In this retrospective cohort study, initiation of GLP-1RAs in people living with T2D treated with metformin was associated with a significantly lower rate of initiation of a subsequent glucose-lowering medication compared with initiation of SGLT2is, DPP-4is, or SUs. Participants treated with GLP-1RAs also had significantly greater improvements in HbA1c levels after 12 months of treatment compared with those treated with other drug classes, with a higher proportion of participants reaching HbA1c level < 7%. Improvements in BMI and the proportion of participants with ≥ 5% body weight loss were also significantly greater among participants treated with GLP-1RAs compared with those treated with DPP-4is and SUs but not SGLT2is. Together, these findings suggest that, when treatment intensification beyond metformin is indicated, GLP-1RAs may be more beneficial regarding subsequent treatment intensification needs and metabolic risk factor control compared with other frequently used drug classes in people living with T2D.

Early treatment intensification for people living with T2D treated with metformin is associated with improved glycemic outcomes and is likely to result in a reduction in microvascular and macrovascular events [8, 9]. However, failure to initiate or intensify treatment in a timely manner, known as therapeutic inertia, remains frequent and is a substantial contributor to suboptimal glycemic control [10]. Significant reductions in HbA1c levels and body weight observed with GLP-1RA treatment are likely to provide the additional benefits of preventing or delaying the onset or progression of T2D-associated comorbidities related to (or exacerbated by) suboptimal glycemic control, overweight or obesity, or weight gain [3, 11]. Although the drivers of therapeutic inertia are multifactorial and require multifaceted solutions, these findings suggest that use of GLP-1RAs as a second therapy in people treated with metformin may mitigate the adverse impact of therapeutic inertia at the individual and population levels. In addition, direct effects of GLP-1RAs on the need for subsequent treatment intensification may reduce concerns around increased treatment complexity that may contribute to therapeutic inertia at the patient and provider levels.

Historically, clinical guidelines for treating T2D have supported an additive style of glycemia management, recommending initial treatment with metformin for all people living with T2D, with consideration for combination therapy if HbA1c targets are not reached after 3 months [1, 3]. For people not considered high risk for cardiorenal events, the choice of a second agent is typically based on efficacy, avoidance of unwanted side effects (such as hypoglycemia and weight gain), cost, and patient preference [1, 3]. In recent years, clinical guidelines have shifted to treatment recommendations that holistically consider the patient’s risk of CV or kidney complications alongside goals related to management of glycemia and body weight. This change in approach was advanced by the results of CV outcome trials that demonstrate improved outcomes for some people (e.g., those with CV disease or chronic kidney disease) living with T2D receiving certain drugs, such as GLP-1RAs and SGLT2is [12, 13]. However, this updated approach does not negate the potential need for additional medications, and the incremental benefits of additional glucose-lowering therapies may be counterbalanced by increased regimen complexity or other challenges associated with polypharmacy [4,5,6,7, 14, 15]. As such, the potential burden of treatment intensification should be considered among the other priorities that inform individualized and timely treatment decisions [1, 3].

There are limited data on time to treatment intensification after initiation of a second glucose-lowering therapy in people living with T2D. The primary analysis from a retrospective cohort study based on data from 10,600 patients in Italy from 2008 to 2015 indicated that time to subsequent treatment intensification was similar with DPP-4is and SUs as the first add-on to metformin monotherapy [16]. In a randomized parallel-group study of 5047 patients in the USA, glargine and liraglutide were more effective at reaching and maintaining HbA1c targets compared with glimepiride or sitagliptin when added to metformin (mean follow-up time of 5 years) [17]. Furthermore, treatment durability of insulin with or without the addition of GLP-1RAs for patients already treated with oral glucose-lowering treatment was investigated in a multicenter phase 3 randomized controlled trial, which found that participants treated with insulin degludec plus liraglutide combination therapy had a significantly longer time until treatment intensification than those treated with insulin glargine alone (p < 0.0001) [18]. Findings from the current analysis using data from routine care support and extend these observations from prospective trials. The comparative effects on HbA1c level observed in this study are mostly consistent with evidence from head-to-head clinical trials and real-world evidence that suggest improved glycemic outcomes with GLP-1RAs compared with SGLT2is, DPP-4is, and SUs when added to metformin monotherapy [17, 19, 20]. This is also the case for weight management, for which GLP-1RAs demonstrated a higher reduction in body weight compared with DPP-4is and SUs when added to metformin monotherapy, consistent with results from other trials [17, 20]. However, when comparing GLP-1RAs with SGLT2is, this study found no significant difference in body weight reduction. Notably, this contrasts with results from other studies [19, 21] and may be related to several factors, including the early generation of GLP-1RA used in this population.

The choice of additional glucose-lowering treatment is not solely based on clinical need, and other factors may be important in influencing treatment preferences or access. In a non-interventional UK study of adults living with T2D treated with metformin, people aged ≥ 30 years and of non-white race were less likely than people aged < 30 and who were white to be prescribed SGLT2is or DPP-4is [22]. Similarly, lower prescription rates of GLP-1RAs and SGLT2is have been observed among adults of non-white race or Hispanic ethnicity living with T2D compared with people who were white living with T2D in the US Veterans Health Administration system [23]. People with higher out-of-pocket treatment costs may also be less likely to switch to SGLT2is or GLP-1RAs [24]. In addition, lower socioeconomic status is associated with a lower probability of SGLT2i or GLP-1RA initiation, even in countries with universal healthcare [25, 26]. These observations emphasize the need for policy efforts targeting equitable access to GLP-1RAs and SGLT2is in global populations living with T2D, particularly when additionally indicated for cardiorenal risk reduction and/or weight management.

Finally, most participants who remained alive during the study follow-up eventually experienced treatment intensification despite combination therapy, even those treated with GLP-1RAs. This provides a unique perspective on the chronic and progressive nature of T2D in real-world populations and highlights that glucose-lowering pharmacotherapy, although a foundational component of care, does not constitute a “magic bullet” for T2D. As such, contemporary strategies for the prevention and treatment of T2D should be individualized, intensive, multifactorial, and target-modifiable root causes. In addition to optimal pharmacotherapy, comprehensive and sustained lifestyle interventions are needed to achieve holistic treatment targets, including glycemic control, reduction in risk of CV, kidney, and metabolic complications, improved health-related quality of life, and healthy BMI. Weight management has been associated with improved T2D control (or remission), reduced CV morbidity and mortality, and a longer life expectancy [27,28,29,30]. Although not investigated in this study, weight management through metabolic surgery has also been associated with a reduced need for glucose- and blood-pressure-lowering therapies [27], suggesting that modification of disease trajectory in obesity may reduce polypharmacy. In addition, the large number of people exposed to multiple glucose-lowering therapies highlights the need for prospective clinical outcomes trials evaluating the safety and efficacy of different combination therapy approaches in T2D.

Key strengths of this study include the large cohort, use of real-world evidence, availability of 12-month follow-up data for HbA1c level and body weight, comparison between standard glucose-lowering therapies, and high relevance of the research question to patient priorities and routine clinical practice. However, there are important limitations. First, although propensity score matching ensured groups were largely similar at baseline, there is potential for residual confounding. Moreover, some factors were not accounted for, such as comorbidities and BMI, and missing data due to incomplete documentation, loss to follow-up, or other factors may have introduced bias. Second, data on dose intensification, duration of therapy, treatment adherence, and safety were not available. Third, treatments were assessed by drug class, rather than by individual agent, so any within-class differences between specific medications were not evaluated. This may be particularly important for GLP-1RAs, which have demonstrated considerable within-class differences in treatment effects on body weight, glycemic control, and other cardiometabolic risk factors in head-to-head trials and meta-analyses [13, 31,32,33]. Fourth, as this analysis included data from 2006 to 2021, it did not include information on therapies approved after this period (e.g., semaglutide 2.0 mg and tirzepatide), which may have impacted estimation of class-wide treatment effects, particularly for GLP-1RAs. Fifth, changes in clinical guideline recommendations and evidence for effects of medications beyond glycemic control over the study period may have impacted the study findings. Finally, these data from the USA may not be generalizable to other countries or global regions.