A Research Study Comparing RYBELSUS® to Other Blood Sugar Lowering Tablets in People Living in America With Type 2 Diabetes (REALYSE): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Research Study Comparing RYBELSUS® to Other Blood Sugar Lowering Tablets in People Living in America With Type 2 Diabetes (REALYSE) (NCT05035082), a randomized phase 4 study. The primary outcome is Change in Glycated Haemoglobin (HbA1c).

2. Clinical Question

REALYSE was a randomized, pragmatic trial in a real-world adult population with type 2 diabetes. Assignment to a treatment was described as being decided by chance. The study compared medicines given in addition to metformin, rather than comparing either tablet strategy with no additional glucose-lowering medicine. Its primary endpoint was change in glycated haemoglobin, or HbA1c, over the period from week 0 to year 1.

The registry record identifies 1,018 participants as enrolled. For the primary outcome, the analysis population is described as the full analysis set of randomized participants, with the number analyzed determined by availability of outcome data. The posted group summaries include 326 participants in the oral semaglutide group and 293 in the other-medication group. These numbers describe participants with available data for this outcome, not the total enrollment.

The comparison is clinically framed around treatment choice in routine care, but the statistical question is narrower: what difference in mean HbA1c change was observed between the randomized groups, and how precisely was that difference estimated? The registry also includes secondary outcomes concerning HbA1c targets, body weight and treatment satisfaction. Those outcomes add context but do not replace the primary comparison.

3. Trial Design

REALYSE is recorded as a completed, randomized, parallel-group Phase 4 trial with no masking. The sponsor is Novo Nordisk A/S. The registry records a start date of 2021-09-01 and a primary completion date of 2025-06-18. The study title characterizes it as a pragmatic randomized trial of once-daily oral semaglutide versus any other oral glucose-lowering medication in adults with type 2 diabetes receiving metformin monotherapy in US-based health care systems.

The two study groups received tablets used in addition to metformin. One group received oral semaglutide; the other received another oral glucose-lowering medication. The brief summary says the treating doctor prescribed the assigned medicine and advised participants how to take it. Because the trial was unmasked, participants and treating clinicians were not blinded to the assigned treatment, according to the registry’s masking classification. The record does not report further details about how randomization was generated or concealed.

The study was intended to follow participants for about one year. The brief summary describes two planned visits with the doctor as part of usual diabetes management: an initial visit when participants entered the study and a one-year follow-up visit. Study personnel could contact participants up to three times during that period to follow up on information from doctor visits. Participants were also asked to complete questionnaires three times, using a personal smartphone or tablet, or paper when they did not have access to a device. The record says the clinic visits were part of routine diabetes management and covered by participants’ health insurance plans.

The registry reports total enrollment of 1,018. The primary outcome analysis includes participants with available outcome data, with 326 analyzed in the oral semaglutide group and 293 in the other oral medication group. The original design’s allocation ratio, sample-size justification, and assumptions used to plan the trial are not reported in the registry record. The equal-allocation assumption described for the planning illustrations is not a report of the trial’s actual allocation plan.

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 were randomized in a 1:1 ratio to receive either oral semaglutide or one other oral glucose-lowering medication, respectively.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full analysis set included all randomised participants. Overall number of participants analyzed = Participants with available data for the outcome measure.

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: Full analysis set included all randomised participants.

6. Primary Endpoint

The registry lists the primary outcome as Change in Glycated Haemoglobin (HbA1c), measured over Week 0, year 1.

The primary endpoint is change in glycated haemoglobin, or HbA1c, from week 0 to year 1. HbA1c is a continuous measure used to assess average blood glucose over time. A change score expresses how the measure differs between the starting and follow-up assessments; the registry specifies that the outcome is presented as a mean with a standard deviation.

In the posted group summaries, the mean change was -1.2 in the oral semaglutide group and -0.99 in the other oral glucose-lowering medication group. Negative changes indicate that the measured HbA1c value was lower at follow-up than at the starting point on average. The standard deviations describe variation in individual changes around each group mean; they do not describe the uncertainty in the group mean itself.

A continuous endpoint retains information about the magnitude of change across participants. This differs from a threshold outcome, which classifies participants according to whether they did or did not reach a specified target. The registry lists several secondary outcomes based on HbA1c thresholds and weight outcomes, but the primary question is about average HbA1c change.

The outcome population is described as the full analysis set of randomized participants, while the number analyzed is defined as participants with available data for the outcome. This distinction matters when reading the group summaries: they reflect the outcome data available for analysis, rather than necessarily every randomized participant. The registry does not report further details about how missing HbA1c measurements were handled.

7. Results Posted in the Registry

• Other Oral Glucose-lowering Medication: mean -0.99 versus -1.20 on control; difference 0.21 (95% CI -0.03 to 0.45), standardized effect d = 0.14; z-test p = 0.080.

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.

Relative 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 deviations.
• Result:
  • Other Oral Glucose-lowering Medication: mean -3.53 versus -4.00 on control; difference 0.47 (95% CI -0.60 to 1.54), standardized effect d = 0.07; z-test p = 0.390.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Other Oral Glucose-lowering Medication and the control. The interval gives the range of effects that remain compatible with the data.

Change in Body Weight (Lbs)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Other Oral Glucose-lowering Medication: mean -7.75 versus -8.68 on control; difference 0.93 (95% CI -1.58 to 3.44), standardized effect d = 0.06; z-test p = 0.468.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Other Oral Glucose-lowering Medication and the control. The interval gives the range of effects that remain compatible with the data.

Diabetes Treatment Satisfaction Questionnaire, Change Version (DTSQc), Relative Treatment Satisfaction Total Score

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Other Oral Glucose-lowering Medication: mean 12.22 versus 12.28 on control; difference -0.06 (95% CI -1.33 to 1.21), standardized effect d = -0.01; z-test p = 0.926.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Other Oral Glucose-lowering Medication and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists secondary outcomes addressing HbA1c thresholds and individualized targets, body-weight change and treatment satisfaction. The posted numerical secondary summaries cover continuous measures of relative body-weight change, change in body weight in pounds and treatment satisfaction. Across these outcomes, the unadjusted group comparisons do not show a clear separation: each reported confidence interval includes zero. The estimates and intervals convey uncertainty, rather than establishing that the groups are identical.

These secondary outcomes provide context around the primary HbA1c comparison, but they do not replace it. Body-weight measures capture a different dimension of outcome, while treatment satisfaction reflects participants’ reported experience. HbA1c target measures, by contrast, classify participants against specified thresholds. Evidence across different outcomes can help describe a trial’s pattern, but consistency or lack of separation in secondary outcomes does not resolve the primary analysis discrepancy.

The registry does not report sponsor analyses for the posted continuous secondary comparisons. The displayed secondary comparisons are unadjusted recalculations, and their p-values are not reported as multiplicity-adjusted. When many outcomes are examined, the chance of at least one apparently notable result can increase. Without a stated multiplicity strategy, isolated secondary findings deserve less weight than the prespecified primary analysis and should be treated as exploratory rather than confirmatory.

These secondary data also have their own available-data denominators, which differ from the primary outcome counts. That means the analyzed participants can vary by outcome, and apparent patterns should be interpreted with awareness that each summary may be based on a different subset of participants. Reasons for outcome-specific data availability and methods for handling missing secondary outcomes are not reported in the registry record.

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

• Participants Achieving HbA1c Less Than (<) 7.0 Percentage (%) (Yes /No)
• Participants Achieving HbA1c Less Than or Equal to (≤) 6.5% (Yes/No)
• Participants Achieving HbA1c <7.0% or at Least 1.0%-Point Reduction in HbA1c (Yes/No)
• Participants Achieving Greater Than or Equal to (≥) 5% Reduction in Body Weight (Yes/No)
• Participant Achieving Individualized HbA1c Target Per Healthcare Effectiveness Data and Information Set (HEDIS) Criteria (Yes/No)
• Participant Achieving HbA1c ≤ Treatment Provider Defined Individualized Target (Yes/No)
• Relative Change in Body Weight (%)
• Change in Body Weight (Lbs)

9. Statistical Interpretation

10. Statistical Methods Explained

Several statistical ideas help distinguish what the posted estimates say from what they do not say. The primary endpoint is continuous, and the registry reports group means and standard deviations. A mean gives the average observed change in a group. A standard deviation describes how spread out individual changes are around that mean. Neither measure alone gives the uncertainty in the comparison between groups.

What is the mean difference?

A mean difference contrasts the average outcome in one group with the average outcome in another. Its sign depends on the order in which groups are subtracted. In the unadjusted recalculation, the reported contrast is expressed as the other-medication group relative to oral semaglutide and is 0.21. The sponsor-reported ANCOVA estimate is -0.24. These values come from different analyses and should not be treated as interchangeable. In particular, the registry record does not report enough about the sponsor analysis’s coding and adjustment details to reconcile the directions or magnitudes.

An unadjusted comparison uses the posted group summaries without accounting for other variables in a model. A recalculation of this kind can be useful for checking the relationship between the group means and the uncertainty implied by their summary statistics. It is not a substitute for the trial’s own analysis when the two disagree. Here, the registry-reported ANCOVA analysis takes precedence because the discrepancy is documented.

What does ANCOVA do?

ANCOVA is a regression-based method that compares groups while adjusting for one or more covariates. In a randomized trial, adjustment can account for selected measured characteristics and may improve precision when those characteristics are related to the outcome. The registry names ANCOVA for the reported primary analysis, but the covariates, model specification, missing-data approach and other implementation details are not reported in the registry record. It is therefore not possible to attribute the difference between the ANCOVA result and the unadjusted calculation to any particular adjustment or model choice.

How should the p-value be read?

A p-value assesses how compatible the observed data are with a specified null model, under the assumptions of the analysis. It is not the probability that the null hypothesis is true, and it is not a measure of the size or clinical importance of an effect. The sponsor-reported p-value is 0.018; the unadjusted recalculation has p = 0.08. Because the estimates arise from different analyses, their p-values answer the null comparison within their respective methods and assumptions. The p-value does not address bias from missing data, unreported model choices or the practical importance of the difference.

Why do confidence intervals matter?

A confidence interval presents a range of values compatible with the data and the analysis at the stated confidence level. Its width reflects precision: narrower intervals indicate less statistical uncertainty, while wider intervals indicate more. For the sponsor-reported estimate, the 95% interval runs from -0.44 to -0.04. For the unadjusted recalculation, the interval runs from -0.025 to 0.445. The first interval excludes zero; the second includes zero. This distinction describes the analyses’ statistical uncertainty, not a guarantee that the true effect lies within a particular interval in this individual trial.

What does variability contribute?

The primary group standard deviations are 1.463 and 1.516. These describe variation among participants’ HbA1c changes. A group mean can be estimated with greater or lesser precision depending on the amount of variation and the number of observations, among other factors. This is why the group means alone are insufficient for evaluating the group contrast. The standard deviations are not confidence intervals and should not be read as such.

How should the secondary outcomes be interpreted?

The registry lists secondary outcomes involving HbA1c target attainment, body weight and treatment satisfaction. Some are continuous measures, for which a mean difference summarizes an average contrast; others are binary outcomes, which classify whether a participant met a defined criterion. These different outcome types require different statistical summaries. The available secondary continuous comparisons are unadjusted recalculations, and the registry does not report sponsor analyses for those comparisons. Their p-values are unadjusted, and the record does not report a multiplicity adjustment. As a result, they are best treated as exploratory context rather than independent confirmation of the primary analysis.

Statistical significance and clinical importance are separate questions. A small p-value does not by itself establish that a difference matters to patients, and a result that is not statistically distinguishable from zero does not prove that the groups are identical. The registry record does not specify a threshold for a clinically meaningful HbA1c difference or report a formal equivalence or noninferiority analysis. Interpretation should therefore remain anchored to the reported estimates, their intervals, the analysis population and the limits of the available registry information.

11. Confidence Intervals and Precision

The sponsor-reported primary estimate is -0.24, with a 95% confidence interval from -0.44 to -0.04. This interval does not include zero. Under the reported ANCOVA analysis, the data are compatible with treatment differences spanning the values within those limits, subject to the model and analysis assumptions. The interval gives a sense of precision around the estimated contrast; it does not describe the range of individual participants’ HbA1c changes.

The unadjusted recalculation gives a mean difference of 0.21, with a confidence interval from -0.025 to 0.445. That interval includes zero and extends to values on either side of it. It therefore conveys appreciable uncertainty about the direction and size of the unadjusted contrast. The two intervals should not be combined or selectively chosen: they belong to distinct analyses, and the discrepancy means the sponsor’s own analysis takes precedence when describing the trial’s reported result.

Precision is affected by the amount of variation in the outcome and by the number of observations contributing to an estimate. The posted standard deviations show notable variation in participant-level HbA1c change, while the primary outcome summaries include 326 and 293 participants with available data. The registry does not report additional details that would permit a fuller assessment of precision, such as the ANCOVA covariates, missing-data methods or sensitivity analyses.

A confidence interval is not a prediction interval for an individual and does not imply that values outside its endpoints are impossible. It quantifies uncertainty in an estimated group contrast under the analysis used. The contrast’s statistical precision also does not resolve whether its magnitude is clinically important; the registry record does not report a clinical-importance threshold for this endpoint.

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 participants per arm for the continuous outcomes in this trial; they are not total event counts. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set as the stated fraction of the observed effect. These are planning illustrations computed from the observed effect, not the trial’s original design assumptions. The original sample-size justification and design assumptions are not reported in the registry record.

The 80% power column describes the number of participants per arm that the planning calculation estimates would give an 80% chance of detecting the specified effect under the stated assumptions, if that effect and the model assumptions were correct. The 90% power column gives the corresponding per-arm planning count for a higher chance of detection. Higher planned power generally requires more participants, all else being equal. The rows for 100%, 75% and 50% of the observed effect illustrate that smaller target effects require more participants per arm to detect.

For example, the primary endpoint illustration uses 789 participants per arm for 80% power and 1,056 participants per arm for 90% power when planning for 100% of the observed effect. For 75% of that effect, the corresponding figures are 1,402 and 1,877 per arm; for 50%, they are 3,154 and 4,223 per arm. These values are not a reconstruction of the completed trial’s design and should not be compared with enrollment as if they were an observed outcome.

This is not observed power. It is a prospective planning illustration based on an observed effect and specified assumptions; it does not measure the probability that the completed trial’s result is correct. The planning calculations also do not account for any features not stated in the assumptions, and the registry record does not report whether these illustrations reflect the considerations used by the trial designers.

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 = 0.210; pooled standard deviation σ = 1.488.
2. n = 2 × (1.96 + 0.842)² × 1.488² / 0.210² = 788.5, rounded up to 789 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 <- 0.21                      # observed difference in means
sigma <- 1.48831792243781                     # 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(5035082)
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 = 789 (80% power), 1,056 (90% power)
 75% of observed effect: n per arm = 1,402 (80% power), 1,877 (90% power)
 50% of observed effect: n per arm = 3,154 (80% power), 4,223 (90% power)
power.t.test: 789.4 
Simulated power at n = 789 per arm: 0.802 (20000 replicates)

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

14. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary analysis as an ANCOVA treatment difference of -0.24, with a 95% confidence interval from -0.44 to -0.04 and p = 0.018. It classifies this as a superiority analysis. Separately, a simple unadjusted recalculation from the posted group means and standard deviations gives a difference of 0.21, a 95% confidence interval from -0.025 to 0.445 and p = 0.08. The reported estimate and the recalculation differ, so they should not be presented as if they were the same analysis.

The recalculation uses only unadjusted summary statistics and is not the sponsor’s model-based analysis. The record does not describe the ANCOVA covariates, the exact treatment-difference coding or other model details, so it is not possible to identify why the estimates differ. The comparison may involve differences in adjustment or analytic specification, but the registry does not establish a specific explanation.

Because the discrepancy is flagged, the sponsor’s own analysis takes precedence in describing the trial’s reported primary result. That does not make the recalculation irrelevant: it shows that a simple comparison of the posted summaries yields a different estimate and level of statistical evidence. The appropriate reading is to keep the analyses distinct, report their methods and uncertainty accurately, and avoid treating the unadjusted result as a correction to the sponsor’s analysis.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies one primary outcome analysis and lists 10 secondary outcomes as posted. The secondary outcomes include several ways of defining HbA1c target attainment, body-weight measures and treatment satisfaction. When multiple outcomes or comparisons are assessed, the chance of obtaining at least one small p-value by chance can increase. A multiplicity strategy may control that risk, but the registry record does not report whether the trial used an adjustment, a hierarchy of testing or another approach to multiplicity.

Consequently, the sponsor-reported primary analysis has a different evidential role from the secondary comparisons. The secondary recalculations are unadjusted, and their results should not be treated as if each had been protected against the consequences of examining multiple outcomes. The record also does not report whether secondary endpoints were prespecified in a formal statistical analysis plan or how confirmatory claims were intended to be ordered.

The primary endpoint population is described as the full analysis set, including all randomized participants, while the number analyzed is the number with available outcome data. The posted primary summaries include 326 participants in the oral semaglutide group and 293 in the other-medication group. These counts are outcome-specific; the registry does not report a detailed accounting of missing measurements, withdrawals, treatment discontinuations or reasons participants did not contribute primary outcome data.

The missing-data method for the primary analysis is not reported in the registry record. The record also does not describe imputation, sensitivity analyses, assumptions about the relationship between missingness and outcome, or analyses comparing participants with and without observed outcomes. Without those details, it is not possible to assess how robust the result is to different missing-data assumptions.

The record describes the full analysis set as including all randomized participants, but also states that the number analyzed depends on available data for the outcome. This makes clear that the reported outcome summaries do not necessarily include all randomized participants. The registry’s brief description of routine visits and follow-up contacts gives context about data collection, but it does not establish how missing outcome information was addressed analytically.

16. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

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: pneumonia (1.0%), sepsis (1.0%), syncope (1.0%), fall (0.8%), chest pain (0.6%).

The registry reports serious adverse events for 500 participants at risk in the oral semaglutide group, with 51 affected, and for 492 participants at risk in the other oral glucose-lowering medication group, with 55 affected. It also reports other adverse events for 500 at risk in the oral semaglutide group, with 81 affected, and for 492 at risk in the other-medication group, with 46 affected. These are the counts as recorded; they should be read alongside the stated at-risk denominators and not as evidence of a causal difference.

The registry reports 3 deaths in the oral semaglutide group and 2 in the other-medication group. The listed serious-event terms and percentages are pneumonia (1.0%), sepsis (1.0%), syncope (1.0%), fall (0.8%), chest pain (0.6%) and nausea (0.6%). The record does not provide a treatment-group breakdown for those listed terms in the information, so they should not be assigned to either group.

Adverse-event counts alone do not establish that a medicine caused an event, and comparisons depend on definitions, follow-up and completeness of event reporting. The registry record does not provide further detail about event adjudication, exposure time or statistical comparisons of safety outcomes. These counts describe the posted safety information, not a conclusion about individual risk.

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

The sponsor-reported ANCOVA result indicates a statistically distinguishable treatment difference in the primary HbA1c analysis: the estimate is -0.24, its 95% confidence interval runs from -0.44 to -0.04, and the p-value is 0.018. The registry classifies the analysis as superiority. This is the trial’s reported primary result, and it is the appropriate result to foreground because the unadjusted recalculation differs from it.

The result does not mean that every participant had the same change, or that the estimated average describes each person’s experience. The group standard deviations show variation in observed change. Nor does statistical significance alone establish whether the magnitude is important in clinical practice. The registry does not report a threshold for a clinically meaningful HbA1c difference, so that judgment cannot be made from the reported p-value alone.

The unadjusted recalculation yields a different estimate and confidence interval that includes zero. That comparison is based on posted summary statistics and does not replace the sponsor’s ANCOVA analysis. The discrepancy underscores why model specification and analysis details matter. Since the record does not report the covariates or full statistical plan, it does not permit a complete explanation of the two results.

The comparison is between oral semaglutide and a broad group of other oral glucose-lowering medications, both used with metformin. It does not establish how semaglutide compares with any one named alternative medicine. The study was unmasked, and the registry describes a pragmatic setting with treatment prescribed by participants’ doctors. Those features are relevant context for understanding what was compared and how closely the trial reflects routine care.

Finally, the outcome summaries are based on available data and do not account in the posted record for every enrolled participant. The registry does not explain the missing outcome data or report sensitivity analyses. The findings therefore describe the posted primary analysis and its precision, while leaving questions about missingness, model choices and the broader applicability of the estimate unresolved.

18. Why This Trial Matters Statistically

This trial illustrates why statistical interpretation requires more than reading a p-value or comparing two group means. Its primary analysis is reported as an ANCOVA treatment difference, while a simple unadjusted recalculation from the posted summaries produces a different estimate and a different confidence interval. Keeping those analyses separate makes the role of the sponsor’s model and the limits of an independent summary-statistic calculation visible.

The example also shows the importance of understanding the analysis population. The registry reports total enrollment of 1,018 but defines the number analyzed for the primary outcome by availability of outcome data. The primary group summaries include 326 and 293 participants. These outcome-specific counts are not interchangeable with enrollment, and missing-data information is needed to assess how the analyzed population relates to the randomized population.

REALYSE is described as a pragmatic, randomized comparison in routine health care settings, with a broad comparator of other oral glucose-lowering medications. Such a design addresses a practical treatment-choice question, but the breadth of the comparator means the aggregate result is not a comparison with one single named medicine. The unmasked design and available-data analysis also matter when considering how the observed estimates were generated.

Finally, the registry posts multiple secondary outcomes alongside the primary endpoint. Without a reported multiplicity strategy, the secondary comparisons are best read as supporting context rather than definitive tests. This trial therefore offers a useful statistical case in interpreting model-based and unadjusted analyses, separating precision from clinical importance, and distinguishing enrollment from the number contributing outcome data.

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 of the primary analysis unspecified. It names ANCOVA but does not report the covariates, model specification, handling of missing data or sensitivity analyses. These omissions limit independent assessment of why the sponsor-reported estimate differs from the unadjusted recalculation and how robust the estimate is to analytic choices.

The outcome population is described as the full analysis set, but the number analyzed is defined by outcome availability. The posted primary summaries include 326 and 293 participants, while total enrollment was 1,018. The record does not provide a complete account of missing HbA1c outcomes, reasons for nonavailability or the relationship between missingness and treatment assignment or outcome. The difference between enrollment and the outcome-specific counts should not be converted into a missing-data rate without further information.

The comparator includes any other oral glucose-lowering medication. The registry does not report the individual medicines or their distribution in the comparator group in the information available for the results, so the group-level comparison cannot be attributed to a particular alternative drug. The trial was also unmasked, which may be relevant to participant and clinician behavior and to subjective outcomes such as treatment satisfaction; the record does not report analyses addressing that possibility.

The registry does not report the original sample-size justification, the trial’s design assumptions, a formal multiplicity strategy, or a threshold for a clinically meaningful primary difference. The planning illustrations are separate from the original trial design. Together, these limitations constrain conclusions about design adequacy, secondary outcome evidence and clinical importance, without changing the fact that the registry reports a primary ANCOVA result and a distinct unadjusted recalculation.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Research Study Comparing RYBELSUS® to Other Blood Sugar Lowering Tablets in People Living in America With Type 2 Diabetes (REALYSE) — NCT05035082.
• Registry record last updated 2026-08-10; status last verified 2026-07.
• Results first posted 2026-08-10.
• 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

REALYSE (NCT05035082) is a completed Phase 4 randomized, parallel-group, unmasked trial sponsored by Novo Nordisk A/S. It compared oral semaglutide with other oral glucose-lowering medication, both used in addition to metformin, in adults with type 2 diabetes. The registry reports enrollment of 1,018 and a primary completion date of 2025-06-18.

The primary endpoint was change in HbA1c from week 0 to year 1. The sponsor-reported ANCOVA treatment difference was -0.24 (95% confidence interval -0.44 to -0.04; p = 0.018), while the unadjusted recalculation from posted summaries differed. Because of that discrepancy, the sponsor’s own analysis takes precedence. The registry lists secondary outcomes for HbA1c targets, body weight and treatment satisfaction, along with adverse-event counts. Details about the original sample-size justification, multiplicity, missing-data handling and full statistical plan are not reported in the registry record.