Effect of semaglutide on liver enzymes and markers of inflammation in subjects with type 2 diabetes and/or obesity
- Design
- Design unknown · Biomarker outcome
- Match to healthy normal-weight adults aged 55–75
- Population unclear[Auto] Review/guidance/regulatory document rather than a primary study; applicability depends on the underlying studies.
- Could weight loss explain it?
- Not applicable[Auto] Not a primary outcome study.
Adjusted for: body weight - Study tier
- Not rated[Auto] Review, guidance, regulatory document or model without new primary outcome data.
- 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] Semaglutide significantly reduced ALT and hsCRP in clinical trials in subjects with obesity and/or type 2 diabetes.
01Findings
What the study reported
- Drugs
- Semaglutide
- Dose
- 1.0 mg/week
- Comparator
- placebo
- Primary outcome
- Not extracted
- Effect
- Not extracted
- 95% confidence interval
- Not extracted
- Follow-up
- Not stated
- Adverse events
- Not extracted
- Limitations
- Auto-classified from abstract only; effect estimates, adverse events and limitations not extracted. Requires manual review.
Who was studied
- Obesity status
- obesity/overweight present (all or most)
- Diabetes status
- type 2 diabetes present (all or most)
- Baseline condition
- MASH / MASLD
Study quality details
- Study design
- Unknown
- Sample size
- not extracted
- Randomization
- no
- Blinding
- not stated
- Comparator
- placebo
- Follow up duration
- not stated
- Outcome type
- biomarker
- Replication
- not assessed (auto)
- Consistency with other evidence
- not assessed (auto)
- Population applicability
- UNKNOWN
- Statistical precision
- not extracted
- Funding conflicts
- partial
- 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, treatment ratio vs placebo for change from baseline (log-transformed hs-CRP, MMRM), NOT adjusted for weight change · week 52 (end of treatment); week 28 also reported
High risk of bias · RoB 2 · Two reviewers agreed on every domain
How the overall was reached: D5 is High and drives the overall judgment: the assessed hs-CRP result is an explicitly post hoc analysis of a trial whose endpoint was weight, reported across multiple doses, timepoints and subgroups without multiplicity adjustment and with the most favourable cell in the headline. D3 adds Some concerns (hs-CRP availability unreported; structural exclusion of discontinuers' data). D1, D2 and D4 are Low, with D1 resting on a presumption rather than a verified concealment mechanism. Under the RoB 2 algorithm, High in any domain gives High overall, and there is no reason to override.
| Domain | Judgment and reasoning |
|---|---|
| D1 · Randomisation process Low risk of bias | The parent trial is identified by registration number and described as randomised, double-blind and placebo-controlled with a stated 6:1 allocation ratio and matched placebo of identical dosing volume and escalation schedule. Under guide ruling 6 (v1.2) the large-industry-trial presumption reaches the parent trial, so allocation concealment is PRESUMED from the parent trial's character, NOT verified: no randomization-sequence or concealment mechanism (IVRS/IWRS, central allocation) is described in any retrieved source, and the parent report (O'Neil 2018) and registry entry were not retrieved. The identifying source is this paper's Methods section 2.2, which names "NCT02453711". Baseline characteristics (Table 1) show no imbalance beyond chance, so the presumption is not rebutted. Randomization sequence generation and concealment are therefore PY on presumption, not on evidence; baseline imbalance N.NCT02453711 was a phase 2, randomised, double‐blind, multinational, placebo‐ and active‐controlled dose‐finding trial of semaglutide in combination with both dietary and exercise counselling. For all active treatment groups (semaglutide or liraglutide), participants were randomised 6:1 to active drug or a matched placebo of identical dosing volume and escalation schedule, and all placebo groups were pooled for analysis. |
| D2 · Deviations from intended interventions Low risk of bias | Double-blind with matched placebo of identical volume and escalation schedule. Per the collection guide, for lab biomarker outcomes awareness alone rarely biases the result; D2 is judged on discontinuation imbalance and ITT handling. The analysis explicitly used data irrespective of whether the subject was on trial medication, i.e. an assignment-based (ITT-style) approach with no per-protocol exclusion for non-adherence. Per-arm discontinuation counts are not reported (overall 19% did not complete 52 weeks of treatment), so a discontinuation imbalance cannot be ruled out; that concern is carried at D3 rather than here, since it affects data availability rather than the ITT principle of the analysis.All analyses of ALT and hsCRP in the weight management trial used data collected during the trial irrespective of whether the subject was on trial medication. participants were randomised 6:1 to active drug or a matched placebo of identical dosing volume and escalation schedule |
| D3 · Missing outcome data Some concerns about risk of bias | Outcome data availability for hs-CRP at week 52 is not reported at all — no per-arm n, no completer counts for the biomarker. What can be established is that 81% of randomized subjects received the full 52 weeks, and that the 12% "retrieved" participants who returned for week-52 weight assessment after early discontinuation had NO ALT or hs-CRP collected under the protocol. Missing hs-CRP is therefore at least ~19% of randomized subjects — just under the guide's 20% prong, but the prong cannot be evaluated precisely and the per-arm difference prong (>5 points) is NI because arm-level discontinuation is not reported. Missingness is structurally tied to treatment discontinuation, which for a GLP-1 RA correlates with drug exposure and GI adverse events, and no sensitivity analysis addresses value-dependent missingness for hs-CRP (the jump-to-reference multiple imputation described applies to the weight endpoint, not to hs-CRP). Not High: neither guide prong is affirmatively tripped. Not Low: the data needed to apply the decisive thresholds are absent.Overall, 81% (777/957) of subjects received the full 52 weeks of treatment, and week 52 weight data were also available for an additional 12% (115/957) of “retrieved” participants who returned for evaluation after early treatment discontinuation. However, data on these parameters were not collected under the study protocol from those retrieved participants who discontinued drug but returned for week 52 weight assessment. |
| D4 · Measurement of the outcome Low risk of bias | hs-CRP is an objective central-laboratory biomarker; per the collection guide, measurement bias is Low unless the assay changed partway through or measurement differed between arms, and there is no evidence of either. Central assessment is stated explicitly for ALT in both trials; for hs-CRP the paper does not state the assay location, but the measurement is a standardised immunoassay applied to a blinded, placebo-matched design, and nothing suggests arm-differential ascertainment.ALT at baseline and during the trial was assessed centrally in both trials Changes in high‐sensitivity C‐reactive protein (hsCRP) from baseline were analysed by baseline ALT subgroup in the weight management trial only, as this parameter was not assayed in the cardiovascular outcomes trial, using the same weight‐adjusted and ‐unadjusted mixed‐model approach and with the same factors as for ALT, but with log‐transformed baseline hsCRP replacing ALT as a covariate. |
| D5 · Selection of the reported result High risk of bias | Guide ruling 7 (v1.3): an author statement that the assessed analysis was post hoc or exploratory is affirmative evidence of non-prespecification, and D5 is High. The paper states in its own words that the whole analysis, hs-CRP included, is post hoc — "we report the results of a post hoc analysis evaluating the effect of semaglutide on levels of ALT and C-reactive protein" and "Data were drawn and analysed post hoc from two published clinical trials". hs-CRP was not an endpoint of NCT02453711 (whose primary endpoint was percentage weight change). No protocol, SAP or registry entry was retrievable to check prespecification, but under ruling 7 the authors' own statement settles it without one. Compounding: multiple selectable framings of the same result exist — two timepoints (weeks 28 and 52), two baseline-ALT subgroups, five dose groups, and further sex and age splits — with no multiplicity adjustment reported, and the headline "up to 43%" selects the most favourable cell. Per ruling 1, D5 judges this result; the framings are also noted in limitations.Herein, we report the results of a post hoc analysis evaluating the effect of semaglutide on levels of ALT and C‐reactive protein in subjects enrolled in two clinical trials of semaglutide treatment for obesity or type 2 diabetes, two conditions related to NASH. Data were drawn and analysed post hoc from two published clinical trials from the semaglutide development programme |
Assessed, not settled (1)
Recorded rather than omitted. Leaving these out would make the appraisal look cleaner than it is, and a disagreement between two careful readers is itself worth knowing.
hs-CRP, treatment ratio vs placebo for change from baseline, ADJUSTED for change in body weight (weight-adjusted MMRM) — the weight-independence result for C-001 · week 52 (end of treatment); week 28 also modelled
Assessed, not settled · RoB 2 · Two reviewers reached different judgments and nothing has settled it
This result has no overall judgment, because the two independent reviews did not reach one. Both readings are shown. The domains they did agree on are below.
| Where they differ | The two readings |
|---|---|
| D2 · Deviations from intended interventions | One reviewer: Low risk of bias The other: Some concerns about risk of bias |
| Domain | Judgment and reasoning |
|---|---|
| D1 · Randomisation process Low risk of bias | Identical to O1, and deliberately so: per guide ruling 4, D1 stays trial-level (orthodox RoB 2) and the non-randomized character introduced by conditioning on a post-randomization variable is NOT written into D1. It is carried at D5 and in the overall judgment below. The parent trial NCT02453711 is identified by registration number in Methods 2.2, and allocation concealment is PRESUMED from the parent trial's character under ruling 6 — presumed, not verified: no concealment mechanism is described in any retrieved source, and the parent report and registry entry were not retrieved. No baseline imbalance rebuts the presumption.NCT02453711 was a phase 2, randomised, double‐blind, multinational, placebo‐ and active‐controlled dose‐finding trial of semaglutide in combination with both dietary and exercise counselling. For all active treatment groups (semaglutide or liraglutide), participants were randomised 6:1 to active drug or a matched placebo of identical dosing volume and escalation schedule |
| D2 · Deviations from intended interventions Low risk of bias | As for O1: double-blind with matched placebo, and analyses used data irrespective of on-treatment status, so the analysis retains an assignment-based population. The weight adjustment does not exclude non-adherers; it conditions on a measured covariate. That conditioning is a different problem — it breaks the randomization balance for the adjusted contrast — but under ruling 4 it is not a D2 deviation-from- intervention question and is carried at D5 and overall.All analyses of ALT and hsCRP in the weight management trial used data collected during the trial irrespective of whether the subject was on trial medication. This weight‐adjusted model used the same fixed factors and log‐transformed baseline ALT covariate as the unadjusted model, plus baseline body weight and change from baseline body weight as additional covariates. |
| D3 · Missing outcome data Some concerns about risk of bias | Same missingness structure as O1, and slightly worse: the weight-adjusted model requires both an hs-CRP value and a body-weight change value at the visit, so a subject missing either drops out of the adjusted contrast. Availability is reported for neither. At least ~19% of randomized subjects did not complete treatment, and hs-CRP was not collected from the retrieved discontinuers at all; per-arm figures are NI. No sensitivity analysis for value-dependent missingness. Neither guide prong is affirmatively tripped, so not High; the thresholds cannot be applied, so not Low.Overall, 81% (777/957) of subjects received the full 52 weeks of treatment However, data on these parameters were not collected under the study protocol from those retrieved participants who discontinued drug but returned for week 52 weight assessment. |
| D4 · Measurement of the outcome Low risk of bias | The measured outcome is unchanged from O1 — the same objective hs-CRP assay in the same blinded design — and the guide rates central-lab hs-CRP Low absent an assay change or arm-differential measurement. The adjustment affects the estimand, not the measurement, so it does not belong here.Changes in high‐sensitivity C‐reactive protein (hsCRP) from baseline were analysed by baseline ALT subgroup in the weight management trial only ALT at baseline and during the trial was assessed centrally in both trials |
| D5 · Selection of the reported result High risk of bias | Three things converge, and each alone would suffice for at least Some concerns. (1) Ruling 7: the authors state the analysis is post hoc, and describe the weight-adjusting analysis specifically as "the exploratory analysis adjusting for change in body weight". That is affirmative evidence of non-prespecification and reaches High. (2) Ruling 4 carry-forward: the adjusted contrast conditions on change in body weight, a POST-RANDOMIZATION variable. The adjusted comparison is therefore no longer a randomized comparison — it is an observational contrast within levels of an on-treatment outcome, open to collider/mediator conditioning. Per ruling 4 this derived character is not written into D1; it is recorded here, in writing, as the reason the selected estimand is not the randomized one. (3) The result is reported only narratively. Nothing quantitative is available, so the reader cannot check whether "clustered around 1.0" describes a genuinely null set of ratios or a materially non-null but under-powered one, and cannot tell which dose, timepoint or baseline-ALT subgroup is being summarised. Selective reporting from the multiple available framings cannot be excluded — it cannot even be examined — and the summary that survives into the text is the one that supports the paper's thesis that the effect is weight-mediated.As with the ALT analysis, statistical significance was lost and all treatment ratios clustered around 1.0 when adjusted for change in body weight (Figure 5 B). suggested by the loss of a significant treatment effect and the clustering of treatment ratios around 1.0 in the exploratory analysis adjusting for change in body weight |
02Funding
Funding and conflicts
- Funding
- NCATS NIH HHS; Novo Nordisk
- Industry funded
- Partial
- Manufacturer
- Novo Nordisk
- Sponsor role
- mixed industry and public/foundation funding
- Author conflicts
- not available in metadata
- Independent replication
- unknown
- Notes
- Authors declare relationships with the drug's manufacturer: Novo Nordisk
Funding is shown on every study and never used to score it.
03Claims
Claims this study bears on
- GLP-1 receptor agonist treatment reduces systemic inflammatory markers beyond what can be explained by treatment-associated weight loss.Contradicts
hs-CRP significance lost on adjustment for weight change; no adjusted estimate or interval published.
04Source
The source, as retrieved
Abstract
[BACKGROUND] Obesity and type 2 diabetes are drivers of non-alcoholic fatty liver disease (NAFLD). Glucagon-like peptide-1 analogues effectively treat obesity and type 2 diabetes and may offer potential for NAFLD treatment. [AIM] To evaluate the effect of the glucagon-like peptide-1 analogue, semaglutide, on alanine aminotransferase (ALT) and high-sensitivity C-reactive protein (hsCRP) in subjects at risk of NAFLD. [METHODS] Data from a 104-week cardiovascular outcomes trial in type 2 diabetes (semaglutide 0.5 or 1.0 mg/week) and a 52-week weight management trial (semaglutide 0.05-0.4 mg/day) were analysed. Treatment ratios vs placebo were estimated for ALT (both trials) and hsCRP (weight management trial only) using a mixed model for repeated measurements, with or without adjustment for change in body weight. [RESULTS] Elevated baseline ALT (men >30 IU/L; women >19 IU/L) was present in 52% (499/957) of weight management trial subjects. In this group with elevated ALT, end-of-treatment ALT reductions were 6%-21% (P<0.05 for doses≥0.2 mg/day) and hsCRP reductions 25%-43% vs placebo (P < 0.05 for 0.2 and 0.4 mg/day). Normalisation of elevated baseline ALT occurred in 25%-46% of weight management trial subjects, vs 18% on placebo. Elevated baseline ALT was present in 41% (1325/3268) of cardiovascular outcomes trial subjects. In this group with elevated ALT, no significant ALT reduction was noted at end-of-treatment for 0.5 mg/week, while a 9% reduction vs placebo was seen for 1.0 mg/week (P = 0.0024). Treatment ratios for changes in ALT and hsCRP were not statistically significant after adjustment for weight change. [CONCLUSIONS] Semaglutide significantly reduced ALT and hsCRP in clinical trials in subjects with obesity and/or type 2 diabetes.
Where this record came from
| Source | Retrieved | Identifier |
|---|---|---|
| pubmed | Sep 13, 2026 | 31246368 first ingestion |