Oral Semaglutide and Change in Cardiovascular Risk Factors in High-Risk Type 2 Diabetes: A Post Hoc Secondary Analysis of the SOUL Randomized Clinical Trial
- Design
- Randomized trial · 9495 participants · Mixed outcome
- Match to healthy normal-weight adults aged 55–75
- Different population[Auto] Participants had type 2 diabetes (age/BMI not reported in abstract).
- Could weight loss explain it?
- Possibly[Auto] Participants had obesity and/or type 2 diabetes and the abstract does not separate direct drug effects from weight loss or glycaemic improvement.
- Study tier
- Study tier 1[Auto] Large randomized trial with clinical outcomes (auto-provisional; risk of bias and consistency not yet assessed).
- Assessment
- Version 1 · automatic, not yet reviewed by a person · Sep 13, 2026
Study facts come from the paper. Population match, weight-loss explanation and study tier are our judgments, made against the reference group of healthy normal-weight adults aged 55–75.
[Auto, unreviewed; quoted from abstract conclusions] In this post hoc secondary analysis of the SOUL randomized clinical trial, oral semaglutide was associated with early and sustained improvements vs placebo in multiple ASCVD risk factors in high-risk participants with T2D and ASCVD and/or CKD, incremental to SoC.
01Findings
What the study reported
- Drugs
- Semaglutide
- Dose
- 14 mg
- Route
- oral
- Comparator
- placebo
- Primary outcome
- Not extracted
- Effect
- Not extracted
- 95% confidence interval
- Not extracted
- Follow-up
- 13 weeks
- Adverse events
- Not extracted
- Limitations
- Auto-classified from abstract only; effect estimates, adverse events and limitations not extracted. Requires manual review.
Who was studied
- Diabetes status
- type 2 diabetes present (all or most)
- Cvd status
- cardiovascular disease present in population (see abstract)
- Ckd status
- chronic kidney disease present in population (see abstract)
- Baseline condition
- chronic kidney disease
- Sample size
- 9495
Study quality details
- Study design
- Randomized controlled trial
- Sample size
- 9495
- Randomization
- yes
- Blinding
- double-blind
- Comparator
- placebo
- Follow up duration
- 13 weeks
- Outcome type
- mixed
- Replication
- not assessed (auto)
- Consistency with other evidence
- not assessed (auto)
- Population applicability
- INDIRECT
- Statistical precision
- CI reported: 95% CI, -0
- Funding conflicts
- unclear
- Peer review status
- yes
01bRisk of bias
How much the result can be relied on
Risk of bias is judged per result, not per paper: the same study can report one marker at low risk and another at high. Two reviewers assess each result independently against a written guide, and every domain judgment below carries its reasoning and the sentences it rests on.
These words are not quality scores. On the RoB 2 scale, High risk of bias is the worst rating a result can receive and Low risk of bias is the best — the opposite of how the same words read on some other scales. Levels are always written out in full here for that reason.
hs-CRP (hsCRP plasma level), change vs placebo, expressed as estimated treatment ratio (ETR) · Week 104 (the latest and only long-term hsCRP timepoint; week 13 also reported)
Some concerns about risk of bias · RoB 2 · Two reviewers differed; settled by a recorded ruling
How the overall was reached: D1-D4 Low; D5 Some concerns drives the overall judgment. The randomized comparison, near-complete follow-up, ITT handling and central-lab assay are all sound; the residual concern is that this is a self-declared post hoc, multiplicity-unadjusted analysis whose prespecification could not be checked against the protocol/SAP from the sources available to this pass.
| Domain | Judgment and reasoning |
|---|---|
| D1 · Randomisation process Low risk of bias | Individually randomized 1:1 in a large multicentre industry phase 3b double-blind placebo-controlled trial; baseline characteristics including hsCRP are reported as well balanced across arms and the baseline table shows identical median hsCRP (0.2 mg/dL) in both arms. The bundle does not describe the sequence-generation mechanism or an IVRS/IWRS, so allocation concealment is taken as probably adequate rather than documented (guide D1 rule for large industry registration trials); nothing in the available sources contradicts adequate randomization.Participants, all receiving standard of care (SoC) for CV risk mitigation and for glucose management, were randomized in a 1:1 ratio to once-daily treatment with either oral semaglutide or matching placebo. Participant demographics and clinical characteristics, including CV risk factor profile, were well balanced at baseline. |
| D2 · Deviations from intended interventions Low risk of bias | Effect of interest is assignment. The assessed estimate is an ITT analysis of the full analysis set including all randomized participants regardless of adherence. Trial completion was near-universal and essentially equal by arm (98.4% overall), so there is no discontinuation imbalance to bias the assignment effect. Per the collection's D2 rule, potential unblinding by GI effects is not itself a downgrade for a central-lab biomarker. No arm-specific treatment-discontinuation figures are reported in the available sources; this is recorded as a limitation, not a downgrade, since the assessed analysis is ITT with imputation fitted regardless of treatment status.The analyses of CV risk factors in the overall trial population were intention-to-treat (ITT) analyses using in-trial data and therefore included all randomized participants (full analysis set) regardless of treatment adherence. 9650 individuals (eFigure 1 in Supplement 2 ) were randomized (4825 in each arm), and 9495 participants (98.4%) completed the trial (attended end-of-trial visit or had died), with a mean (SD) follow-up of 47.5 (10.9) months. |
| D3 · Missing outcome data Low risk of bias | Neither prong of the guide's D3 rule trips. Missingness is reported as approximately 10.1% vs 11.5% by arm (body weight at week 156), i.e. well under 20%, with a 1.4-point arm difference, well under 5 points; the paper states other risk factors, which includes hsCRP at week 104, had similar proportions missing. Missingness is attributed mainly to COVID-19 visit conversion/omission, a cause unrelated to the true hsCRP value. Under calibration ruling 2, the absence of an MNAR sensitivity analysis is a limitations note rather than a downgrade. Caveat recorded in limitations: hsCRP-specific per-arm missingness is not given numerically.Data were collected from most patients for most end points at most time points, although during the COVID-19 pandemic, some site visits were converted to telephone visits or missed, accounting for the majority of the missing data. As an example, 10.1% of data for body weight at week 156 were missing for the oral semaglutide arm, with 11.5% missing for the placebo arm. Similar proportions of data were missing for the other risk factors at either week 156 or week 104. Before analyses, missing data were imputed as follows: the imputation model (linear regression) was done separately for each treatment arm and included baseline value as a covariate and was fitted to all participants with a measurement regardless of treatment status at week 156 (week 104 for hsCRP plasma level). |
| D4 · Measurement of the outcome Low risk of bias | hsCRP is an objective central-laboratory assay, measured identically in both arms by a single central laboratory; no indication of an assay change during the trial or of arm-differential measurement. Guide D4 rule: Low.All protocol-required blood tests, including HbA 1 c and plasma levels of lipids and hsCRP, were collected at study sites and analyzed in a central laboratory to ensure consistency and accuracy of the measurements. plasma levels of hsCRP and lipids (total cholesterol [TC], non–high-density lipoprotein cholesterol [non–HDL-C], HDL-C, low-density lipoprotein cholesterol [LDL-C], and triglycerides; all measured in a nonfasting state) |
| D5 · Selection of the reported result Some concerns about risk of bias | The paper affirmatively labels the hsCRP analysis post hoc and states it was not adjusted for multiplicity; hsCRP is listed among "post hoc analyses" rather than among the supportive secondary end points (which are only HbA1c and body weight). The guide's D5 rule makes not-prespecified or unverifiable-prespecification at least Some concerns. It does not reach High under calibration ruling 3: the protocol and SAP (Supplement 1) were not available to this pass, so prespecification cannot be affirmatively excluded, the paper makes no prespecification claim, and the result is reported at both available timepoints (weeks 13 and 104) with a single consistent ETR framing rather than a menu of transformations selected for emphasis. Some concerns is carried because the timepoint and model choice cannot be verified against the SAP.First, the analyses were post hoc, not adjusted for multiplicity, and thus should be used as a prompt for further validation studies. Supportive secondary end points included change in HbA 1 c and change in body weight. Post hoc analyses included BP (systolic BP [SBP], diastolic BP [DBP], pulse pressure), pulse, plasma levels of hsCRP and lipids |
Inflammatory (hsCRP) change not explained by weight change — assessed as the paper's only weight-relevant framing: the temporal contrast between the early hsCRP reduction and the gradual body-weight reduction · Week 13 hsCRP (−18.08%; ETR 0.81, 95% CI 0.78-0.84) set against week 13 body weight (−2.54%; ETD −2.28 percentage points) with weight nadir at week 52
High risk of bias · RoB 2 · Two reviewers differed; settled by a recorded ruling
How the overall was reached: D1-D4 Low but D5 High, and the derived contrast is descriptive rather than an estimated weight-independent effect: no weight-matched comparator, no mediation estimate, no weight-adjusted hsCRP model, no statistical test of the temporal dissociation, and no prespecification. The result can suggest but cannot establish that any part of the hsCRP reduction is independent of weight loss.
| Domain | Judgment and reasoning |
|---|---|
| D1 · Randomisation process Low risk of bias | As O1 — the underlying contrast is the randomized trial-level comparison. Per calibration ruling 4, the non-randomized character of a derived temporal contrast is carried in D5 and the overall judgment, not in D1.Participants, all receiving standard of care (SoC) for CV risk mitigation and for glucose management, were randomized in a 1:1 ratio to once-daily treatment with either oral semaglutide or matching placebo. Participant demographics and clinical characteristics, including CV risk factor profile, were well balanced at baseline. |
| D2 · Deviations from intended interventions Low risk of bias | Both component trajectories come from the ITT full analysis set including all randomized participants regardless of adherence, with near-complete trial completion and no reported discontinuation imbalance.The analyses of CV risk factors in the overall trial population were intention-to-treat (ITT) analyses using in-trial data and therefore included all randomized participants (full analysis set) regardless of treatment adherence. 9495 participants (98.4%) completed the trial (attended end-of-trial visit or had died) |
| D3 · Missing outcome data Low risk of bias | Same missing-data structure as O1 for both components (hsCRP and body weight); neither the 20% nor the 5-point prong trips (10.1% vs 11.5%), and the stated cause (COVID-19 visit disruption) is not value-dependent. Calibration ruling 2 applies.As an example, 10.1% of data for body weight at week 156 were missing for the oral semaglutide arm, with 11.5% missing for the placebo arm. Similar proportions of data were missing for the other risk factors at either week 156 or week 104. |
| D4 · Measurement of the outcome Low risk of bias | hsCRP is a central-laboratory assay; body weight is a routine objective measurement taken identically in both arms. No assay change or arm-differential measurement reported.All protocol-required blood tests, including HbA 1 c and plasma levels of lipids and hsCRP, were collected at study sites and analyzed in a central laboratory to ensure consistency and accuracy of the measurements. |
| D5 · Selection of the reported result High risk of bias | The weight-independence framing is load-bearing for C-001 and is affirmatively not a prespecified analysis: the paper states the analyses were post hoc and not adjusted for multiplicity, and the Outcomes section lists no weight-independence, mediation or weight-adjusted analysis at all. The contrast rests on comparing two separately estimated trajectories at a timepoint chosen after the data were seen, with no test statistic and no reported prespecified hypothesis; the hsCRP schedule itself (baseline, weeks 13 and 104 only) constrains which contrast can be shown. Under calibration ruling 3, absent prespecification on a load-bearing derived analysis is High rather than Some concerns. Note the paper does not itself claim prespecification, and the protocol/SAP were not available to this pass; High is carried on the paper's own affirmative post hoc declaration plus the complete absence of any weight-independence analysis from the Outcomes and Statistical Analysis sections.First, the analyses were post hoc, not adjusted for multiplicity, and thus should be used as a prompt for further validation studies. ETDs and ETRs were determined using analysis of covariance (ANCOVA) models with treatment as a fixed factor and baseline value as a covariate. |
02Funding
Funding and conflicts
- Funding
- not reported in abstract
- Industry funded
- Unclear
- Manufacturer
- None identified
- Sponsor role
- not reported in abstract
- Author conflicts
- not available in metadata
- Independent replication
- unknown
Funding is shown on every study and never used to score it.
04Source
The source, as retrieved
Abstract
[IMPORTANCE] Individuals with type 2 diabetes (T2D) are at high risk of atherosclerotic cardiovascular disease (ASCVD). In the SOUL randomized clinical trial, once-daily oral semaglutide reduced risk of major adverse cardiovascular (CV) events by 14% vs placebo in people with T2D and ASCVD and/or chronic kidney disease (CKD) receiving standard of care (SoC); however, whether oral semaglutide modifies recognized CV risk factors in the long term is unclear. [OBJECTIVE] To investigate whether treatment with oral semaglutide was associated with changes in ASCVD risk factors vs placebo. [DESIGN, SETTING, AND PARTICIPANTS] This secondary analysis comprises post hoc intention-to-treat analyses of the SOUL (A Heart Disease Study of Semaglutide in Patients With Type 2 Diabetes) double-blind multicenter randomized clinical trial (randomization 1:1 to oral semaglutide or placebo) among adults with T2D and ASCVD and/or CKD receiving SoC. Participants underwent randomization from June 2019 to March 2021, with a mean (SD) of 47.5 (10.9) months of follow-up, and data were analyzed from February to December 2025. [INTERVENTION(S)] Participants were treated with either once-daily oral semaglutide (maximum dose, 14 mg) or placebo, in addition to standard care. [MAIN OUTCOMES AND MEASURES] The primary outcome was the association of oral semaglutide vs placebo with glycated hemoglobin (HbA1c), body weight, and blood pressure (BP) using estimated treatment differences (ETDs) and with high-sensitivity C-reactive protein (hsCRP) and lipid plasma levels using estimated treatment ratios (ETRs). [RESULTS] Of 9650 randomized participants (mean [SD] age, 66.1 (7.6) years; 2790 female participants [28.9%]), 9495 participants (98.4%) completed the trial. Early (13 weeks) improvements in HbA1c (-0.87 percentage points), body weight (-2.54%), systolic BP (SBP, -3.84 mm Hg), pulse pressure (-3.81 mm Hg), hsCRP (-18.08%), total cholesterol (TC, -7.00%), non-high-density lipoprotein cholesterol (non-HDL-C, -8.02%), HDL-C (-4.49%), and triglycerides (-8.15%) were observed with oral semaglutide vs placebo and sustained over the trial duration. Body weight reductions were gradual across both groups. At week 156, in favor of oral semaglutide were ETDs for HbA1c (-0.47 percentage points; 95% CI, -0.52 to -0.42), body weight (-3.26 percentage points; 95% CI, -3.55 to -2.98), SBP (-1.83 mm Hg; 95% CI, -2.47 to -1.18), and pulse pressure (-2.17 mm Hg; 95% CI, -2.72 to -1.61) and ETRs for hsCRP (0.77; 95% CI, 0.74-0.81), TC (0.99; 95% CI, 0.98-1.00), non-HDL-C (0.98; 95% CI, 0.97-0.99), HDL-C (1.01; 95% CI, 1.01-1.02), and triglycerides (0.94; 95% CI, 0.93-0.96). No significant treatment differences were observed for low-density lipoprotein cholesterol or diastolic BP. [CONCLUSIONS AND RELEVANCE] In this post hoc secondary analysis of the SOUL randomized clinical trial, oral semaglutide was associated with early and sustained improvements vs placebo in multiple ASCVD risk factors in high-risk participants with T2D and ASCVD and/or CKD, incremental to SoC. [TRIAL REGISTRATION] ClinicalTrials.gov Identifier: NCT03914326.
Where this record came from
| Source | Retrieved | Identifier |
|---|---|---|
| pubmed | Sep 13, 2026 | 41879791 first ingestion |