This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for A Study in China Evaluating the Safety and Efficacy of Adding Sitagliptin to Stable Therapy With Sulfonylurea With or Without Metformin in Participants With Type 2 Diabetes Mellitus (T2DM) (MK-0431-253) (NCT01590771), a randomized phase 3 study. The primary outcome is Change From Baseline in A1C Levels at Week 24 in Participants Receiving Sitagliptin and Sulfonylurea Alone or in Combination With Metformin.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT01590771 |
| Phase | Phase 3 |
| Condition | Type 2 Diabetes Mellitus |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Triple |
| Enrollment | 498 |
| Arms in the primary outcome | 2 |
| Lead sponsor | Merck Sharp & Dohme LLC |
| Results posted | Yes |
2. Clinical Question
Population
Participants in China with type 2 diabetes mellitus whose glycemic control was inadequate on sulfonylurea therapy, either alone or combined with metformin. Participants continued pre-study gliclazide or glimepiride, with or without metformin. The primary analysis population included participants who received at least one dose of study drug and had a baseline and at least one post-randomization measurement.
Intervention
Sitagliptin 100 mg once daily for 24 weeks, added to the participant’s ongoing sulfonylurea therapy, with or without metformin. The registry describes a placebo run-in period for all participants.
Comparator
Matching placebo once daily for 24 weeks, added to the same kinds of ongoing sulfonylurea-based therapy. Participants continued their pre-study treatment during the study.
Primary question
The primary question was whether adding sitagliptin, compared with adding placebo, produced a greater reduction in A1C after 24 weeks. The registered endpoint was change from baseline, defined as the Week 24 A1C percent minus the Week 0 A1C percent. The reported comparison therefore concerns the estimated difference between groups in that change, rather than a comparison of final A1C values alone.
3. Trial Design
This was a Phase III, multicenter, randomized, double-blind, placebo-controlled, parallel-group clinical trial conducted in China. The registry describes triple masking. Enrollment was 498, and the record lists the trial as completed. The recorded start date was 2012-07-09 and the primary completion date was 2014-06-10.
The study compared sitagliptin with matching placebo when each was added to stable sulfonylurea therapy. Participants continued pre-study gliclazide or glimepiride, alone or with metformin. The treatment descriptions specify once-daily study medication for 24 weeks. The sitagliptin description specifies a dose of 100 mg; the placebo was matching placebo. The registry also describes a placebo run-in period for all participants.
Randomization and masking are design features intended to reduce the influence of participant characteristics, expectations, and study conduct on the comparison. A parallel-group design assigns participants to a study group and compares outcomes between groups over the study period. These are general features of trial design; the registry record does not report further details about the randomization procedure, allocation concealment, or how masking was maintained.
The primary endpoint was change in A1C from baseline to Week 24. The posted arm estimates use 243 participants for sitagliptin and 236 for placebo. Those endpoint-analysis counts are distinct from the enrollment total. The registry defines the primary analysis population by receipt of at least one dose, a baseline measurement, and at least one post-randomization measurement. More detailed participant-flow information, including reasons for exclusion from that analysis population, is 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
| Stage | Sitagliptin | Placebo |
|---|---|---|
| STARTED | 249 | 249 |
| COMPLETED | 230 | 210 |
| NOT COMPLETED | 19 | 39 |
Reasons for not completing this period:
| Reason | Sitagliptin | Placebo |
|---|---|---|
| Other Protocol Specified Criteria | 5 | 14 |
| Withdrawal by Subject | 8 | 15 |
| Protocol Violation | 1 | 2 |
| Lost to Follow-up | 1 | 1 |
| Lack of Efficacy | 1 | 0 |
| Adverse Event | 3 | 7 |
Analysis population for the primary outcome, as defined in the registry: Full analysis set consists of all participants who received at least one dose of study drug, have a baseline measurement, and have at least one post-randomization measurement.
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.
| Characteristic | Unit | Sitagliptin | Placebo | Total |
|---|---|---|---|---|
| Participants analysed | 249 | 249 | 498 | |
| Age, Continuous | Years | 57.5 (SD 9.5) | 56.5 (SD 9.3) | 57.0 (SD 9.4) |
| Sex: Female, Male: Female | Participants | 132 | 117 | 249 |
| Sex: Female, Male: Male | Participants | 117 | 132 | 249 |
| Hemoglobin A1c (A1C) | Percent of glycosylated hemoglobin (A1C) | 8.61 (SD 1.06) | 8.48 (SD 0.91) | 8.55 (SD 0.99) |
| 2-hour post meal glucose (2-hr PMG) | mg/dL | 296.5 (SD 68.0) | 290.2 (SD 72.4) | 293.4 (SD 70.2) |
| Fasting plasma glucose (FPG) | mg/dL | 181.5 (SD 40.9) | 179.8 (SD 40.7) | 180.6 (SD 40.8) |
6. Primary Endpoint
The registry lists the primary outcome as Change From Baseline in A1C Levels at Week 24 in Participants Receiving Sitagliptin and Sulfonylurea Alone or in Combination With Metformin, measured over Baseline and Week 24.
| Element | Registry entry |
|---|---|
| Result type | Least Squares Mean |
| Measure of spread | 95% Confidence Interval |
| Unit | Percent of glycosylated hemoglobin (A1C) |
| Endpoint classification | Continuous Endpoint |
The primary endpoint was change from baseline in A1C at Week 24, measured as percent of glycosylated hemoglobin. The registry defines the change as the A1C percent at Week 24 minus the A1C percent at Week 0. Under this convention, a negative change represents a decrease from baseline.
A1C was selected as the primary measure because the registered study question concerned glycemic control and specifically asked whether adding sitagliptin produced a greater reduction than adding placebo. Treating the outcome as a continuous measure allows the comparison to focus on estimated average changes in the two groups. This is a general statistical explanation of a continuous endpoint; it does not establish that every participant had the same change or that an average describes each individual’s outcome.
The primary outcome is not a count of participants reaching a threshold. It is a model-based mean change, reported as a least-squares mean for each arm, accompanied by a 95% confidence interval. The registry reports an adjusted comparison between the arms as well. These quantities address related but different questions: the arm estimates describe mean change within each group, while the adjusted contrast estimates the difference between groups under the reported analysis.
The endpoint timeframe is baseline and Week 24. Results for the additional registered outcomes, including fasting plasma glucose and 2-hour post-meal glucose, are not described in detail in the primary result summarized in the registry record.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
In the posted primary results, the least-squares mean change in A1C was -0.88 for sitagliptin, with a 95% confidence interval from -0.99 to -0.76. For placebo, the estimate was -0.27, with a 95% confidence interval from -0.38 to -0.15. Both estimates are negative, consistent with decreases from baseline under the registry’s definition of change.
The sponsor-reported adjusted comparison was -0.61, with a 95% confidence interval from -0.77 to -0.44 and p < 0.001. The negative contrast is in the direction of a larger decrease in the sitagliptin group relative to placebo, given the endpoint’s subtraction convention and the listed group order. The confidence interval gives a range of values compatible with the estimated contrast under the statistical model and its assumptions. It does not describe the range of individual participants’ changes.
The p-value expresses how incompatible the observed data are with a null comparison under the analysis assumptions. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of the contrast. The estimate and confidence interval are therefore important alongside the p-value: they convey both the direction and the precision of the group comparison. The registry does not report a threshold for clinical importance, so the result cannot be classified against such a threshold from this record alone.
The arm-specific confidence intervals overlap. That visual feature should not replace the reported adjusted comparison: the uncertainty for a between-group contrast is not determined by simply inspecting whether separate arm intervals overlap. The registered comparison uses ANCOVA with robust regression, and its reported interval and p-value are the relevant inferential results for the primary comparison.
The primary analysis uses the full analysis set defined in the registry, rather than every participant counted in enrollment. The arm counts for the endpoint are 243 and 236. The record does not provide a participant-flow account explaining the difference between those counts and total enrollment. It also does not describe the handling of missing endpoint measurements in enough detail to assess how alternative missing-data assumptions might affect the estimate.
The result is an estimate of the average contrast for the analysis population and study conditions. It does not show that every participant experienced a reduction, establish a particular individual outcome, or answer every safety or long-term question. The trial’s primary comparison is limited to the specified 24-week endpoint and the background therapies described in the registry.
9. Statistical Methods Explained
The primary endpoint is continuous: it is a numerical change in A1C, not a binary outcome such as whether a participant crossed a cutoff. The registry reports least-squares means for the arms and an ANCOVA comparison using robust regression. The following explanations describe general statistical ideas relevant to those methods; details beyond the method names and results recorded are not reported in the registry record.
What does a least-squares mean represent?
A least-squares mean is a model-based estimate of an average outcome. It is often used when an analysis adjusts for other variables or when the model provides a structured estimate rather than relying only on a raw arithmetic average. Here, the registry identifies the arm results as least-squares means and reports their confidence intervals. The adjustment variables used to produce those means are not reported in the registry record, so the estimates should not be described as adjusted for any particular baseline characteristic.
The arm estimates are -0.88 for sitagliptin and -0.27 for placebo. Each is a mean change from baseline in that arm, not the outcome of a typical or specifically identified participant. Variation among participants is not represented by the mean itself. The confidence interval quantifies uncertainty around the estimated mean under the model; it is not a range within which most individual outcomes must fall.
What does ANCOVA do in a group comparison?
ANCOVA is a regression-based method for comparing an outcome between groups while accounting for covariates included in the model. In general, adjustment can improve precision or address imbalances in variables that are included and measured. The registry names ANCOVA but does not identify the covariates, model specification, or whether baseline A1C was included as a covariate. Those details should not be inferred from the method label alone.
The reported contrast is -0.61, with a 95% confidence interval from -0.77 to -0.44. This is the sponsor’s adjusted estimate for the between-group comparison. It is not necessary for that adjusted contrast to be reconstructed by subtracting the displayed arm estimates: adjusted estimates can depend on model terms and the analysis population. No recalculated contrast is reported in the record summarized for this analysis.
Why use robust regression?
Robust regression is a modeling approach designed to reduce sensitivity to some departures from standard regression assumptions or to influential observations, depending on the specific method used. The exact robust procedure is not described in the registry record. Robustness is not a guarantee that a model is correct, nor does the label show which sources of misspecification or unusual observations were addressed. The registry’s method description should therefore be read as identifying the general analysis family, not as a complete reproducible analysis plan.
For a continuous outcome, variability can differ across participants or groups, and the distribution of individual changes may not perfectly follow a simple model. In general, robust methods can be useful when conventional model-based uncertainty estimates are sensitive to such features. Whether that concern applied in this trial, and how it was assessed, is not reported in the registry record.
How should the confidence interval and p-value be read?
A confidence interval expresses uncertainty about an estimated quantity under the statistical model. The reported interval for the contrast, -0.77 to -0.44, is compatible with a negative difference throughout the interval. Its width communicates precision in the context of the chosen model, while the point estimate -0.61 is the single best estimate reported by that analysis. The interval is not a probability statement about a fixed parameter and is not a prediction interval for an individual participant.
The reported p-value is p < 0.001. In general, a p-value assesses the compatibility of data with a specified null hypothesis, conditional on the model and analysis assumptions. It does not establish the probability that the result is due to chance, and it does not measure clinical importance. The registry does not report a prespecified decision threshold or a clinical-importance threshold for interpreting the endpoint.
What do the planning illustrations mean?
Any sample-size illustrations based on the observed effect are planning calculations, not trial results. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the arms, and a normal approximation, with the target effect expressed as a fraction of the observed effect. Such illustrations show how a hypothetical study might be planned under those assumptions. They are not observed power for this completed trial, do not establish the trial’s original design rationale, and do not reveal the sample-size justification used by the investigators. That original justification is not reported in the registry record.
10. Analyses Reported by the Sponsor
| Method | Estimate | 95% CI | P-value |
|---|---|---|---|
| ANCOVA | Robust regression -0.61 | -0.77 to -0.44 | < 0.001 |
The sponsor’s reported primary analysis is an ANCOVA using robust regression. It gives an adjusted sitagliptin-versus-placebo contrast of -0.61, with a 95% confidence interval from -0.77 to -0.44 and p < 0.001. The arm estimates are least-squares mean changes, reported as -0.88 for sitagliptin and -0.27 for placebo. Their confidence intervals are -0.99 to -0.76 and -0.38 to -0.15, respectively.
The registry record does not report a separate recalculated comparison from the arm-level summary statistics. Consequently, there is no recalculated estimate to set against the sponsor’s model-based result. The sponsor’s own analysis takes precedence for describing the inferential result because it is the analysis explicitly reported for the primary comparison. A simple unadjusted calculation, if available, would answer a less model-adjusted question and would not replace the registered analysis.
The registry does not provide the covariates or full model specification needed to reproduce the ANCOVA result independently. The sponsor’s contrast, interval, and p-value can be reported as posted, while the limits on independently verifying the model details should remain clear.
11. Multiplicity, Missing Data and Analysis Populations
The registry lists 11 outcome measures and 8 posted secondary outcomes. The secondary outcomes include changes in 2-hour post-meal glucose, fasting plasma glucose, and A1C, with some outcomes specified for participants receiving sulfonylurea alone and others for sulfonylurea in combination with metformin. The primary result is the overall registered A1C comparison at Week 24.
When a study examines multiple outcomes or makes multiple comparisons, the chance of at least one apparently notable result can increase if every test is interpreted independently. General methods for addressing this include prespecifying a testing sequence, adjusting significance levels, or treating secondary analyses as exploratory. The registry record does not report whether a multiplicity procedure was planned or used, nor does it describe a hierarchy for testing the secondary outcomes. The reported primary p-value should therefore be read as the result of the stated primary analysis, without assuming an adjustment across outcomes.
The primary analysis population is defined as participants who received at least one dose of study drug, had a baseline measurement, and had at least one post-randomization measurement. This definition identifies the population included in the analysis, but it does not explain how missing Week 24 measurements were handled. The registry record does not report the amount or pattern of missing primary endpoint data, the imputation method, or sensitivity analyses under alternative assumptions.
The lack of these details limits assessment of how robust the result is to missing outcome data and to the choice of analysis population. It also prevents a full reconstruction of the analysis from the registry record alone. Whether interim analyses were conducted, and whether any statistical plan addressed them, is not reported in the registry record.
12. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Sitagliptin | 7 | 248 | 2.8% |
| Placebo | 7 | 249 | 2.8% |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 5% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Sitagliptin | 54 of 248 |
| Placebo | 40 of 249 |
Most frequent serious adverse event terms
| Event term | Organ system | Sitagliptin | Placebo |
|---|---|---|---|
| Angina pectoris | Cardiac disorders | 1 of 248 | 0 of 249 |
| Erysipelas | Infections and infestations | 1 of 248 | 0 of 249 |
| Lung infection | Infections and infestations | 1 of 248 | 0 of 249 |
| Mastitis | Infections and infestations | 0 of 248 | 1 of 249 |
| Pneumonia | Infections and infestations | 0 of 248 | 1 of 249 |
| Blood glucose increased | Investigations | 1 of 248 | 0 of 249 |
| Diabetes mellitus | Metabolism and nutrition disorders | 0 of 248 | 1 of 249 |
| Hypokalaemia | Metabolism and nutrition disorders | 0 of 248 | 1 of 249 |
| Rhabdomyolysis | Musculoskeletal and connective tissue disorders | 0 of 248 | 1 of 249 |
| Cervix carcinoma | Neoplasms benign, malignant and unspecified (incl cysts and polyps) | 1 of 248 | 0 of 249 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Sitagliptin | Placebo |
|---|---|---|---|
| Hypoglycaemia | Metabolism and nutrition disorders | 35 of 248 | 17 of 249 |
| Urinary tract infection | Infections and infestations | 14 of 248 | 11 of 249 |
| Hyperlipidaemia | Metabolism and nutrition disorders | 8 of 248 | 15 of 249 |
Serious events with the highest arm-level rate: angina pectoris (0.4%), erysipelas (0.4%), lung infection (0.4%), blood glucose increased (0.4%), cervix carcinoma (0.4%).
The safety figures use different denominators from the primary endpoint analysis. For serious adverse events, the registry lists 248 participants at risk in the sitagliptin group and 249 in the placebo group; 7 participants in each group were affected. The record separately lists other adverse events for the same at-risk counts, with 54 affected in the sitagliptin group and 40 in the placebo group.
These are participant counts, not counts of every event occurrence. They describe what was recorded for the defined safety groups and do not by themselves establish whether a treatment caused an event. The registry lists Angina pectoris, Erysipelas, Lung infection, Blood glucose increased, Cervix carcinoma, and Asthma among the top serious-event terms, each at 0.4%. The listed terms are not assigned to a specific arm in the record, so they should not be interpreted as arm-specific comparisons.
Deaths are not reported in the registry record. The safety figures are descriptive and do not provide the detail needed to assess event timing, duration, severity beyond the serious-event classification, or causal attribution. They should be read separately from the primary efficacy analysis and within the limits of the registry’s reporting.
13. What the Result Does, and Does Not, Mean
The posted primary analysis estimates a more negative mean change in A1C for sitagliptin than for placebo over the registered 24-week timeframe. The sponsor-reported adjusted contrast is -0.61, with a 95% confidence interval from -0.77 to -0.44 and p < 0.001. The estimate indicates the direction and size of the average between-group contrast under the reported model; the interval communicates its statistical precision.
This result does not mean that every participant’s A1C decreased, or that every participant had the same response. Group averages can coexist with substantial individual variation. Nor does the p-value establish the probability that the treatment difference is real or clinically important. The registry record does not state a threshold for clinical importance, so statistical evidence and clinical importance cannot be equated from the posted result alone.
The result concerns adding sitagliptin or placebo to the stable sulfonylurea-based regimens described in the trial, in participants meeting the study’s population criteria. It does not directly answer questions about other background treatments, populations, endpoints, or timeframes. The primary endpoint is change in A1C at Week 24; it is not a measure of long-term outcomes.
The analysis population excludes anyone not meeting the recorded full analysis set definition, and the registry’s reported arm counts differ from total enrollment. The record does not describe missing-data handling in sufficient detail to evaluate sensitivity to assumptions about unobserved Week 24 outcomes. Additional registered outcomes and safety information may inform other questions, but they do not change the definition of the primary comparison.
These results are statistical findings from a clinical trial record, not individualized guidance. They describe the posted group-level estimates and their uncertainty, and should not be extended beyond the study conditions without further evidence.
14. Why This Trial Matters Statistically
This trial illustrates the distinction between an outcome measured within each group and the comparison that addresses the randomized study question. The sitagliptin and placebo arms each have a least-squares mean change, but the primary inferential target is the adjusted difference between those changes. The sponsor-reported contrast and its confidence interval are therefore central to interpreting the primary result.
It also shows why endpoint definitions matter. Change is defined as Week 24 minus Week 0, so decreases appear as negative values. Without attending to that convention, the direction of a negative estimate could be misread. The continuous endpoint and the model-based estimates also make clear that an average is not a statement about every participant.
The registry names ANCOVA with robust regression but does not report the full model specification. This is a useful reminder that method labels do not make an analysis fully reproducible: covariates, missing-data handling, and sensitivity analyses can affect interpretation, and those details are not reported in the registry record.
Finally, efficacy and safety results use different analysis counts, and the primary outcome result does not answer the safety question. Reading each set of figures with its own population and denominator helps prevent inappropriate direct comparisons. Together, these features make the record a practical example of interpreting an adjusted estimate, a confidence interval, an analysis population, and descriptive safety counts without claiming more than the registry reports.
15. 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 describe the randomization sequence, allocation concealment, or practical procedures used to maintain triple masking. It identifies the primary analysis method as ANCOVA with robust regression but does not specify covariates, model terms, or the robust-regression procedure. These omissions limit independent reproduction of the reported contrast.
The primary analysis population is defined by treatment receipt and availability of baseline and post-randomization measurements, but participant flow and reasons for missing data are not reported in the registry record. The method for handling missing endpoint observations, the extent of missingness, and sensitivity analyses are also not reported. It is therefore not possible to assess from this record whether the reported estimate changes under alternative missing-data assumptions.
The registry does not report a multiplicity strategy for the listed outcomes, an interim-analysis plan, or the original sample-size justification. Planning illustrations based on the observed effect do not establish what assumptions informed the completed trial’s design. The record also does not state a clinical-importance threshold for the primary endpoint.
The primary result applies to the specified Week 24 A1C change and the full analysis set described in the registry. It should not be generalized automatically to other populations, regimens, outcomes, or durations. The safety information is presented as counts and selected serious-event terms; deaths are not reported in the registry record, and the available summary does not detail event timing or causal attribution. These reporting limits are specific to what can be assessed from the registry record and do not establish that unreported procedures or analyses were absent.
16. Related Tutorials
17. Related Statistical Calculators
18. Sources
- ClinicalTrials.gov: A Study in China Evaluating the Safety and Efficacy of Adding Sitagliptin to Stable Therapy With Sulfonylurea With or Without Metformin in Participants With Type 2 Diabetes Mellitus (T2DM) (MK-0431-253) — NCT01590771.
- Registry record last updated 2018-08-17; status last verified 2018-07.
- Results first posted 2015-05-04.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
19. Record Summary
NCT01590771 is a completed Phase III randomized, double-blind, placebo-controlled, parallel-group trial in China, with triple masking recorded. Enrollment was 498. The study compared sitagliptin 100 mg once daily with matching placebo for 24 weeks, added to stable sulfonylurea therapy with or without metformin.
The primary endpoint was change in A1C from baseline to Week 24. The sponsor reported an ANCOVA robust-regression contrast of -0.61, with a 95% confidence interval from -0.77 to -0.44 and p < 0.001. The registry defines the primary analysis population and provides arm-level estimates, but does not report full model details, missing-data methods, multiplicity handling, or the original sample-size justification.
Serious adverse events affected 7 participants in each group. The registry record does not report deaths.