What randomized evidence exists on cardiovascular outcomes after planned discontinuation of GLP‑1 therapy?

1. What the randomized trials actually tested and what they reported

Large cardiovascular outcome trials (CVOTs) for GLP‑1 receptor agonists were primarily designed and randomized to test active drug versus placebo during treatment, demonstrating reductions in MACE and other benefits while therapy continued, but they were not designed to randomize patients to “stop versus continue” after treatment and therefore offer limited randomized insight into post‑cessation cardiovascular risk ^{[1] [4]}. Some RCTs include off‑treatment observation windows or report post‑treatment biomarker changes, and meta‑analyses of those trials show cardiovascular benefit during exposure, but randomized contrasts that isolate the effect of planned discontinuation on incident CV events are largely absent from the CVOT literature ^{[1] [4]}.

2. Randomized evidence on post‑discontinuation physiology and its cardiovascular implications

A 2025 systematic review and meta‑analysis of randomized trials that mandated at least 12 weeks of post‑discontinuation follow‑up pooled 18 RCTs (3,771 participants) and found substantial metabolic rebound after stopping GLP‑1 RAs—mean weight regain and modest rises in HbA1c—outcomes that plausibly erode cardiovascular risk factor improvements obtained on therapy ^{[2] [3]}. That randomized evidence is robust for anthropometric and glycemic endpoints (weight, HbA1c) but the analysis emphasizes physiological rebound rather than showing randomized increases in hard cardiovascular events attributable to discontinuation itself ^[2].

3. Where randomized evidence ends and observational evidence begins

Because RCTs seldom randomize discontinuation strategies, the literature relies on observational and registry studies to examine outcomes after stopping GLP‑1 therapy; these real‑world analyses report higher rates of coronary disease, heart failure, and other events after discontinuation, but they carry confounding and selection bias that randomized trials avoid ^{[5] [6]}. The observational signal—that early discontinuers experience higher event rates—is plausible and consistent with physiological rebound, yet cannot substitute for randomized causal evidence and therefore should be interpreted as suggestive rather than definitive ^{[5] [6]}.

4. Implications for clinicians, patients, and policy given the evidence gap

The randomized evidence establishes that GLP‑1 RAs reduce cardiovascular risk while taken, and that stopping the drugs causes reproducible metabolic regression toward baseline within months—facts that together create a mechanistic rationale for increased long‑term cardiovascular risk after discontinuation even if randomized trials directly measuring post‑cessation MACE are lacking ^{[1] [7] [2]}. Practical drivers of discontinuation—cost, access, side effects and adherence—are well documented and likely bias who stops treatment in real life, complicating observational inferences and highlighting the need for randomized discontinuation trials or protocolized maintenance studies to guide policy ^{[8] [9]}.

5. Bottom line and research priorities

Randomized data show clear cardiovascular benefit during GLP‑1 RA exposure and randomized post‑treatment follow‑up demonstrates consistent metabolic rebound after stopping therapy, but randomized trials that specifically assign continuation versus planned discontinuation and are powered for cardiovascular endpoints do not yet exist in the public record; therefore causation between discontinuation and increased MACE remains biologically plausible and supported by observational studies but not definitively proven by randomized discontinuation trials ^{[1] [2] [5]}. Priority next steps are randomized discontinuation or maintenance trials with adequate follow‑up for MACE, and policy solutions to reduce involuntary discontinuation driven by cost and access ^{[8] [9]}.