A Study Assessing Saxagliptin Treatment in Subjects With Type 2 Diabetes Who Are Not Controlled With Diet and Exercise: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study Assessing Saxagliptin Treatment in Subjects With Type 2 Diabetes Who Are Not Controlled With Diet and Exercise (NCT00374907), a randomized phase 3 study. The primary outcome is Insulin Secretion Rate Area Under the Curve (AUC) During Intravenous (IV)-Oral Hyperglycemic Clamp - Percent Change From Baseline at Week 12.

2. Clinical Question

The result addresses a physiological measure of insulin secretion, rather than a direct summary of blood sugar control or a clinical outcome. The registry’s description frames the study as examining whether saxagliptin could improve the body’s ability to make insulin and lower blood sugar, but the primary endpoint reported is the insulin secretion rate measure. The posted result therefore answers a narrower question: how that measured rate changed from baseline by Week 12 in the analyzed participants.

The record describes a randomized, double-masked, parallel phase 3 study and lists an overall enrollment of 156. The arm-level primary outcome estimates are based on 16 participants in the saxagliptin group and 15 in the placebo group. These analysis counts should not be confused with overall enrollment. Why these outcome counts differ from enrollment, or whether other participants lacked eligible measurements, is not reported in the registry record.

3. Trial Design

This was a randomized, double-masked, parallel phase 3 trial sponsored by AstraZeneca. The record lists type 2 diabetes as the condition and describes the study as enrolling people not controlled with diet and exercise. It reports enrollment of 156 and a completed study. The recorded start was 2006-09 and primary completion was 2008-01.

The treatment descriptions include saxagliptin 5 mg and placebo in the short-term period. The record also lists metformin, both blinded and open-label. Its arm descriptions refer to open-label metformin 500-1500 mg as needed for rescue in the long-term period, and identify placebo in the short-term period followed by metformin in the long-term period for the comparator arm. The primary outcome result presented is for the short-term comparison at Week 12. The registry record does not give enough detail to reconstruct the full sequence, timing, or use of rescue treatment for individual participants.

Random allocation is intended to make treatment groups comparable on average at the start of a trial, while masking is intended to limit the influence of treatment knowledge on care, measurement, and assessment. These are general design principles; the registry entry does not describe the exact allocation procedure or masking procedures. A parallel design assigns participants to groups that are followed in their assigned treatment groups rather than describing a crossover comparison.

The primary outcome population is described as randomized participants with both baseline and post-baseline values up to Week 12. The posted arm estimates list 16 participants for saxagliptin and 15 for placebo. These are the numbers attached to the primary outcome estimates, not a replacement for the enrollment figure. The registry record does not explain the participant flow, the reasons measurements may be unavailable, or whether the analyzed population followed any additional analysis rules.

There is one listed secondary outcome, insulin secretion rate AUC during an IV hyperglycemic clamp as percent change from baseline at Week 12. The record reports four outcome measures and two analyses overall, but it does not provide a complete statistical analysis plan. Details such as allocation concealment, prespecified adjustment covariates, analysis of protocol deviations, and interim analyses 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.

Before assignment: 156 participants were enrolled in the study; 110 participants failed screening; 10 subjects entered lead-in and discontinued prior to randomization.

12-Week Short-term Period

Reasons for not completing this period:

116-Week Short-term + Long-term Period

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Randomized participants with both a baseline and post-baseline value (up to Week 12).

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 Insulin Secretion Rate Area Under the Curve (AUC) During Intravenous (IV)-Oral Hyperglycemic Clamp - Percent Change From Baseline at Week 12, measured over Baseline, Week 12.

The primary endpoint is the adjusted percent change from baseline at Week 12 in insulin secretion rate area under the curve during an intravenous-oral hyperglycemic clamp. The clamp combines an intravenous procedure with an enteral glucose load; the described measurement window is 180-480 minutes. C-peptide deconvolution was used to estimate insulin secretion rate. These are the measurement features stated in the registry record.

An area under the curve represents accumulated measured activity over a defined interval, rather than a single reading at one moment. In this setting, the endpoint summarizes insulin secretion rate during the clamp procedure. Percent change expresses the Week 12 measure relative to baseline. The registry reports group-level geometric means for the arm estimates and an adjusted percent difference for the comparison.

As a physiological endpoint, this measure can describe a biological response under the conditions of the clamp. It is not itself a direct measure of a participant’s everyday blood glucose, long-term health, or symptoms. The registry’s description connects the research question to the body’s ability to produce insulin and blood sugar, but the primary endpoint reported is specifically the clamp-based insulin secretion rate measure.

The endpoint is useful for investigating a mechanism of action because it focuses on insulin secretion under a controlled glucose challenge. That general rationale does not establish that the measurement is a validated substitute for a clinical outcome, nor does it show how the result relates to outcomes not measured in this record. Such validation or linkage is not reported in the registry record. Interpretation should therefore stay within the endpoint’s defined scope: the posted change in a clamp-derived physiological measure through Week 12.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry reports an ANCOVA analysis for the comparison of saxagliptin and placebo. ANCOVA is a general statistical approach for comparing an outcome between groups while adjusting for one or more covariates. It can account for baseline measurements or other prespecified characteristics, depending on the model. The registry record does not identify the covariates, model form, or diagnostics, so those details should not be inferred.

What does the adjusted comparison represent?

The reported adjusted percent difference is 18.5. Its 95% confidence interval is 1.3 to 38.7, and the reported p-value is 0.035. This comparison is distinct from the two arm-level geometric mean adjusted percent changes, which are 15.9 for saxagliptin and -2.2 for placebo. The comparison comes from the registry-reported ANCOVA analysis; it should not be reconstructed by subtracting the arm estimates because the reported quantities are not stated to be the same unadjusted scale or calculation.

In general, adjustment aims to compare groups after accounting for modeled covariate information. The meaning of the adjusted estimate depends on the variables included and the model specification. Since these details are not reported in the registry record, readers can interpret the reported contrast and its uncertainty but cannot independently assess which covariates shaped it from the posted summary.

How should the confidence interval be read?

A confidence interval expresses sampling uncertainty around an estimated effect under the model and its assumptions. The interval from 1.3 to 38.7 is compatible with a range of positive adjusted differences, rather than identifying a single known value. It does not describe how much participants vary individually, and it is not a probability statement about where the fixed underlying effect lies. Wider intervals generally indicate less precision than narrower intervals, all else equal; this is a general statistical principle.

The arm-level intervals also communicate uncertainty: 4.2 to 29.0 for saxagliptin and -12.4 to 9.3 for placebo. Those intervals are reported around the group estimates, not as individual response ranges. The analysis population and observed group sizes matter for precision, but the posted summary does not provide enough information to attribute the interval width to any particular source.

What does the p-value say?

In general, a p-value measures how incompatible the observed data are with a specified null model, assuming that model and the analysis assumptions hold. The reported p-value of 0.035 is not the chance that the result occurred randomly, the chance that the null is true, or a measure of effect size. It does not by itself establish whether an effect is important in clinical practice. The registry record does not state whether the analysis was designated as confirmatory or how the result fits into a complete prespecified testing strategy.

Why do analysis populations matter?

The primary population is defined as randomized participants with a baseline and post-baseline value up to Week 12. This is a measurement-availability definition in the registry description. It does not tell readers how participants with incomplete measurements were handled, whether observations were imputed, or whether sensitivity analyses were conducted. Those missing-data methods are not reported in the registry record.

The arm counts attached to the primary estimates are 16 and 15, while the overall enrollment is 156. These numbers describe different aspects of the trial and should not be conflated. Without a participant flow or explanation of analysis eligibility, the reasons for the difference are unknown. The registry record also does not report whether the primary analysis retained all randomized participants in their assigned groups or used other exclusions.

What can planning calculations show?

Sample-size illustrations are planning exercises based on assumptions, not observed results and not a measure of the trial’s observed power. The stated assumptions for such illustrations are a two-sided significance level of 0.05, equal allocation, a normal approximation, and an effect set to a stated fraction of the observed effect. For non-time-to-event outcomes, the planning quantity is participants per arm for 80% power or another stated target. These assumptions do not establish that the original trial was designed using the same effect, allocation, or calculation. The trial’s own sample-size justification is not reported in the registry record.

More generally, a planning estimate depends on the chosen effect size, variability, significance threshold, allocation, and model. A smaller target effect typically requires more information to detect with comparable power, but that is a general relationship, not a new calculation for this trial. The planning assumptions should not be mistaken for the protocol’s rationale or a guarantee of success in a future study.

What remains unknown about the statistical plan?

The registry record does not report the full analysis plan, covariate list, model diagnostics, treatment of missing values, multiplicity procedure, interim analyses, or sensitivity analyses. These omissions limit independent assessment of robustness and of how the primary analysis was selected. The reported ANCOVA result remains the sponsor’s posted analysis, while the registry summary alone does not support a reconstruction of every modeling decision.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s comparison as an ANCOVA adjusted percent difference of 18.5, with a 95% confidence interval from 1.3 to 38.7 and p = 0.035. It also posts separate arm-level geometric mean adjusted percent changes: 15.9 for saxagliptin and -2.2 for placebo. The adjusted comparison is the sponsor-reported inferential analysis, while the arm estimates describe each group. They are related summaries but are not interchangeable calculations.

The registry record does not provide the model covariates or enough detail to reproduce the ANCOVA estimate from the arm summaries. No independent recalculated comparison is stated in the registry record. A difference that might be derived informally from displayed group summaries would not replace the sponsor’s adjusted model result. The posted comparison should therefore be read as reported rather than treated as a directly reproducible arithmetic contrast.

The record does not flag a discrepancy between the reported analysis and any recalculation. This does not fill in the missing model details or establish that all analyses agree; it simply leaves the sponsor-reported estimate as the principal comparison available in the record. Its confidence interval and p-value describe statistical uncertainty under the reported analysis, not clinical importance or effects on outcomes beyond the primary endpoint.

11. Multiplicity, Missing Data and Analysis Populations

The primary outcome population is described as randomized participants with both a baseline and post-baseline measurement up to Week 12. The registry record does not explain how participants without both measurements were handled, why their values were unavailable, or whether any imputation or sensitivity analysis was used. It also does not provide a participant-flow account linking overall enrollment to the endpoint counts. The arm-level outcome estimates are based on 16 participants for saxagliptin and 15 for placebo, but the reasons for those analysis counts are not reported.

The distinction between enrollment and analysis population is important. Enrollment describes the trial’s overall recorded size; an outcome estimate may include only people meeting the endpoint-specific availability rule. The record does not establish whether missing measurements were related to treatment, participant characteristics, or the endpoint itself. Without that information, the direction or magnitude of any possible missing-data bias cannot be assessed.

One secondary outcome is listed, while the record reports four outcome measures and two analyses overall. The full outcome and analysis structure is not available from these counts alone. The registry record does not report whether the primary result was protected against multiplicity, whether secondary outcomes were tested in a hierarchy, or whether any adjustment to significance thresholds was applied. Multiple testing can increase the chance of obtaining at least one apparently notable result when many hypotheses are examined; this is a general statistical concern, not a claim about what happened in this trial.

The reported p-value of 0.035 should therefore be interpreted as the p-value for the posted comparison, without assuming that it is multiplicity-adjusted or that it accounts for every analysis conducted. The record also does not state whether interim analyses were performed or whether any stopping rules were used. Those aspects of the statistical plan are not reported in the registry record, so their influence on interpretation cannot be evaluated.

12. 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: angina unstable (6.2%), cholecystitis (6.2%), colon cancer (6.2%), chest discomfort (5.0%).

The registry lists serious adverse event counts for placebo/metformin and saxagliptin 5 mg groups. For placebo/metformin, 2 of 16 were recorded as affected by serious adverse events; for saxagliptin 5 mg, 1 of 20 were recorded as affected. These are the counts posted for the safety groups and should be read separately from the smaller primary-endpoint analysis groups. The registry does not explain the difference between the safety and primary-outcome denominators.

The record also lists other adverse events for 14 of 16 in the placebo/metformin group and 17 of 20 in the saxagliptin 5 mg group. These counts are not the same category as serious adverse events. The record lists serious-event terms with percentages, including angina unstable, cholecystitis, colon cancer, and chest discomfort. It does not make clear in the posted registry entries which group or denominator applies to each listed term, so those percentages should not be assigned to a treatment group without further information.

Deaths are not reported in the registry record. The posted counts alone do not establish comparative safety: interpretation would require context on event definitions, timing, exposure, completeness of collection, and the trial’s safety analysis plan. These details are not reported in the registry record. The figures describe events recorded in the trial, not advice about treatment.

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

The result supports a limited statistical statement: the registry’s ANCOVA comparison for adjusted percent change in clamp-derived insulin secretion rate at Week 12 is positive, with an estimate of 18.5 and a 95% confidence interval from 1.3 to 38.7. The reported p-value is 0.035. The interval and p-value quantify aspects of uncertainty under the reported model; neither conveys the full range of possible clinical consequences.

The result does not show that every participant’s insulin secretion increased, nor does it describe the distribution of individual responses. Group estimates summarize participants and can coexist with substantial variation within groups. The registry does not post individual-level data, so variation and the influence of unusual observations cannot be assessed from these summaries.

The endpoint is a physiological measurement during an intravenous-oral hyperglycemic clamp. It is not itself a measure of routine blood glucose, long-term diabetes outcomes, symptoms, or quality of life. The study description mentions insulin production and blood sugar, but the primary result directly addresses only the stated clamp-based endpoint. Links between this measure and other clinical outcomes are not reported in the registry record.

The analysis also has defined population and timing boundaries: participants with baseline and post-baseline measurements up to Week 12, in the short-term comparison. The record does not provide enough information to infer longer-term effects from this result. Metformin is described in relation to the long-term period and rescue use, but the posted primary comparison should not be read as a long-term comparison.

Finally, statistical evidence of a positive difference is not a decision rule about treatment. The registry result alone does not establish the importance of the size of the effect, the balance of benefits and harms, or how the finding would apply to people outside the analyzed population. Those questions require evidence beyond this endpoint estimate and the limited safety summaries in the record.

14. Why This Trial Matters Statistically

This trial is statistically informative because its primary outcome is a mechanistic, physiological measure rather than a broad clinical endpoint. It illustrates why the exact outcome definition matters: a clamp-based insulin secretion rate AUC measures a specific response under a controlled procedure, and its interpretation should not be silently extended to different outcomes.

The record also illustrates the distinction between arm summaries and an adjusted group comparison. The posted geometric mean percent changes provide a description of each arm, while the ANCOVA result is the reported comparison between groups. These quantities carry different statistical roles. An adjusted contrast cannot necessarily be recovered by applying simple arithmetic to separate arm estimates, particularly when the model and scale are not fully described.

Uncertainty is central to reading the result. The 95% confidence interval for the adjusted comparison ranges from 1.3 to 38.7. That interval communicates that the estimate is imprecise enough to include a range of possible positive values under the analysis assumptions. A p-value gives another summary of evidence against a specified null model, but it does not replace interval estimates or establish clinical relevance.

The difference between enrollment and endpoint analysis counts highlights the importance of defining who contributed to each result. The primary population requires both baseline and post-baseline measurements, while the overall trial enrollment is 156. Without participant flow and missing-data details, readers cannot assess how the analyzed group relates to all enrolled participants or how missing measurements may affect the estimate.

Finally, the registry record makes clear that a statistical result sits within a wider evidential context. The available summary does not establish multiplicity handling, model covariates, or detailed safety methods. Recognizing those unknowns is part of sound interpretation: it keeps conclusions aligned with the outcome, population, period, and analysis that are actually reported.

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 available record gives a concise outcome summary but not the complete protocol or statistical analysis plan. It does not state the covariates used in the ANCOVA, how the model was specified, or whether its assumptions were assessed. Without these details, the adjusted estimate cannot be independently reconstructed from the registry summary.

The primary population requires a baseline and post-baseline measurement up to Week 12, while the endpoint estimates are attached to 16 saxagliptin participants and 15 placebo participants. The overall enrollment is 156. No participant flow or explanation of the difference is reported. The record does not specify how missing measurements were handled or describe sensitivity analyses, making it difficult to judge the robustness of the result to missing data.

The outcome is a clamp-based physiological measure at Week 12. The record does not establish its relationship to longer-term clinical outcomes or provide a longer-term primary comparison. Although metformin is included in the intervention descriptions for the long-term period and as rescue treatment, the posted primary result is the short-term comparison and should not be generalized to a different treatment period.

Multiplicity procedures, interim analyses, and the trial’s sample-size justification are not reported in the registry record. The presence of multiple outcome measures and analyses makes the missing multiplicity information relevant to interpretation of the reported p-value, but the record does not establish whether adjustment was or was not performed.

Safety information is also limited: the registry posts serious and other adverse event counts, but deaths are not reported and the listed serious-event terms do not have a clearly specified group denominator in the record details. Event definitions, collection periods, and exposure context are not described in the summary. These limitations constrain conclusions about both the primary estimate and the wider trial results.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study Assessing Saxagliptin Treatment in Subjects With Type 2 Diabetes Who Are Not Controlled With Diet and Exercise — NCT00374907.
• Registry record last updated 2015-05-07; status last verified 2015-03.
• Results first posted 2011-01-21.
• 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.

19. Record Summary

NCT00374907 is a completed randomized, double-masked, parallel phase 3 study sponsored by AstraZeneca. It lists type 2 diabetes as the condition and records enrollment of 156. The primary result concerns adjusted percent change from baseline at Week 12 in insulin secretion rate AUC during an intravenous-oral hyperglycemic clamp.

The arm-level geometric mean estimates are 15.9 for saxagliptin 5 mg and -2.2 for placebo. The sponsor-reported ANCOVA comparison is an adjusted percent difference of 18.5, with a 95% confidence interval from 1.3 to 38.7 and p = 0.035. The outcome population required baseline and post-baseline measurements up to Week 12.

The registry also lists one secondary outcome and safety counts. Important details—including missing-data handling, multiplicity procedures, analysis covariates, and the full statistical plan—are not reported in the registry record. Interpretation is therefore best limited to the posted Week 12 physiological endpoint and its reported adjusted comparison.