1136

The Long QT Syndrome

Prospective Longitudinal Study of 328 Families
Arthur J. Moss, MD; Peter J. Schwartz, MD; Richard S. Crampton, MD; Dan Tzivoni, MD;
Emanuela H. Locati, MD; Jean MacCluer, PhD; W. Jackson Hall, PhD;
Lowell Weitkamp, MD; G. Michael Vincent, MD; Arthur Garson Jr., MD;
Jennifer L. Robinson, MS; Jesaia Benhorin, MD; and Sungsub Choi, PhD

Background. The Long QT Syndrome (LQTS) is an infrequently occurring familial disorder

in which affected individuals have electrocardiographic QT interval prolongation and a
propensity to veiltricular tachyarrhythmic syncope and sudden death. We prospectively
investigated the clinical characteristics and the long-term course of 3,343 individuals from 328
families in which one or more members were identified as affected with LQTS (QTc>0.44
sec"12).
Methods and Results. The first member of a family to be identified with LQTS, the proband,
was usually brou,ght to medical attention because of a syncopal episode during childhood or
teenage years. Probands (n=328) were younger at first contact (age 21±+15 years), more likely
to be female (69%o), and had a higher frequency of preenroilment syncope or cardiac arrest with
resuscitation (80%), congenital deafness (7%), a resting heart rate less than 60 beats/min
(31%), QTc.0.50 sec"2 (52%), and a history of ventricular tachyarrhythmia (47%) than other
affected (n=688) and unaffected (n=1,004) family members. Arrhythmogenic syncope often
Downloaded from http://ahajournals.org by on January 7, 2020

occurred in association with acute physical, emotional, or auditory arousal. The syncopal
episodes were frequently misinterpreted as a seizure disorder. By age 12 years, 50% of the
probands had experienced at least one syncopal episode or death. The rates of postenrollment
syncope (one or more episodes) and probable LQTS-related death (before age 50 years) for
probands (n=235; average follow-up 54 months per patient) were 5.0% per year and 0.9% per
year, respectively; these event rates were considerably higher than those observed among
affected and unaffected family members.
Conclusions. Among 232 probands and 1,264 family members with prospective follow-up, three
factors made significant independent contributions to the risk of subsequent syncope or probable
LQTS-related death before age 50 years, whichever occurred first (Cox hazard ratio; 95%
confidence limits): 1) QTc (1.052; 1.017, 1.088), 2) history of cardiac event (3.1; 1.3, 7.2), and 3)
heart rate (1.017; 1.004, 1.031). The findings from this prospective longitudinal study highlight
the clinical features, risk factors, and course of LQTS. (Circulation 1991;84:1136-1144)

T he idiopathic long QT syndrome (LQTS) is an
infrequently occurring disorder in which af-
fected individuals have an unusual electro-
cardiographic repolarization abnormality (QT pro-
longation) and a propensity to syncope and fatal
ventricular arrhythmias.1,2 The genetics of the syn-
drome are not completely understood, but autosomal
dominant3,4 and autosomal recessive5,6 modes of in-
heritance are suggested by the pattern of occurrence
of this disorder in families. In addition, sporadic
cases without familial involvement occasionally oc-
cur. This disorder primarily affects children and

From the Department of Medicine (A.J.M., L.W.), the Heart
Research Follow-up Program of the Department of Preventive
and Community Medicine (J.L.R.), and the Division of Biostatis-
tics (W.J.H., S.C.), University of Rochester School of Medicine
and Dentistry, Rochester, N.Y.; Instituto Clinica Medica 2
(E.H.L., P.J.S.), University of Milan, Milan, Italy; Department of
Medicine (R.S.C.), University of Virginia School of Medicine,
Charlottesville, Va.; Bikur Cholim Hospital and the Hebrew
University of Jerusalem (D.T., J.B.), Jerusalem, Israel; Southwest
Foundation for Biomedical Research (J.M.), San Antonio, Tex.;
Department of Medicine (G.M.V.), University of Utah School of
Medicine, Salt Lake City, Utah; and the Pediatric Cardiology
Section (A.G.), Texas Children's Hospital and the Baylor College
of Medicine, Houston, Tex.
Supported in part by research grant HL-33843 from the Na-
tional Institutes of Health and by funds from the Daisy-Marquis-
Jones Foundation.
Address for correspondence: Arthur J. Moss, MD, PO Box 653,
University of Rochester Medical Center, Rochester, NY 14642- 8653.
Received May 18, 1990; revision accepted May 7, 1991.

young adults with structurally normal hearts. The
lethality of the condition in terms of the development
of life-threatening ventricular tachyarrhythmias, in-
cluding torsades de pointes, may be quite variable
among affected individuals. Antiadrenergic therapy
with ,B-blocking drugs7 and/or left cervicothoracic
sympathetic ganglionectomy89 is usually prescribed
to prevent recurrent syncope and cardiac death in
symptomatic patients with LQTS. Sympathetic gan-
glionectomy may be effective in preventing recurrent
syncope that is refractory to medical management.10
Permanent cardiac pacing has been used to prevent
bradycardia-dependent ventricular tachyarrhythmias
and syncope in this disorder.1'
Beginning in 1979, our research group established
an international, prospective, longitudinal study of
patients affected with LQTS. The purpose of this
ongoing investigation is to elucidate the clinical as-
pects of this disorder. A preliminary report of the
first 196 patients with LQTS enrolled in this program
was published in 1985.12 To date, we have enrolled
3,343 individuals from 328 families in which one or
more members were identified as having LQTS. The
clinical findings and long-term course of this popula-
tion are the subject of this report.
Methods
Identification of Probands
Physicians who were caring for patients with sus-
pected LQTS contacted one of the LQTS physician-
Downloaded from http://ahajournals.org by on January 7, 2020
investigators at our internationally dispersed clinical
centers (Rochester, N.Y.; Charlottesville, Va.; Salt
Lake City, Utah; Houston, Tex.; Milan, Italy; and
Jerusalem, Israel) for consultation advice regarding
diagnosis and management of their patient. The
referring physician was queried regarding the occur-
rence of syncope or cardiac arrest and the use of
prescription medication in the patient under consid-
eration, and a prospectively designed data form was
filled out. A 12-lead electrocardiogram was for-
warded to the LQTS physician-investigator, and the
QT and R-R intervals were measured in lead II (lead
I or III if the QT interval was not measurable in lead
II). The observed QT interval was corrected for heart
rate (QTc) according to Bazett's formula,13 where
QTc is in the peculiar time dimension sec"12. The first
living referred patient identified independently of
relatives with OTc>0.44 was defined as the proband
and served as the index case (propositus) for family
evaluation. Index patients receiving medication
known to prolong the QT interval were excluded
from enrollment.
Population
As of July 1, 1990, 328 probands have been en-
rolled in this longitudinal study. Family pedigrees
were constructed and attempts were made to enroll
in the study all living immediate family members
(siblings, children, parents, aunts and uncles, and
grandparents) of each proband and to identify family
Moss et al The Long QT Syndrome 1137
members who had died. Attempts were also made to
obtain 12-lead electrocardiograms on all living family
members and copies of electrocardiograms, if avail-
able, on family members who had died. Before
formal enrollment, each patient, a parent in the case
of a minor, provided informed consent for participa-
tion in this longitudinal study. The enrolled patient
population consists of 3,343 individuals; 328
probands and 3,015 family members (688 with
QTc>0.44 [affected], 1,004 with QTcs0.44 [unaf-
fected], and 1,323 individuals without a recorded
electrocardiogram [undetermined]).
Follow-up
For each individual enrolled in the study we at-
tempted to gather yearly follow-up clinical informa-
tion through phone or mail contact with the individ-
ual and by contact with the identified personal
physician to evaluate the patient's clinical course and
medication usage. Sixty-four percent of the follow-up
information was obtained from the patient or family
member, 32% from the attending physician, and 4%
from other sources.
End Points
Syncopal events (fainting spell with transient loss
of consciousness) and deaths identified during pedi-
gree construction and during follow-up were specifi-
cally recorded. The logistics of this international
study did not permit an independent, uniform review
of all the terminal events. However, family members
were queried about the clinical circumstances sur-
rounding the death of their relative as to the pres-
ence or absence of preexisting medical problems and
the chronological abruptness of the terminal event.
Unexpected, sudden (abruptly and without warning
as described by the family), natural death before age
50 years, exclusive of a known cause, was categorized
as a probable LQTS-related death. Age 50 years was
used as an arbitrary cut-off to exclude older patients
whose death might be due to more common cardio-
vascular disorders such as coronary disease.
Data Management and Analysis
Clinical data were recorded on prospectively de-
signed forms and included demographic, historical,
family history, electrocardiographic, therapeutic, and
follow-up information. Quality control procedures
were in place throughout the study to ensure internal
consistency of the recorded data on printed data
forms and to minimize missing data when they could
be retrieved. Family pedigrees were constructed by a
trained geneologist with a master's degree in genet-
ics. The study electrocardiograms were interpreted
by four physicians, and all R-R and QT intervals
were rechecked by a second physician. The study
data were maintained in a relational database system
(INGRES) on a Digital Equipment Corporation (VAX
8350) computer. The reported analyses used the
analytic database released July 1, 1990.
 1138 Circulation Vol 84, No 3 September 1991
TABLE 1. Clinical Characteristics of Long QT Syndrome Probands and Affected and Unaffected Family Members
Family members*
Probands Affected Unaffected
Characteristics (n =328) (n =688) (n= 1,004)
Demographics
Mean age at first contact (yr) 21+ 15 34+24 30+21
Females 69 60 46
Caucasian 93 99 99
History
Cardiac eventt 80 16 4
Mean age at first cardiac event 14±+12 16±14 17±12
Congenital deafness 7 1 <1
Ventricular tachyarrhythmiat 47 2
Probands with family history of:
QTc>0.44 sec 85 ... ..

Syncope/cardiac arrest 58 ... ...

Probable LQTS-related death

before age 50 45 ...

Parent consanguinity 2.
Electrocardiographic findings§
Mean heart rate (beats/min) 74±21 80±21 77±21
Heart rate<60 beats/min 31 16 18
Mean QTc (sec'/2) 0.52±0.07 0.48+0.04 0.41±0.02
QTc>0.50 sec`12 52 22 ...
Sinus pause>1.2 sec 2 <1 <1
LQTS therapy at enrollment
,l-Blockers 78 23 11
Left cervicothoracic sympathetic
Downloaded from http://ahajournals.org by on January 7, 2020

ganglionectomy 12 <1 0
Pacemakers 13 1 0
No LQTS therapy 16 76 89
Figures are percentages unless otherwise indicated; values are mean+SD.
LQTS, long QT syndrome.
*Affected family members had QTc>0.44 sec112 whereas unaffected family members had QTc.0.44 sec1'2.
tlncludes syncope or cardiac arrest with resuscitation.
tIncludes nonsustained or sustained ventricular tachycardia, torsades de pointes, or ventricular fibrillation
documented on an electrocardiogram or Holter recording before enrollment. Only 89 probands and 79 affected family
members had Holter recordings performed before enrollment.
§Electrocardiographic findings on the first recorded tracing.

The cumulative age-related probability of experi-
encing a first known cardiac event (syncope or prob-
able LQTS-related death, whichever occurred first)
using both birth and enrollment date as the time
origin was determined for the various subgroups of
the population using the life-table method of Kaplan
and Meier.14
The joint contribution of pertinent variables from
enrollment to the occurrence time of first follow-up
cardiac events was analyzed by the Cox proportional
hazards regression model.15 Analyses were carried
out on the total population stratified into probands
and family members; separate proband and family
member analyses were also performed. The analyses
identified the risk factors that made an independent
contribution to the Cox model. Treatment factors
were then added to the risk factor model one at a
time to evaluate the contribution that each of these
therapies made to the full model. The contributions
of identified factors to outcome cardiac events are
presented in terms of the hazard ratio, the risk of
having an outcome event per unit of time during
follow-up for each extra unit of the risk factor.
Attempts were made to identify any clustering effects
within families; if nothing significant was found,
family members were then analyzed without regard
to the identifying proband.
Results
Population Characteristics
The pertinent clinical characteristics of the 328
probands, the 688 affected, and 1,004 unaffected fam-
ily members are presented in Table 1. The population
was nearly equally divided between those with
QTc>0.44 (n=1,016) and those with QTc.0.44
(n = 1,004). The probands were younger, more likely to
be female, and had a higher frequency of prior cardiac
 Moss et al The Long QT Syndrome 1139

TABLE 2. Ten Representative Long QT Syndrome Patients Initially Misdiagnosed as Having Benign Syncope or
Seizure Disorder
Age at Duration of
Patient LQTS dx (yr) Gender misdiagnosis (mo) Clinical features QTc (sec"'2)
01-0072 30 M 228 Recurrent syncope 0.47-0.51
Rx: anticonvulsants
01-0055 34 F 144 Recurrent syncope 0.48-0.57
Rx: anticonvulsants
TdP with near fatal
event
01-0193 38 F 84 Recurrent syncope 0.45-0.49
VF with resuscitation
01-9277 21* F 84 Recurrent syncope 0.50-0.65t
Rx: anticonvulsants
Sudden death on arousal
01-0592 39 F 54 Recurrent syncope 0.50-0.58
Rx: anticonvulsants
VF with resuscitation
01-9627 19* F 53 Recurrent syncope 0.59 in
Rx: anticonvulsants brother
Sudden death
01-1678 11 M 48 Recurrent syncope 0.45-0.49
Rx: anticonvulsants
VF with resuscitation
01-1062 6 F 46 Recurrent syncope 0.45-0.48
Rx: anticonvulsants
Sudden death in sister
01-1921 15 F 24 Recurrent syncope 0.62
Rx: anticonvulsants
01-8036 15* F 21 Recurrent syncope 0.45-0.59 in
Sudden death on arousal father and
four siblings
LQTS, long QT syndrome; dx, diagnosis; Rx, treatment; TdP, torsades de pointes; VF, ventricular fibrillation.
Downloaded from http://ahajournals.org by on January 7, 2020

*Posthumous diagnosis.
tPremortem electrocardiogram interpreted as normal but postmortem review revealed these QTc values.
Note: Recurrent syncopal episodes in this group of patients were considered to be seizure disorders by the attending
physicians despite normal electroencephalograms.

events (syncope or cardiac arrest with resuscitation),
congenital deafness, bradycardia, QTc2 0.50, and ven-
tricular tachyarrhythmia than the affected and unaf-
fected family members. Eighty-five percent of the
probands had one or more family members with
QTc>0.44. Only 10% of the probands had no family
members with QTc>0.44 and no family history of
syncope or death before age 50 years. This small group
of probands might represent the so-called sporadic
form of LQTS.
Almost 80% of the probands were receiving ,3-block-
ing medication at the time of enrollment, and 12%
already had left cervicothoracic sympathetic gangli-
onectomy. Among those with available information
about factors associated with syncopal events prior to
enrollment, 47% (77 of 163) of the patients had one or
more syncopal episodes occurring in association with
intense emotions (anger or fright), 41% (68 of 165) with
vigorous physical activity exclusive of swimming, 19%
(32 of 166) on awakening, 15% (21 of 143) while
swimmning, and 8% (14 of 166) on arousal by auditory
stimuli such as the ringing of an alarm clock or a
telephone or the sound of thunder.
Misdiagnosis
During the course of this 10-year prospective fol-
low-up study, we became aware of several LQTS pa-
tients in whom the diagnosis of LQTS was missed
despite unexplained recurrent syncope. The pertinent
findings in 10 representative patients misdiagnosed as
benign syncope or seizure disorder are presented in
Table 2. These patients were, in fact, experiencing
arrhythmogenic syncope, and the correct diagnosis of
LQTS was established only after a tragic cardiac arrest
occurred in a family member or when an electrocardio-
gram was fortuitously obtained on the patient for some
unrelated reason. On a few occasions, an electrocardio-
gram had been taken earlier but QT prolongation was
not appreciated by the physician interpreting the trac-
ing. Once the diagnosis of LQTS was established, a
family history of premature sudden cardiac death or
unexplained syncope, or both, extending back a gener-
ation or two, became quite obvious.
Mortality Events
In the LQTS study population, 147 patients died
before age 50 years. One hundred eight patients
identified retrospectively during family pedigree con-
struction had little if any clinical information avail-
able other than age at death, gender, and whether or
not the patient had ever been treated for LQTS
before death; in the remaining 39 patients, clinical
 1140 Circulation Vol 84, No 3 September 1991

TABLE 3. Selected Clinical Findings in 147 Patients Clinical Course

From the Long QT Syndrome Study Population Who Died
Before Age 50 Years The cumulative age-related probability of experi-
Variable Percent or encing a first known cardiac event (syncope or
(n with available data) mean + SD probable LQTS-related death) from birth until age
Female (%) (n= 147) 56 50 years is presented in Figure 1 for the entire
History of syncope (%) (n=36) 31
LQTS population subdivided into probands and
Mean age at death (yr) (n=147) 21±14
affected, unaffected, and undetermined family
0-10 yrs
members. The probands were usually identified
24
after a syncopal event, and this partly explains the
11-20 yrs 33
high probability of experiencing a cardiac event in
21-30 yrs 18 this group compared with the other groups. In the
31-40 yrs 12 undetermined group, the identified events were
41-50 yrs 13 probable LQTS-related deaths, as no information
QTc (sec 12) (n = 39) 0.52±0.07 was available regarding possible preceding syncope.
Antiadrenergic therapy* (%) (n=116) 9 By age 12 years, 50% of the probands, 8% of the
*Antiadrenergic therapy includes fl-blocker medication at any affected family members, and 2% of undetermined
time in the past and/or left cervicothoracic sympathetic gangli- and unaffected family members experienced a first
onectomy. cardiac event.
The cumulative probability of experiencing a fol-
information including an electrocardiogram was low-up cardiac event (syncope or probable LQTS-
available. The clinical characteristics of patients on related death) after enrollment for probands and the
whom recorded data are available are presented in affected and unaffected family members is presented
Table 3. Mortality events occurred predominantly in in Figure 2. After 10 years of follow-up, 37% of the
young patients (57% by age 20 years) who had a probands, 5% of the affected family members, and
history of syncope and marked QT prolongation but less than 1% of the unaffected family members had
were not receiving antiadrenergic therapy. experienced a cardiac event.

z 1.0
Downloaded from http://ahajournals.org by on January 7, 2020

0 Probands (QTc > 0.44)

< 0.9 (n = 328)
a
cc
0

0.8
0
z
v 0.7
cc
< 0.6-
z
Z 05
Z 0.5 S
a-
Wx
LL
0.4-
0
i 0.3
Affected (QTc > 0.44)
(n = 688)
mcc
o
............
.............
W
0.1
Unaffected (QTc < = 0.44)
(n = 1004)
t 0.0
0 5 10 15 20 25 30 35 40 45 50
AGE IN YEARS
FIGURE 1. Graph showing cumulative probability of experiencing afirst known cardiac event (syncope orprobable LQTS-related
death) using birth as the time origin forprobands and affected, unaffected, and undetermined family members enrolled in the LQTS
study.

! 0.5
0
0
C3
a. 0.4
0 Probands (QTc > 0.44)
z
03
iz
W
'°~50.2.
0.31
0.
0.
D Unaffected (QTc
0 0.01
0 1 2 3 4 5
FOLLOW-UP IN YEARS
FIGURE 2. Graph showing cumulative probability of experiencing a follow-up cardiac event (syncope orprobable LQTS-related
death) using enrollment as the time origin for probands and affected and unaffected family members.
Downloaded from http://ahajournals.org by on January 7, 2020
Prospective Follow-up
Follow-up information was obtained on 235
probands, 549 affected family members, and 756
unaffected family members. The clinical characteris-
tics, therapy, and cardiac event rates for each of the
three groups are presented in Table 4. The rates of
syncope and death for probands were considerably
higher than those observed among affected and un-
affected family members. In addition, probands with
a history of a cardiac event before enrollment had
more than twice the rate of syncope during follow-up
than did probands with a negative prior history (5.5
versus 2.6 syncopal events per year). The death rates
during follow-up were similar in the two proband
subgroups. Of note, among the probands with fol-
low-up events, a majority (84%) were receiving
/-blockers at the time of the latest recorded event. In
those with events, the QTc was quite prolonged in
the probands and the affected family members.
The clinical, electrocardiographic, and treatment
characteristics of the 12 affected family members and
the three unaffected family members who developed
cardiac events during the prospective follow-up after
enrollment are presented in Table 5. One half of the
patients had a prior cardiac event, and 29% were
receiving /3-blockers.
Separate Cox regression survival analyses were car-
ried out for: 1) the entire population, stratified into
probands and family members, 2) the probands, and 3)
Moss et al The Long QT Syndrome 1141
(n =235)
Affected (QTc > 0.44)
......................... (n 549)
< = 0.44)
(n =
756)
6 7 8 9 10
the affected and unaffected family members. These
time-dependent analyses used enrollment as the time
origin and first follow-up cardiac events (syncope or
probable LQTS-related death, whichever occurred
first) for the end points. The findings from the three
analyses are presented in Table 6. For the entire study
population, three risk variables were identified: QTc,
history of a cardiac event, and heart rate (Table 6). For
the probands, two risk variables were identified: QTc
and heart rate (Table 6). For the family members, the
risk variables were history of a cardiac event and female
gender (Table 6). For continuous variables, the hazard
ratio is expressed per unit increase in the value of the
variable. Thus, a proband with QTc=0.60 would have a
2.2-fold greater risk of having a follow-up cardiac event
per unit of time than a proband with QTc=0.45 (haz-
ard ratio= 1.05515=2.2). Since there were no interac-
tions in any of the models, the hazard ratios in each
model in Table 6 permit quantification of the joint risk
when two or more factors coexist. The combined risk is
simply the product of the hazard ratios of the coexisting
factors. For example, a proband with QTc=0.60 and a
heart rate of 80 beats/min would have a 3.2-fold greater
risk of having a follow-up cardiac event than a proband
with QTc=0.45 and a heart rate of 60 beats/min
(combined hazard ratio= 1.05515 x 1.01820=3.2). Of
note, none of the treatment factors (/3-blockers, left
cervicothoracic sympathetic ganglionectomy, or pace-
makers) made a significant contribution to any of the
three risk factor models.
 1142 Circulation Vol 84, No 3 September 1991

TABLE 4. Follow-up Duration and Cardiac Events Among Probands and Affected and Unaffected Family Members
Family members
Probands Affected Unaffected
Number of patients with follow-up 235 549 756
Average follow-up per patient in months 54 38 31
Number of patients with events
Syncope (one or more episodes) 53 9 3
Probable LQTS-related death before age 50 years 10 3 0

Baseline variables in patients with events (n=74)

Heart rate<60 beats/min (%) 27 25 67
Mean QTc (sec" ) 0.54±0.08 0.50+0.04 0.40±0.03
QTc>0.50 sec1!2 (%) 61 33 0

Therapy at time of last event (n)* 59 11 3

p-Blockers 46 3 1

Left cervicothoracic sympathetic ganglionectomy 19 0 0

Pacemakers 15 1 0

Event rate (percent per year)

Syncope 5.3 [5.5/2.61t 0.5 0.2
Probable LQTS-related death before age 50 years 0.9 [0.8/1.3]t 0.2 .

LQTS, long QT syndrome.

*Number of patients with follow-up events in each group on whom therapy data are available.
tNumbers in the brackets reflect the follow-up event rates for those with/without a history of a cardiac event before
enrollment.

Discussion because of syncopal episode during childhood or

a
This prospective long-term follow-up study of pa- teenage years. Probands had a more malignant course
tients with LQTS indicates that this disorder is largely than the secondarily identified affected family members
familial (85% of probands had family members with in terms of syncope and death, both retrospectively
Downloaded from http://ahajournals.org by on January 7, 2020

QTc>0.44) with variable severity of expression among using birth as the time origin and prospectively after
those with QTc>0.44. The first member of a family to enrollment in this study.
be identified with LQTS, the proband, was usually The findings from the Cox survival analyses (Table
brought to the attention of the medical profession 6) are of interest. Different risk variables were iden-

TABLE 5. Clinical, Electrocardiographic, and Treatment Characteristics Among Affected Family Members and
Unaffected Family Members Who Developed Cardiac Events During Follow-up
Age at History Heart rate QTc Therapy*
Patient enrollment (yr) Sex of cardiac event (beats/min) (sec112) pf-blockers Follow-up event
Affected family members
1 10 M N 57 0.58 N S
2 57 M N 50 0.51 Y S
3 8 F N 80 0.47 N S
4 59 F Y 94 0.45 N S
5 7 F N 77 0.50 N S
6 35 F N 63 0.48 N S
7 59 F N 56 0.46 N S
8 58 F Y 94 0.46 N S
9 6 F Y 71 0.49 Y S
10 18 M ? 86 0.53 ? D
11 10 F Y 82 0.52 Y D
12 37 M N 115 0.52 N D
Unaffected family members
13 47 F Y 65 0.41 N S
14 13 F Y 58 0.37 Y S
15 20 F Y 48 0.41 N S
M, male; F, female; N, no; Y, yes; S, syncope; D, death.
*None of the patients had sympathetic ganglionectomy or pacemaker therapy.
 Moss et al The Long QT Syndrome 1143

TABLE 6. Cox Model Regression Analyses of Risk Factors for Subsequent Cardiac Events
Risk factor* p Hazard ratiot (95% confidence limits)
Model 1: Total population (n=1,496; 72 events)
QTc (per 0.01 units) <0.01 1.052 (1.017, 1.088)
History of cardiac event <0.01 3.1 (1.3, 7.2)
Heart rate (beats/min) 0.01 1.017 (1.004, 1.031)
Model 2: Probands (n=235; 58 events)
QTc (per 0.01 units) <0.01 1.055 (1.017, 1.094)
Heart rate (beats/min) 0.02 1.018 (1.003, 1.033)
Model 3: Family members (n=1,264; 14 events)
History of cardiac event <0.01 6.2 (2.1, 18)
Female 0.04 3.9 (1.1, 15)
*Other baseline variables (congenital deafness, repetitive ventricular arrhythmias, family history of death before age
50 years, and age at enrollment) and interaction effects were dropped from the models (p>0.10). None of the
treatment factors (j-blockers, left cervicothoracic sympathetic ganglionectomy, or pacemakers) made significant
contributions to any of the three risk factor models.
tRatio of the risk of experiencing a cardiac event (syncope or probable long QT syndrome-related death before age
50 years, whichever occurred first) per unit time during follow-up for patients with the factor to those without the factor.
For continuous variables, the hazard ratio is expressed per unit increase in the value of the variable.

tified in the separate analyses performed on the
probands and on the family members. The reason for
this is unclear, but it may relate to the fact that 84%
of the probands were receiving some form of LQTS
therapy at enrollment, whereas only 20% of family
members were under treatment. In the entire popu-
lation and in the proband group, QTc and heart rate
were significant risk factors. That is, patients with
longer QTc and faster heart rates were at greater risk
of having events than patients with shorter QTc and
Downloaded from http://ahajournals.org by on January 7, 2020
have resulted from this error in diagnosis. Our data-
base does not permit us to quantitate the overall
frequency of this problem in the study population.
However, misdiagnosis is not an infrequent occur-
rence in this relatively rare disorder. Several articles
describing this phenomenon have been published,18-21
but the potentially fatal consequences of missing
the diagnosis of LQTS in patients with syncope are
still not adequately appreciated by practicing phy-
sicians. An electrocardiogram with careful analysis

slower heart rates. This is the first time that the
duration of QTc has been identified as a risk factor in
LQTS, and it makes sense with what we know about
this disorder. Heart rate as a risk factor is more
difficult to explain. It may be that the patients with
slower heart rates were receiving ,B-blockers, and this
therapy may provide protection even though 13-block-
ers did not enter the Cox models as a beneficial
factor. Patients not receiving ,3-blockers and young
children would surely have faster heart rates, and
both these factors may be masked by the heart rate
variable.
We used the conventional definition of a pro-
longed QT interval (QTc>0.44 sec'2) to identify
patients as being affected with LQTS.13 This electro-
cardiographic cut point is surely associated with some
misclassification of patients as being affected and
unaffected with LQTS, especially among patients
with QTc values close to the 0.44 interval. Multidi-
mensional quantitation of ventricular repolariza-
tion16 may in the future improve the diagnostic
identification of patients with LQTS.17 At present,
the QTc cut point of 0.44 has diagnostic utility since
very few cardiac events occurred during prospective
follow-up in patients with QTc<0.44.
We have highlighted the serious problem of misdi-
agnosis by describing 10 representative LQTS patients
in whom the occurrence of syncopal episodes before
enrollment in this study were misinterpreted as a
seizure disorder (Table 2). Several tragic outcomes
and measurement of QT and QTc intervals should
be included in the workup of all patients with
unexplained syncope.
We cannot answer the question of the efficacy of
,B-blocker therapy, surgical left cervicothoracic sym-
pathetic ganglionectomy, or pacemakers in the man-
agement of this population of patients with LQTS.
This investigation was designed as an observational
study, and we did not control the therapy rendered to
individual patients. None of these therapies was
identified in the univariate or multivariate analyses as
showing a significant beneficial effect on outcome.
However, the high use of ,8-blockers (78%) and the
low frequency of ganglionectomy or pacemakers (12-
13%) in the probands compromise the power of
detecting a clear-cut effect with any of these thera-
pies. More important, therapy was generally used in
patients perceived to be at higher risk.
This current investigation of 328 LQTS families
involving 1,016 patients with OTc>0.44 extends our
earlier findings on the 196 LQTS patients reported in
1985. Syncope continues to be associated with acute
physical, emotional, or auditory arousal. Historical
data regarding syncope and ventricular ectopy re-
main important risk factors for subsequent cardiac
events. In our earlier study,12 females had a border-
line significant increased risk of experiencing a car-
diac event relative to males (relative risk 2.2, 95%
confidence interval 1.0, 4.8); an increased risk for
females was seen among family members in the
 1144 Circulation Vol 84, No 3 September 1991
current study (relative risk 3.9, 95% confidence in-
terval 1.1, 15). The current rates of syncope (5.0%
per year) and cardiac death (0.9% per year) for the
probands are somewhat less than those reported in
the earlier study (8.6% per year and 1.3% per year,
respectively). These reductions in event rates may
reflect the ongoing enrollment of patients with less
severe forms of LQTS as suggested by Schwartz,22
possibly a beneficial effect from the increased use of
antiadrenergic therapy during the past 5 years (78%
of the probands were on ,B-blockers at the time of
enrollment; Table 1), or both.
There are several limitations inherent in this ob-
servational study of a probable genetic disorder. An
ascertainment bias was present in that we required
the probands to be alive at the time of enrollment. In
genetic studies, probands are identified by a pheno-
typic characteristic whether they are alive or dead at
the time of ascertainment.23 The probands were
usually brought to the attention of the LQTS study
because of a major cardiac event, and this would
certainly bias the reporting of cumulative cardiac
events in favor of the probands as presented in
Figure 1. Selection bias probably also accounts for
the increased proband event rate after enrollment
because LQTS patients frequently experience multi-
ple episodes of syncope. Although LQTS predomi-
nantly affects the younger age group, only one of our
nine physician-investigators was a pediatric cardiolo-
gist. Thus, there was probably an inadvertent selec-
tion bias in favor of older probands. Follow-up
Downloaded from http://ahajournals.org by on January 7, 2020
information was available on only 76% of the
probands and family members. The missing follow-up
information contributes to an incomplete data set
with some reduced precision in the quantitated event
rates and the risk factor hazard ratios. Despite these
limitations, this prospective longitudinal study of
LQTS patients and family members is the largest yet
reported with this disorder. The findings from this
study enhance our understanding of the natural
history and the clinical features of this familial,
prolonged repolarization syndrome.
Acknowledgments
The authors are grateful to the physicians who
referred their patients to this study and thank Shmuel
Gottlieb, MD, for assisting in the coding of the baseline
electrocardiograms; Eric Carleen, Mark Andrews, and
Patricia Severski for their programming proficiency;
Aartje deWaard for her commitment to patient follow-
up, Susan ErkenBrack, Cheryl Freese, Aharon Medina,
MD, Sergio Sala, MD, and Katherine Timothy for
identifying and enrolling LQTS patients and family
members in this study; and Julie Gross and Nancy
Kellogg for preparation of the manuscript.
References
1. Schwartz PJ, Periti M, Malliani A: The long QT syndrome. Am
Heart J 1975;89:378-390
2. Moss AJ: Prolonged QT syndromes. JAMA 1986;256:
2985-2987
3. Romano C, Gemme G, Pongiglione R: Aritmie cardiache rare
dell'eta pediatrica. Clin Pediatr (Phila) 1963;45:656-683
4. Ward OC: New familial cardiac syndrome in children. J Irish
Med Assoc 1964;54:103-106
5. Jervell A, Lange-Nielsen F: Congenital deaf mutism, func-
tional heart disease with prolongation of the QT interval, and
sudden death. Am Heart J 1957;54:59-68
6. Fraser GR, Froggatt P, Murphy T: Genetical aspects of the
cardio-auditory syndrome of Jervell and Lange-Nielsen (con-
genital deafness and electrocardiographic abnormalities). Ann
Hum Genet 1964;28:133-157
7. Schwartz PJ, Locati E: The idiopathic long QT syndrome:
Pathogenetic mechanisms and therapy. Eur Heart J 1985;6:
103-114
8. Moss AJ, McDonald J: Unilateral cervicothoracic sympathetic
ganglionectomy for the treatment of long QT interval syn-
drome. N Engl J Med 1970;285:903-904
9. Schwartz PJ, Locati E, Priori SG, Zaza A: The idiopathic long
QT syndrome, in Zipes DP, Jalife J (eds): Cardiac Electrophys-
iology: From Cell to Bedside. Philadelphia, WB Saunders Co,
1990, pp 589-605
10. Bhandari AK, Scheinman MM, Morady F, Svinarich J, Mason
J, Winkle R: Efficacy of left cardiac sympathectomy in the
treatment of patients with the long QT syndrome. Circulation
1984;70:1018-1023
11. Eldar M, Griffin JC, Abbott JA, Benditt D, Bhandari A, Herre
JM, Benson DW, Scheinman MM: Permanent cardiac pacing
in patients with the long QT syndrome. J Am Coll Cardiol
1987;10:600-607
12. Moss AJ, Schwartz PJ, Crampton RS, Locati E, Carleen E:
The long QT syndrome: A prospective international study.
Circulation 1985;71:17-21
13. Bazett HC: An analysis of the time relations of electrocardio-
grams. Heart 1920;7:353-367
14. Kaplan EL, Meier P: Nonparametric estimation from incom-
plete observations. J Am Stat Assoc 1958;53:457-481
15. Cox DR: Regression models and life-tables. J R Stat Soc [B]
1972;34:187-220
16. Merri M, Benhorin J, Alberti M, Locati E, Moss AJ: Electro-
cardiographic quantitation of ventricular repolarization. Cir-
culation 1989;80:1301-1308
17. Benhorin J, Merri M, Alberti M, Locati E, Moss AJ, Hall WJ,
Cui L: The long QT syndrome: New electrocardiographic
characteristics. Circulation 1990;82:521-527
18. Singer PA, Crampton RS, Bass NH: Familial Q-T prolonga-
tion syndrome: Convulsive seizures and paroxysmal ventricular
fibrillation. Arch Neurol 1974;31:64-66
19. Wahl RA, Dick M II: Congenital deafness with cardiac
arrhythmias: The Jervell Lange-Nielsen Syndrome. Am Ann
Deaf 1980;125:34-37
20. Schwartz PJ: Stress and sudden cardiac death: The role of the
autonomic nervous system. J Clin Psych Mon 1984;2:7-13
21. Gospe SM, Choy M: Hereditary long QT syndrome presenting
as epilepsy: Electrocardiography laboratory diagnosis. Ann
Neurol 1989;25:514-516
22. Schwartz PJ: Idiopathic long QT syndrome: Progress and
questions. Am Heart J 1985;109:399- 411
23. Morton NE: Outline of Genetic Epidemiology. Basel/New York,
S Karger, AG, 1982, pp 47-67
KEY WORDS * delayed repolarization * ventricular arrhythmias
* stellectomy * syncope * sudden cardiac death