Potential Lives Saved Through Widespread Global Availability of GLP-1 Receptor Agonists: A Modeling Study

Introduction

Obesity and T2DM are major risk factors for cardiovascular disease, the leading cause of mortality globally. GLP-1 receptor agonists, like semaglutide and tirzepatide, have been shown to effectively reduce body weight and improve glycemic control. Semaglutide has also been shown to reduce mortality (1, 2).

Recently, Pandey et al estimated the number of lives that could be saved by expanded access to weight loss drugs in the United States (3). Health equity advocates have also highlighted access disparities in obesity treatments for African Americans (4). However, to our knowledge there is no estimate of the number of lives that could potentially be saved globally if GLP-1s were more broadly available, nor an advocacy movement for global equitable access.

In this study, we developed a model to estimate the number of lives that could be saved globally through the widespread and equitable availability of GLP-1s. In developing this model, we drew on the most recent peer-reviewed sources to estimate the prevalence and impact on mortality of obesity, T2DM and cardiovascular disease, noting the disparate geographic scope, timespan and populations of these studies. We estimated reductions in all-cause mortality based on the findings of the SELECT and FLOW trials of semaglutide (1,2). These limitations introduce a level of uncertainty in the findings, however, our intention in building this model is not to provide a precise measure of potential benefits, but rather to better understand their magnitude. We also wanted to publicly post a simple and transparent model that others could improve, and update as new data becomes available.

Methods

Discussion

Our model estimates that widespread global availability of GLP-1 receptor agonists could save approximately 943,000 lives annually. This figure underscores the profound public health impact these drugs could have, particularly in reducing all-cause mortality associated with cardiovascular disease (CVD) deaths, obesity and T2DM. Currently, however, the benefits of GLP-1s are concentrated in high-income countries, with limited access in low- and middle-income countries (LMICs).

This estimate of lives saved from GLP-1s falls between the estimated lives saved from childhood vaccines of 154 m lives over 50 years (3 million per year), and HIV drugs of 16.5 m lives over 20 years (0.825 million per year) (5, 6). As such, equitable access to GLP-1s represents an opportunity to save lives on a large scale, emphasizing the need for urgent action to expand access to these treatments.

Countries with the most potential lives saved are populous countries with high numbers of people with diabetes and obesity: China, India, US, Pakistan, and Brazil (7, 8). Those with the greatest potential reduction in prevalence are Pacific small island developing states with obesity prevalence up to 70% (8). Focusing on both is essential if we want equity among countries.

From a public health policy perspective, our model also shows that targeting those with both T2DM and obesity could be impactful, because of their high baseline event rates. And that developing evidence on the group with CVD but no T2DM or obesity could also be impactful, because of the large population in this subgroup.

While our model provides a useful estimate of the potential impact of GLP-1s on all-cause mortality in target populations, it has limitations. The mortality impacts of obesity, T2DM, and CVD are estimated based on the best available peer-reviewed evidence, but further refinement will be needed as more data becomes available. Additionally, rates of reductions in mortality are based on publicly available data from a limited number of trials of semaglutide, which do not provide data on its impact on all combinations of co-morbidities. All-cause mortality was a secondary outcome in these trials. Finally, the distribution of imputed risk in the population may not precisely match the participants in the clinical trials. This is a global model, and a country-by-country analysis may provide greater precision and added insights.

We used conservative figures in our estimation of mortality benefit. For example, the FLOW trial was a randomized trial of people with diabetes (and chronic renal failure) which showed a reduction in all-cause mortality of 20% (2). A retrospective cohort study of use of GLP1s in Taiwan showed a reduction in all-cause mortality of 52% (9). We used the more conservative figure of 20%; had we used the higher figure our lives saved estimate would be larger. At this point, it’s unclear as to why the two values differ to the extent that they do. In the absence of more rigorous analysis, we made the conservative decision to use the lower figure from the randomized trial in our model.

Newer GLP-1s such as tirzepatide, often in combination with GIP and Glucagon receptor effects, have been shown to have even more significant effect on weight loss, although whether this leads to greater reductions in all-cause mortality is currently unknown. Moreover, GLP-1s are also being studied (and show preliminary but promising results) in the treatment of a wider range of conditions beyond CVD. (10)

Our goal was not to provide a comprehensive model of reducing mortality; that has been done recently in an outstanding analysis (11). From a policy perspective, some will argue that existing interventions such as taxes on sugar sweetened beverages, front of package labelling, and regulation of ultra-processed foods are more acceptable and/or cost-effective approaches to reduce mortality from obesity, T2DM and CVD in LMICs. There is evidence that these approaches are effective but, in many countries, they are already being used and obesity and T2DM rates continue to rise. This suggests the need for further interventions to curb and, ideally, reverse these rates.

Others will argue that availability of less expensive cardiovascular drugs such as anti-hypertensives should be prioritised instead of GLP-1s. Since the mechanisms of all these treatments are independent, and since currently available approaches are not stemming the rise in obesity and T2DM (and presumably the corresponding mortality) in most countries, we suggest the broader and more equitable GLP-1 access should also be pursued. Indeed, there is early evidence that in the US, the introduction of GLP1s has led to “peak obesity,” reversing this trend for the first time (12).

Modelling the effects of widespread global access is not the same as providing it. There are multiple barriers to access that would need to be overcome, including manufacturing, affordability, finance, regulatory, and intellectual property. While beyond the scope of this paper, we note the dynamic aspects of this market, including the expiry of the patent on semaglutide in India and China in 2026 (13, 14), with domestic firms preparing to manufacture it.

The precautionary principle suggests that, even in the face of uncertainty, immediate action should be taken to expand access to GLP-1s. Delaying broader access to these drugs risks could lead to the unnecessary loss of up to a million lives each year. While future trials and research may refine our understanding of these drugs’ impact and, hence, their most impactful deployment, the urgency to act now is clear.

Conclusion

Expanding global access to GLP-1 receptor agonists offers a unique opportunity to reduce mortality associated with obesity, T2DM and CVD. Our model, which estimates 943,000 lives saved annually, highlighting the critical importance of overcoming barriers to access.

Through innovative financing, advocacy, and global cooperation, the world can seize this opportunity to save lives on an almost unprecedented scale.

Annex 1: All-World Model - Potential Lives Saved from GLP1 (CVD only) – Oct 2024 Preprint

Disclosures

David Brook works part-time with Gilbert’s Solutions (gilbertslaw.ca) that has received advisory fees from Novo Nordisk but not for the development of this global model. Peter Singer owns shares of Eli Lilly and Novo Nordisk.

An initial draft of this paper was developed using ChatGPT-4.