Long term treatment of atrial fibrillation in elderly patients: a decision analysis

Atrial fibrillation (AF) is more frequent with age but it is not clear whether, and how, older age should influence therapeutic choice.

We developed a Markov decision analytic model simulating the long term effectiveness of 4 therapeutic strategies (rate control (RateC) or rhythm control (RhythmC) using amiodarone, each combined with aspirin or warfarin) in two hypothetical cohorts of patients with persistent AF, 60 and 80 years old at baseline. Two different base risks of stroke, low and moderate/high, were analysed. Outcomes studied were: predicted mortality, quality-adjusted years (QALYs), stroke, and disability. Time horizon was 10 years.

All results applied similarly to patients 60 and 80 years old at baseline. RateC + warfarin obtained in all cases the lowest predicted mortality (0.5% to 3.9% absolute reduction). RateC + warfarin also gained the more cumulated QALYs in patients at moderate/high risk of stroke, but RateC + aspirin obtained better results in QALYs in patients at low risk of stroke. Differences between strategies in terms of QALYs were limited (0.07 to 0.25 QALY of difference). Sensitivity analysis identified four variables, the same in younger and in older patients, that could change which strategy was optimal: impact on quality of life provoked by AF and by warfarin treatment, baseline risk of stroke and risk of major bleeding on warfarin.

No important difference in the decision making between patients 60 and 80 years old was found. Several individual variables influenced the optimal choice of long term treatment of AF, but not age by itself.