Table 2.

Association between generative topographic mapping–derived atrial fibrillation phenotypes and clinical outcomes in the Murcia Atrial Fibrillation Project III cohort. Adjusted Fine-Gray competing-risk models were used for nonfatal outcomes (thromboembolic events, major bleeding, and MACE^a), with death treated as a competing event, whereas adjusted Cox proportional hazards models were used for cardiovascular death and all-cause death.

Outcome		Adjusted modelb
		HRc (95% CI)	P value
Any thromboembolic event
	Phenotype 1 (n=619; reference)
	Phenotype 2 (n=677)	0.78 (0.52-1.17)	.23
	Phenotype 3 (n=353)	0.61 (0.37-1.02)	.06
	Phenotype 4 (n=1610)	0.70 (0.50-0.98)	.04
Major bleeding
	Phenotype 1 (n=619; reference)
	Phenotype 2 (n=677)	0.95 (0.63-1.43)	.08
	Phenotype 3 (n=353)	0.80 (0.48-1.33)	.38
	Phenotype 4 (n=1610)	0.61 (0.42-0.89)	.01
MACE
	Phenotype 1 (n=619; reference)
	Phenotype 2 (n=677)	0.68 (0.45-1.03)	.07
	Phenotype 3 (n=353)	0.61 (0.37-1.01)	.05
	Phenotype 4 (n=1610)	0.67 (0.47-0.94)	.02
Cardiovascular death
	Phenotype 1 (n=619; reference)
	Phenotype 2 (n=677)	0.52 (0.33-0.85)	.008
	Phenotype 3 (n=353)	0.57 (0.32-0.99)	.048
	Phenotype 4 (n=1610)	0.49 (0.33-0.72)	<.001
All-cause death
	Phenotype 1 (n=619; reference)
	Phenotype 2 (n=677)	0.66 (0.50-0.87)	.003
	Phenotype 3 (n=353)	0.44 (0.30-0.65)	<.001
	Phenotype 4 (n=1610)	0.56 (0.44-0.71)	<.001

^aMACE: major adverse cardiovascular events.

^bAdjusted model by concomitant treatment (oral anticoagulation type [direct-acting oral anticoagulants and vitamin K antagonists], antiarrhythmics, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, calcium channel blockers, antilipemic agents, beta-blockers, diuretics, oral hypoglycemic agents, insulins and antiplatelet therapy).

^cHR: hazard ratio; subdistribution HRs are reported for nonfatal outcomes, while adjusted HRs are reported for cardiovascular and all-cause death.