Optimising prediction of mortality, stroke, and major bleeding for patients with atrial fibrillation: external validation of the GARFIELD-AF model in UK primary care electronic records

Background: The GARFIELD-AF tool is a novel risk tool that simultaneously assesses the risk of all-cause mortality, stroke or systemic embolism, and major bleeding in patients with atrial fibrillation (AF).

Aim: To validate the GARFIELD-AF tool in UK primary care electronic records. Design and setting: Retrospective cohort study using Clinical Practice Research Datalink (CPRD) linked with Hospital Episode Statistics data and Office for National Statistics mortality data.

Method: Discrimination was evaluated using the area under the curve (AUC) and calibration was evaluated using calibration-in-the-large regression and calibration plots.

Results: 486,818 patients aged ≥18 years with incident diagnosis of non-valvular AF between 2 January 1998 and 31 July 2020 were included; 50.6% received anticoagulation at diagnosis. The GARFIELD-AF models outperformed the CHA2DS2VASc and HAS-BLED tools in discrimination ability of death, stroke, and major bleeding at all the time points. The AUC (95%CI) for events at 1 year for the 2017 model were: death 0.747 (0.744 to 0.751) vs 0.635 (0.631 to 0.639) for CHA2DS2VASc; stroke 0.666 (0.663 to 0.669) vs 0.625 (0.622 to 0.628) for CHA2DS2VASc; and major bleeding 0.602 (0.598 to 0.606) vs 0.558; (0.554 to 0.562). Calibration between predicted and Kaplan-Meier observed events was inadequate.

Conclusions: The GARFIELD models were superior to the CHA2DS2VASc score for discriminating stroke and death and to the HAS-BLED score for discriminating major bleeding. The models consistently under-predicted the level of risk, suggesting that a recalibration is needed to optimise its use in the UK population.