Semaglutide carbon model: what it found

A 2026 modelling study estimated that adding once-weekly semaglutide 2.4 mg to diet and exercise could slightly reduce lifetime healthcare carbon emissions in two UK obesity populations, while increasing them in a third population with established cardiovascular disease. These were computer-model projections built from earlier trials and other inputs. The researchers did not measure emissions or health outcomes in a new group of patients^[1].

What the study set out to estimate

The authors asked whether obesity management with semaglutide 2.4 mg, added to diet and exercise, would change a patient's modelled lifetime carbon footprint compared with diet and exercise alone. They used the Core Obesity Model, a Markov state-transition cohort model. In this design, a hypothetical cohort moves between health states over time according to specified probabilities. The output depends on those transition assumptions and on the data used to populate them^[1]. It is not a direct environmental measurement of semaglutide treatment.

Clinical inputs came from the STEP and SELECT trial programmes. Carbon emissions were estimated through resource-based and cost-based methods. The outcomes included life-years, quality-adjusted life-years and an incremental carbon-footprint effectiveness ratio^[1]. The broader policy setting is real: NHS England has published modelling and trajectories for reducing the health service's carbon footprint^[2]. That policy goal does not validate this particular disease model or its semaglutide assumptions.

The three populations were not interchangeable

The first population included people with a body mass index of at least 30, or at least 27 with one or more obesity-related complications. A second group required a BMI of at least 27 plus a complication and included type 2 diabetes. Participants modelled in the third group had a BMI of at least 27 with established cardiovascular disease^[1]. These definitions matter because the model produced different environmental results in the cardiovascular group. The semaglutide evidence summary should be read with the same population-level care.

The comparison was semaglutide 2.4 mg plus diet and exercise versus diet and exercise alone. That is a modelled treatment strategy, not advice about dose or eligibility. Semaglutide is a prescription medicine, and UK access has clinical and regulatory conditions. Our UK semaglutide regulation page covers that separate question.

Lower modelled emissions in two groups

For the broad BMI30+/BMI27+C population, the model projected 0.30 additional life-years and 0.51 additional quality-adjusted life-years with semaglutide. Lifetime emissions were 9,850 kg of carbon-dioxide equivalent, compared with 10,030 kg for diet and exercise alone. The difference was reported as a 1.8% reduction^[1]. These figures describe model output per patient, not observed emissions recorded during STEP or SELECT. The semaglutide clinical evidence and this carbon calculation answer different questions.

In the BMI27+C population, the model projected 0.25 additional life-years and 0.46 additional quality-adjusted life-years. Estimated lifetime emissions were 9,792 kg CO2e with semaglutide and 9,983 kg with diet and exercise alone, a reported reduction of 1.9%^[1]. The authors attributed the net decrease to fewer obesity-related complications offsetting emissions from manufacturing semaglutide. That explanation comes from the model structure and its inputs; it was not verified by tracking individual supply chains and patient journeys. The UK prescription context does not supply that missing environmental evidence.

Why emissions increased in the CVD group

The cardiovascular-disease population moved in the other direction. Here, the model projected 0.56 additional life-years and 0.59 additional quality-adjusted life-years, but lifetime emissions rose to 14,700 kg CO2e from 14,444 kg with diet and exercise alone^[1]. The paper's abstract attributes the increase to longer survival and more monitoring. This is not presented as a safety harm. It shows that a lifetime carbon total can rise when the model projects that people live longer and use healthcare for longer. Clinical outcomes remain covered separately in our semaglutide research guide.

That result is the most useful check on a simple green-versus-not-green headline. A treatment can be associated with projected health gains while adding lifetime emissions in one population. Carbon totals also depend on the time horizon, care pathway and which downstream services the model counts. The authors report that scenario and sensitivity analyses supported their results^[1], but the PubMed record does not give enough detail to reproduce each scenario. For prescription context, see the semaglutide regulation overview.

The funding and transparency limits matter

Two authors were Novo Nordisk employees and shareholders. Another sat on the company's Sustainability Advisory Council under contract. Several authors worked for IQVIA, while the York Health Economics Consortium author reported that the organisation had received Novo Nordisk funding during the previous three years^[1]. A conflict disclosure does not make a result false. It does make independent scrutiny of model choices especially important, because small differences in emission factors and transition assumptions can shape a small percentage change.

The PubMed record says supporting data are available from the corresponding author on request but are not publicly available because of privacy or ethical restrictions. It also says code is not available^[1]. That limits independent reproduction from the public record. NHS England's net-zero report describes system-wide carbon modelling and the pace of change^[2], but it does not independently confirm the semaglutide estimates. Readers can review the semaglutide safety record separately from environmental claims.

What the result can and cannot support

The study shows one way to include greenhouse-gas estimates in health-economic disease modelling. It suggests that avoided complications may offset treatment-related emissions under some assumptions, while survival and monitoring can raise lifetime emissions in another population. It does not establish the measured carbon footprint of every semaglutide prescription, clinic or supply chain. Nor does it compare semaglutide with every other obesity treatment. The drug's established boundaries are summarised on the semaglutide page.

The model should not drive an individual's treatment choice. Clinical suitability depends on approved indications, contraindications, possible adverse effects and a person's health goals. Environmental assessment is a health-system question that needs transparent assumptions and comparisons. Anyone considering a prescription should discuss the clinical evidence with a qualified professional and use the semaglutide overview as educational background only.

What we do not yet know

The public abstract does not provide each resource quantity, emission factor, transition probability or scenario result. It does not show how the estimates would change with different discontinuation patterns, medicine wastage, delivery systems, energy mixes or future decarbonisation of NHS services. The projected reductions in the first two populations were under 2%, so close examination of those assumptions matters^[1].

Independent replication using disclosed code and inputs would strengthen the finding. Direct product life-cycle data and real-world healthcare use could then test whether the model's direction holds outside its chosen assumptions. For now, the careful conclusion is mixed: the model estimated modestly lower lifetime emissions in two UK populations and higher emissions in the established-CVD population. It is a modelling result, not a measured environmental benefit attached to every use of semaglutide.