CARMELINA - Linagliptin in type 2 diabetes

in

Rosenstock J, et al. Effect of linagliptin vs placebo on major cardiovascular events in adults with type 2 diabetes and high cardiovascular and renal risk: The CARMELINA randomized clinical trial. JAMA

Bottom line: In patients with type 2 diabetes with additional risk for CV or renal adverse events, linagliptin did not reduce the risk of CV events, nephropathy or retinopathy vs placebo over 2.2 years. Linagliptin may increase the risk of acute pancreatitis (NNH 500).

This is consistent with all other trials of DPP-4 inhibitors showing no clinical benefit from this class of medications.

Patients (n=6991)

  • Included

    • Adults with T2DM with HbA1c 6.5-10.0%

    • + antihyperglycemic meds stable for at least 2 months

    • + either high CV or renal risk

      • High CV risk: Existing CAD/stroke/PAD, or micro/macroalbuminuria (urinary albumin:creatinine ratio [UACR] >30 mg/g)

      • High renal risk: eGFR 15-45, or eGFR 45-75 + UACR >200 mg/g

  • Key exclusion criteria

    • ACS/PCI/CABG in last 2 months or PCI/CABG planned

    • Stroke/TIA in last 3 months

  • Typical baseline characteristics

    • 66 y/o, male (62%), white (80%)

    • Diabetes mean duration 15 y, HbA1c 8.0%

    • Ischemic heart disease (27%), HF (27%)

    • eGFR mean 55 mL/min/1.73 m^2: 30-45 (28%), <30 (15%)

    • Meds:

      • Antihyperglycemics: Insulin (58%), metformin (54%), sulfonylurea (32%)

      • ASA (62%), statin (72%), ACEI/ARB (81%)

Intervention & control

  • I: Linagliptin 5 mg PO daily

  • C: Matching placebo

  • Co-interventions: Glycemic control using most available antihyperglycemics (no DPP-4i, SGLT2i or GLP-1 receptor agonists)

Results @ median 2.2 years

Efficacy

  • No reduction in the original primary outcome (“4-point MACE” a composite of CV death, MI, stroke, or hospitalization for unstable angina): Linagliptin 13.3%, placebo 13.2%

    • Hazard ratio (HR) 1.00, 95% confidence interval (CI) 0.88-1.13

    • CV death: 7.3% vs 7.6%

    • Non-fatal MI: 4.5% vs 3.9%

    • Non-fatal stroke: 1.9% vs 2.1%

    • UA hospitalization: 1.2% vs 1.4%

    • Consistent lack of benefit in all subgroups (if anything, HR 1.20 in pts with baseline HbA1c >8%)

  • No effect on death from any cause: 10.5% vs 10.7% (HR 0.98, 95% CI 0.84-1.13)

  • No reduction in kidney outcomes (composite of death due to kidney disease, end-stage renal disease, or sustained -50% eGFR): 6.6% vs 6.5% (HR 0.98, 0.82-1.18)

  • No reduction in retinopathy: 1.0% vs 1.4% (HR 0.73, 95% CI 0.47-1.12)

Safety

  • Serious adverse events: 37.0% vs 38.5%

  • D/C due to adverse event: 10.3% vs 11.5%

  • Possible increase in acute pancreatitis: 0.3% vs 0.1% (NNH 500)

  • No increase in HF hospitalizations: 6.0% vs 6.5% (HR 0.90, 0.74-1.08)

Surrogate outcomes vs placebo

  • HbA1c: At month 3: -0.5%; over entire trial: -0.36%

  • No difference in weight, SBP, DBP, LDL-C, HDL-C

Internal validity

  • Low risk of allocation, performance, detection bias:

    • Computer-generated random sequence;

    • Block-randomized by interactive phone/web system;

    • Participants, clinicians & investigators blinded;

    • Central adjudication of CV & renal events by committee unaware of treatment allocation.

  • Low risk of attrition bias for CV outcomes & death, but unclear for renal outcomes:

    • Low loss-to-follow-up (LTFU) of 0.3% for mortality & 1.3% for CV events, but high (12%) LTFU for kidney outcomes;

    • Modified intention-to-treat (mITT) including all patients as randomized who received study drug for at least 1 dose.