(The) cardiovascular implications of glucagon-like peptide-1 receptor agonists discontinuation and potential role of sodium-glucose cotransporter-2 inhibitors

다국어 초록 (Multilingual Abstract)

Background: Given considerable cardiovascular and kidney benefits and significant weight loss effects, there has been an explosion in the use of glucagon-like peptide-1 receptor agonists (GLP-1RAs) recently. However, high rates of GLP-1RA discontinuation raise concerns, yet the impact of GLP-1RA discontinuation on cardiovascular outcomes remains unclear. Furthermore, whether there are mitigating ways to reduce this risk also remains to be explored.
Objective: To evaluate the cardiovascular risk in individuals with type 2 diabetes (T2D) who discontinue GLP-1RAs therapy compared to those who continue it, and the cardiovascular risk associated with early initiation versus late initiation or no initiation of sodium-glucose cotransporter-2 (SGLT2) inhibitors following the discontinuation of GLP-1RA therapy.
Methods: We conducted 2 target trial emulation studies utilizing the nationwide healthcare database of South Korea from 2012 to 2023. The first study compared the treatment strategy of discontinuing GLP-1RA therapy within 6 months in the absence of toxicity with continuing GLP-1RA therapy in patients with T2D who received GLP-1RA for at least 12 weeks. The second study compared the following 4 treatment strategies following GLP-1RA discontinuation: 1) initiate SGLT2 inhibitors within 90 days of GLP-1RA discontinuation, 2) within 91-180 days, 3) within 181-360 days, 4) delayed initiation for at least 360 days. The primary outcomes of interest were major adverse cardiovascular events (MACE, a composite of myocardial infarction, stroke, or cardiovascular death), and secondary outcomes were each individual component of MACE and all-cause death. Absolute risks, risk differences, and risk ratios with 95% confidence intervals (CIs) over 3 years, respectively, were estimated through the clone-censor-weight method.
Results: The first study included 34,703 eligible patients, and the second study included 50,002 eligible patients. Compared to continuing therapy, discontinuing therapy was associated with a 34% increased risk of MACE over 3 years (risk ratio, 1.34 [95% CI, 1.10 to 1.63]; risk difference, 0.30 [95% CI, 0.10 to 0.49]). Compared to delayed initiation for > 360 days, early initiation within 90 days was linked to a 36% lowered risk of MACE over 3 years (risk ratio, 0.69 [95% CI, 0.60 to 0.80]; risk difference, -1.84 [95% CI, -2.42 to -1.26]); initiation within 91-180 days was associated with a 12% lowered risk of MACE over 3 years (risk ratio, 0.88 [95% CI, 0.79 to 0.99]; risk difference, -0.71 [95% CI, -1.27 to -0.15]; and initiation within 181-360 days was associated with a 10% lowered risk of MACE over 3 years (risk ratio, 0.90 [95% CI, 0.82 to 0.98]; risk difference, -0.59 [95% CI, -1.06 to -0.12]. The results remained generally consistent across sensitivity analyses and subgroup analyses. The traditional “from threshold” method introduced bias and yielded inconsistent results.
Conclusions: In conclusion, our findings suggested an increased risk for MACE when discontinuing GLP-1RA therapy in patients with T2D. Furthermore, we found that SGLT2 inhibitor initiation, especially within 90 days after GLP-1RA discontinuation, could mitigate the increased risk of MACE, compared to delayed or no initiation. Replication in further studies is required, and if consistent findings emerge, healthcare providers and policymakers should implement strategies to help patients adhere to GLP-1RA therapy. If discontinuation of GLP-1RA becomes necessary, prompt initiation of SGLT2 inhibitors should be strongly considered.

국문 초록 (Abstract)

GLP-1 수용체 작용제(GLP-1RA)는 심혈관 및 신장 보호 효과와 체중 감량 효과로 인해 최근 사용이 급증하고 있다. 그러나 높은 중단율이 보고되고 있음에도 불구하고, 중단이 심혈관 질환 위험에 미치는 영향은 명확히 밝혀지지 않았다. 더불어, 이로 인한 위험을 완화할 수 있는 전략 역시 규명되지 않은 실정이다. 본 연구는 제2형 당뇨병(T2D) 환자에서 GLP-1RA 치료 중단이 주요 심혈관 사건(MACE) 위험에 미치는 영향을 평가하고, GLP-1RA 중단 이후 SGLT2 억제제를 조기에 시작하는 것이 이 위험을 완화할 수 있는지를 분석하고자 하였다.
본 연구는 2012년부터 2023년까지의 한국 전국 단위 건강보험 데이터를 활용하여 두 개의 타겟 트라이얼 에뮬레이션 연구를 수행하였다. 첫 번째 연구는 GLP-1RA를 최소 12주 이상 복용한 환자를 대상으로, 부작용 없이 6개월 이내 치료를 중단한 군과 지속한 군을 비교하였다. 두 번째 연구는 GLP-1RA 중단 이후 SGLT2 억제제 시작시점에 따라 네 가지 전략을 비교하였다: 1) 90일 이내 시작, 2) 91–180일 이내 시작, 3) 181–360일 이내 시작, 4) 360일 이상 지연 시작 또는 비시작. 주요 결과지표는 주요 심혈관 사건(MACE) 이었으며, 보조 지표로는 MACE 구성요소 각각(심근경색, 뇌졸중, 심혈관 사망) 및 전체 사망률을 포함하였다. Clone-censor-weight 방법을 활용하여 3년간의 절대 위험도, 위험 차이, 위험 비(RR, 95% CI)를 추정하였다.
분석 결과, 첫 번째 연구에서는 34,703명이, 두 번째 연구에서는 50,002명이 분석에 포함되었다. GLP-1RA 치료를 지속한 환자에 비해, 중단한 환자에서는 3년 내 MACE 위험이 34% 증가하였다 (RR: 1.34 [95% CI: 1.10–1.63], 위험 차이: 0.30% [95% CI: 0.10–0.49]). 360일 이내, SGLT2 억제제를 시작하지 않은 사람과 비교했을 때, GLP-1RA 중단 후 SGLT2 억제제를 90일 이내 시작한 경우, MACE 위험이 31%감소하였다 (RR: 0.69 [95% CI: 0.60–0.80], 위험 차이: –1.84% [95% CI: –2.42 to –1.26]). 91–180일 또는 180-360 이내 시작한 경우에도 각각 12% (RR: 0.88 [0.79–0.99]), 10% (RR: 0.90 [0.82–0.98])의 위험 감소 효과를 보여주었으며 , 이러한 결과는 다양한 하위군 분석에서도 일관되게 나타났다.
GLP-1RA 치료 중단은 제2형 당뇨병 환자의 주요 심혈관 사건 위험 증가와 연관되며, 중단이 불가피한 경우에는 90일 이내 SGLT2 억제제의 조기 시작이 이 위험을 줄일 수 있는 효과적인 전략이 될 수 있다. 본 연구의 결과는 다른 데이터환경에서의 추가적인 확인이 필요하며, 재현된다면 환자에게 GLP-1RA 치료지속을 지원하고, 불가피한 경우 조기 대체 전략을 마련하는 것이 정책적으로도 중요할 것이다.

목차 (Table of Contents)

• Chapter I. Introduction 1
• 1. Overview 1
• 2. GLP-1RAs in Type 2 Diabetes 3
• A. Clinical benefits of GLP-1RA 4
• B. Increased use of GLP-1RAs 12
• C. High discontinuation rate of GLP-1RAs 16
• 3. Study Objectives 19
• Chapter II. Literature Review 20
• 1. The evidence from Randomized Clinical Trial 20
• 2. The evidence from Observational Studies 25
• 3. Remaining Evidence Gaps and Statement of Originality 27
• Chapter III. Methods 29
• 1. Data Source 29
• 2. Target Trial Specification 33
• A. The risk for cardiovascular outcomes of discontinuing GLP-1RA versus continuing GLP-1RA therapy 33
• B. The risk for cardiovascular outcomes of different timing of SGLT2 inhibitors initiation 37
• 3. Target Trial Emulation 41
• A. The risk for cardiovascular outcomes of discontinuing GLP-1RA versus continuing GLP-1RA therapy 41
• B. The risk for cardiovascular outcomes of different timing of SGLT2 inhibitors initiation 48
• 4. Clone-Censor-Weight Method 56
• 5. Sensitivity Analyses 60
• A. The risk for cardiovascular outcomes of discontinuing GLP-1RA versus continuing GLP-1RA therapy 60
• B. The risk for cardiovascular outcomes of different timing of SGLT2 inhibitors initiation 62
• 6. Sample Size Calculation 63
• 7. Ethic approval 64
• Chapter IV. Results 65
• 1. The risk for cardiovascular outcomes of discontinuing GLP-1RA versus continuing GLP-1RA therapy 65
• A. Identification of Study Cohort 65
• B. Post-discontinuation Patterns of Antidiabetic Medication Use 69
• C. Baseline Characteristics of Study Cohort 71
• D. Primary Analyses 78
• E. Secondary Analyses 82
• F. Subgroup Analyses 83
• G. Sensitivity Analysis 85
• 2. The risk for cardiovascular outcomes of different timing of SGLT2 inhibitors initiation 88
• A. Identification of Study Cohort 88
• B. Baseline Characteristics of Study Cohort 90
• C. Primary Analyses 95
• D. Secondary Analyses 99
• E. Subgroup Analyses 100
• F. Sensitivity Analysis 102
• Chapter V. Discussion 104
• 1. Main Findings 104
• 2. Comparison with Previous Studies 105
• 3. Implications 107
• 4. Strengths and Limitations 109
• Chapter VI. Conclusion 111
• List of Abbreviations 112
• References 114
• 논문요약 131