ATTR-ACT - Tafamidis for transthyretin amyloid cardiomyopathy

Maurer MS, et al. Tafamidis treatment for patients with transthyretin amyloid cardiomyopathy. NEJM 2018;379:1007-16.

Bottom line:

  • In patients with transthyretin amyloid cardiomyopathy (ATTR-CM) and NYHA functional class 1-2, tafamidis at a dose of 20-80 mg/d reduced the risk of death (NNT 8) & cardiovascular hospitalization (NNT 13) over 30 months.

  • Tafamidis does not appear to reduce mortality, & increases mortality, in patients with NYHA functional class 3 HF at baseline.

  • Tafamidis did not increase overall or any specific adverse effects.

Patients (n=441)

  • Included

    • 18-90 y/o

    • ATTR, wild-type (ATTRwt) or due to a mutation (ATTRm), confirmed by cardiac or extra-cardiac biopsy

    • ATTR cardiomyopathy confirmed by

      • Echo - End-diastolic interventricular septal wall thickness >12 mm

      • Hx of heart failure (HF) with 1+ HF hospitalizations or clinical evidence of volume overload or NT-proBNP 600+ pg/mL

    • 6-minute walk-test (6MWT) distance >100 m

  • Excluded

    • NYHA functional class 4

    • Cause of HF other than ATTR CM, light-chain amyloidosis

    • Receiving other proven/potential therapy for ATTR

      • Hx of liver or heart transplant

      • Diflunisal, doxycycline, tauroursodeoxycholate

    • eGFR <25; ALT/AST >2x ULN; Severe malnutrition (quantified as serum albumin [g/L] * BMI <600)

    • Use of calcium-channel blockers or digoxin

  • Baseline characteristics

    • 74 y/o; male (90%); white (80%), black (14%)

    • ATTRwt (76%), ATTRm (24%)

    • NYHA functional class (FC): 1 (~8%), 2 (~59%), 3 (~33%)

    • Kansas City Cardiomyopathy Questionnaire (KCCQ) 66/100

    • 6MWT distance 350 m

    • BP 115/70 mm Hg (supine & standing)

    • Echo: LVEF 48%, interventricular wall thickness 16.2-16.2 mm

    • NT-proBNP median ~3000

    • Meds: RAAS inhibitor (27%), beta-blocker (29%), diuretic (66-70%), antithrombotic (40%)

Interventions & control

  • I: Tafamidis 20 or 80 mg PO once daily

    • If adverse effect, those randomized to 80 mg/d could be reduced to 40 mg/d

  • C: Matching placebo

  • Adherence: 97% from each group took at least 80% of their doses (assessed by pill count at follow-up visits)

Results @ 30 months

  • Primary outcome of death or CV hospitalization: Win ratio: 1.70 (95% confidence interval [CI] 1.26-2.29)

    • “Win ratio” calculated using the Finkelstein-Schoenfeld method

      • Non-parametric test that compares every patient from the intervention group to every patient in the control group in pairwise fashion;

      • First, pairs are compared based on whether each is alive or dead at last common follow-up (e.g. if patient A followed for 2 years, but patient B dropped out at 1 year, then both are compared at year 1);

      • Second, if both are alive or both dead, then there is a draw, & they are compared based on the rate of CV hospitalizations;

      • The win ratio is a tally of all of these comparisons, where higher scores are better. Because this is a ratio, a confidence interval that does not include 1.00 is statistically significant

  • Death: Tafamidis 29.5%, placebo 42.9% (NNT 8)

    • Hazard ratio (HR) 0.70 (95%CI 0.51-0.96)

    • No significant subgroup difference based on TTR genotype or NYHA FC

  • CV hospitalizations: 52.3% vs 60.5% (NNT 13, relative risk 0.68, 95%CI 0.56-0.81)

    • Rate per patient/year: 0.30 vs 0.46

    • Subgroup based on NYHA class: p<0.001 for interaction; INCREASE with tafamidis in NYHA FC 3

    • Subgroup based on TTR genotype: p=0.11 for interaction'\

  • 6MWT distance vs placebo: +76 m

  • KCCQ vs placebo: +13.6 points (range 0-100, higher scores are better)

  • No difference in overall, serious, or any individual adverse effects

Internal validity

  • Low risk of allocation, performance, detection & attrition bias

    • Computer-generated randomization stratified by TTR status (wild-type or mutant variant), baseline NYHA class

    • Allocation concealed: Allocation by using interactive web-response system

    • Blinding by use of matching placebo

    • Loss to follow-up <0.5% & modified intention-to-treat (mITT) analysis