Effect of Detemir and Sitagliptin on Blood Glucose Control in Type 2 Diabetes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Effect of Detemir and Sitagliptin on Blood Glucose Control in Type 2 Diabetes (NCT00789191), a randomized phase 3 study. The primary outcome is HbA1c (Glycosylated Haemoglobin A1c).

2. Clinical Question

The question is about the average HbA1c result at a specified follow-up, not about whether every participant reached a glycaemic target. The trial also recorded secondary outcomes that include target attainment, target attainment without symptomatic hypoglycaemia, body measures, fasting plasma glucose and hypoglycaemic episodes. Those outcomes address related but distinct aspects of glycaemic control and treatment experience.

The comparison should be understood as a contrast between treatment strategies rather than as a test of detemir in isolation. Sitagliptin was part of both strategies, while the stated background treatment differed: metformin for the combination group, and metformin and/or sulphonylurea for the comparator group. The registry record does not provide further detail about how background treatment was managed during follow-up.

Random allocation supports a between-group comparison, but the record does not give enough detail to assess all features of implementation or analysis. The posted outcome is a mean at Week 26. It does not, by itself, describe the distribution of individual responses, the course of HbA1c throughout follow-up, or the balance between benefits and harms across all outcomes.

3. Trial Design

This was a completed phase 3, randomized, parallel-group trial with no masking reported. The registry record gives an enrollment of 222 and describes the study as taking place in Asia, Europe and North America. The recorded start is 2008-11, and primary completion is recorded as 2009-08.

One group received insulin detemir once daily plus sitagliptin added to participants’ own pre-trial metformin treatment. The other received sitagliptin once daily added to participants’ own pre-trial metformin and/or sulphonylurea treatment. Both strategies therefore included sitagliptin, but they were not identical apart from the addition of insulin detemir: the background treatment descriptions also differ. The registry record does not report further detail about treatment adjustments, adherence, or how background medication differences were addressed in the analysis.

The primary endpoint was mean HbA1c at Week 26. The posted arm summaries include 103 participants for the combination group and 106 for the sitagliptin group for this outcome. Counts differ for some secondary outcomes, so the outcome-specific denominators should be read separately rather than treated as a single analysis population for every result.

The record identifies a full analysis set as randomized participants exposed to at least one dose of trial product, but the remaining wording of its definition is incomplete in the registry record. The details of any additional analysis populations, handling of missing observations, and deviations from the intended treatment assignment are not reported in the registry record. Likewise, the original rationale for the trial’s enrollment target beyond the stated HbA1c sample-size calculation is not reported in the registry record.

Novo Nordisk A/S is listed as sponsor. The design makes the primary result an open-label, randomized comparison of the assigned treatment strategies. Randomization is relevant to interpreting between-group differences, while lack of masking may matter for outcomes affected by participant or investigator expectations, particularly subjective outcomes.

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: Between screening and randomisation, eligible subjects were to continue their usual pre-trial oral anti-diabetic drug (OAD) dose and dosing frequency. Subjects on sulphonylurea (SU) treatment randomised to the insulin detemir group had their SU discontinued

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Sample size calculation: Assuming a standard deviation of 1.0 for HbA1c, 100 subjects in each treatment arm would be required to obtain a power of 80% for detecting a HbA1c difference of 0.4%. FAS (Full Analysis Set) is all randomised subjects exposed to at least one dose of trial product with a post baseline observation.

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 HbA1c (Glycosylated Haemoglobin A1c), measured over Week 26.

The primary endpoint is HbA1c, or glycosylated haemoglobin, measured as a percentage at Week 26. The posted statistic is the group mean. HbA1c is used as a measure of glycaemic control because it reflects glucose exposure over time rather than only the glucose value at a single measurement. For a trial comparing glucose-lowering treatment strategies, a mean HbA1c outcome offers a continuous measure that can capture differences across the full range of observed values.

A group mean is useful for comparing average outcomes, but it does not show how many individuals achieved a particular target or how widely results varied among participants. The registry lists separate secondary outcomes for achieving HbA1c thresholds, including outcomes that also require no symptomatic hypoglycaemia. Those responder outcomes complement the continuous primary endpoint, but they answer different questions.

The primary endpoint is reported at Week 26. It therefore describes the recorded outcome at that time point, not the full trajectory of HbA1c during the trial or longer-term outcomes. The registry record does not report a detailed measurement schedule or additional endpoint definition beyond the stated measure and timeframe. Interpretation should stay with the mean difference and its uncertainty, while recognizing that the endpoint alone does not summarize all aspects of glycaemic control or safety.

7. Results Posted in the Registry

• Sita: mean 7.64 versus 7.08 on control; difference 0.56 (95% CI 0.31 to 0.81), standardized effect d = 0.61; z-test p = <0.001.

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 in BMI (Body Mass Index)

• 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 errors.
• Result:
  • Sita: mean -0.58 versus -0.30 on control; difference -0.28 (95% CI -0.68 to 0.12), standardized effect d = -0.19; z-test p = 0.172.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Sita and the control. The interval gives the range of effects that remain compatible with the data.

Change in Body Weight

• 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 errors.
• Result:
  • Sita: mean -1.66 versus -0.81 on control; difference -0.85 (95% CI -2.04 to 0.34), standardized effect d = -0.19; z-test p = 0.162.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Sita and the control. The interval gives the range of effects that remain compatible with the data.

FPG (Fasting Plasma Glucose)

• 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 errors.
• Result:
  • Sita: mean 8.52 versus 6.08 on control; difference 2.44 (95% CI 1.79 to 3.09), standardized effect d = 1.02; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Sita than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes broaden the picture beyond mean HbA1c. Their overall pattern is mixed: fasting plasma glucose follows the direction of the primary endpoint, while the reported adjusted body-measure comparisons are less conclusive. This contrast is plausible because different outcomes measure different aspects of treatment response and have different variability. Secondary findings should complement the primary result rather than be treated as interchangeable measures of the same effect.

The registry lists outcomes for HbA1c target attainment, target attainment without symptomatic hypoglycaemia, body mass index, body weight, fasting plasma glucose, and hypoglycaemic episodes. It reports 11 secondary outcomes, while numerical results in the record include three continuous secondary outcomes. Numerical results for the other listed secondary outcomes are not reported in the registry record. Their absence means the listed outcomes cannot all be assessed from the posted numerical results.

Secondary analyses generally carry less weight than the primary endpoint because they may be numerous and because the registry record does not report a multiplicity procedure or a hierarchy for interpreting them. Without such information, nominal evidence from an individual secondary comparison can overstate how persuasive it is when considered among many outcomes. The record also does not give enough detail to establish which secondary outcomes were prespecified for formal testing or how missing measurements were handled.

The secondary results can provide context and suggest whether different measures tell a consistent story, but they do not replace the primary comparison. In particular, outcomes that are not numerically reported cannot be assumed to agree with or contradict the HbA1c result. A balanced interpretation recognizes the supportive pattern for fasting plasma glucose, the less decisive body-measure analyses, and the incomplete numerical reporting for other listed outcomes.

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

• Number of Subjects Achieving HbA1c Less Than or Equal to 7.0%
• Number of Subjects Achieving HbA1c Less Than or Equal to 7.0% Without Symptomatic Hypoglycaemia
• Number of Subjects Achieving HbA1c Less Than or Equal to 6.5%
• Number of Subjects Achieving HbA1c Less Than or Equal to 6.5% Without Symptomatic Hypoglycaemia
• Change in BMI (Body Mass Index)
• Change in Body Weight
• FPG (Fasting Plasma Glucose)
• Hypoglycemic Episodes

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous, so the posted arm summaries are means with standard errors. A mean describes the central value of a group; a standard error describes the precision of that estimated mean. It is not the same as the spread of individual participant measurements. The standard errors for the two primary arm means are both 0.09, and the posted sample counts are 103 and 106. These summaries permit a simple unadjusted comparison of the group means.

What does the re-calculation represent?

The re-calculated mean difference is formed by comparing the posted means without adjustment for other variables. For the primary outcome, the contrast is expressed as sitagliptin minus combination, giving 0.56. A positive value in that ordering means the sitagliptin group’s mean was higher. This is a descriptive re-calculation from summary statistics, not a replacement for the registry’s own analysis and not an adjusted treatment effect.

The associated interval expresses the range of values compatible with the summary-based estimate under the calculation’s assumptions. Its position entirely above zero indicates that the simple comparison is not compatible with no mean difference at the interval’s stated precision. The unadjusted p-value is <0.001. A p-value measures how unusual data at least as discrepant as those observed would be under a specified null model; it is not the probability that the null hypothesis is true, nor does it measure clinical importance.

What does ANCOVA add?

The registry reports an ANCOVA analysis for the primary endpoint. ANCOVA compares outcomes between groups while adjusting for covariates included in the model. Adjustment can account for measured baseline features and can improve precision when those features are related to the outcome. The registry record does not identify the covariates used, the model specification, or the details needed to reproduce the analysis. Consequently, the ANCOVA estimate should be described as the sponsor-reported adjusted comparison, without guessing which adjustments were made.

The sponsor-reported primary estimate is -0.55, with a 95% confidence interval from -0.77 to -0.33 and p=0.001. Its sign follows the registry’s group order, which differs from the sitagliptin-minus-combination ordering used for the simple re-calculation. Signs that point in opposite directions can therefore describe the same substantive ordering when the subtraction order is reversed. The re-calculation and ANCOVA estimate are close in magnitude and agree in direction after accounting for that ordering.

How should uncertainty be read?

A confidence interval gives a range of estimates consistent with the data and the method used, under repeated use of that method. It conveys both direction and precision. An interval that excludes zero is evidence against a zero difference under the corresponding model, but it does not show the probability that a particular value is true. A narrow interval indicates more precision than a wide interval, while neither interval width nor a p-value alone determines whether a difference matters clinically.

The registry reports 95% confidence intervals for its adjusted analyses. The simple re-calculations use posted means and standard errors and are unadjusted. They cannot account for baseline covariates, analysis-population choices, or model features that are not available in the summary statistics. Their purpose is to give a transparent comparison of posted group summaries, not to recreate the sponsor’s full model.

Why can outcomes differ in their apparent certainty?

Each outcome has its own measurement variability, sample count, and estimate. The fasting plasma glucose result and the body-measure results therefore need not have the same degree of precision as HbA1c. For body-mass index and body weight, the registry’s adjusted intervals include no difference; that means their estimates are less decisive in this record, not that the groups have been shown to be identical. A broad interval can remain compatible with differences in either direction.

Over-dispersion is a general concern when outcome variability is greater than a model assumes. In that situation, standard errors can be too small and intervals too narrow if the model does not accommodate the extra variation. The registry record does not report a specific over-dispersion assessment for these continuous outcomes, so there is no basis to claim that it occurred or that a particular correction was used. For count outcomes such as episodes, the appropriate statistical model can depend on the distribution and frequency of events; the registry’s listed outcome does not provide enough detail to identify such a model.

Finally, statistical evidence across several outcomes requires attention to multiplicity. When many comparisons are examined, the chance of at least one apparently unusual result can increase if all are interpreted without accounting for the number of tests. The registry record does not report whether an adjustment, hierarchy, or other multiplicity strategy was used. The primary endpoint should be distinguished from the secondary analyses, and the latter interpreted with greater caution when their testing context is unavailable.

11. Confidence Intervals and Precision

The primary mean difference is 0.56 for sitagliptin minus combination, based on the posted Week 26 means. The re-calculated interval is entirely above zero, so the summary-based comparison supports a difference in the direction of a lower mean HbA1c in the combination group. The registry’s ANCOVA estimate is -0.55, with a 95% confidence interval from -0.77 to -0.33. The negative sign reflects its group ordering; it does not indicate a different direction from the re-calculated contrast.

Confidence intervals are more informative than a point estimate alone because they show the range of values compatible with the analysis, rather than implying that the estimated difference is known exactly. The sponsor-reported interval excludes zero, and its endpoints are separated by a limited range. That supports a fairly well-defined direction in the adjusted comparison, although it does not address every source of uncertainty, such as missing outcome data or unreported model choices.

Precision is outcome-specific. The registry’s adjusted intervals for body-mass index and body weight include zero, so those results are compatible with no between-group difference as well as with differences in either direction. This is a different pattern from the primary HbA1c result. It does not prove equivalence: an interval that includes zero can still allow meaningful effects, depending on its width and on what differences would matter. The registry record does not state a threshold for clinical importance.

Fasting plasma glucose has a distinct estimate and interval, and should be interpreted on its own scale and in its own clinical context. Confidence intervals do not capture all limitations of the evidence. They do not resolve the open-label design, differences in stated background treatment, multiplicity, or incomplete reporting of analysis populations and missing data. Precision describes statistical uncertainty under the analysis; it is not a guarantee that the analysis is free from bias.

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 calculation. The continuous-outcome entries give participants per arm. They are not total event counts. The assumptions stated for these illustrations are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each planning row uses the stated fraction of the observed effect as the effect to detect.

The 80% power column gives the participants per arm estimated to provide an 80% chance of detecting the specified effect under those assumptions, if that effect is real and the model assumptions hold. The 90% power column uses a higher probability of detecting the same specified effect, which generally requires more participants per arm. The values are hypothetical planning requirements, not outcomes from the completed trial.

Rows using a smaller fraction of the observed effect represent a smaller target difference. Smaller effects are harder to distinguish from random variation and therefore require more participants per arm under the same assumptions. This pattern is a general feature of power calculations, not evidence that a larger or smaller effect would necessarily occur in a future study.

The trial’s own primary endpoint description states a planning assumption of a standard deviation of 1.0 for HbA1c and 100 subjects in each treatment arm to obtain 80% power for detecting a HbA1c difference of 0.4%. Those are the assumptions recorded for that calculation. Details of the broader original design rationale are not reported in the registry record. The planning illustrations based on the observed result must not be interpreted as observed power.

Because the illustrations depend on the observed effect, they should be read as exploratory scenarios rather than as a recommendation for a new study. The calculations do not address attrition, unequal allocation in a future design, multiplicity, covariate adjustment, or other design choices unless those are among the stated assumptions. Their purpose is to show how required group size changes as the target effect changes under a fixed set of assumptions.

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 = 0.560; pooled standard deviation σ = 0.920.
2. n = 2 × (1.96 + 0.842)² × 0.920² / 0.560² = 42.4, rounded up to 43 participants per arm for 80% power.

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

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- 0.56                      # observed difference in means
sigma <- 0.920122865141673                     # 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(789191)
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 = 43 (80% power), 57 (90% power)
 75% of observed effect: n per arm = 76 (80% power), 101 (90% power)
 50% of observed effect: n per arm = 170 (80% power), 227 (90% power)
power.t.test: 43.4 
Simulated power at n = 43 per arm: 0.794 (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’s own primary analysis is an ANCOVA comparison of final values. It reports a mean difference of -0.55, a 95% confidence interval from -0.77 to -0.33, and p=0.001. The sign is based on the registry’s group order. The simple re-calculation compares sitagliptin minus combination and gives 0.56, with an unadjusted p-value of <0.001. Reversing the subtraction order reverses the sign, so the two estimates point in the same direction.

The estimates are close, but they are not identical methods. The re-calculation uses only the posted arm means and standard errors and does not adjust for covariates. ANCOVA can account for measured covariates, but the registry record does not report its full model specification. The two analyses therefore should not be treated as interchangeable or as independent confirmations.

The record does not flag a discrepancy between the recalculation and the sponsor-reported result. The registry analysis remains the result of the trial’s stated analysis, while the simple calculation offers a transparent, unadjusted check based on the posted summaries. It does not establish why the estimates differ slightly or permit a reconstruction of the adjusted model.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists multiple outcomes and reports 17 analyses. When many outcomes or comparisons are examined, some results can appear statistically notable by chance even if there is no underlying difference. Statistical approaches to this issue include adjusting significance thresholds, organizing outcomes into a testing hierarchy, or identifying a limited primary analysis and treating others as exploratory. The registry record does not report whether any such approach was used, so the secondary p-values should not be interpreted as if their multiplicity had been addressed.

The primary endpoint is clearly identified as HbA1c at Week 26, but the full statistical analysis plan is not reported in the registry record. Details about interim analyses, covariates in ANCOVA, prespecified analysis populations, and procedures for deviations from the assigned treatment are also not reported in the registry record. Their absence prevents a complete audit of how the posted analysis followed the planned methods; it does not establish that a particular procedure was or was not used.

The record describes the full analysis set as randomized subjects exposed to at least one dose of trial product, but the remainder of the definition is incomplete in the registry record. The exact analysis population used for each posted endpoint is therefore not fully clear from the available wording. Outcome-specific participant counts vary, and the record does not explain the reasons for those differences.

Missing-data methods are not reported in the registry record. It is not possible to tell from the posted summaries whether missing outcomes were handled by imputation, complete-case analysis, a model-based method, or another approach. Missingness can affect estimates when participants with unavailable outcomes differ systematically from those with observed outcomes. The available record does not establish whether that concern arose in this trial, or how it was assessed.

These reporting gaps limit assessment of robustness, but they do not erase the reported results. They mean that interpretation should distinguish the clearly presented group summaries and sponsor ANCOVA estimate from details of analysis conduct that are not documented 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: fibromatosis (0.9%), alcohol withdrawal syndrome (0.9%), atrial fibrillation (0.9%), optic ischaemic neuropathy (0.9%), bile duct obstruction (0.9%).

The serious adverse event counts are reported with their at-risk denominators: 2 of 107 in the combination group and 4 of 110 in the sitagliptin group. Other adverse events are reported as 76 of 107 and 72 of 110, respectively. These are counts in the posted safety summaries; they should not be combined with outcome-analysis denominators, which differ by endpoint.

The registry lists serious-event terms including fibromatosis, alcohol withdrawal syndrome, atrial fibrillation, optic ischaemic neuropathy, bile duct obstruction, and bronchopneumonia, each recorded as 0.9%. The record does not assign these listed terms to a particular treatment group in the information, so they should not be attributed to one arm. Death counts are not reported in the registry record.

These figures describe reported events, not proof that a treatment caused an event or that the groups’ safety profiles are equivalent. The record does not provide event timing, severity beyond the serious-event classification, exposure-adjusted rates, or detailed adjudication information. Interpretation should therefore remain descriptive and should not rely on the small differences in counts as a definitive comparison of safety.

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

The posted primary result shows lower mean HbA1c in the combination group than in the sitagliptin group at Week 26. The simple summary-based contrast and the sponsor’s adjusted ANCOVA analysis agree in direction once their different subtraction orders are taken into account. The sponsor-reported interval excludes zero, and the unadjusted re-calculated p-value is <0.001. Together, these results support a difference in the recorded group means under the analyses reported.

That conclusion is narrower than a claim about every participant or every outcome. A mean difference does not show how many people improved, remained unchanged, or worsened. The trial includes separate target-attainment outcomes, but numerical results for several listed secondary measures are not reported in the registry record. The body-measure analyses are also less decisive than the primary endpoint, and safety counts require cautious descriptive interpretation.

The trial compared treatment strategies with different stated background regimens, not insulin detemir added to an otherwise identical regimen in both groups. It was randomized but unmasked. The registry record does not provide enough information to assess all possible effects of open-label conduct, covariate adjustment, missing outcomes, or multiplicity. These qualifications shape how confidently the result can be extended beyond the reported comparison.

The result does not by itself establish long-term outcomes, a complete safety profile, or the effect in settings and populations not described in the registry record. Nor does statistical evidence alone establish clinical importance: the registry record does not state a threshold for a clinically important HbA1c difference. The most defensible reading is that the reported Week 26 mean HbA1c differed between the assigned strategies, with uncertainty and scope bounded by the available analysis and reporting details.

18. Why This Trial Matters Statistically

This trial illustrates why a result should be read at several levels. The primary endpoint is a continuous measure, so its mean difference gives a direct summary of average separation between groups. The confidence interval adds information about precision, while the p-value addresses compatibility with a zero-difference model. None of these statistics alone describes individual variation, clinical importance, or the full range of outcomes.

It also demonstrates the value of distinguishing an adjusted analysis from a simple re-calculation. The registry’s ANCOVA result accounts for covariates included in its model, although those details are not reported. A calculation from posted means is easier to interpret and reproduces the general direction, but it cannot stand in for the full adjusted analysis. The close directional agreement is informative, while the methodological distinction remains important.

The stated treatment strategies highlight another central issue in trial interpretation: the comparison is defined by what each group received, including background treatment. Because the registry describes different background regimens, the result should be attached to the strategies as recorded, not reduced to an isolated component. This helps prevent causal claims that go beyond the actual randomized contrast.

Finally, multiple secondary outcomes and incomplete reporting make hierarchy and transparency important. The pattern across outcomes offers context, but several numerical results and analysis details are not reported in the registry record. This limits assessment of consistency, multiplicity, missing data, and robustness. Reading the primary result alongside these qualifications gives a more useful statistical account than relying on a single p-value.

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 leaves important details unavailable for assessing the analysis. Although the primary endpoint and sponsor ANCOVA result are reported, the ANCOVA covariates and full model specification are not. The full analysis set description is incomplete, and the record does not explain the outcome-specific differences in participant counts or report how missing outcomes were handled.

The trial was unmasked. That feature is particularly relevant when interpreting outcomes that may be affected by participant or investigator expectations, although the record does not establish that such bias occurred. The comparator also differs from the intervention group in the stated background treatment: the combination group was described with pre-trial metformin, while the sitagliptin group could have pre-trial metformin and/or sulphonylurea. The registry does not give further information about these regimens or how their differences were addressed.

The record lists numerous secondary outcomes and analyses but does not report a multiplicity strategy or a detailed testing hierarchy. Numerical results are not reported for all listed secondary outcomes. This makes it difficult to assess the totality of evidence across the planned outcome set and increases the need for caution with individual secondary findings.

The primary endpoint is reported at Week 26. The registry record does not provide longer-term results or enough detail to evaluate the course of outcomes over time. Safety reporting is also limited: death counts are not reported, and the serious-event terms are not linked to treatment groups in the available information. These limitations constrain interpretation of durability and the broader safety comparison without negating the posted primary outcome.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Effect of Detemir and Sitagliptin on Blood Glucose Control in Type 2 Diabetes — NCT00789191.
• Registry record last updated 2017-03-14; status last verified 2017-02.
• Results first posted 2010-12-10.
• 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 record describes a completed, randomized, open-label phase 3 parallel-group trial sponsored by Novo Nordisk A/S, with 222 enrolled participants. It compared insulin detemir plus sitagliptin on pre-trial metformin with sitagliptin on pre-trial metformin and/or sulphonylurea. The primary outcome was mean HbA1c at Week 26.

The combination group had a lower posted mean HbA1c than the sitagliptin group. The simple unadjusted re-calculation and the sponsor-reported ANCOVA analysis agree in direction after accounting for their different group orderings. Secondary findings vary by outcome, and numerical results for several listed outcomes are not reported in the registry record.

Interpretation is limited by unmasked treatment, differences in stated background therapy, and missing detail on multiplicity, analysis populations, model specification, and missing-data handling. The registry’s safety counts are descriptive, and death counts are not reported.