Identification of sinus node dysfunction by use of

P-wave signal-averaged electrocardiograms in

paroxysmal atrial fibrillation: A prospective study
Takahisa Yamada, MD,a Masatake Fukunami, MD,a Tsuyoshi Shimonagata, MD,a Kazuaki Kumagai, MD,a
Yoshihiro Asano, MD,a Akio Hirata, MD,a Mitsutoshi Asai, MD,a Masatsugu Hori, MD,b and Noritake Hoki, MDa
Osaka, Japan

Background In patients with paroxysmal atrial fibrillation (Paf), the identification of the coexistence of sinus node dys-
function (SND) has therapeutic implications. This study sought to prospectively determine whether SND in patients with Paf
would be identified by use of atrial early potential (EP), low-amplitude potentials early in signal-averaged P wave.
Methods The study population consisted of 149 patients with Paf. Signal-averaged electrocardiography was recorded
with the P-wave–triggering technique. The root mean square voltage for the initial 30 MS and the duration of initial low-
amplitude signals <4 µV of signal-averaged P wave were measured in the vector magnitude. The criteria of EP were defined
as “the root mean square voltage for the initial 30 MS <3.0 µV and the duration of initial low-amplitude signals <4 µV >22
MS.” SND was diagnosed by use of the conventional 12-lead electrocardiography, 24-hour Holter monitoring, and bed-
side electrocardiographic monitoring.
Results Thirty-eight of 149 patients with Paf had EP. Eighteen (47%) of 38 patients with Paf and EP had SND, whereas
SND was found in only 5 (5%) of the other 111 patients with Paf without EP (P < .0001). EP gave a sensitivity of 78% and a
specificity of 84% for the detection of SND in patients with Paf.
Conclusion EP would be useful for the identification of SND in patients with Paf. (Am Heart J 2001;142:286-93.)

Some patients with paroxysmal atrial fibrillation (Paf)
have the coexistence of sinus node dysfunction (SND).1-4
In cases in which the preventive antiarrhythmic drugs
for Paf were administered to these patients with occult
SND, the termination of Paf might be followed by unpre-
dictable longer pauses, leading to presyncope or syn-
cope. Thus the identification of the coexistence of SND
in patients with Paf has therapeutic implications.
We5-8 and other investigators9-11 reported that the elec-
trophysiological abnormalities of the atrial muscle in
patients with Paf could be detected noninvasively by P-
wave signal–averaged electrocardiography. In the patho-
logic studies of SND, a lesion such as degeneration and
fibrosis has been shown not only in the sinus node but
also in the perinodal atrial muscle.12-15 Recently, in a ret-
rospective study, we reported that low-amplitude poten-
tials early in the signal-averaged P wave, atrial early poten-
tials (EP), which might reflect the conduction
From the aDivision of Cardiology, Osaka Prefectural General Hospital, and bDe-
partment of Internal Medicine and Therapeutics, Osaka University Medical School.
Presented in part at the 48th Scientific Sessions of the American College of Cardiol-
ogy, New Orleans, La, March 9, 1999.
Submitted August 24, 2000; accepted March 28, 2001.
Reprint requests: Takahisa Yamada, MD, Division of Cardiology, Osaka Prefectural
Hospital, 3-1-56, Mandai-Higashi, Sumiyoshi-ku, Osaka 558-8558, Japan.
Copyright 2001 by Mosby, Inc.
0002-8703/2001 $35.00 + 0 4/1/116474
doi:10.1067/mhj.2001.116474
abnormalities in the perinodal atrial muscle, would be
characteristic of SND.16 Therefore the purpose of this
study was to prospectively determine whether the
involvement of SND in patients with Paf would be identi-
fied by EP.
Methods
Study patients
One hundred fifty-six consecutive patients with Paf, who
underwent P-wave signal–averaged electrocardiography
between January 1995 and December 1998 in Osaka Prefec-
tural General Hospital, were screened for this study. One hun-
dred eighteen of 156 study patients were admitted because of
the examination results and for treatment of arrhythmias in 51
patients, heart catheterization in 37 patients, and the manage-
ment of diseases other than heart disorder (diabetes mellitus,
brain infarction, etc) in 30 patients, whereas the remaining 38
patients were in the outpatient clinical setting. Paf was defined
as an arrhythmia of supraventricular origin associated with a
grossly irregular ventricular rhythm and no visible P or flutter
waves that lasted for >1 minute and did not persist for 6
months. Seven patients were excluded because (1) 6 patients
were taking antiarrhythmic agents that could affect the results
of P-wave signal–averaged electrocardiography, and (2) the
mean noise level was >1 µV in the composite lead of P-wave
signal–averaged electrocardiogram in the remaining 1 patient.
A total of 149 patients were enrolled in this study. Each patient
gave informed consent to participate in this study, which was
American Heart Journal
Volume 142, Number 2
approved by the Osaka Prefectural General Hospital Review
Committee.
The mean age of 149 patients was 61 ± 14 years. There
were 97 men and 52 women. Eighty-six patients had no
organic heart disease, and the remaining 63 patients had
organic heart diseases or another disease that possibly caused
Paf (24 patients with ischemic heart disease, 14 with pericar-
dial or myocardial diseases, 11 with valvular heart disease, 9
with hypertension or hypertensive heart disease, 4 with con-
genital heart disease, and 1 with pulmonary embolism). One
hundred forty of 149 patients had symptomatic episodes. Sev-
enty-seven patients experienced only palpitation, and the
remaining 63 patients had other symptoms besides palpita-
tion. Eighteen patients had syncope, 15 presyncope, 5 light-
headedness, 17 exertional dyspnea, 7 chest pain, and 1 easy
fatigability. The interval from the first symptomatic episode,
such as palpitation, to entry, defined as the time of P-wave sig-
nal-averaged electrocardiogram, was 26 ± 49 months, and the
interval from the last symptomatic episode to entry was 1.2 ±
0.8 months in patients with symptomatic Paf attacks.
As a control group, we also studied 15 patients with SND
without Paf in our previous prospective study.17 The mean
age of 15 patients was 66 ± 15 years. There were 8 men and 7
women. Ten patients had no organic heart disease, and the
remaining 5 patients had organic heart disease (2 patients
with ischemic heart disease, 2 with hypertension or hyperten-
sive heart disease, and 1 with chronic myocarditis). Thirteen
of 15 patients had symptomatic episodes. Six patients experi-
enced syncope, 3 had presyncope, 2 had light-headedness, 1
had chest pain, and 1 had palpitation. The sinus pause docu-
mented by Holter monitoring was 5.3 ± 2.5 seconds.
P-wave signal–averaged electrocardiography
In all patients, P-wave signal-averaged electrocardiography
was performed at the entry in a blinded fashion to the patients’
clinical data. The method of P-wave signal-averaged electrocar-
diographic recording and analysis has been described previ-
ously.5-7,16,18 No patients in this study had received antiarrhyth-
mic drugs for at least 1 week before undergoing P-wave
signal-averaged electrocardiography. They had never received
amiodarone before the entry. Eight patients received digitalis,
and 9 patients received calcium antagonist at the recording of
P-wave signal-averaged electrocardiography. In an electrically
shielded room, which minimized noise, P-wave signal-averaged
electrocardiography was recorded from a modified X-, Y-, and
Z- lead system by use of the VCM-3000 (Fukuda Denshi, Ltd,
Tokyo, Japan), which was recently developed for P-wave–trig-
gered signal averaging. The X lead was between the right and
left shoulders (standard lead I). The aVF lead was used as the Y
lead. The precordial V1 lead was used as the Z lead. The signal
from each lead was amplified up to 5 µV/cm and passed
through a unidirectional Butterworth filter of 40 Hz (the slope;
18 decibels per octave [dB/oct]) to 300 Hz (the slope; 12
dB/oct) and was then converted from analog to digital data to a
12-bit accuracy at a sampling rate of 1 kHz.
A specially filtered P wave derived from the selected domi-
nant sinus P wave of the standard II lead served as a reference
signal (template) for all processing. The specially filtered P
wave was obtained with a band-pass filter of 10 to 30 Hz for
making the P wave spiky as a trigger signal. After passing
through a P-wave recognition program to eliminate ectopic
Yamada et al 287
atrial beats, the signals of more than 200 beats were averaged
on a trigger point within a specially filtered P wave. Noise lev-
els were measured every 1 millisecond in the last 20 millisec-
onds of TP segment on the filtered lead of a vector magnitude,
the square root of X2 + Y2 + Z2. Signal averaging was contin-
ued until the noise amplitude at all points in this interval was
reduced to less than 1 µV (peak noise level). The root mean
square noise value was 0.3 ± 0.2 µV.
The signals for the X, Y, and Z leads were combined into the
vector magnitude. The signal-averaged P wave in the vector
magnitude was defined as signal within the interval showing a
persistent level >1 µV. The onset and offset of signal-averaged
P waves were manually determined without knowledge of the
patient’s clinical data. We measured the root mean square volt-
ages for the initial 30 milliseconds of signal-averaged P wave
(EP30) and the shortest duration from the onset up to 4 µV of
signal-averaged P wave (ED4). In our previous study,16 the cri-
teria of atrial early potentials (EP) had been defined as EP30
<3.0 µV and ED4 >22 MS. Furthermore, the duration and the
root mean square voltage for the last 20 MS of signal-averaged
P wave were also measured.
Diagnosis of the presence of SND
SND was diagnosed by use of conventional 12-lead elec-
trocardiography, 24-hour ambulatory Holter monitoring,
bedside monitoring with continuous electrocardiographic
telemetry, or electrophysiologic study, which were per-
formed within 1 month after P-wave signal-averaged electro-
cardiography was recorded. All of the 149 patients under-
went 24-hour Holter monitoring or bedside
electrocardiographic monitoring. In 136 of 149 patients, 24-
hour ambulatory Holter electrocardiography was recorded
and was analyzed by use of a Marquette Electronics 8000
Holter monitoring system (Marquette Electronics, Milwau-
kee, Wis). In 118 of 149 patients, bedside monitoring with
continuous electrocardiographic telemetry was also per-
formed for at least 2 days (5 ± 3 days). Patients were consid-
ered to have SND if they had an episode of sinus arrest or
sinoatrial block (with a pause >2 seconds), persistent and
unexplained sinus bradycardia (<40 beats/min), prolonged
corrected sinus recovery time (CSRT >600 milliseconds), or
sinoatrial conduction time (SACT >150 milliseconds).4,16-19
Electrophysiologic study
We performed electrophysiologic study within several days of
the recording of P-wave signal-averaged electrocardiography in
40 patients for the further examination of sinus node function in
21 patients and coexistence of arrhythmias other than SND in
19 patients (ventricular tachycardia in 8, atrioventricular tachy-
cardia in 4, atrioventricular nodal reentrant tachycardia in 3,
paroxysmal atrial flutter in 2, intraatrial reentrant tachycardia in
1, and ventricular fibrillation in 1 patient). As described in our
previous study,16 high right atrial electrograms were recorded
through a band pass filter of 30 to 500 Hz, and atrial stimulation
was performed by use a programmable stimulator (Nihon
Koden, Ltd, Tokyo, Japan) used at twice-diastolic threshold and
2 MS in duration. SACT was obtained according to Narula’s
method.20 Furthermore, CSRT was obtained by overdriving
method (pacing high right atrium for a 30-second period at 80 to
180 beats/min).21

288 Yamada et al
Figure 1
Representative P-wave signal-averaged electrocardiograms in patients with Paf with and without atrial EP. Crite-
ria of EP was defined as “EP30 < 3.0 µV and ED4 >22 milliseconds.” Dotted lines indicate beginning and end of
signal-averaged P wave. Note that initial portion of signal-averaged P wave is lower in amplitude and longer in
duration in patient with Paf with than without EP.
Echocardiographic measurement
Echocardiography was performed by a Toshiba SSH-160A
recorder equipped with 2.5- to 3.5-MHz transducers in 135 of
149 patients at the entry. The standard technique22 was used
for sizing the left ventricle and atrium. Left ventricular end-
diastolic and end-systolic dimensions and left atrial dimension
were measured, and the left ventricular ejection fraction was
calculated by Gibson’s method.
Statistical analysis
Data are presented as mean ± SD. The Student t test and a
Fisher exact test were used to compare differences of continu-
ous and discrete variables, respectively, in patients with Paf with
and without EP. These analyses were also used for the compari-
son between patients with Paf with and without SND and the
comparison between patients with SND with and without Paf.
The statistical significance was detected at P < .05.
Results
Clinical characteristics in patients with Paf with and
without EP
Thirty-eight (26%) of 149 patients with Paf had EP.
Figure 1 shows the representative tracings of P-wave
American Heart Journal
August 2001
signal-averaged electrocardiograms in patients with Paf
with and without EP. Of note, the initial portion of the
signal-averaged P wave is lower in amplitude and longer
in duration in patients with than without EP. Table I
shows clinical characteristics in patients with Paf with
and without EP. There were no significant differences
in age, sex, the presence of organic heart disease, or
resting heart rate between the two groups.
Study characteristics in patients with Paf with and
without EP
EP30 and ED4 were 2.02 ± 0.47 µV and 37.2 ± 14.7
milliseconds in patients with Paf with EP and 4.14 ±
1.44 µV and 12.8 ± 6.1 milliseconds in patients with Paf
without EP, respectively. Table II shows study charac-
teristics of patients with Paf with and without EP.
There were no significant differences in the percentage
of patients who underwent Holter monitoring (97% vs
89%) or bedside electrocardiography monitoring (84%
vs 78%) or the period of the bedside monitoring (4.8 ±
3.0 vs 5.1 ± 3.0 days) between patients with Paf with
and without EP. In Holter or bedside monitoring, sinus
American Heart Journal
Volume 142, Number 2 Yamada et al 289

Table I. Clinical characteristics of patients with Paf with and Table II. Study characteristics of patients with Paf with and
without EP without EP
Paf with EP Paf without EP Paf with Paf without
EP EP
n 38 111
Age (y) 63 ± 17 61 ± 14 Holter/bedside monitoring (n) (38) (111)
Sex (% male) 27 (71%) 70 (63%) Sinus arrest or sinoatrial block 17 (45%) 5 (5%)*
Symptoms (syncope/ 11 [6/5] (29%) 22 [12/10] (20%) Maximum pause time (sec) 5.5 ± 2.1 3.7 ± 1.9
presyncope) Persistent sinus bradycardia 1 (3%) 0
Heart rate (beats/min) 63 ± 13 67 ± 12 Electrophysiologic test (n) (18) (22)
Organic heart disease (n) 15 (40%) 49 (44%) Sinoatrial conduction time (ms) 176 ± 67 105 ± 41*
Ischemic 5 19 Corrected sinus recovery time (s) 2.16 ± 2.08 0.70 ± 0.88†
Valvular 3 8 Echocardiography (n) (34) (101)
Hypertension 3 6 LV end-diastolic dimension (mm) 50 ± 7 48 ± 6
Myocardial 2 10 LV ejection fraction (%) 65 ± 12 68 ± 10
Congenital 2 2 Left atrial dimension (mm) 41 ± 8 40 ± 7
Others 0 4
Coexistent arrhythmia LV, Left ventricular.
AVNRT 1 5 *P < .001.
AVRT 0 4 †P < .01.
IART 0 1
VT 5 5
Vf 0 1 Identification of patients with Paf with SND by use
None of data show significant differences between the two groups. AVNRT, Atri-
of EP
oventricular nodal reentrant tachycardia; AVRT, atrioventricular reentrant tachycar-
dia; IART, intraatrial reentrant tachycardia; VT, ventricular tachycardia; Vf, ventricu-
Eighteen (47%) of 38 patients with EP had SND,
lar fibrillation. whereas SND was found in only 5 (5%) of the other 111
patients without EP. EP had a sensitivity of 78%, a
specificity of 84%, a positive predictive value of 47%,
arrest or sinoatrial block was significantly more fre- and a negative predictive value of 95% for the identifi-
quently observed in patients with Paf with than with- cation of SND in patients with Paf. Patients with Paf
out EP, although there was no significant difference in with EP had an increased likelihood of the involvement
maximum pause time between the two groups. In elec- of SND (odds ratio 19.1; 95% confidence interval [CI]
trophysiologic study, SACT and CSRT were significantly 6.4 to 57.3; P < .0001).
longer in patients with Paf with than without EP.
Echocardiographically, there were no differences in the Comparison between patients with Paf with and
left ventricular end-diastolic dimension, ejection frac- without SND
tion, or left atrial dimension between the two groups. Table III shows the clinical and study characteristics of
patients with Paf with and without SND. Although there
Clinical characteristics in patients with Paf with SND were no significant differences in sex, resting heart rate,
Twenty-three of 149 patients with Paf were diag- or the presence of organic heart disease between
nosed as having SND. All but two patients with SND patients with Paf with and without SND, patients with
experienced symptoms related to the arrhythmia, in SND were significantly older and had the episode of syn-
addition to palpitation. Seven patients experienced syn- cope or presyncope significantly more frequently than
cope, 7 had presyncope, 3 had light-headedness, 2 had those without SND. In patients with Paf with SND, the
chest pain, 1 had exertional dyspnea, and 1 had easy fati- 95% CIs of EP30 and ED4 were 1.90 to 2.63 µV and 28.2
gability. The mean of maximum pause time was 5.1 ± 4.7 to 45.4 milliseconds, respectively. In patients with Paf
(2.2 to 9) seconds. The pause was documented at the without SND, the 95% CI of EP30 and ED4 were 3.57 to
termination of Paf in 13 patients and was also observed 4.12 µV and 14.1 to 17.4 milliseconds, respectively.
at the occurrence of sinoatrial block in 11 patients. EP30 was significantly lower, and ED4 was longer in
Eleven (61%) of 18 patients with Paf with SND and EP patients with than without SND (Figure 2).
experienced the pause because of sinoatrial block,
whereas sinoatrial block was documented in none of 5 Comparison between patients with SND with and
patients with Paf with SND but without EP. The pause without Paf
documented at sinoatrial block was significantly more Table III also shows the clinical and study charac-
frequently observed in patients with Paf with SND and teristics of patients with SND with and without Paf.
EP than those without EP (P < .05), whereas there was There was significant difference in neither EP30 nor
no significant difference in the maximum pause time ED4 between SND patients with and without Paf (Fig-
between the patients with and without EP. ure 2). The incidence of EP in patients with SND with
 American Heart Journal
290 Yamada et al August 2001

Table III. Clinical and study characteristics of patients with Paf with and without SND, and patients with SND without Paf

Paf with SND Paf without SND SND without Paf

n 23 126 15
Age (y) 68 ± 14* 60 ± 15 66 ± 15
Sex (% male) 15 (65%) 83 (66%) 8 (53%)
Symptoms (syncope/presyncope) 14 [7/7] (60%)† 19 [11/8] (15%) 9 [6/3] (60%)†
Heart rate (beats/min) 62 ± 11 67 ± 13 63 ± 13
Organic heart disease (n) 8 (35%) 55 (44%) 5 (33%)
P-wave signal-averaged electrocardiograms
EP 18 (78%)† 20 (16%) 12 (80%)†
EP30 (µV) 2.26 ± 0.85† 3.84 ± 1.54 2.17 ± 1.01†
ED4 (MS) 36.7 ± 20.4† 15.8 ± 9.5 41.7 ± 23.7†
Ad (MS) 150.5 ± 17.8* 138.6 ± 18.5 140 ± 14.7
LP20 (µV) 2.84 ± 1.32 2.89 ± 1.14 3.25 ± 0.96
Electrophysiologic test
Sinoatrial conduction time (MS) 174 ± 68† 101 ± 30 206 ± 33†
Corrected sinus recovery time (sec) 2.33 ± 1.95† 0.37 ± 0.11 1.92 ± 1.12†
Echocardiography (n = 20) (n = 20) (n = 12)
Left atrial dimension (mm) 40 ± 8 40 ± 7 38 ± 4

Ad, The duration of signal-averaged P wave; LP20, the root mean square voltage for the last 20 milliseconds of the signal-averaged P wave.
*P < .05 vs Paf without SND.
†P < .001 vs Paf without SND.

Paf was similar to that in patients with SND without
Paf. This result indicates that EP would not be spe-
cific to patients with Paf but might be a general
marker of SND.
Follow-up study
After the initial examination, all of the patients in this
study were followed up for 18 ± 15 months. In 19 of 23
patients with Paf who had been diagnosed as having
SND at the initial examination, permanent pacemakers
were implanted according to American College of Car-
diology/American Heart Association guidelines.23 Two
years after the recording of P-wave signal-averaged elec-
trocardiography, SND was newly involved in one of
126 patients (20 with and 106 without EP) who had
been diagnosed as having no coexistence of SND at the
initial examination. One (5%) of 20 patients with EP
had the involvement of SND, whereas SND coexisted in
none of 106 patients without EP during the follow-up
period.
Discussion
Signal-averaged electrocardiography had been
developed to detect ventricular and atrial late poten-
tials from the body surface and provided a useful
approach to identify patients at risk for ventricular
tachycardia24-26 and paroxysmal atrial fibrillation.5-11
Recently, we reported that patients with SND had
long, low-amplitude signals in the initial portion of
signal-averaged P wave.16 However, the previous
study was performed in a retrospective fashion.
Therefore this prospective was done, which demon-
strated that EP could be useful for identifying the
involvement of SND in patients with Paf.
Pathophysiologic consideration about EP
Electrophysiological study with intraatrial catheter
mapping of the right atrium showed that the area of dis-
eased atrial muscle, where fractionated atrial electro-
grams were recorded, was more extensive in patients
with both Paf and SND than that in patients with Paf
alone.3 Irrespective of the presence of Paf, patients with
SND had the fractionated atrial electrograms mainly
recorded in the high right atrium in the vicinity of the
sinus node27 and the widening of fragmented activity
zone, compared with those with normal sinus node
function.19,28 In the pathologic studies of SND, abnor-
malities such as degeneration and fibrosis have been
shown not only in the sinus node but also in the atrial
muscle, especially the perinodal portion.12-15 These
findings show that some pathologic and electrophysio-
logic abnormalities in the atrial muscle besides the sinus
node itself might be involved in SND.
In this study SACT was significantly prolonged in
patients with than without EP. In our previous study,
EP30 significantly inversely correlated and ED4 also
significantly correlated with SACT.16 Moreover, in this
study the duration of atrial electrography at the high
right atrium was significantly correlated with EP30 (r
= 0.39, P = .03 [n = 30]) and ED4 (r = 0.55, P = .002).
We believe that EP might reflect the conduction
abnormalities in the perinodal sinoatrial conducting
cells or perinodal atrial muscle. In SND the conduc-
tion of excitation from sinus node through perinodal
American Heart Journal
Volume 142, Number 2
Figure 2
Plot of each parameter of atrial EP in patients with Paf with and without SND and patients with SND without Paf.
Boxes indicate 95% CIs of each parameter, and bold lines indicate mean value. Irrespective of presence or
absence of Paf, EP30 was significantly lower and ED4 was significantly longer in patients with than without SND.
tissue might become slower because perinodal atrial
muscles are widely separated by fibrous tissue,12-15
which might make the amplitude and duration in the ini-
tial portion of signal-averaged P wave lower and longer.
Patients with both Paf and SND often show the sub-
stantial pause documented not only at the termination
of Paf attacks but also at the occurrence of sinoatrial
block. In this study the pause resulting from sinoatrial
block was significantly more frequently observed in
patients with SND with than without EP. This finding
also suggests that EP might reflect the conduction
abnormalities of the excitation from sinus node
through perinodal atrial tissue.
Clinical implication of EP
In this study some patients with Paf had coexisting
SND. If the preventive antiarrhythmic drugs for Paf are
administered to these patients with occult SND, the termi-
nation of Paf might be followed by unpredictable longer
pauses leading to presyncope or syncope. The presence
of EP in patients with Paf might suggest sinoatrial or atrial
conducting disease as the underlying problem. Before the
preventive antiarrhythmic drugs are administered to
patients with Paf and EP, we should investigate whether
SND might be involved in these patients. Thus the identifi-
cation of coexistence of SND in patients with Paf by use
of EP would have therapeutic implication.
In this study permanent pacemakers for SND were
implanted in 19 of 149 patients with Paf. EP30 was signifi-
cantly lower (2.10 ± 1.08 vs 3.82 ± 1.54 µV; P < .0001),
Yamada et al 291
and ED4 was significantly longer (40.8 ± 19.7 vs 16.0 ±
9.7 milliseconds, P < .0001) in patients with Paf with than
without pacemaker implantation. Eighteen (47%) of 38
patients with Paf and EP had undergone the implantation
of pacemaker, whereas a pacemaker was implanted in
only 3 (3%) of 111 patients with Paf but without EP. EP
gave a sensitivity of 84% and a specificity of 75% for the
identification of patients with Paf with pacemaker implan-
tation for coexistent SND. These results suggest that EP
would also be a marker to identify patients with Paf with
the requirement of pacemaker implantation for SND.
In patients diagnosed with syncope, it is clinically
important to rapidly clarify the cause. We previously
reported that EP would be also useful for the identifica-
tion of SND in patients with syncope.18 In this study,
18 of 149 patients with Paf experienced syncope, and
EP was observed in 6 of the 18 patients with Paf with
syncope. All (100%) of the 6 patients with Paf with syn-
cope and EP had SND, whereas SND was found in only
1 (8%) of the other 12 patients with Paf with syncope
without EP (P = .001; sensitivity 87%, specificity 100%).
This result also indicates that EP could be a rapid and
accurate marker to identify the involvement of SND in
patients with Paf and syncope.
Identification of SND in patients with Paf by the
combination of EP and the episode of syncope or
presyncope
In this study SND was significantly more frequently
found in patients with Paf with than without the epi-

292 Yamada et al
sode of syncope or presyncope (43% [14/33] vs 8%
[9/116]; odds ratio 8.9; 95% CI 3.3 to 23.1; P < .0001).
Furthermore, patients with Paf with both EP and the
episode of syncope or presyncope had the involvement
of SND significantly more frequently than those without
both EP and the episode of syncope or presyncope (92%
[10/11] vs 9% [13/138]; odds ratio 96.1; 95% CI 11.3 to
812; P < .0001). In other words, the likelihood of the
involvement of SND was increased 96-fold if the patients
with Paf with EP had the episode of syncope or presyn-
cope. Thus the combination of EP and the episode of
syncope or presyncope would identify the highest risk
subset of patients with Paf with SND.
Limitations of this study
In this study we carefully examined the involvement of
SND by conventional electrocardiography, bedside moni-
toring, or 24-hour Holter monitoring. However, it cannot
be ruled out that some cases of SND may remain undocu-
mented, because there are variations in the result of the
electrocardiographic monitorings. Furthermore, the rate
of invasive evaluation for SND was low (27%), because
the acceptance of electrophysiological study as a diag-
nostic tool in patients with SND remains limited.4 How-
ever, in this study, SND was significantly more frequently
found in patients with EP (n = 17) than in those without
EP (n = 23) who underwent electrophysiologic study
(82% vs 23%; P = .0003). SND was also significantly more
frequently found in patients with EP (n = 21) than those
without EP (n = 89) who did not undergo electrophysio-
logic study (19% vs 0%; P = .001).
In this study, the low amplitude of the initial portion
of the signal-averaged P wave was not always found in
all patients with sick sinus syndrome. SND with pro-
longed CSRT without prolonged SACT was significantly
more frequently observed in patients without than with
EP (80% vs 11%; P < .01). The initial low-amplitude sig-
nals might be lacking in patients whose disease is
purely limited to depressed automaticity of sinus node
without sinoatrial conduction disease or atrial disease.
Conclusion
This study revealed that EP on P-wave signal-averaged
electrocardiography would be useful for the identifica-
tion of the involvement of SND in not only patients
with Paf but also patients without Paf, indicating that
EP would be specific to SND.
We thank Ms S. Ishida and Ms H. Maekawa for the
technical assistance.
References
1. Rubenstein JJ, Schulman CL, Yurchak PM, et al. Clinical spectrum of
the sick sinus syndrome. Circulation 1972;46:5-13.
2. Gomes JAC, Kang PK, Matheson M, et al. Coexistence of sick sinus
rhythm and atrial flutter-fibrillation. Circulation 1981;63:80-6.
American Heart Journal
August 2001
3. Tanigawa M, Fukatani M, Konoe A, et al. Prolonged and fraction-
ated right atrial electrograms during sinus rhythm in patients with
paroxysmal atrial fibrillation and sick sinus syndrome. J Am Coll
Cardiol 1991;17:403-8.
4. Benditt DG, Milstein S, Goldstein M, et al. Sinus node dysfunc-
tion: pathophysiology, clinical features, evaluation and treatment.
In: Zipes DP, Jalife J, editors. Cardiac electrophysiology: from cell
to bedside. 2nd ed. Philadelphia: WB Saunders; 1994. p. 708-
34.
5. Fukunami M, Yamada T, Ohmori M, et al. Detection of patients at
risk for paroxysmal atrial fibrillation during sinus rhythm by P wave-
triggered signal-averaged electrocardiogram. Circulation 1991;
83:162-9.
6. Yamada T, Fukunami M, Ohmori M, et al. Characteristics of fre-
quency content of atrial signal-averaged electrocardiograms during
sinus rhythm in patients with paroxysmal atrial fibrillation. J Am Coll
Cardiol 1992;19:559-63.
7. Abe Y, Fukunami M, Yamada T, et al. Prediction of transition to
chronic atrial fibrillation in patients with paroxysmal atrial fibrillation
by signal-averaged electrocardiography: a prospective study. Cir-
culation 1997;96:2612-6.
8. Yamada T, Fukunami M, Shimonagata T, et al. Dispersion of signal-
averaged P wave duration on precordial body surface in patients with
paroxysmal atrial fibrillation. Eur Heart J 1999;20:221-20.
9. Stafford PJ, Turner I, Vincent R. Quantitative analysis of signal-aver-
aged P waves in idiopathic paroxysmal atrial fibrillation. Am J Car-
diol 1991;68:751-5.
10. Guidera SA, Steinberg JS. The signal-averaged P wave duration: a
rapid and noninvasive marker of risk atrial fibrillation. J Am Coll
Cardiol 1993;21:1645-51.
11. Steinberg JS, Zelenkofske S, Wong SC, et al. Value of the P-wave
signal-averaged ECG for predicting atrial fibrillation after cardiac
surgery. Circulation 1993;88:2618-22.
12. Thery C, Gosselin R, Lekieffre J, et al. Pathology of sinoatrial node:
correlations with electrocardiographic findings in 111 patients. Am
Heart J 1977;93:735-40.
13. Kaplan BM, Landendorf R, Lev M, et al. Tachycardia-bradycardia
syndrome (so-called “sick sinus syndrome”). Am J Cardiol 1973;
31:497-508.
14. Bharati S, Nordenberg A, Bauernfiend R, et al. The anatomic sub-
strate for the sick sinus syndrome in adolescence. Am J Cardiol
1980;46:163-72.
15. Sugiura M, Ohkawa S. A clinicopathological study on sick sinus
syndrome with histological approach to the sinus node. Jpn Circ J
1980;44:497-504.
16. Yamada T, Fukunami M, Kumagai K, et al. Detection of patients
with sick sinus syndrome by use of low amplitude potentials early in
filtered P wave. J Am Coll Cardiol 1996;28:738-44.
17. Yamada T, Fukunami M, Kim J, et al. Identification of patients with
sick sinus syndrome by use of atrial early potentials: a prospective
study [abstract]. J Am Coll Cardiol 1996;27:372A.
18. Yamada T, Fukunami M, Kumagai K, et al. Usefulness of atrial early
potential for the identification of sick sinus syndrome in patients with
syncope [abstract]. Circulation 1997;96:I-652.
19. Ohe T, Matsuhisa M, Kamakura S, et al. Relation between the
widening of the fragmented atrial activity zone and atrial fibrilla-
tion. Am J Cardiol 1983;52:1219-22.
20. Narula OS, Shantha N, Vasquez M, et al. A new method for mea-
surement of sinoatrial conduction time. Circulation 1978;58:706-14.
21. Narula OS, Samet P, Javier RP. Significance of the sinus-node re-
covery time. Circulation 1972;45:140-58.
American Heart Journal
Volume 142, Number 2
22. Feigenbaum H. Echocardiography, 4th ed. Philadelphia: Lea and
Febiger; 1986. p. 50-126.
23. ACC/AHA guidelines for implantation of cardiac pacemakers and
antiarrhythmic devices: a report of the American College of Cardiol-
ogy/American Heart Association Task Force on practice guidelines
(Committee on Pacemaker Implantation). J Am Coll Cardiol 1998;
31:1175-209.
24. Simson MB. Use of signals in the terminal QRS complex to identify
patients with ventricular tachycardia after myocardial infarction. Cir-
culation 1981;64:235-42.
25. Gomes JA, Winters SL, Stewart D, et al. A new noninvasive index to
predict sustained ventricular tachycardia and sudden death in the
Yamada et al 293
first year myocardial infarction based on signal-averaged radionu-
clide ejection fraction and Holter monitoring. J Am Coll Cardiol
1987;10:349-57.
26. Cain ME, Ambos HD, Markham J, et al. Diagnostic implications of
spectral and temporal analysis of the entire cardiac cycle in patients
with ventricular tachycardia. Circulation 1991;83:1637-48.
27. Centurion OA, Fukatani M, Konoe A, et al. Different distribution of
abnormal endocardial electrograms within the right atrium in
patients with sick sinus syndrome. Br Heart J 1992;68:596-600.
28. Shimizu A, Fukatani M, Tanigawa M, et al. Intra-atrial conduction
delay and fragmented atrial activity in patients with paroxysmal
atrial fibrillation. Jpn Circ J 1989;53:1023-30.