on Insulin (MK-0431-260) (NCT01462266), a randomized phase 3 study. The primary outcome is Change From Baseline in Daily Insulin Dose at Week 24.

2. Clinical Question

3. Trial Design

This was a Phase III, multicenter, randomized, double-blind, placebo-controlled, parallel-group clinical trial. The registry lists an enrollment of 660 and identifies Merck Sharp & Dohme LLC as the sponsor. The record gives the study start as 2012-01-13, primary completion as 2013-06-06, and status as completed.

Participants with type 2 diabetes mellitus and inadequate glycemic control on insulin alone or insulin with metformin were assigned to sitagliptin or placebo. The sitagliptin regimen was 100 mg once daily; the placebo was taken once daily. The stated study purpose was to examine the insulin-sparing effect over 24 weeks. Both groups are reported with 329 participants for the primary outcome. Outcome-level analysis populations can differ from the enrollment figure, and the registry record does not explain the relationship between those figures.

Randomization is intended to make treatment groups comparable on average at the start of a trial, while double masking is intended to limit the influence of treatment knowledge on participants and study personnel. These design features support interpretation of the between-group comparison, but do not by themselves establish that all possible sources of bias were eliminated. The record identifies the design features but does not report the randomization procedure, allocation-concealment details, or masking procedures.

The primary outcome was change from baseline in daily insulin dose at Week 24, measured in International Units (IU). The registry also lists changes in hemoglobin A1c and fasting plasma glucose, achievement of a fasting glucose target, and time to achieve that target as secondary outcomes. Detailed posted results are available for some, but not all, of those secondary measures. The original sample-size justification, trial-specific power assumptions, interim analyses, and complete statistical analysis plan are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set (FAS) population included all randomized participants who took at least one dose of study medication and had at least one measurement either at baseline or post-randomization.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: One participant in both the Sitagliptin group and Placebo group was randomized in error and did not receive study medication.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Daily Insulin Dose at Week 24, measured over Baseline and Week 24.

The primary endpoint is the change from baseline in daily insulin dose at Week 24, reported in International Units (IU). Change is defined as the daily dose at Week 24 minus the daily dose at baseline. With this direction of subtraction, a positive change represents a higher dose at Week 24 than at baseline, while a negative change represents a lower dose.

The registry reports least-squares means for the two groups and a mean difference for sitagliptin relative to placebo. The posted group estimates were 19.0 for sitagliptin and 23.8 for placebo. The recalculated unadjusted difference was -4.8 IU, with a 95% confidence interval from -8.371 to -1.229. These are estimates of group-average changes, not individual participant changes. They indicate that the average increase in dose was smaller in the sitagliptin group than in the placebo group over the stated period.

Insulin dose is the direct measure of the trial’s stated insulin-sparing question. It does not, on its own, describe glycemic control, the reasons for dose adjustments, or the experience of each participant. The registry lists glycemic outcomes separately, including hemoglobin A1c and fasting plasma glucose. A full interpretation therefore considers the dose outcome alongside those measures and the limits of the reported analysis. The registry record does not report the detailed dose-adjustment rules or the clinical context of individual dose changes.

7. Results Posted in the Registry

• Sitagliptin: mean 19.00 versus 23.80 on control; difference -4.80 (95% CI -8.37 to -1.23), standardized effect d = -0.21; z-test p = 0.008.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Change From Baseline in Hemoglobin A1c (A1C) at Week 24

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Sitagliptin: mean -1.31 versus -0.87 on control; difference -0.44 (95% CI -0.60 to -0.28), standardized effect d = -0.41; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Sitagliptin than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change From Baseline in Fasting Plasma Glucose (FPG) at Week 24

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Sitagliptin: mean -55.50 versus -44.80 on control; difference -10.70 (95% CI -18.62 to -2.78), standardized effect d = -0.21; z-test p = 0.008.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Sitagliptin than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percent of Participants Achieving Fasting Glucose Target at Any Time During the Study

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Sitagliptin: 77.4% (253/327) versus 74.2% (242/326) on control; risk difference 3.1 percentage points (95% CI -3.4 to 9.7), risk ratio 1.04 (0.96–1.14); unadjusted two-proportion z-test p = 0.349.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Sitagliptin and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes broadly address glycemic measures alongside the primary insulin-dose question. The posted estimates for hemoglobin A1c and fasting plasma glucose move in a direction that is numerically more favorable to sitagliptin, while the fasting glucose target proportions are also numerically higher for that group. These results offer context for the primary outcome but do not redefine it: insulin-dose change remains the direct measure of the stated insulin-sparing hypothesis.

The secondary measures differ in scale and meaning. Hemoglobin A1c and fasting plasma glucose are continuous outcomes measured as change from baseline, whereas achievement of a fasting glucose target is a binary outcome summarized as a proportion. The listed time-to-achieve-target outcome asks a different question about timing. A difference in one measure should not be treated as a direct substitute for another, and the registry does not report a detailed result for time to achieve the target.

The registry record lists multiple secondary outcomes, but the posted secondary comparisons are unadjusted recalculations. The record does not report whether the study used a multiplicity procedure, a prespecified testing sequence, or a hierarchy for interpreting these endpoints. When several comparisons are considered, the chance of at least one apparently notable result can increase if each is assessed separately without adjustment. For that reason, secondary findings carry less confirmatory weight than the primary outcome unless a prespecified analysis plan supports a stronger interpretation.

The overall secondary pattern can be considered supportive context, but it cannot establish a unified treatment effect across distinct endpoints. The uncertainty for each measure, the endpoint definitions, missing-data handling, and multiplicity all matter. The registry record does not report enough analytic detail to determine how these issues were handled in the sponsor’s secondary analyses.

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

• Change From Baseline in Hemoglobin A1c (A1C) at Week 24
• Change From Baseline in Fasting Plasma Glucose (FPG) at Week 24
• Percent of Participants Achieving Fasting Glucose Target at Any Time During the Study
• Time to Achieve the Fasting Glucose Target

9. Statistical Interpretation

10. Statistical Methods Explained

The registry reports the primary outcome as a least-squares mean change from baseline, with a 95% confidence interval. It also gives the sponsor’s longitudinal data analysis and a mean difference in final values. The recalculated between-group estimates use posted summary statistics and are explicitly unadjusted; they are useful as a straightforward comparison but are not substitutes for the sponsor’s trial analysis.

What does a mean difference represent?

A mean difference compares the average outcome in one group with the average in another. For the primary outcome, the comparison is sitagliptin minus placebo. The estimate of -4.8 IU therefore represents a lower average change in daily insulin dose in the sitagliptin group. The sign depends on the order of subtraction, so it should always be read alongside the stated group order and the endpoint definition. Here, change is Week 24 dose minus baseline dose, and both group-level changes are positive.

A group mean is not a description of every participant. Individuals may have had changes that differed from the group average, and the registry summary does not provide the distribution of individual changes or reasons for dose adjustment. The listed spread values are not needed to identify the direction of the group comparison, and the interval around the difference is more directly relevant to precision than comparing the separate group intervals by eye.

What does the confidence interval tell the reader?

General statistical explanation: a confidence interval expresses uncertainty around an estimate under the assumptions of the method used to calculate it. It is not a guarantee that the true value lies within the observed interval, nor does it describe the range of individual participant outcomes. For the primary recalculation, the 95% interval extends from -8.371 to -1.229 IU. Its position below zero is consistent with a lower average change in the sitagliptin group under this unadjusted calculation. The sponsor-reported interval is based on the sponsor’s longitudinal data analysis and is reported separately.

Why do the sponsor analysis and recalculation differ?

The registry identifies the sponsor’s method as longitudinal data analysis and reports its mean difference as -4.7, with a 95% confidence interval from -8.3 to -1.2 and p = 0.009. The recalculation uses posted summary values, an approximate pooled standard deviation, and an unadjusted comparison. A model-based longitudinal analysis can use information differently from a calculation based on group summary statistics. The record does not provide enough detail to reproduce every element of the sponsor’s model, such as its covariates or handling of repeated measurements. The results are close, but the sponsor’s own analysis is the relevant reported trial analysis.

The unadjusted p-value is a measure of how incompatible the observed estimate would be with a specified null comparison under the calculation’s assumptions. It does not measure the probability that the null hypothesis is true, nor does it measure the size or importance of the estimated difference. The registry reports p = 0.009 for the sponsor analysis and the recalculated primary comparison has p = 0.008. Those values should not be conflated: one is the trial’s reported analysis and the other is the unadjusted recalculation.

What does randomization contribute?

General statistical explanation: random allocation is designed to balance measured and unmeasured baseline characteristics between groups on average. This supports causal interpretation of a between-group difference when follow-up, outcome measurement, and analysis are also appropriately conducted. Randomization does not ensure exact balance in a particular trial, and the registry record does not provide baseline balance details or the randomization implementation. Double masking can reduce the opportunity for expectations to influence behavior or assessment, but the record does not describe how masking was maintained or assessed.

The primary analysis population is the Full Analysis Set, defined in the record as randomized participants who took at least one dose of study medication and had at least one measurement at baseline or after randomization. This definition includes a measurement at either time, rather than requiring the record’s wording to specify both. The registry does not give a full account of missing outcome data, exclusions after randomization, or sensitivity analyses. Those omissions matter because analysis populations and missing observations can affect estimates.

Why can variability matter?

General statistical explanation: greater variability in an outcome makes a group difference harder to estimate precisely, because observed values are more dispersed around their averages. The standard error reflects both variability and the amount of information available. Confidence intervals combine the estimated difference with its uncertainty. Over-dispersion is especially relevant for count or binary outcomes when observed variation exceeds what a simple model assumes; methods that account for it may be needed. The primary endpoint is continuous, while the fasting glucose target outcome is binary. The registry record does not report a detailed model description for the recalculated secondary analyses, so they should be treated as simple unadjusted comparisons rather than as the sponsor’s complete inferential plan.

The recalculated p-values and confidence intervals are unadjusted. When several outcomes are examined, chance alone can produce apparently notable findings across the set. General statistical explanation: multiplicity procedures can control the overall false-positive risk across planned comparisons, but may affect intervals or thresholds. The registry record does not state whether such procedures were used, so no adjustment should be presumed. This is why the primary endpoint, which is directly tied to the stated hypothesis, carries a different interpretive role from secondary outcomes.

11. Confidence Intervals and Precision

The primary recalculated mean difference is -4.8 IU, with a 95% confidence interval from -8.371 to -1.229. This interval communicates uncertainty about the estimated difference in average change in daily insulin dose. Every value in the interval is negative, so the interval is consistent with a lower average dose change for sitagliptin than for placebo under the unadjusted calculation. The interval does not mean that every participant had a lower dose change, and it does not provide a range for individual responses.

The sponsor-reported interval is -8.3 to -1.2 around a mean difference of -4.7. It has a similar direction and scale to the recalculated interval, but it comes from the sponsor’s longitudinal data analysis. The two analyses should remain distinct when interpreting precision. The recalculated interval is based on posted summary statistics; the sponsor’s interval reflects the trial’s reported method.

General statistical explanation: a confidence interval’s width reflects uncertainty under the method’s assumptions. More precise estimates tend to have narrower intervals, while less precise estimates tend to have wider intervals. The interval should be interpreted with its point estimate and the endpoint’s units, not as proof that all values within it are equally likely or that values outside it are impossible.

The secondary outcomes have their own intervals and scales. For the fasting glucose target comparison, the recalculated risk-difference interval extends from -0.034 to 0.097, so it includes zero. This indicates that the posted summary data do not give a precise direction for that unadjusted difference at the interval level. The continuous secondary comparisons have intervals reported in their respective units. Their precision does not alter the definition or result of the primary insulin-dose outcome.

Confidence intervals do not address every source of uncertainty. They do not resolve possible missing-data effects, multiplicity, departures from model assumptions, or limitations in generalizing beyond the studied population. The registry record does not report all of those analytic details.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size figures are planning illustrations based on the observed effect, not the trial’s own original design assumptions. The calculations use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. For continuous outcomes, the stated effect to detect is a specified fraction of the observed effect. The displayed participant counts are per arm, not totals. The assumptions for the trial’s original sample-size design are not reported in the registry record.

The 80% power column gives an illustrative number of participants per arm for a design that would have an 80% chance of detecting the chosen effect under the stated assumptions, if that effect and the assumptions were correct. The 90% power column describes the corresponding planning target with a higher chance of detection. These are design calculations, not measures of how much power the completed trial actually had. This is not observed power.

The planning figures are shown at 100%, 75%, and 50% of the observed effect. A smaller target effect requires more participants because a modest difference is harder to distinguish from sampling variability. For example, for the primary outcome, planning for 100% of the observed effect gives 373 participants per arm for 80% power and 499 participants per arm for 90% power. Planning for 50% of that observed effect gives 1,489 participants per arm for 80% power and 1,993 participants per arm for 90% power.

These illustrations should not be read as evidence that the original trial was designed around the observed effect. The registry record does not report the original design effect, sample-size assumptions, or calculation method. The planning values are conditional on the stated assumptions and are not forecasts of what a future study would necessarily require.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -4.800; pooled standard deviation σ = 23.369.
2. n = 2 × (1.96 + 0.842)² × 23.369² / -4.800² = 372.1, rounded up to 373 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 <- -4.8                      # observed difference in means
sigma <- 23.3685909277768                     # 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(1462266)
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 = 373 (80% power), 499 (90% power)
 75% of observed effect: n per arm = 662 (80% power), 886 (90% power)
 50% of observed effect: n per arm = 1,489 (80% power), 1,993 (90% power)
power.t.test: 373 
Simulated power at n = 373 per arm: 0.800 (20000 replicates)

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

14. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary analysis as a longitudinal data analysis, with a mean difference in final values of -4.7, a 95% confidence interval from -8.3 to -1.2, and p = 0.009. This is the trial’s own reported inferential result and should be used when describing the sponsor analysis.

The recalculated unadjusted comparison from the posted group summaries estimates a mean difference of -4.8 IU, with a 95% confidence interval from -8.371 to -1.229 and p = 0.008. Its purpose is to provide a simple check based on the reported summary statistics, not to duplicate the longitudinal model. The estimates are close in direction and magnitude, but differences in analysis methods can produce small differences in point estimates, intervals, and p-values.

The recalculation uses a pooled variability estimate described as approximate. It does not incorporate undocumented model details or substitute for participant-level analysis. The registry record does not give enough information to reproduce the sponsor’s full longitudinal analysis. There is no recorded discrepancy flag, but that does not make the methods interchangeable: the sponsor’s reported result remains the primary description of the trial’s own analysis, while the recalculation is an unadjusted supporting comparison.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies a primary endpoint and several secondary outcomes. It reports a sponsor analysis for the primary outcome and provides posted secondary outcome estimates, but does not describe a multiplicity adjustment or a prespecified hierarchy for testing outcomes. The recalculated p-values are unadjusted. They therefore should not be read as if each were protected against the possibility of chance findings arising from examination of multiple outcomes. Whether the sponsor used a formal multiplicity strategy is not reported in the registry record.

The primary outcome is the most direct test of the stated hypothesis because the trial’s purpose centered on insulin dose. Secondary outcomes address related but separate questions about glycemic change and target achievement. Without reported multiplicity procedures, secondary results are best treated as contextual or exploratory rather than as independent confirmation. The absence of reported detail does not show that a procedure was not used; it means the registry record does not document it.

The primary analysis population is described as the Full Analysis Set: all randomized participants who took at least one dose of study medication and had at least one measurement either at baseline or post-randomization. The registry does not report how many randomized participants were excluded from this population, how many lacked a Week 24 measurement, or the reasons for missing measurements. Detailed missing-data methods, including any imputation or sensitivity analyses, are not reported in the registry record.

Missing data can affect a comparison if the availability of outcomes is related to treatment group, participant characteristics, or the outcome itself. General statistical explanation: analyses that rely on complete observations can be biased under some missingness patterns, while model-based or imputation approaches require assumptions about the unobserved values. No particular missing-data mechanism or method can be inferred from the endpoint summaries alone.

The secondary outcome counts differ slightly across groups for the fasting glucose target analysis, and the record reports the analyzed denominators for that measure. The reasons for those denominators are not reported. The primary group counts are 329 in each arm, but those outcome counts do not by themselves describe all randomized participants or all missing observations. Conclusions about missing-data robustness are therefore limited by the information in the registry record.

16. 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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: thrombocytopenia (0.3%), acute coronary syndrome (0.3%), acute myocardial infarction (0.3%), angina pectoris (0.3%), coronary artery disease (0.3%).

The registry reports serious adverse events affecting 13 of 329 participants in the sitagliptin group and 12 of 329 in the placebo group. It also reports other adverse events affecting 121 of 329 and 168 of 329 participants, respectively. These are counts of participants reported as affected, not a complete description of event severity, timing, duration, or causal relationship to study treatment.

The listed serious-event terms include thrombocytopenia, acute coronary syndrome, acute myocardial infarction, angina pectoris, coronary artery disease, and myocardial infarction, each reported as 0.3%. These terms are registry entries and should not be interpreted as proof that a treatment caused an event. The registry record does not report deaths for these groups.

Adverse-event counts are descriptive. A simple comparison of counts does not account for exposure time, event recurrence, differences in participant characteristics, or uncertainty around rare events. The record does not provide enough detail to infer a causal safety difference, and these figures alone do not establish the full safety profile of either study regimen.

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

The primary comparison describes a lower average change in daily insulin dose in the sitagliptin group than in the placebo group over 24 weeks. The group means themselves were positive, so the result is a smaller average increase, not an average reduction from baseline. The recalculated mean difference is -4.8 IU; the sponsor-reported estimate is -4.7. The sponsor’s longitudinal data analysis is the trial’s reported analysis, and the recalculation is an unadjusted check based on group summaries.

The confidence intervals indicate uncertainty around the average between-group estimates. They do not describe the range of individual participant outcomes or show that each participant had a change in the same direction. Group averages can conceal substantial variation between participants, and the registry does not provide individual-level outcome distributions or detailed dose-adjustment histories.

The secondary glycemic outcomes provide additional context, but they are distinct from the primary endpoint and the registry does not report a multiplicity strategy. Their results should therefore not be treated as equally confirmatory to the primary analysis. The registry lists time to achieve the fasting glucose target, but its detailed result is not reported in the registry record.

These findings apply to the trial population and the treatment comparison described in the record. The record concerns participants with type 2 diabetes mellitus whose glycemic control was inadequate on insulin alone or with metformin, and reports an outcome at Week 24. It does not establish how results would generalize to other populations, treatment settings, durations, or regimens not described there. Nor does a group-level estimate determine an individual participant’s treatment response.

The registry results do not explain all clinical decisions behind insulin dosing or the reasons for changes. They should be understood as statistical evidence about the reported group comparison, not as medical advice or a recommendation for an individual.

18. Why This Trial Matters Statistically

The trial is statistically relevant because its primary outcome directly measures the insulin-sparing question stated in the registry: change in daily insulin dose rather than a surrogate for dose use. The endpoint’s subtraction direction is important. Positive group estimates indicate increases from baseline, while the negative between-group difference indicates a smaller increase in the sitagliptin group relative to placebo.

The record also illustrates why a point estimate should be read with its uncertainty and analysis method. The sponsor’s longitudinal data analysis and a simple unadjusted recalculation produce close estimates, but they are distinct analyses. Reporting both makes clear that summary-statistic calculations can provide a useful check while not replacing the trial’s own model-based result.

The secondary outcomes illustrate a further statistical distinction: different endpoints address different questions and use different scales. Glycemic measures can provide context for an insulin-dose comparison, but they do not answer the same question. Their interpretation also depends on uncertainty and multiplicity, and the registry does not state whether secondary testing was adjusted.

Finally, the record highlights the difference between planning calculations and the design of a completed trial. The planning illustrations use observed effects and explicit assumptions, whereas the trial’s original sample-size justification is not reported. Treating those as separate prevents a retrospective calculation from being mistaken for evidence about the design decisions or operating characteristics of the actual trial.

19. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record does not report the original sample-size justification, the effect assumed for planning, or the trial’s design power. Planning illustrations based on observed effects cannot fill that gap. The record also does not give the full statistical analysis plan, detailed model specification, randomization procedure, allocation-concealment process, or masking procedures.

Missing-data information is limited. The Full Analysis Set definition is reported, but the number and reasons for missing Week 24 measurements, exclusions from the analysis population, and any imputation or sensitivity analyses are not. The registry record does not establish whether missingness could have influenced the reported results.

The sponsor’s primary result is identified as a longitudinal data analysis, but further details needed to reproduce the model are not reported. The recalculation is unadjusted and relies on summary statistics, including an approximate pooled standard deviation. It is not equivalent to participant-level reanalysis and should not be treated as the sponsor’s analysis.

Multiplicity handling for the several outcomes is not reported. This limits the strength of conclusions drawn from the secondary results. Detailed results for time to achieve the fasting glucose target are also not reported in the registry record, despite that outcome being listed. Safety information is presented as participant counts and selected serious-event terms; the record does not supply enough detail to assess exposure time, event recurrence, causality, or the full safety profile.

The study outcome is reported at Week 24. The record therefore does not establish longer-term patterns. It also does not provide individual-level data, so the distribution of responses and possible differences among participant subgroups cannot be assessed from these summaries.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Study of Sitagliptin for the Treatment of Type 2 Diabetes Mellitus With Inadequate Glycemic Control on Insulin (MK-0431-260) — NCT01462266.
• Registry record last updated 2018-08-17; status last verified 2018-07.
• Results first posted 2014-04-08.
• 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.

23. Record Summary

The registry describes a completed, Phase III, multicenter, randomized, double-blind, placebo-controlled, parallel-group trial of sitagliptin 100 mg once daily versus placebo in participants with type 2 diabetes mellitus and inadequate glycemic control on insulin alone or with metformin. The stated primary question concerned the change in daily insulin dose at Week 24.

The sponsor-reported longitudinal data analysis estimated a mean difference of -4.7, with a 95% confidence interval from -8.3 to -1.2 and p = 0.009. The unadjusted recalculation estimated -4.8 IU, with a 95% confidence interval from -8.371 to -1.229 and p = 0.008. The group-average changes were positive in both arms, so the comparison indicates a smaller average increase in the sitagliptin group.

Secondary glycemic outcomes and adverse-event counts provide additional context, but multiplicity handling, detailed missing-data methods, the original sample-size justification, and several analysis details are not reported in the registry record.