In this post hoc analysis of SUSTAIN 6 and PIONEER 6, treatment with semaglutide versus placebo reduced cardiovascular outcomes and body weight, with a similar safety profile, across all age subgroups.

The effect of semaglutide versus placebo on MACE consistently favored semaglutide across age groups for the composite endpoint and its three components. Furthermore, HRs generally decreased with decreasing age. This suggests that younger participants experienced a greater treatment effect of semaglutide versus placebo than older participants, despite having a higher risk of MACE based on the trial inclusion criteria for established CV disease or CKD for people aged < 60 years, while people aged ≥ 60 years needed only to have cardiovascular risk factors [15, 16].

This analysis is consistent with a meta-analysis of eight GLP-1RAs CVOTs (including SUSTAIN 6 and PIONEER 6), which showed that GLP-1RAs reduced MACE irrespective of age [9]. Specifically, that analysis showed that GLP-1RAs, versus placebo, reduced the risk of MACE both in people with T2D aged < 65 years and in those aged ≥ 65 years (pinteraction between age subgroups, 0.78) [9]. The meta-analysis included only two age groups, unlike the present analysis that spans four age subgroups, and therefore provides a more comprehensive overview.

Statistically significant heterogeneity in HbA1c reduction was observed, with a larger ETD for semaglutide versus placebo in the youngest age subgroup (≤ 60 years) compared with other age subgroups. However, this difference was small and of limited clinical significance. Despite HbA1c being significantly greater in older people than in younger people at baseline, this was corrected for and the actual change from baseline in HbA1c was similar in all age subgroups with semaglutide, and always greater than with placebo. It should be noted that the actual change from baseline in HbA1c in those receiving placebo was greater in older versus younger people and is likely to contribute to the heterogeneity observed in ETDs across the subgroups.

There was no difference in serious AEs across age subgroups for both semaglutide and placebo. However, this was not specifically analyzed in this combined analysis of SUSTAIN 6 and PIONEER 6; therefore, this hypothesis requires further study to substantiate. The rate of severe hypoglycemia was lower with semaglutide than placebo in most age subgroups. Furthermore, in the placebo groups, the highest incidence of severe hypoglycemia occurred in the oldest subgroup. Additional glucose-lowering medications were used to maintain glucose control according to local guidelines in those who received placebo, suggesting that semaglutide allows target HbA1c values to be achieved without the need for additional glucose-lowering medications or at the expense of hypoglycemia.

Results from this analysis are consistent with observations from a previous post hoc analysis of SUSTAIN 1–5, in which OW s.c. semaglutide (0.5 mg and 1.0 mg) had similar efficacy (HbA1c and body weight reduction) and safety profiles in non-elderly (< 65 years) and elderly (≥ 65 years) people with T2D [17]. Semaglutide 1.0 mg (versus comparators) decreased mean HbA1c from baseline by 1.5–1.9% (versus 0–0.9%) in those aged < 65 years, and by 1.2–1.8% (versus 0.2–1.0%) in those aged ≥ 65 years. Similarly, body weight was reduced by − 4.6 to − 6.4 kg (vs + 1.1 to − 2.1 kg) in those aged < 65 years, and – 4.1 to – 6.7 kg (vs + 1.5 to − 1.7 kg) in those aged ≥ 65 years. Similar proportions of people experienced AEs across age subgroups and there was no increased risk of severe or blood glucose-confirmed hypoglycemia with semaglutide versus comparators, between age subgroups [18]. Again, however, this previous analysis included only two age groups, unlike the present analysis that spans four age subgroups and indicates consistent metabolic and safety outcomes with semaglutide for older adults including for those in the higher age range (subgroups > 65 to  ≤ 70 and > 70 years, which have not been included in previous analyses).

Our findings are particularly relevant to the management of T2D in older adults with long-duration diabetes, as semaglutide reduces MACE, improves HbA1c, and reduces body weight without increasing serious AEs. As a result, HbA1c targets may not need to be adjusted for older adults, as is currently advised by the ADA [6], when using glucose-lowering treatment with a low intrinsic risk of hypoglycemia. Furthermore, there is suggestion that tight glycemic control in older adults has benefits on microvascular complications and minimizes the risk of other geriatric syndromes [19]. However, taken together, the above highlights the importance of individualized treatment approaches.

Despite recommendations that medications that improve cardiovascular and kidney outcomes in people with T2D should not differ for older adults (unless the person is assessed as frail) [5], the particular challenges of managing T2D in this population can mean that treatment is not always intensified and, in some cases, the paternalistic practice of treatment de-escalation is used [20, 21]. This strategy denies older adults from receiving MACE-reduction benefits associated with GLP-1RAs. Additionally, lower HbA1c targets when achieved with GLP-1RAs are not harmful in older adults, unless accompanied by increased risk of hypoglycemia.

The findings from this post hoc analysis of SUSTAIN 6 and PIONEER 6 reinforce the recommendation in the guidelines that GLP-1RAs should be used to reduce the risk of MACE across the spectrum of age, and offer an alternative approach to treatment de-escalation, as is currently advised by the ADA in adults ≥ 65 years [6], which should lead to better outcomes for older adults [1, 18, 22]. Glucose-lowering therapies that cause hypoglycemia in the older population should be avoided, as semaglutide offers an alternative with a low intrinsic risk of hypoglycemia. By prescribing GLP-1RAs, it may be possible to reduce the risk of cardiovascular events, maintain glycemic control, and avoid complications/comorbidities with a low risk of hypoglycemia even in elderly people; these data may help overcome the treatment inertia that this population can experience [23] and suggest that de-escalation recommendations may require refinement, considering the availability of new medications with a lower risk of hypoglycemia.

Limitations of this analysis include the relatively short durations of SUSTAIN 6 and PIONEER 6 trials compared with CVOTs such as LEADER (trial duration 60 months) [24], making it difficult to draw long-term conclusions. Data for HbA1c and body weight were combined for visits that occurred at different weeks in the two trials, with a variation of up to 6 weeks. This could introduce variability, as both HbA1c and body weight can change over short periods in response to treatment and different phases of the treatment response may be observed in different weeks. Safety findings were based on analysis of serious AEs only and do not incorporate other AEs that may be meaningful to an elderly population, such as non-serious gastrointestinal AEs. Additionally, differences between trial populations and real-world practice in terms of age, evident by the low numbers of participants aged ≥ 65 years included in clinical trials, may affect the ability to extrapolate clinical trial results to a real-world population. It is also important to note that during the time of these trials, sodium–glucose cotransporter 2 inhibitors (SGLT2is) were not considered standard of care; therefore, the number of participants treated with SGLT2is within this post hoc analysis is low, and higher number of patients were treated with α-glucosidase inhibitors, sulfonylureas, and other drugs that do not reflect current real-world practice. As SGLT2is also have proven benefits on metabolic and CV outcomes [25], this should be taken into consideration when interpreting these results. Furthermore, very elderly and frail people may not have been randomized in SUSTAIN 6 and PIONEER 6 (although they were not specifically excluded); hence, these findings may not be applicable to all people with T2D and high cardiovascular risk. While the inclusion of four age subgroups is an extension of previous analyses [9, 18], it limits the number of cardiovascular events within each subgroup, reducing statistical power. This analysis is not powered to determine whether there are significant treatment effects within each subgroup; pinteraction < 0.05 suggests a lack of heterogeneity in treatment effect between subgroups. Finally, there are inherent limitations with post hoc compared with prospective analyses from randomized clinical trials; hence, these findings should be interpreted as hypothesis generating. However, the similarity in trial designs, populations, and outcome measures in SUSTAIN 6 and PIONEER 6 supports the scientific validity of post hoc analyses of pooled data from these trials. In addition, the high retention rates (98.0–99.7% completing the trials) and recording of vital status (known for 99.6–100% of the participants at the end of the trials) support the validity of the data and the good conduct of the original trials [15, 16]. Further studies investigating the effects of semaglutide, including those on sarcopenia and its sequelae, across the spectrum of ages would be of interest.