The Maze procedure has been developed as a surgical approach to the ma
nagement of patients with atrial fibrillation refractory to medical tr
eatment. The recent modification of the technique (Maze 3) achieves go
od rate control with coordinated AV contractions. However, the procedu
re involves cuts that completely isolate a block of left atrial (LA) w
all, including the four ostia of the pulmonary veins. The electrical a
nd mechanical activity of this isolated LA block are dissociated from
the rest of the atrium, and the area may, in fact, continue to fibrill
ate. This may provide a nidus for the development of mural thrombus. T
he weight and endocardial surface area of the LA block and of the enti
re LA were estimated in ten formalin fixed hearts from trauma victims
with no evidence of cardiac disease, in these samples, the LA block re
presented 35% of the endocardial surface area of the entire LA and 29%
of the weight. The LA block is of sufficient size to allow macroreent
rant circuits to form and has the potential to fibrillate if isolated
from the rest of the atrium. We modified the Maze 3 procedure to recru
it the otherwise isolated LA block by using two additional cuts around
each pair of pulmonary veins as they enter the LA, The first patient
who underwent the modified procedure demonstrated sinus rhythm on Holt
er monitoring postoperatively and remained in sinus rhythm following b
urst atrial pacing at 300 and 420 beats/min each for 30 seconds. In ad
dition, atrial contractions were found to contribute 29% of the cardia
c output. The majority of the atrial wall and, in particular, the recr
uited area between the pulmonary veins contracted well, as demonstrate
d by transesophageal echocardiography. We suggest that this modificati
on of Maze 3 has a potential advantage over the standard procedure by
recruiting the entire LA without leaving any dyskinetic endocardial su
rface for thrombus formation. This should in turn reduce the risk of t
hromboembolic complications.