1015

Specific Electrocardiographic Features of Manifest

Coronary Vein Posteroseptal Accessory Pathways
ATSUSHI TAKAHASHI. M.D., DIPEN C. SHAH, M.D.. PIERRE JAIS, M.D.,
MELEZE HOCINI, M.D., JACQUES CLEMENTY, M.D.,
and MICHEL HAISSAGUERRE, M.D.
From Service d'Electrophysiologie Cardiaque, Hopital Cardiologique du Haut-L^v£que, Bordeaux-Pessac, France

Coronary Vein Accessory Pathways. Introduction: Some posteroseptal accessory path-

ways {APs) can be successfully ablated by radiofrequency current only from inside tbe coro-
nary sinus (CS) or its brancbes, because of an absolute or relatively epicardial location. Tbe
aim of this study was to identify ECG features of manifest posteroseptal APs requiring abla-
tion in the CS or the middle cardiac veins (MCVs).
Methods and Results: One hundred seventeen consecutive patients witb manifest posterosep-
tal APs successfully ablated: (1) > 1 cm deep inside tbe MCV (group MCV: n = 13); (2) inside
the CS, including tbe area adjacent to tbe MCV ostium (group CS: n = JO); (3) at tbe rigbt
(group R: n = 60); or (4) the left posteroseptal endocardial region (group L: n - 34) were in-
cluded. We reviewed delta wave polarity (initial 40 msec) and QRS morphology during sinus
rhythm and atrial pacing as well as electrogram characteristics in tbese patients. Tbe local tar-
get site electrogram in groups MCV and CS was cbaracterized by a longer atrial to ventricular
electrogram interval, suggesting a longer course of the pathway and more frequent recording of
a presumptive AP potential compared to the group ablated at tbe rigbt or left endocardium. The
most sensitive ECG feature for group CS or group MCV was a negative delta wave in lead II in
sinus rbythm (87%), but specificity (79%) and positive predictive value (50%) were relatively
low. A steep positive delta wave in aVR during maximal preexcitation possessed tbe highest
specificity and positive predictive value (98% and 88%, sensitivity 61%) wbicb increased to
99% and 91%, respectively, when combined with a deep S wave in V^ (R wave < S wave).
Conclusion: These data suggest that posteroseptal APs ablated inside the coronary venous sys-
tem bave bigbly specific features, including tbe combination of a steep positive delta wave in lead
aVR and a deep S wave in lead V,, (R wave < S wave) during maximal preexcitation. The bigbest
sensitivity is provided by a negative delta wave in lead II. These findings may be belpful for antici-
pating and planning an epicardial ablation strategy. (J Cardiovasc Electrophy.siol, Vol. 9. pp. 1015-
1025, October 1998)

posteroseptal accessory pathway, catheter ablation, middle cardiac vein, coronary sinus

Introduction ablation from inside the coronary sinus (CS) or

Radiofrequency catheter ablation of accessory
pathways (APs) from the endtx:ardial approach pro-
vides a definite cure lor patients with Wolff-
Parkinson-White syndrome.'^ However, successful
Address for correspondence: Alsushi Takahashi. M.D.. Service
d'Eleirirophysioiogie Cardiaque. Hopital t'ardiologique du Haut-
L^veque. Avenue de Magellan, 33604 Bordeaux-Pessac. France.
Fax: 33-5-56-55-6K-96,
Miinu.SLTipt received 14 April 1998; Accepted for publication 16
July 1998.
its branches has been reported in some patients
whose AP cannot be ablated from the endcx:ardiurn,
suggestive of an epicardial or deep intramural
course.^'" Posteroseptal APs have been recognized
as the most trequent AP with an epicardial course,
with an incidence of approximately 9% to 24% of
all posteroseptai pathways.^•'*'"" Ablation of pos-
teroseptal APs can be difficult, requiring a larger
number of ^plicatioas of radiofrequency energy and
long procedure times. Recognition oF APs that
cannot be ablated from the endocardium would be
1016 Journal of Cardiovascular Electrophysiology Vol. 9, No. JO, October 1998
important to anticipate the appropriate therapeutic
approach for ablation. A single recent article pro-
posed that palicnLs with a negative delta wave in lead
II during sinus rhythm require ablation from middle
Ciirdiac veins (MCVs) or anomalies of the CS.'-
We reviewed surface ECG features of mani-
fest posteroseptal APs to distinguish pathways ab-
lated epicardially from those ablated endocardially.
Methods
Patients
Between December 1993 and December 1997,
160 consecutive patients with a posteroseptal ac-
cessory AV connection underwent radiofre-
quency catheter ablation at our institution's labo-
ratory. Posteroseptal APs were defined as path-
ways located around the proximal CS over the
inferoposterior third of the triangle of Koch; within
1 cm adjacent to the CS ostium over the infero-
medial aspect of the right atrium, or within 2 cm
to the left of the CS ostium at the crux area and
over the posteromedial mitral annulus.' * '•* Of these
160 patients. 121 had manifest preexcitation on a
baseline 12-lead ECG. Two patients with multiple
pathways were excluded from this study, and the
remaining 119 patients were the subject of this
study. The group comprised 38 women and 81 men
with a mean (+ SD) age of 35 ± 16 years {range
8 to 82).
Electrophysiologic Study
After informed consent, an electrophysiologie
study was performed on all patients in a fasting
state and at least five half-lives after discontinua-
tion of antiarrhythmic drugs. One or two 6-French
quadripolar electrode catheters with an interelec-
trode distance of 5 mm (Bard Eiectrophysiology,
Tewksbury. MA, USA) and a 7-French quadripo-
lar deflectable catheter with a 4-mm tip electrode
(Cordis-Webster, Miami, FL, USA or Medtronic,
San Jose, CA, USA) were introduced percuta-
neously into the right femoral vein or left subcia-
vian vein (one case only) and placed in the right
atrium, the CS, and the right ventricle. Surface
electrograms (leads I, II. [II. and V,). unipolar elec-
trograms filtered at I to 500 Hz, and bipolar in-
tracardiac electrograms filtered at 30 to 500 Hz
and amplified at high gain (O.I mV/cm) were si-
multaneously recorded with a polygraph (model
Midas, PPG Biomedical Systems, Overland Park,
KS, USA) at paper speeds of 100 mm/sec. Incre-
mental pacing and programmed stimulation were
performed using a programmable stimulator (Car-
diostimulateur Orthorhythmique, Savita. Paris,
France) with a 2-msec output pulse width and four
times threshold amplitude.
Mapping of Posteroseptal Pathway
The AP was primarily localized and ablated
by mapping the earliest bipolar and unipolar ven-
tricular potential relative to the delta wave onset
associated with a QS pattern of the unipolar ven-
tricular waveform during sinus rhythm, which lo-
calizes the ventricular insertion site. A sharp, high-
frequency potential verified by correlation with the
unipolar electrogram to occur between the atrial
and ventricular intrinsic deflections was consid-
ered as a presumptive AP potential.'"^ If favorable
electrograms were recorded from the right poste-
rior septum, ablation was attempted from the tri-
cuspid annulus adjacent to the CS ostium, the po-
sition of the catheter outside the CS being con-
firmed by fluoroscopy and unsuccessful
advancement- However, if right-sided endocar-
dial mapping appeared suboptimal. the ablation
catheter was introduced into the left side using
either the trans-septal or transaortic approach to
map the posteroseptal region of the mitral annu-
lus. The electrograms obtained from the left side
were compared to those recorded from inside the
CS, and ablation was attempted at the site with
better electrograms. If the femoral approach to the
CS was unsuccessful, the ablation catheter was in-
troduced through the left subclavian vein to search
for an optimal ablation site deep inside a venous
branch of the CS. In cases without previous abla-
tion, the right or left endocardium was targeted
preferentially. In patients witb a prior unsuccess-
ful ablation, the CS and MCV were mapped and
optimal sites were targeted, as described below.
Ablation Procedure
Radiofrequency energy was delivered as a con-
tinuous unmodulated sine wave output from a Stock-
ert (Stockert Gnibh, Freiburg, Gemiany) generator
that delivers current at 550 kHz between the distill
electrode of the ablation catheter and a 575-cm- cu-
taneous patch electixxle placed over the left scapula.
A target temperature of 70°C for the i ight and left
posteroseptal endocai'dium or 55° to 60°C for the
CS or MCV was set and temperature-controlled ra-
diofrequency energy applied for 6() to 90 seconds
if AP conduction was blocked within 20 seconds.
 Takahashi, el al. Coronary Vein Accessory Pathways
but was immediately discontinued in the event of
an increase in impedance or dislocation of the elec-
trode. Tlie successful ablation site was identified ra-
ciiographiciilly in the left oblique projection in which
the catheter recording the His-bundle potential
pointed directly toward the observer; in this pro-
jection, the CS ostium was identified by a point on
the shaft of the catheter placed within the CS in the
same vertical plane ixs tlie tip of ihe His-bundle cath-
eter. Unipolar impedance was monitored continu-
ously and noted just before radiofrequency delivery
at each site. The automatic high-impedance (> 250
O) radiofrequency cutoff was used. The course of
the CS wiis deriiiurated by catheter withdrawal from
the distal to the proximal CS. Tlie localization of
the catheter inside the MCV was confirmed by CS
angiography (five patients) or inferred from a trapped
catheter below the CS catheter position with a
high impedance (> 150 SI). The distance from the
ostium of the CS or the MCV to the successful
ablation site was measured using the interelectrode
distance of the CS catheter as a reference.
Successfiilly ablated APs were classified in four
groups: > I cm within the CS; > I cm inside the
MCV; and from the right (including terminal CS)
oi' leit endtx'iudial surface of the posteroseptal space.
At the successful ablation site, the timing of the lo-
cal peak ventricular potential relative to the delta
wave, the local AV interval, and the ratio of the
amplitude of local atrial and ventricular potentials
were measured, and the unfiltered unipolar elec-
trogram morphology pattem was analyzed.
Electrocardiographic Analysis
A preexcited 12-!ead ECG during sinus rhythm
and during overdrive high right atrial pacing was
recorded at a paper speed of 25 mm/sec. In all pa-
tients, a positive delta wave and full positive QRS
morphology were characteristically recorded in lead
I iuid aVL during sinus rhythm and maximal pre-
excitation. The initial 40 msec of the preexcited
QRS during sinus rhythm and maximal preexci-
tation was taken as the delta wave. The polarity
of the delta wave was categorized as positive, neg-
ative, or isoelectric, as previously described,'^'**
and the angle of the delta wave with the isoelec-
Iric line (>: 45"" [.steepj or < 45") was analyzed
according to Figure I. Moreover, the morphology
of QRS complexes was analyzed, notably the pres-
ence of a deep S wave defined as an R/S ratio <
I in lead V,,. The relevant analyzed chaj'acteristics,
i.e., delta wave in lead II, lead aVR. and S wave
in lead Vf,,'-are shown in Figure 2.
1017
The characteristics of ECG patterns were
compared between the four groups of posterosep-
tal APs. Furthemiore, the specificity of derived cri-
teria were compared in a blind manner to 58 pa-
tients with a single manifest AP inserting at loca-
tions bordering tlie defined postei'oseptaJ areas (right
posterior: 23 patients; left posterior: 27 patients;
midseptal: 8 patients) having both a positive delta
wave and a fully pt;)sitive QRS morphology in lead
I and aVL during maximal preexcitation and
having undergone successful ablation during the
last 2 years.
Statistical Analysis
Continuous data were expressed as mean ± SD.
Statistical comparisons were performed using the
Student's /-test. Fisher's exact test, or Chi-square
test, as appropriate. A P value < 0.05 was con-
sidered to be statistically significant. Sensitivity,
specificity, positive predictive value, and negative
predictive value for predicting the AP ablated from
within the CS and the MCV were calculated.
Results
Ablation Results
One hundred and seventeen out of 119 APs were
successfully ablated without complication. One of
the two failures later underwent successful
+ve A -ve A Iso A
U
t45' 40m8
Figure I. Examples of .steeply positive, steeply negative,
ami isoelectric delta waves (first 40 m.iec of the QRS). Left
panel: A steeply positive delta wave with an angle > 45°.
Middle panel: A steeply negative delta wave with an angle
> 45". Right panel: Example of on isoelectric delta wave. A
= delta wave: l.so = isoelectric: ms = milliseconds: -ve =
negative; +ve = positive.
1018 Journal of Cardiovascular Electrophysiology Vol. 9, No. 10, October 1998

^ 1 : Group L
• I : Group CS
I I : Group MCV

steep -ve A Steep +ve A Deep S (RiS)

Figure 2. Typical examples of stirface ECG features pre-

dicting coronary sinus and middle cardiac vein ablation
during atrial pacing. A = delta wave: -ve = negative: +ve
= positive.
cryosurgery where the preexcitation disappeared
during cryoapplication to the MCV. The other
patient refused reablation.
The successful ablation site was the right pos-
teroseptal endocardial region in 60 patients (group
R) and the left posteroseptal endocardial region
in 34 patients (group L). Of the remaining 23
patients, successful ablation was achieved from
inside the CS in 10 patients (group CS) and > 1
cm inside the MCV (1.7 ± 0.4 cm; range: 1.3 to
2 cm) in 13 patients (group MCV), including a
MCV diverticulum (3-cm diameter) in 1 patient
(Fig. 3). The unipolar impedance before ra-
diofrequency delivei'y was higher in the MCV and
intermediate witbin the CS (MCV: 162 ± 29 O,
CS: 128 ± 7.9 H; right endocardium: 101 ± 10 H;
Figure 3. The location and distribution of Ji7 patients with
posteroseptal accessory pathways. This .schema was drawn
as viewed in the left anterior oblique projection. CS = coro-
nary sinus; MCV = middle cardiac vein: MV = mitral valve:
TV = tricuspid valve. Numbers represent .successful ablation
site distribution. Gtoup R. L, CS. or MCV = patients ablated
from the right or left endocardial space or within the coro-
nary .viniis or the middle cardiac vein, respectively.
favorable, and radiofrequency energy was only
delivered inside the CS or inside the MCV with
a mean number of 2 ± 2 radiofrequency pulses.
Only one patient required a left subclavian ap-
proach to catbeterize a small MCV for successful
ablation (Fig. 5).
IMPEDANCE MAX POWER OUTPUT
250 (watts)
P-ftOOOL

P = 0.001. P - 0.0001, respectively). On the other (ohms)

' P=OOOOI PMI.OOI •
70
hand, the maximal power output was lower in the
60
MCV and intermediate within the CS (MCV: 9.5 ZOO
± 4.7 W. CS: 18 ± 6.8 W; right endocardium: 36 50

± 17 W; P = 0.001, P = 0.001, respectively) 40

(Fig. 4). Tbe three patients with the highest im- 150
30
pedance (180 to 238 0) of ablation catheter in the
MCV had small veins confirmed by CS angiog- 20

rapby in two patients and a completely trapped 100

catheter (which could not be advanced deeper)
in one patient. In 13 (57%) of the 23 patients,
radiofrequency energy was initially delivered
unsuccessfully at the right or left posteroseptal en-
docardial region (mean number of applications:
3.5 ±2.1). A mean number of 1.1 ± 0.5 radiofre-
quency pulses was delivered inside the CS or tbe
MCV to successfully abolish AP conduction.
For the remaining 10 patients who had previously
unsuccessful ablation in otber hospitals, electro-
grams on the right and left endocardium were not
10
0
CS MCV R CS MCV
Figure 4. The unipolar impedance before ablation and the
majitnal power output during radiofrequency application
at right endocardial space (Rj. within the coronary sinus
(CS) atui the middle cardiac vein (MCV). The unipolar im-
pedance before radiofrequency application was higher in
the MCV ititertnediate within the CS (left panel). On the
otiier hatid. the maximal power output achieved was lower
in the MCV and intermediate within the CS (right panel).
The strippted rectangles .show the range of values.
 Takahashi, et al. Coronary Vein Accessory Pathways 1019

SR ABL SITE RAO

Figure 5. A typical acces.mry pathway ablated deep inside the tniddle cardiac vein. The delta wave potarity is tiegative in
lead II and positive in lead aVR with a steep angle of > 45° during .sinus rhythm (SR) and atriat pacing (AP). However, a
deep S wave (R < S) in lead V^ becomes evident only during atrial pacing. The bipolar electrogram at the succes.tful ablation
site (ABL SITE) .shows a long AV interval (50 msec) and a sharp accessory pathway potential (K) with an amplitude of 0.05
mV. The unipolar etectrogram exhibits a P-QS pattern. Right panel: Coronary sinus venography showing that the successful
ahlation site was 2 cm deep inside a middle cardiac vein arising frotn near the coronaiy sinus ostium. ABL = ablation site
bipolar electrogram: AP = atrial pacing; LAO = left atiterior oblique projection: RAO = right anterior oblique projection;
SR = .sinus rhythm; UNI = ahlation site unipolar etectrogram.

iMcal Electrogram Characteristics AP potentials (Fig. 5) were recorded more fre-

Local electrogram characteristics are detailed in
Table 1. There were no significant differences
between groups MCV and CS except for the AV
ratio. In both groups MCV and CS, the local peak
ventricular potential relative to the delta wave
onset occurred later than in group R or L {-0.9 ±
7.5 msec and -0.5 ± 6.4 vs -15 ± 8.6 msec and -
5 ± 5.3 msec; P < 0.05) and the local AV inter-
val was longer (47 ± 16 msec and 46 ± 18 msec
v.s 32 ± 12 msec and 35 ± 8.4 msec, respectively;
P < 0.01).
Although the QS pattem of the unipolar ven-
tricular waveform was evenly distributed, a fusion
of P and QS, i.e., a PQS pattern (short AV delay)
was recorded in 12% of group R, 10% of group
L, and 0% of both groups MCV and CS.
quently in both group MCV (11%) and group CS
(80%) than in group R (37%) or in group L (41 %)
(P < 0.05). There was no significant statistical dif-
ference in the interval between the onset of appli-
cation of radiofrequency energy and the interrup-
tion of AP conduction.
Electrocardiographic Patterns During Sinus
Rhythm
Thefindingsof ECG pattems during sinus rhythm
are detailed in Table 2. A negative delta wave in
lead II and a positive delta wave in lead aVR
were significantly more frequent in group MCV,
group CS, or both, than in group R and group L
(Figs. 5 to 7). A negative delta wave in V, was pms-
ent only in group R. A positive delta wave and QRS
1020 Journal of Cardiovascular Klectrophjsiology Vol. 9, No. 10. October 1998

complex were significantly more frequent in group

u L than in group R. There was no significiuitly dif-
y H- 2 2 2 ferent parameter between groups MCV and CS.
(-' •*-* 2 2 * ' Z Z
E-E Electrocardiographic Patterns During Maximal
Preexcitation (Table 3)
11 There was no significant difference in the neg-
ative delta wave in lead II between sinus rhythm
and atrial pacing in contrast to ECG pattems in
aVR and V,,. A steep negative delta wave in lead
II and a positive delta wave as well as a steep pos-
itive delta wave in lead aVR were significantly
more frequent in group MCV (85%) than in
other groups, and a steep positive delta wave in
u lead aVR was always ass(x:iated with a steep neg-
on c/1 VI
u c/i2 ative delta wave in lead II. The QRS morphology
2 2
^- +-^ * 2 • included a deep S wave in V^ significantly more
22 frequently in groups MCV (85%). CS (50%), or
both, than in group R and group L (Figs. 5 to 7).
> ad
A negative delta wave in V, was present in only
group R. A positive delta wave and QRS complex
were significantly more frequent in group L than
+1 +1 +1 00 in group R.
in n OS ^
— f*! _ ; n-i

V! t Sensitivity, Specificity, and Positive and Negative

U o Predictive Values
a.
TAB
eristic

> The sensitivity, specificity, and positive and neg-

c/5 on c/5 on
ative predictive values of ECG patterns during si-
>s 2222 nus rhythm and atrial pacing tor both groups MCV
I/)
OH
U •C II ^ and CS are detailed in Table 4. In lead II, a nega-
b II "u
tive delta wave yielded a high sensitivity (87%),
e but specificity (79%) and positive predictive values
q ri
(50%) were relatively low. In contrast, a steep pos-
rd ri itive delta wave in aVR and deep S wave in V,, dur-
x: D.
u &••-
ing preexcitation were less sensitive but had higher
specificity (98% and 87%) and positive predictive
> E value (88% and 57%). The combination of ihe two
2 latter parameters dramatically increased the speci-
ficity to 99% and the positive predictive value to
91%, but with a sensitivity of only 43%. The sen-
y II +1 +1 +1 +1 +1 sitivity of this combination increased to 78% with
^z rn
d - o q ^•
o t
the same specificity and positive predictive value
- - O when analyzed only for group MCV (Table 5).
In the 58 patients with a manifest single nearby
AP. none had a combination of a steep positive
delta wave in lead aVR and a deep S wave in lead
S I o V^ during maximal preexcitation. A negative delta
wave in V, had high specificity (100%) and pos-
ego itive predictive value (1(K)%) (sensitivity: 27%) for
Si s : :
C/5 the right endocardial approach during atrial pac-
C- ing. When the epicardial group was excluded from
<1 > analysis, a positive delta wave and QRS complex
> <
had a high sensitivity (85% and 79%) but rela-
 Takatia.stu, et at. Coronary Vein Accessory Pathways 1021

TABLE 2
Prevalence of Selected Surface Electrocardiographic Parameters in Patients with ji Manifest Posteroseptal
Accessory Pathway Related to the Site of Successful Ablation During Sinus Rhyihrn

MCV CS P Value
(N = 13) (N = tO) (N = 34/60) (End vs MCV. CS, Epi)
11 vcA 11 (85%) y(y()9{) 7(21%)/I3(22'7r) t. t . +
Steep - v e A 10 (77%) 4 (40%) 2(6%V6(IO%) t,t,|
aVR + veA 9 (69%) 4 (40%) 3 (9%)/5 (8%) t,t.t
Steep +veA 6 (46%) 4(40%) 1 (3%)/2 (3%) t,t,t
V,, Deep S wave 1 (8%) 1 (10%) 0 (0%V3 (5%) NS, NS, NS
V, — veA 0(0%) 0(0%) 0(0%)/! 7 (28%) t.t.t
+ veA to (77%) 8 (80%) 24{7I7P)/23(38%) NS. NS. *
+ ve QRS 7 (54%) 5 (50%) 20(59%)/18(30%) NS, NS. NS
aVR = lead aVR; CS = accessory patliway ablated inside CS including MCV ostium; A - delta wave; End = both L and R;
Epi = both MCV and CS; II = lead II; L, R=accessory pathway ablated from right or lelt posterosepta! endocardium; MCV
= accessory pathway ablated deep inside middle cardiac vein; —ve = negative; +ve = positive; V, = lead Vi; V,. = lead V^.
* = P<0.()5; t ^ p<O.OI.

tively low specificity (72% and 64%) and positive
predictive value (62% and 54%) for the left en-
docardial approach during atrial pacing.
Discussion
The results of this study indicate that patients
witb posteroseptal APs requiring radiofrequency
catbeter ablation from the CS and notably inside
the MCV pt)ssessed specific features derived from
tbe delta wave polarity and QRS morphology dur-
ing maximal preexcitation, which distinguish them
from APs requiring ablation from the endocardial
space. Only one surface ECG feature — a nega-
tive delta wave in lead V, — identified right en-
docai'dial AP.
Delta Wave Analysis in Sinus Rhythm and Maximal
Preexcitation
Arruda et al.'^ recently reported ECG criteria
during sinus rbytbm to identify APs related lo the
MCV or anomalies of the CS. A negative delta
wave in lead II was found to be I(X)% sensitive
as well as pathognornonic of a subepicardial site
of the AP. In the present study, a negative delta
wave in lead 11 had a high sensitivity of 87%; bow-
ever, the specificity was relatively low. resulting

TABLE 3
Prevalence of Selected Surface Electrocardiographic Parameters in Patients with a Manifest Posteroseptal
Accessory Pathway Related to the Site of Successful Ablation During Atrial Pacing
MCV CS E/R P Value
(N - t3) (N = tO) (N = 34/60) (End v.sMCV, CS, Epi)
11 -vcA 12(92%) 8 (80%) 7(2170/14(23%) t, 1.1
*
Steep - v e A 11 (85%) 4 (40%) 2(6%)/6(10%) t.t.t
aVR + veA 11 (85%) 2 (20%) 4(12%)/5(8%) t.NS.t
Steep +veA 1 1 (85%) 3 (30%) t (3%)/l (2%) t.t.t
V. Deep S wave I 1 (85%) 5 (50%) 6(t8%)/6(10%) tt.t
V, -veA 0(0%) 0(0%) 0(0%)/! 9 (32%) t.t.t
+veA to (77%) 8 (80%) 27 (79%)/23 (38%) NS. NS.*
+ ve QRS 10 (77%) 8 (80%) 29(85%)/18(30%) NS. NS.*
aVR = lead aVR; CS = accessory pathway ablated inside CS including MCV ostium; A = delta wave; End = both L and R;
Epi = both MCV and CS: II = lead II: L. R^accessory palhway ablated Irom right or left posteroseptal endocardium; MCV
= accessory pathway ablated deep inside middle cardiac vein; —vc = negative; +ve = positive; Vi = lead V,: V^ = lead
V.. * = P < 0.05; t == P <().()!.
1022 Journal of Cardiovascular Electrophysiology Vol. 9, No. 10, October 1998

TABLE 4
Sensitivity, Specificity and Predictive Values of the Delta Wave Pattern and QRS Morphology for
Successful Coronary Sinus and Coronary Vein Ablation Site

Sensitivity Specificity PPV NPV

SR/AP SR/AP SR/AP SR/AP

Smgle parameters
II -veA 87/87 79/78 50/49 96/96
Steep —veA 61/65 91/91 64/65 91/91
aVR +veA 52/52 91/90 62/62 90/89
Steep -HveA 43/61 97/98 77/88 88/91
*Vh Deep S wave -/70 -/87 -/57 -/93
[Combination of parameters
paVR +veA
& -/43 /99
'-Vf. Deep S wave
&
-II Steep —veA
&
aVR +veA -/43 -/99 -/91 -/88
&
-Vf, Deep S wave
AP = atrial pacing; aVR = lead aVR; II = lead II; NPV = negative predictive value; PPV = Positive predicti\e value; SR =
sinus rhythm; V^ = lead V^. A = delta wave; —ve = negative; +ve = po.sitive; ""Because only five patients had this finding
in sinus rhythm, the sensitivity, specificity, PPV and NPV of thi.s parameter were calculated during atrial pacing alone.

in low positive predictive values. The difference
may be due to different definitions of delta wave
analysis; the initial 40 msec versus 20 msec, or the
difference in ablation sites based on the earliest
bipolar and unipolar ventricular waveform in the
present study and AP activation potential in the
previous study.'^ In another study, Scaglione et al.'"*
reported the usefulness of lead aVR in predicting
left posteroseptal pathways (including both the CS
and left endocardium). In the present study, a steep
positive delta wave in lead aVR during maximal
preexcitation had a high specificity and predic-
tive value only for ablation inside the CS or MCV.
This criterion (found in 61% of patients) was the
strongest single predictor for identifying such epi-
cardial APs.

TABLE 5
Sensitivity, Specificity and Predictive Values of tbe Delta Wave Pattem and QRS Morphology for a
Successful Middle Cardiac Vein Ablation Site

Sensitivity Specificity PPV NPV

SR/AP SR/AP SR/AP SR/AP

Single parameters
veA 85/92 72/72 28/29 97/99
Steep -veA 77/85 88/88 45/50 97/98
aVR Steep +veA 69/69 88/89 43/54 96/98
Steep +veA 46/85 93/95 46/65 93/98
"V,, Deepp S wave -/85 -/39 -/98
Combination of parameters
Steep + veA
& -/78 -/99 -/91 -/97
v. Deep&S wave
t-II Steep —veA
&
aVR Steep +veA -/69 -/99 -/9I -/96
&
Deep S wave
AP = atrial pacing; aVR = lead aVR; II = lead II; NPV = negative predictive value; PPV = Positive predictive value; SR =
sinus rhythm; V,, = lead Vh. A = delta wave; —ve = negative; +ve = positive; *Because only live palients had this finding
in sinus rhythm, the sensitivity, specificity. PPV and NPV of this parameter were calculated during atrial pacing alone.
 Takahashi, et al.
YYYY
Figure 6. The H-lead ECG during sinus rhythm (left) and
atrial pacing (right) in a patient wilh a manifest accessory
A V connection ablated deep inside the middle cardiac vein.
The delta wave polarity is negative in lead ll and positive
in lead aVR with a steep angle of > 45° during sinus
rhythm ami atrial pacing. However, a deep S wave (R < S)
in lead V^ hecomes evident only during atrial pacing. AP =
alrial pacing; SR = sinus rhythm.
QRS Morphology Analysis in Sinus Rhythm and
Maximal Preexcitation
A deep S wave in lead V^, appeared during
atrial pacing in 85% or 50% of patients whose
AP was ablated from deep inside the MCV or the
CS compared to 5% ablated from the endocar-
dial surface, while this ECG feature was inap-
parent during sinus rhythm except in five patients.
These findings indicate that analysis of complete
QRS morphology in maximal preexcitation in-
cluding terminal vectors is useful. Similarly, the
latest activation in the QRS morphology was used
to identify the origin of ventricular tachycardia.-"
This ECG feature is immediately obvious whereas
the assessment of delta wave polarity in aVR
requires conelation with delta wave onset in other
simultaneous ECG leads. The appearance of a
deep S wave in lead V,, during preexeitation may
be due to epicardial activation and/or ventricular
activation funher away from the mitral annulus.
The combination of a deep S wave in lead V,,
with a positive delta wave in lead aVR increased
the specificity and positive predictive value for
epicardiai approach to a maximum of 99% and
91%, respectively, but sensitivity was still rela-
tively low (43%). When analyzed only for group
MCV, the sensitivity increased to 78% with the
same specificity and positive predictive values.
This combination appears to be the most specific
Coronary Vein Accessory Pathways 1023
feature of posteroseptal APs requiring radiofre-
quency catheter ablation from deep inside the
MCV
Different Electrogram Characteristics of
Posteroseptal AP
Effective target site electrograms of APs ablated
from inside the MCV or CS were characterized by
a clearly longer local AV interval and a more fre-
quently recorded AP potential compared to those
ablated from the endocaidial surface of the right
or left posterior septum. In our experience, the AV
time for these pathways was actually the longest
of all APs, including those Inserted in the free
wall,'^ suggesting that posteroseptal APs ablated
from inside the MCV or CS have a different sub-
strate, arelativelylong course — obvious for MCV
pathways — or a long conduction time in associ-
ation with their epicardial location. The higher
incidence of recorded AP potential within the MCV
or CS has also been shown in previous reports^"*'^
and may be due to longer atrial to ventricular ac-
tivation intervals allowing separation from local
atrial and ventricular potentials or a closer posi-
tion of the recording electrode to the pathway.
Study Limitations
In this study, we did not prospectively test
the ECG features in a new cohort of patients with
manifest posteroseptal APs; however, the sample
size in this study was relatively large, and all pa-
tients with a single pathway were included. We
performed CS angiography in only five patients,
which is an important limitation. However, the
combination of fluoroscopie catheter course and
impedance monitoring allowed us to anticipate
the actual catheter location even in these five pa-
tients, and this method was therefore used for
imaging confirmation. No patients had compli-
cations, including those with radiofrequency ap-
plication inside the CS or MCV, although no spe-
cific precautions other than temperature control
with a low target temperature and automatic high-
impedance (> 250 iT) radiofrequency cutoff were
used. The resultant maximal power output was
low, particularly in small veins, therefore avoid-
ing perforation. The ablation strategy was based
on targeting the ventricular insenion site of pos-
teroseptal APs and, because of oblique orienta-
tion, the ECG correlation with epicardial inser-
tions may be different if the atrial insertion site
were to be targeted.
1024 Journal of Cardiovascular Electrophysiology Vol. 9, No. 10, October 1998

SR AP

Figure 7, The 12-lead ECG dtiring sinus rhythm (left) and atriat pacing {right) in a patient with manifest acces.sory AV con-
nection ablated deep inside the middle cardiac vein. The delta wave polarity is negative in lead II and positive in lead aVR
with a steep angle of > 45° during sinus rhythm and atrial pacing. A deep S wave (R < S) in lead V^ wa.s evident during atrial
pacing as well as sinus rhythm. AP = atriat pacing: SR = sinus rhythm.

Conclusion
The results of this study confirm the high sen-
sitivity of negative delta wave in lead II for pos-
teroseptal APs requiring radiofrequency ablation
inside the CS and MCV. These pathways, no-
tably those inserting deep inside the MCV,
have specific features in the form of a combi-
nation of a steep positive delta wave in lead aVR
and a deep S wave in lead V^. during maximal
preexcitation.
References
Wariti JF, Haissaguene M, Le Metayer P. et al: Cathe-
ter ablation of accessory pathways with a direct ap-
proach. Circulation 1988;78:8OO-815.
Jackman WM, Wang X, Friday KJ, et al: Catheter ab-
lation of accessory atriovenlHcular pathways (Wolff-
Park in son-White syndrome) by radiofrequency current.
N Engl J Med 1991 ;324:1605-1611.
Kuck KH. Schluter M. Geiger M. et al: Radiofre-
quency curren! catheter ablation of accessory atrioven-
tricular pathways. Lancet 1991:337:1557-1561.
 Takahashi. et al.
4. Calkins H, Sousa J. El-Atassi R. el at: Diagnosis and
cure of Wolff-Parkinson-While syndrome or paroxys-
mal supra ventricular tachycardias during a single elec-
trophysiologic lest. N Engl J Med 1991:324:1612-1618,
5. Lesh MD, Van Hare G. Schamp DJ, et ai: Curative
percutaneous catheter ablation using radiofrequency
energy for accessory pathways in all locations: Results
in 100 consecutive patienls. Am J Cardiol 1992;
70:114-116.
6. Kuck KH. Schluter M. Chiladakis I: Accessory path-
ways analoinically related to the coronary sinus. (Ab-
stract) Circulation 1992;86:I-782.
7. Wang X. Jackman WM. McClelland J. el al: Sites of
successful radiofrequency ablaiion of posteroseptal ac-
cessory pathways. (Abstract) PACE 1992;15:121.
8. Oren J, McClelland J, Beckman K. et al: Epicardial
posteroseptal accessory pathways requiring ablation
from the middle cardiac vein. (Abstraci) PACE 1992;
15; 124.
9. Langberg JJ, Ching Man Do K, Vorperian VR. et al:
Recognition and catheler ablation of subepicardial acces-
sory pathways. J Am Coll Cardiol 1993;22; 1100-1104.
10. Arruda MS, Beckman KJ. McClelland J. et al: Coro-
nary sinus anatomy and anomalies in patients with pos-
teroseptal accessory pathway requiring ablation within
a venous branch of the coronary sinus. (Abstract) J Am
Coll Cardiol I994;23:224A.
11. Dhala AA. Deshpande SS. Bremer S. ct al: Transcathe-
ter ablation of posteroseptal accessory pathways using
a venous approach and radiofrequency energy. Circula-
tion I994;9O;I799-1SIO,
12. Arruda MS. McClelland JH. Wang X, et al: Develop-
ment and validation of an ECG algorithm for identify-
ing accessory pathway ablation site in Wolff-Parkin-
Coronary Vein Accessory Pathways 1025
son-White syndrome. J Cardiovasc Electrophysiol
1998;9;2-I2.
13. Davis LM. Byth K. Ellis P, et al: Dimensions of ihe
human posterior septal space and coronary sinus. Am J
Cardiol 1991:68:621-625.
14. Cox JL: Analomy of the 'posterior septal space.' Am J
Cardiol l991;68;675-677.
15. HaVssaguerre M, Gaita F, Marcus FI. et al; Radiofre-
quency catheter ablation of accessory pathways; A
contemporary review. J Cardiovasc Blecirophysiol
1994:5:532-552.
16. Lemery R. Hammill SC. Wood DL, ct al; Value ot ihe
resting 12-lead electrocardiogram and veclorcardio-
gram for locating the accessory pathway in patients
with Wolff-Parkinson-White syndrome. Br Heart J
I987;5S:324-332.
17. Fitzpatrick AP. Gonzales RP, Lesh MD. et al: New al-
gorithm for the localization of accessory atrioventricu-
lar conneclions using a baseline electrocardiogram. J
Am Coll Cardiol 1994;23:107-l 16.
18. Chiang CE, Chen SA, Teo WS. et al; An accurate step-
wise elecirocardiographic algorithm for localization of
accessory pathways in patient.s with Wolff-Parkinson
While syndrome from a comprehensive analysis of
delta waves and R/S ratio during sinus rhythm. Am J
Cardiol 1995:76:40-46.
19. Scaglione M, Riccardi R, Socchierdo M, ei al: ECG
pattern predicting right sided ablation of left pos-
teroseptal accessory pathways. (Abstract) Eur J Car-
diac Pacing Eiectrophysiol I994;4:839.
20. Josepbson ME. Horowitz LN. Waxman HL. ct al; Sus-
tained venlricular tachycardia: Roie of the 12-lead
electrocardiogram in localizing site of origin. Circula-
tion 1981:64:257-272.