Glucagon-like peptide 1 receptor agonists switching patterns in type two diabetes: A retrospective real-world study

Abstract

BACKGROUND

Glucagon-like peptide-1 receptor agonists (GLP-1RAs) play a key role in managing type 2 diabetes mellitus (T2DM). Transitioning between different GLP-1RA has been proposed as a treatment strategy.

AIM

To investigate switching patterns between GLP-1RA and their impact on glycemic control.

METHODS

A retrospective study involving patients with T2DM who initiated GLP-1RA therapy between 2009 and 2021 and transitioned to another GLP-1RA. Baseline glycated hemoglobin (HbA_1c) was defined as the most recent measurement within 1 year prior to switching, and follow-up HbA_1c was the first measurement recorded 4-15 months post-switch.

RESULTS

Among 70654 patients initiating GLP-1RA therapy, 18047 (25.5%) switched regimens. In the 13970 patients with available HbA_1c, levels decreased from 8.5% ± 1.6% to 7.6% ± 1.4% (P < 0.001). HbA_1c decreased in 78.3% (10943/13970) of these patients, with the most frequent improvement observed in those switching from daily to weekly GLP-1RA (81%, 5582/6890).

CONCLUSION

Switching between GLP-1RAs can serve as a practical alternative to treatment intensification for effectively managing T2DM.

Key Words: Type 2 diabetes mellitus; Glucagon-like peptide-1 receptor agonists; Glycated hemoglobin; Glycemic control; Switching; Real world

Core Tip: This retrospective study analyzed the effects of switching between glucagon-like peptide-1 receptor agonists (GLP-1RAs) in patients with type 2 diabetes mellitus. It found that 25.5% of over 70000 patients changed GLP-1RA regimens, with most experiencing significant glycated hemoglobin reductions, from 8.5% to 7.6%. Notably, switching from daily to weekly GLP-1RA was associated with the highest improvement rate (81%), demonstrating that such transitions are a practical and effective alternative to treatment intensification for better glycemic control.

INTRODUCTION

The management of type 2 diabetes mellitus (T2DM) has evolved significantly over the past two decades, particularly with the advent of glucagon-like peptide-1 receptor agonists (GLP-1RAs). Different GLP-1RAs are available for managing T2DM with distinct pharmacokinetic and pharmacodynamic properties, dosing schedules, administration frequencies, and clinical outcomes. This diversity among GLP-1RAs allows for tailored therapy based on patient needs, preferences, and clinical characteristics[1].

Switching between GLP-1RA is common in T2DM management[2]. However, data regarding switching patterns and glycemic control after switching are scarce. The current study provides a real-life analysis of GLP-1RA switching patterns and the glycemic control after switching, using a large retrospective database.

MATERIALS AND METHODS

This study analyzed medical records from Clalit Health Services (CHS), the largest healthcare maintenance organization (HMO) in Israel, focusing on T2DM patients aged over 18 who initiated treatment with any GLP-1RA between 2009 and 2021. Switching was defined as the dispensing of a different GLP-1RA within 1 year following the last prescription of the initial GLP-1RA. The index day (ID) was designated as the date the patient dispensed the new GLP-1RA for the first time. The CHS HMO is in an excellent position to perform such long-term follow-up studies because of the relatively low rate of member drop-out of approximately 2.5% per year.

Baseline glycated hemoglobin (HbA_1c) was defined as the most recent test result within the year prior to the ID, while post-switch HbA_1c was taken as the first test result recorded between 4 and 15 months after the ID.

Statistical analyses

Continuous variables are reported as the means and standard deviations, while categorical variables are presented as counts and proportions. Pre-switching and post-switching HbA_1c levels were compared using a paired t-test when the differences were normally distributed; otherwise, the Wilcoxon signed-rank test was applied. Statistical analyses were conducted using IBM Statistical Package for the Social Sciences Statistics 28.0 (IBM, Armonk, NY, United States), with a significance threshold of P < 0.05 for two-tailed tests.

RESULTS

Of the 70654 patients with T2DM receiving GLP-1RA therapy, 18047 (25.5%) switched regimens. Switchers were significantly younger (58.7 years vs 60.9 years; P < 0.001), more often female (53.6% vs 50.1%; P < 0.001), and less likely to be of low socioeconomic status (40.0% vs 47.5%; P < 0.001). Additionally, a significantly higher proportion of patients initially treated with daily GLP-1RA was observed in the switchers group (56.2% vs 39.1%; P < 0.001).

Switching from daily GLP-1RA was more frequent than from weekly GLP-1RA (33% vs 19.8%). Table 1 provides detailed GLP-1RA switching patterns.

Table 1 Patterns of switching between different glucagon-like peptide-1 receptor agonists, n (%).

	Total (70645)	Switched [18047 (25.5)]	Exenatide BID	Liraglutide	Lixisenatide	Exenatide ER	Dulaglutide	Semaglutide
Daily GLP-1RA	30699	10144 (33)
Exenatide BID	1624	775 (47.7)	-	673	13	71	18	0
Liraglutide	27596	8905 (32.2)	199	-	553	1148	4044	2961
Lixisenatide	1479	464 (31.4)	0	161	-	85	201	17
Weekly GLP-1RA	39955	7903 (19.8)
Exenatide ER	4896	1755 (35.8)	-	338	69	-	1187	161
Dulaglutide	30062	5783 (19.2)	0	584	13	113	-	5073
Semaglutide	4997	365 (7.3)	0	18	0	0	347	-

BID: Twice daily; ER: Extended-release; GLP-1RA: Glucagon-like peptide-1 receptor agonist.

Of the 18047 patients who switched GLP-1RA therapy, 13970 (77.4%) had HbA_1c measurements available before and after the switch within the defined timeframes. These patients experienced significant reduction in mean HbA1c, from 8.5% ± 1.6% to 7.6% ± 1.4% (P < 0.001). This overall decrease was mirrored in subgroups switching from both daily and weekly GLP-1RA regimens, as outlined in Table 2.

Table 2 Glycemic control (glycated hemoglobin) before and after switching between glucagon-like peptide-1 receptor agonists.

Exenatide BID	HbA1c pre	HbA1c post	P value	Exenatide ER	HbA1c pre	HbA1c post	P value
Liraglutide (n = 616)	8.9 ± 1.5, 8.7 (7.9-9.8)	8.6 ± 1.7, 8.2 (7.3-9.5)	< 0.001	Liraglutide (n = 310)	9.2 ± 1.5, 8.9 (8-10)	8.4 ± 1.7, 7.9 (7-9)	< 0.001
Lixisenatide (n = 12)	9.9 ± 1.7, 10 (8.7-11.2)	9.9 ± 1.9, 9.9 (8.5-11.5)	0.665	Lixisenatid (n = 60)	9.2 ± 1.5, 8.9 (8.1-10)	9.2 ± 1.9, 8.8 (7.8-10)	0.940
Exenatide ER (n = 62)	8.5 ± 1.7, 8.1 (7.3-9.3)	7.8 ± 1.4, 7.3 (6.7-8.5)	< 0.001	Dulaglutide (n = 1046)	8.5 ± 1.5, 8.3 (7.5-9.3)	8.0 ± 1.5, 7.7 (7-8.7)	< 0.001
Dulaglutide (n = 18)	8.7 ± 1.7, 8.9 (7.5-9.8)	7.5 ± 0.94, 7.3 (6.9-8.2)	0.005	Semaglutide (n = 117)	7.9 ± 1.4, 7.6 (6.8-8.8)	7.2 ± 1.3, 6.9 (6.3-7.7)	< 0.001
Liraglutide				Dulaglutide
Exenatide BID (n = 159)	9.1 ± 1.6, 8.8 (7.9-9.8)	8.8 ± 1.7, 8.5 (7.5-10.2)	0.085	Liraglutide (n = 479)	8.7 ± 1.4, 8.4 (7.7-9.6)	8.1 ± 1.6, 7.9 (7.1-9)	< 0.001
Lixisenatide (n = 464)	9.4 ± 1.4, 9.3 (8.4-10.3)	9.4 ± 1.7, 9.1 (8-10.5)	0.343	Lixisenatid (n = 11)	9.3 ± 1.9, 8.9 (8.4-10.2)	9.0 ± 1.8, 8.7 (8.1-10)	0.504
Exenatide ER (n = 956)	9.1 ± 1.6, 9 (8.1-10)	8.5 ± 1.6, 8.2 (7.4-9.4)	< 0.001	Exenatide ER (n = 100)	8.9 ± 1.4, 8.6 (7.8-9.9)	8.4 ± 1.5, 8.3 (7.3-9.3)	0.004
Dulaglutide (n = 3392)	8.8 ± 1.6, 8.5 (7.8-9.7)	8.1 ± 1.5, 7.8 (7-8.8)	< 0.001	Semaglutide (n = 3425)	8.0 ± 1.4, 7.8 (7-8.7)	7.4 ± 1.4, 7.2 (6.5-8.1)	< 0.001
Semaglutide (n = 2192)	7.9 ± 1.4, 7.7 (7-8.6)	7.2 ± 1.2, 6.9 (6.4-7.7)	< 0.001
Lixisenatide				Semaglutide
Liraglutide (n = 149)	9.1 ± 1.4, 9 (8.1-10)	8.1 ± 1.4, 7.8 (7-9)	< 0.001	Liraglutide (n = 8)	9.3 ± 2.0, 8.5 (7.5-11)	8.0 ± 1.8, 7.6 (6.6-9.4)	0.183
Exenatide ER (n = 79)	9.7 ± 1.4, 9.4 (8.7-10.5)	8.6 ± 1.4, 8.3 (7.6-9.6)	< 0.001	Dulaglutide (n = 124)	8.4 ± 1.5, 8.1 (7.6-9.2)	7.8 ± 1.5, 7.8 (6.7-8.7)	< 0.001
Dulaglutide (n = 183)	9.3 ± 11.5, 9.2 (8-10)	8.2 ± 1.5, 7.9 (7-8.8)	< 0.001
Semaglutide (n = 8)	7.8 ± 1.2, 7.6 (6.9-8.5)	6.3 ± 0.55, 6.3 (6-6.6)	0.009

BID: Twice daily; ER: Extended-release; HbA_1c: Glycated hemoglobin.

Overall, HbA_1c reduction was observed in 78.3% (10943/13970) of patients, with the highest proportion found in those who transitioned from daily to weekly GLP-1RA (81%, 5582/6890), compared to 76.7% (3690/4812) in the group switched within weekly GLP-1RA and 75% (652/868) in patients who switched from weekly to daily GLP-1RA. Additionally, 72.6% (1019/1400) of patients who transitioned between daily GLP-1RA showed an improvement in HbA_1c.

DISCUSSION

This real-world study demonstrates that switching between GLP-1RA is a relatively common practice among patients with T2DM and is frequently associated with improved glycemic control. Our findings show a strong preference for switching from daily to weekly GLP-1RA regimens, this likely reflects challenges related to adherence to daily injections and the potential for enhanced efficacy with the less frequent dosing. This is consistent with our previous research, which showed enhanced adherence to GLP-1RA among patients receiving weekly formulations, older age, female gender, and individuals of higher socioeconomic status[3].

It is important to acknowledge the inherent variability in pharmacologic and pharmacodynamic properties, as well as differing adverse event profiles, among various GLP-1RAs[4]. Daily GLP-1RAs are absorbed more quickly, have a shorter half-life, resulting in more frequent fluctuations in plasma drug levels. By contrast, weekly GLP-1RAs are absorbed more slowly, have an extended half-life, and provide more stable plasma concentrations. These pharmacokinetic differences impact therapeutic outcomes and side effects[5]. Short-acting (daily) agents produce greater postprandial glucose lowering but may be associated with more variable glycemic control and potentially more frequent gastrointestinal (GI) side effects due to higher peak plasma concentrations. Long-acting (weekly) agents provide more consistent glycemic control, enhance adherence, and reduce the risk of plasma concentration-related side effects. However, GI symptoms such as nausea, vomiting, and diarrhea remain common, particularly during dose escalation[6,7]. Therefore, a gradual titration is recommended to help minimize these adverse effects and improve tolerability. These differences often influence both physician and patient preferences regarding the most suitable GLP-1RA for each individual. This aligns with the American Diabetes Association (ADA) guidelines advocating for individualized treatment plans that consider efficacy, side effects, and patient preferences[8].

The ADA and the European Association for the Study of Diabetes recommend that the daily GLP-1RA should be stopped when the first dose of the weekly agent is administered; there is no need for an overlap period due to differences in half-life and pharmacokinetics[5]. Weekly agents should be started at the lowest available dose and titrated upward every 4 weeks to minimize GI side effects. The American Gastroenterological Association emphasizes that gradual dose titration is essential to reduce GI adverse effects and that clinical judgment should guide the titration schedule based on individual tolerance and response[9]. If a patient experiences significant side effects, dose reduction to the previously tolerated dose or slower titration is recommended[10]. There is no evidence supporting the concurrent use of two GLP-1RAs, and combination therapy should be avoided.

Given the differences between daily and weekly GLP-1RAs, it might be tempting to consider phasing out the daily formulations. Our data show that 10% of patients switching from dulaglutide were transitioned to liraglutide, compared to 88% who switched to semaglutide. Although not formally tested, we believe most transitions to liraglutide occurred before semaglutide was available in our healthcare system. Therefore, it may be reasonable to consider gradually replacing daily GLP-1RAs with their weekly counterparts to improve convenience and adherence.

Although exenatide twice daily (BID) was the first GLP-1RA introduced, it was less frequently used than liraglutide, possibly due to the lack of healthcare basket coverage for exenatide, as its subsidy was approved at the same time as for liraglutide. Moreover, patients initially treated with exenatide BID were often transitioned to liraglutide. This likely reflects liraglutide's superior efficacy, once-daily dosing, and reduced adverse event profile[11]. The LEAD-6 trial supports this observation, demonstrating enhanced glycemic control and fewer side effects when switching from twice-daily exenatide to once-daily liraglutide[12]. Furthermore, a review by Madsbad[13] highlighted that liraglutide is associated with greater reductions in HbA_1c compared to exenatide formulations and has a lower incidence of injection-site reactions. Although the transition resulted in a statistically significant reduction in HbA_1c (P < 0.001), the clinical significance was modest, with only a 0.3% decrease. In contrast, transitioning to weekly preparations yielded clinically significant reductions (defined as a decrease of at least 0.5%), including exenatide extended-release (ER), which led to a 0.7% reduction, and dulaglutide, which achieved a 1.2% reduction. Similarly, switching from liraglutide to other daily GLP-1RA did not produce clinically significant HbA_1c improvements – only a 0.3% reduction with exenatide BID and no change with lixisenatide. However, transitions to weekly GLP-1RA resulted in both statistically and clinically meaningful HbA_1c reductions, as seen with exenatide ER (0.6%) and dulaglutide and semaglutide (each 0.7%). These findings reinforce the benefit of switching from daily to weekly GLP-1RA in patients who do not achieve glycemic targets.

When liraglutide was introduced, it was considered an intermediate-to-long-acting GLP-1RAs[14], while lixisenatide, a short-acting GLP-1RA that primarily targeted postprandial hyperglycemia[15]. This difference in action may have influenced healthcare providers' confidence in switching treatments to liraglutide that targets both, fasting and postprandial glucose levels[16]. While our study showed improved glycemic control with any switch from daily GLP-1RAs, the most substantial HbA_1c reduction (approximately 1.5%) occurred in patients switching from lixisenatide to weekly GLP-1RAs. This suggests that the limited efficacy of lixisenatide even compared to other daily GLP-1RAs, rather than adherence issues, may be a primary driver of these differences.

Among patients using long-acting GLP-1RAs, switching to any other GLP-1RA, including daily formulations, generally resulted in significant HbA_1c reductions. The exception was the switch from dulaglutide to lixisenatide, where the limited number of patients precluded a meaningful analysis. This real-world data underscores the need for personalized treatment approaches, acknowledging that some patients may prefer daily GLP-1RA regimens, such as liraglutide.

Similarly, switches from exenatide ER to semaglutide or dulaglutide resulted in statistically significant HbA_1c reductions, consistent with the EXPERT study's findings[17] showing improved HbA_1c with semaglutide switches. These findings suggest that switching within the GLP-1RA class, particularly from daily to weekly regimens, offers therapeutic advantages by improving glycemic control and potentially delaying the need for more intensive treatments.

Our findings are consistent with previous randomized controlled trials and real-world studies[18,19]. However, this study stands out due to its large sample size, extended data collection period covering nearly the entire duration of GLP-1RA use, and comprehensive evaluation of therapeutic switching across nearly all currently available GLP-1RA.

This study was subject to several limitations inherent to real-world data analysis, including the lack of data on diabetes duration, concomitant antidiabetic medications as the observed HbA_1c reduction could be influenced by simultaneous changes in other diabetes medications, and pre-switch and post-switch adherence/persistence. Furthermore, defining the most recent HbA1c within the year prior to switching GLP-1RA as the baseline glycemic measure may be challenging, as in some cases it could be a very recent measurement, while in others it could be up to 1 year old. Additionally, the timing of HbA1c assessment can influence the evaluation of baseline glycemic control, especially since the decision to switch GLP-1RA is often driven by poor glycemic management. Future research should focus on addressing these limitations, especially the simultaneous use or cessation of other antidiabetic medications to more comprehensively evaluate switching efficacy and should also look to determine the specific drivers and predictors for switching decisions.

CONCLUSION

Switching between GLP-1RA represents a potentially valuable strategy to optimize glycemic control before escalating to more intensive treatment options.

ACKNOWLEDGEMENTS

The authors thank Dr. Benjamin Glaser, Hadassah Medical Center, for his critical review of the manuscript.