Table 3

Risk of non-persistence by oral anticoagulant

nUnivariable
HR (95% CI)
Multivariable
HR (95% CI)
36 63634 109
OAC (effect on first prescribing)
 VKA1.00 (–)1.00 (–)
 Dabigatran1.02 (0.93 to 1.11)1.24 (1.08 to 1.42)
 Rivaroxaban0.71 (0.67 to 0.74)0.85 (0.77 to 0.93)
 Apixaban0.53 (0.50 to 0.57)0.53 (0.46 to 0.60)
Time-dependent effect of OAC (per year of prescriptions)
 VKA1.00 (–)
 Dabigatran0.75 (0.65 to 0.86)
 Rivaroxaban0.69 (0.62 to 0.77)
 Apixaban0.91 (0.78 to 1.06)
CHA2DS2VASc (effect on first prescribing)
 0–11.00 (–)
 20.71 (0.66 to 0.76)
 3–40.66 (0.62 to 0.71)
 5–90.69 (0.63 to 0.76)
Time-dependent effect of CHA2DS2VASc (per year of prescriptions)
 0–11.00 (–)
 21.05 (0.99 to 1.12)
 3–41.12 (1.07 to 1.18)
 5–91.13 (1.06 to 1.20)
HASBLED-1
 0–21.00 (–)
 3–91.04 (0.99 to 1.08)
Number of drugs1.00 (0.99 to 1.01)
Townsend quintile
 11.00 (–)
 20.99 (0.95 to 1.03)
 30.95 (0.91 to 1.00)
 40.97 (0.93 to 1.02)
 50.97 (0.92 to 1.03)
Date of first prescription* (years after 1 January 2011)1.03 (1.02 to 1.05)
  • *Time difference (in years) between the date of the first ever OAC prescription for each patient and the start date of the study (1 January 2011) .

  • CHA2DS2-VASc, Congestive heart failure, Hypertension, Age≥75 years, Diabetes mellitus, Stroke, Vascular disease, Age 65-74 years, Sex category; HASBLED-1, Hypertension; Abnormal renal/liver function; Stroke; Bleeding; Labile INR; Elderly; Drugs or alcohol; OAC, oral anticoagulant; VKA, vitamin K antagonist.