Background The maze procedure failed to abolish atrial fibrillation (A
F) in 14% of patients with underlying organic cardiac lesions. Identif
ication of contributing risk factors will improve results either by tr
eatment of such factors or by avoidance of high-risk patients. Methods
and Results We analyzed 192 consecutive patients with AF undergoing t
hree variations of the maze procedure performed simultaneously with co
rrection of valvular diseases (n=165), congenital anomalies (n=19), is
olated AF (n=7), and ischemic disease (n=1). Twenty-six preoperative f
actors and two postoperative cardiac size parameters were analyzed by
multivariate analyses between patients with successfully ablated AF (n
=165) and those who remained in AF (n=27) after the maze procedure. Am
ong all factors, postoperative left atrial dimension was the most pote
nt in predisposing patients to persistent AF. Duration of AF, left atr
ial dimension, and cardiothoracic ratio were identified as preoperativ
e risk factors, whereas modifications of the maze procedure and pathog
enesis and location of underlying disease failed to have a significant
prognostic impact on rhythm after surgery. Individual risk analysis u
sing the three preoperative variables revealed that left atrial size r
eduction to normalize its dimension played a pivotal role in determini
ng rhythm after the maze procedure.Conclusions Results favor earlier p
erformance of the procedure before these risk factors develop, after w
hich omission of the procedure or extensive left atrial plication may
be appropriate. This requires further study.