SELECT: Semaglutide in patients with CV disease & overweight/obesity

in ,

SELECT. N Engl J Med 2023; 389:2221-2232

Bottom line: In overweight/obese patients with existing CV disease, semaglutide reduced the risk of death & cardiovascular events and reduced weight, but increased the risk of discontinuation due to GI intolerance. For every 1000 patients treated for 3.3 years, 9 deaths and 10 non-fatal MIs would be avoided, but 84 more patients receiving semaglutide would stop the drug due to GI intolerance.

Patients (n=17,604 randomized)

  • 41 countries, Oct 2018-March 2021

  • Included:

    • Age >=45 years

    • BMI >=27

    • Established cardiovascular disease: Prior MI or stroke (ischemic or hemorrhagic), or symptomatic PAD

  • Key exclusions:

    • Diabetes: Prior diagnosis, A1c >=6.5% at screening, or treatment with GLP1 RA or any other glucose-lowering drug

    • NYHA 4 HF

    • ESRD/dialysis

  • Baseline:

    • Age 62 y, male 72%

    • White 84%, Asian 8%, Black 4%

    • Weight mean 97 kg, BMI mean 33.3 (71.5% BMI >=30)

    • MI 76%, stroke 23%, symptomatic PAD 9%, HF 24%

    • EQ-5D-VAS 77/100

    • ASA 78%, P2Y12i 33-34%, statin 90%, beta-blocker 70%

Intervention: Semaglutide subcutaneous once weekly

  • Starting dose: 0.24 mg q1w

  • Titration: Uptitrated every 4 weeks (to 0.5 -> 1.0 -> 1.7 -> 2.4 mg q1w) as tolerated

  • Target dose: 2.4 mg q1w (reached after 16 weeks)

    • 77% of those still taking the drug at year 2 received the target dose

Comparator: Matching placebo

Outcomes @ median 3.3 years

All % present semaglutide first, then placebo

  • Death: 4.3% vs 5.2%, HR 0.81 (0.71-0.93), i.e. -0.3%/y

  • Primary outcome: Composite of CV death, non-fatal MI, or non-fatal stroke:

    • 6.5% vs 8.0%, HR 0.80 (95% CI 0.72-0.90), -1.5% or ~ -0.5%/y

    • CV death: 2.5% vs 3.0%, HR 0.85 (0.71-1.01)

    • Non-fatal MI: 2.7% vs 3.7%, HR 0.72 (0.61-0.85)

    • Non-fatal stroke: 1.7% vs 1.9%, HR 0.93 (0.74-1.15)

    • Subgroup: Visually fairly consistent results across all subgroups (including sex, age, BMI, with vs without HF), though p-values for interaction by subgroup not provided

  • HF composite (CV death, HF hospitalization or urgent medical visit for HF: 3.4% vs 4.1%, HR 0.82 (0.71-0.96)

  • Renal composite (renal death, dialysis, transplantation, eGFR <15, persistent eGFR reduction >=50%, or persistent uACR >300 mg/g): 1.8% vs 2.2%, HR 0.78 (0.63-0.96)

  • Safety

    • Serious adverse events: 33.4% vs 36.4% (-3%)

    • Treatment discontinuation: 26.7% vs 23.6% (+4.1%)

    • Adverse event leading to discontinuation: 16.6% vs 8.2% (+8.4%)

    • Gallbladder-related disorder: 2.8% vs 2.3% (+0.5%)

    • Acute pancreatitis: 0.2% vs 0.3%

  • Mean difference in %change in weight at 2 years: -8.5% vs placebo

Internal validity

  • Computer-generated random sequence generation

  • Allocation concealment by centralized interactive web-based response system

  • Blinding by matching placebo & titration schedule

  • Intention-to-treat analysis

  • 2.2% lost to follow-up

Generalizability

  • All patients in this trial had some form of prior CVD, mostly atherosclerotic/ischemic disease, and were inherently at “high” risk of recurrent CV event. “Primary prevention” patients, or those with CAD or cerebrovascular disease without prior MI or stroke would inherently be at lower risk & therefore may experience lower absolute benefit.

  • Subgroup based on history of HF & HF subtype:

    • Consistent reductions in MACE & HF events regardless of history of HF (with vs without) or HF subtype (HFrEF vs HFmrEF/pEF); all p-interaction >0.10

  • Details on the subgroup of patients with HF are currently sparse, and it is unclear if this benefit extends across the spectrum of LVEF