A Phase 4, Randomized, Double-blind, Parallel-group, Comparative Study and a Phase 4, Open-label, Long-term Study of SYR-472 (100 mg) in Combination With Insulin in Patients With Type 2 Diabetes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Phase 4, Randomized, Double-blind, Parallel-group, Comparative Study and a Phase 4, Open-label, Long-term Study of SYR-472 (100 mg) in Combination With Insulin in Patients With Type 2 Diabetes (NCT02324569), a randomized phase 3 study. The primary outcome is Change in HbA1c From Baseline at the End of Treatment Period I (End of Treatment Period I - End of the Screening Period).

2. Clinical Question

The clinical question is about a short-term laboratory measure of glycemic control in a particular treatment context: participants already receiving insulin, with diet and/or exercise also part of their care. It is not a comparison of insulin therapy with no insulin, nor does the registry record establish that the results apply to people with other forms of diabetes or different background treatment.

The registry describes a comparative, double-blind study followed by an open-label, long-term study. The primary endpoint is limited to Treatment Period I, while some posted secondary outcomes extend to later timeframes. These distinct periods and outcome windows matter when interpreting which question an individual result addresses. The posted primary analysis concerns HbA1c change; it does not by itself answer every question about longer-term outcomes or safety.

3. Trial Design

The registry describes a randomized, parallel-group study with quadruple masking. Its title identifies a Phase 4 randomized, double-blind comparative study and a Phase 4 open-label long-term study. The registry’s phase field, however, lists PHASE3. That difference in phase labeling is present in the record and should not be silently reconciled.

The study record lists 240 participants enrolled. For the primary outcome, the Full Analysis Set is defined as all randomized participants who received at least one dose of study drug. The posted primary arm counts are 116 in Treatment Group I and 124 in Treatment Group II. The adverse-event denominators are not identical across groups: 116 and 119 participants are listed as at risk for the serious-event summaries. These are analysis-specific counts, not interchangeable descriptions of one common analysis population.

The study was intended to compare SYR-472, 100 mg once weekly, added to insulin therapy with placebo, and also to evaluate longer-term use of SYR-472 in an open-label study. Treatment Group I is described as receiving one SYR-472 tablet weekly before breakfast. Treatment Group II is described as receiving SYR-472 or one placebo tablet weekly before breakfast. The registry record does not provide further detail on how participants moved between the comparative and long-term periods.

The primary outcome timeframe runs from the end of the screening period, identified as Week 0, to the end of Treatment Period I, up to Week 12. The registry lists the study as completed. The original sample-size rationale, allocation ratio used in the trial design, detailed randomization procedures, and details of any 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: Participants with a historical diagnosis of type 2 diabetes mellitus with inadequate glycemic control despite treatment with insulin therapy were enrolled in 1 of the 2 groups: Group I: 100 milligram (mg) SYR-472 in Period I followed by 100 mg SYR-472 in Period II, Group II: SYR-472 Placebo in Period I followed by 100 mg SYR-472 in Period II.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set: All randomized participants who received at least one dose of study drug.

5. Baseline Characteristics

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

Population: Randomized set: Randomized set included all randomized participants.

6. Primary Endpoint

The registry lists the primary outcome as Change in HbA1c From Baseline at the End of Treatment Period I (End of Treatment Period I - End of the Screening Period), measured over End of the screening period (Week 0) and End of Treatment Period I (Up to Week 12).

The primary endpoint is change in HbA1c from baseline to the end of Treatment Period I. HbA1c is reported as a percentage, and the registry defines the change as the value at the end of Treatment Period I minus the value at the end of the screening period. The stated timeframe is Week 0 through up to Week 12.

HbA1c is a commonly used continuous measure of glycemic control. In general, a change-from-baseline endpoint uses each participant’s starting value as a reference, so the comparison concerns how the measure shifted during the treatment period rather than only comparing final values. A negative change indicates a lower value at the endpoint than at baseline; a positive change indicates a higher value. This is a general explanation of the measure’s direction, not a claim about the clinical meaning of a particular change for an individual.

The primary outcome is reported as a least-squares mean, with standard errors for the arm estimates. The sponsor’s analysis reports an adjusted least-squares mean difference from an ANCOVA model. In general, ANCOVA compares groups while accounting for specified covariates, but the covariates and full model specification are not reported in the registry record. The use of HbA1c provides a focused, quantitative outcome for the comparative period; it does not alone describe long-term outcomes, all aspects of diabetes control, or the balance of benefits and harms.

7. Results Posted in the Registry

• Treatment Group II: mean 0.07 versus -0.56 on control; difference 0.63 (95% CI 0.44 to 0.82), standardized effect d = 0.84; 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.

Number of Participants With Markedly Abnormal Values of Laboratory Parameters (Total Bilirubin >2.0) Before Start of Treatment Period II

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Treatment Group II: 0.8% (1/124) versus 0.9% (1/116) on control; risk difference -0.1 percentage points (95% CI -2.4 to 2.2), risk ratio 0.94 (0.06–14.78); unadjusted two-proportion z-test p = 0.962.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Treatment Group II and the control. The interval gives the range of effects that remain compatible with the data.

Number of Participants With Total Hypoglycaemia After 1st Dose of Study Drug and Before Start of Treatment Period II

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Treatment Group II: 12.9% (16/124) versus 14.7% (17/116) on control; risk difference -1.8 percentage points (95% CI -10.5 to 7.0), risk ratio 0.88 (0.47–1.66); unadjusted two-proportion z-test p = 0.694.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Treatment Group II and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists a range of secondary outcomes, including HbA1c, fasting plasma glucose, meal-tolerance-test plasma glucose, vital signs, ECG parameters, laboratory measures, hypoglycaemia, and self-monitoring of blood glucose. Only a subset of secondary results is detailed in the posted numerical summaries described in the record. The two available recalculated binary comparisons have unadjusted p-values that do not indicate a clear statistical difference. Their estimates are imprecise, so they do not establish that the groups had equivalent outcomes.

Secondary outcomes can add context to the primary HbA1c result by covering other measures and timeframes, but they answer different questions. They should not be treated as independent confirmations of the primary finding, particularly when the registry does not post sponsor analyses for the available secondary comparisons. Nor should an unreported secondary result be interpreted as showing no effect: its result is not reported in the registry record.

Multiplicity is a concern when many outcomes or comparisons are examined. In general, each additional statistical test creates another opportunity for a chance finding, so unadjusted p-values across numerous outcomes carry less weight as confirmatory evidence. The registry lists ten outcome measures and eight posted secondary outcomes, but a multiplicity adjustment or hierarchy for interpreting the secondary tests is not reported in the registry record. The number of outcomes alone does not reveal which were prespecified as confirmatory, how they were prioritized, or whether any correction was applied.

The overall secondary pattern is therefore best treated as exploratory context rather than a second definitive conclusion. Interpretation should give priority to the prespecified primary question and sponsor analysis, while recognizing that different secondary measures may capture different aspects of the study. No broader conclusion about unreported outcomes can be drawn from the numerical results that are available.

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

• Change From Baseline in HbA1c
• Change From Baseline in Fasting Plasma Glucose
• Change From Baseline in Plasma Glucose Measured by the Meal Tolerance Test in Treatment Period I
• Number of Participants With Markedly Abnormal Values of Vital Signs Before Start of Treatment Period II
• Number of Participants With Markedly Abnormal Values of ECG Parameters Before Start of Treatment Period II
• Number of Participants With Markedly Abnormal Values of Laboratory Parameters (Total Bilirubin >2.0) Before Start of Treatment Period II
• Number of Participants With Total Hypoglycaemia After 1st Dose of Study Drug and Before Start of Treatment Period II
• Change From Baseline in Self-Monitoring of Blood Glucose Before Breakfast

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is continuous: each participant has a numerical change in HbA1c, and the arm summaries are means rather than counts of participants crossing a threshold. The registry reports least-squares means and standard errors. A standard error describes the statistical uncertainty in an estimated mean; it is not the spread of individual participant outcomes. In general, a smaller standard error indicates a more precisely estimated mean, though its interpretation depends on the analysis and data.

What does ANCOVA contribute?

The sponsor reports an ANCOVA analysis and a least-squares mean difference. In general, analysis of covariance compares outcomes between groups while adjusting for covariates included in the model. Adjusting for baseline or other prespecified variables can improve precision or account for differences in those variables. The registry record does not report the covariates, model details, missing-data handling, or other assumptions used in this ANCOVA. It is therefore not possible to reconstruct the sponsor’s adjusted analysis from the brief result alone.

The sponsor’s reported contrast is Treatment Group I minus Treatment Group II: -0.63 percentage points. The unadjusted recalculation instead uses Treatment Group II minus Treatment Group I and reports 0.63. These values express the same direction of contrast under reversed subtraction. The recalculation is derived from the posted group summaries and is not a replacement for the sponsor’s adjusted analysis.

Why are the recalculation and reported result not identical in every detail?

For a continuous outcome, a basic recalculation can compare the group means and estimate uncertainty from their standard errors. The posted recalculation uses an approximate pooled standard deviation. In general, pooling estimates a common within-group spread from the two groups; this can be useful for a simple standardized comparison, but it does not reproduce covariate adjustment or every feature of the sponsor’s model. The registry marks the approximation as applicable. The posted record does not indicate a discrepancy between the recalculation and sponsor result, and the sponsor’s analysis remains the reported trial analysis.

The p-value is calculated under a null hypothesis of no group difference, with the method’s assumptions. In general, a small p-value indicates that data at least as incompatible with that null as the observed data would be relatively unusual if the null and model assumptions held. It is not an effect-size measure, a probability that the treatment caused the result, or a measure of clinical importance. The sponsor reports p < 0.001, while the simple recalculation reports p < 0.001. Those expressions arise from different analyses and should not be treated as interchangeable model outputs.

How should secondary binary outcomes be read?

The posted secondary outcomes include counts of participants with an event, such as a markedly abnormal laboratory value or total hypoglycaemia. For binary outcomes, the basic quantities are the number of participants with the event and the number included in the group denominator. A risk difference compares the observed event proportions between groups. In general, a risk difference near zero means the observed proportions are similar, but its uncertainty must also be considered. A confidence interval that includes zero is compatible with a range of possible group differences in either direction.

The secondary comparisons are recalculated from counts and denominators, and their p-values are unadjusted. The registry does not report sponsor analyses for these outcomes. Such simple comparisons do not account for other variables unless an adjusted model is separately reported. The available record gives risk differences and intervals, as well as other effect summaries, but does not establish that any one measure was designated as the main analysis for each secondary endpoint.

For a rare event, a small count can make estimates particularly unstable. In general, when few participants experience an event, changing the event count by a small amount can substantially affect the estimated relative effect. A confidence interval can consequently be wide, reflecting uncertainty about both the direction and size of a difference. Failure to find a statistically significant difference is not evidence that the groups are equivalent; equivalence requires a suitable design, margin, and analysis, which are not reported for these secondary comparisons.

What does over-dispersion mean in general?

For count outcomes, over-dispersion means that the observed variability is greater than a statistical model expects. If it is ignored, standard errors may be too small and confidence intervals too narrow. This is a general statistical concern, not a finding established by the registry record. The primary endpoint is a continuous change measure, and the posted secondary results are participant counts for binary outcomes; the record does not report an over-dispersion analysis.

The registry reports a Full Analysis Set definition for the primary endpoint: randomized participants who received at least one dose of study drug. In general, an analysis population definition affects which participants contribute to an estimate and how missing observations are handled. The record does not state how many participants had missing primary outcome data, how such data were handled, or whether additional analysis populations were used. Those details are important for judging the robustness of a result but cannot be inferred from the reported means and p-value.

11. Confidence Intervals and Precision

The sponsor’s 95% confidence interval for the adjusted primary difference is -0.826 to -0.443 percentage points, with the contrast defined as Treatment Group I minus Treatment Group II. It lies entirely below zero, matching the negative point estimate of -0.63. Under the ANCOVA model and its assumptions, the interval describes a range of values compatible with the observed data and estimation procedure.

A confidence interval is not a guarantee that the true value lies inside the particular interval, and it is not a range containing a specified proportion of individual participants’ changes. In general, a confidence interval reflects uncertainty in an estimated group contrast. Its width depends on the amount and variability of information, the analysis model, and other assumptions. The registry’s 95% interval is relatively informative about direction, but it does not by itself determine whether the estimated difference is clinically important.

The unadjusted recalculation reports its own interval for Treatment Group II minus Treatment Group I, in the positive direction. That recalculation uses posted summary statistics and an approximate pooled standard deviation. It is not the same as the sponsor’s adjusted ANCOVA interval and should not be used to substitute for it. Reversing which group is subtracted changes the sign of a contrast, not the underlying group comparison.

For the posted secondary binary outcomes, the intervals are broader relative to the estimates, including values on either side of no difference. This indicates limited precision for those comparisons. In general, an interval that includes no difference does not establish that the groups are identical; it means the data are compatible with a range that includes no difference, as well as some nonzero differences. The registry does not report adjusted secondary analyses or a multiplicity strategy that would further define how these intervals should be interpreted.

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 planning figures are illustrations based on the observed effects, not the trial’s original design calculations. For these outcomes, the entries are participants per arm, not total events. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The target effect is a stated fraction of the observed effect. The original sample-size assumptions and justification for the trial are not reported in the registry record.

The 80% power column gives the illustrative number of participants per arm for a design that would have an 80% chance of detecting the specified effect under the stated assumptions, if that effect and model assumptions were correct. The 90% power column gives the corresponding number per arm for a higher chance of detecting that effect. Power is a property of a proposed design under specified assumptions; it is not a measure of whether a completed trial has found a true effect.

Each planning row uses a different target relative to the observed effect: the full observed effect, 75% of the observed effect, or 50% of the observed effect. Smaller effects require more participants to detect with the same target power and significance level. The very large illustrative figures for some secondary outcomes reflect the small observed differences in those data under this planning calculation; they are not recruitment targets set by the trial investigators.

These figures are not observed power. They are sensitive to the estimated effect used as the planning basis, and an observed effect can be an uncertain estimate. They should be read as conditional planning illustrations, not as proof that the original study was adequately or inadequately powered. The registry record does not report its own design assumptions, target effect, or original power calculation.

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 = 0.630; pooled standard deviation σ = 0.750.
2. n = 2 × (1.96 + 0.842)² × 0.750² / 0.630² = 22.2, rounded up to 23 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 <- 0.63                      # observed difference in means
sigma <- 0.749880180905113                     # 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(2324569)
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 = 23 (80% power), 30 (90% power)
 75% of observed effect: n per arm = 40 (80% power), 53 (90% power)
 50% of observed effect: n per arm = 89 (80% power), 120 (90% power)
power.t.test: 23.2 
Simulated power at n = 23 per arm: 0.796 (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 sponsor’s reported primary analysis is an ANCOVA comparison of least-squares means. It estimates Treatment Group I minus Treatment Group II as -0.63 percentage points, with a 95% confidence interval from -0.826 to -0.443 and p < 0.001. This is the analysis explicitly reported in the registry and is the principal inferential result for the primary endpoint.

The simple recalculation from posted summary statistics expresses the contrast in the reverse order, Treatment Group II minus Treatment Group I, and gives 0.63 with p < 0.001. The apparent sign difference follows from reversing the subtraction order. The recalculation is unadjusted and uses an approximate pooled standard deviation; it does not recreate the covariate-adjusted ANCOVA model.

The two results are directionally consistent: both indicate a difference favoring a greater decrease in the Treatment Group I mean change relative to Treatment Group II. Their estimates and p-values should not be compared as if they were outputs from the same model. Because the registry’s sponsor analysis is the reported adjusted analysis, it takes precedence over the unadjusted recalculation. The record does not give enough model detail to independently reproduce the ANCOVA estimate.

15. Multiplicity, Missing Data and Analysis Populations

The registry record identifies a primary endpoint and lists multiple secondary outcomes. It reports one sponsor analysis for the primary comparison and no sponsor analyses for the secondary outcomes with numerical comparisons. The record does not report whether secondary hypotheses were tested in a prespecified sequence, whether family-wise error was controlled, or whether any multiplicity adjustment was used. Without those details, unadjusted secondary p-values should be treated cautiously and should not be given the same confirmatory weight as the primary result.

Multiplicity arises when several statistical tests are considered together. In general, testing more outcomes increases the chance that at least one result may appear unusual by chance alone, even if no corresponding underlying effect exists. Adjustment procedures can address this issue, but they may also change the threshold or interpretation of individual results. Which approach is suitable depends on the trial’s objectives and analysis plan; that plan is not reported in the registry record.

The primary analysis population is the Full Analysis Set, defined as all randomized participants who received at least one dose of study drug. The registry does not report how many participants were excluded from the primary endpoint analysis, whether data were missing, the reasons for missingness, or the method used to handle missing values. It also does not report sensitivity analyses assessing how different missing-data assumptions might affect the result.

Missing outcome data can affect estimates if participants with missing measurements differ in relevant ways from those with observed measurements. In general, the implications depend on why data are missing and on the assumptions used in the analysis. The arm counts and estimates alone do not identify those reasons or establish that missingness had no effect.

The adverse-event denominators differ from the primary analysis counts, and the record does not provide a full account of population definitions for each safety summary. Detailed analysis plans, interim-analysis procedures, and additional analysis populations are not reported in the registry record. These reporting gaps limit assessment of the full inferential framework, but they do not erase the reported primary estimate and its stated uncertainty.

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: cataract (1.7%), acute myocardial infarction (0.9%), atrial fibrillation (0.9%), cardiac failure congestive (0.9%), mitral valve incompetence (0.9%).

The serious adverse event summary lists 16 affected participants among 116 at risk in Treatment Group I and 9 among 119 at risk in Treatment Group II. These are participant counts, not a count of every event, and the at-risk denominators differ from the primary outcome counts. The registry also lists serious-event terms, including cataract, acute myocardial infarction, atrial fibrillation, congestive cardiac failure, mitral valve incompetence, and ischaemic colitis, with the percentages recorded for those terms.

These figures describe reported serious adverse events in the registry. They do not establish that a study treatment caused an event, and they do not provide a formal adjusted comparison between groups. The record does not give a deaths count in the safety summary. A complete interpretation would also depend on details such as event timing, duration of exposure, definitions, and participant-level context, which are not reported in the registry record.

Serious adverse events are distinct from all adverse events, and this summary should not be read as a complete inventory of symptoms or harms. The counts should be considered alongside the study design and analysis populations, without drawing a causal or clinical recommendation from them alone.

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

The posted primary result shows a statistically clear difference in change in HbA1c between the two treatment groups through the end of Treatment Period I. The sponsor’s adjusted estimate is -0.63 percentage points for Treatment Group I minus Treatment Group II, and its 95% confidence interval lies below zero. The observed group means are also in opposite directions: -0.56 for Treatment Group I and 0.07 for Treatment Group II.

This result answers a focused question about a continuous laboratory endpoint during the comparative period, among the defined Full Analysis Set. It does not by itself establish how participants felt or functioned, what happened over longer follow-up, or how every safety outcome compared. The record’s long-term study is described as open-label, while the primary endpoint is measured through up to Week 12; those features define different evidentiary contexts.

Statistical evidence of a difference does not automatically establish that the difference is clinically important. The registry record does not state a threshold for clinical importance, and the confidence interval should be interpreted as uncertainty around the group contrast rather than a prediction for an individual participant. The result also does not establish that every participant had a change like the group average.

The secondary results do not resolve all questions raised by the primary finding. The available secondary comparisons are unadjusted, imprecise, and subject to the broader issue of multiple outcomes. Their lack of a statistically significant difference is not proof of equivalence. Similarly, the serious adverse event counts are descriptive and do not alone support a causal safety conclusion.

Accordingly, the registry supports a description of a primary HbA1c difference in this study’s treatment groups and timeframe. It does not support extending that result to unreported endpoints, other populations, individual treatment choices, or conclusions about overall benefit and risk.

18. Why This Trial Matters Statistically

This trial illustrates why an endpoint, analysis population, and timeframe must be read together. The primary result concerns HbA1c change from baseline through up to Week 12 in participants who were randomized and received at least one dose. It is not a generic estimate of long-term diabetes outcomes. Keeping that scope explicit prevents a short-term laboratory comparison from being treated as an answer to a much broader question.

It also shows why the reported analysis and a simple recalculation are not interchangeable. The sponsor reports an adjusted ANCOVA contrast, while the recalculation uses posted summary statistics and is unadjusted. The direction of the numerical contrast changes when the subtraction order is reversed, but the interpretation of which group had the greater reduction remains consistent. Distinguishing model-based results from summary-statistic checks is central to transparent statistical reading.

The record further demonstrates that statistical significance, precision, and clinical importance are separate considerations. The primary interval is entirely on one side of no difference, while the registry does not provide a clinical-importance threshold. Secondary comparisons are less conclusive and illustrate why small counts and unadjusted testing require restraint. A non-significant result does not demonstrate equality.

Finally, the registry contains limits on what can be assessed: it does not report the original sample-size rationale, full ANCOVA specification, multiplicity plan, or missing-data methods. A careful reading can still describe the principal result while recognizing that these omissions constrain independent evaluation of its assumptions and of the wider evidence.

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.

Several limitations specific to this record constrain interpretation. The phase designation is inconsistent: the title describes Phase 4 studies, while the registry phase field lists PHASE3. The record does not explain that difference. The arm descriptions also state that Treatment Group II received SYR-472 or placebo, while the study summary describes the comparison as SYR-472 versus placebo. The summary supports the intended comparative question, but the record does not explain the arm wording in more detail.

The primary analysis is reported as ANCOVA, but covariates, model specification, and missing-data handling are not reported in the registry record. The sponsor’s estimate therefore cannot be independently reconstructed from the summary. The simple recalculation is unadjusted and is not a substitute for those missing details.

The record does not report the original sample-size justification, the trial’s design assumptions, interim analyses, or a multiplicity plan. It lists multiple outcomes, but only some have numerical results in the available record. Secondary comparisons are unadjusted and have no sponsor analyses posted. This limits their confirmatory interpretation.

For safety, the at-risk denominators are not the same as the primary arm counts, and deaths are not reported in the safety summary. The serious-event counts alone do not describe event timing, duration, exposure, or causality. The primary endpoint timeframe is up to Week 12, whereas the record describes a separate open-label long-term study and gives a secondary hypoglycaemia timeframe up to Week 53. These different windows should not be conflated.

Finally, the registry record does not provide clinical-importance criteria for the HbA1c difference. Statistical evidence can be described, but its clinical significance cannot be settled from the posted estimate alone.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Phase 4, Randomized, Double-blind, Parallel-group, Comparative Study and a Phase 4, Open-label, Long-term Study of SYR-472 (100 mg) in Combination With Insulin in Patients With Type 2 Diabetes — NCT02324569.
• Registry record last updated 2023-12-12; status last verified 2023-12.
• Results first posted 2019-02-11.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

23. Record Summary

The registry describes a randomized, quadruple-masked, parallel-group comparison of SYR-472 and placebo as add-ons to insulin therapy in people with type 2 diabetes and inadequate glycemic control, along with an open-label long-term study. The record lists 240 enrolled participants and identifies the study as completed.

The primary endpoint is change in HbA1c from baseline to the end of Treatment Period I, up to Week 12. The sponsor reports an ANCOVA least-squares mean difference of -0.63 percentage points for Treatment Group I minus Treatment Group II, with a 95% confidence interval from -0.826 to -0.443 and p < 0.001. The posted analysis indicates a statistical difference for this primary endpoint.

Secondary results and serious adverse event counts provide additional descriptive context, but the record does not report a multiplicity plan, full model details, or missing-data methods. The secondary comparisons are unadjusted, and the safety counts alone do not establish causality. The registry phase field and study title also differ in their phase labels.