This study examined the relationship between heart rate (RR interval) and QT interval (measured as QaT interval) in rats, guinea pigs, rabbits, and primates. The researchers systematically varied heart rate using various procedures and measured the resulting QT intervals. They found that in rats, QT interval did not change with heart rate. In guinea pigs, rabbits, and primates, QT interval decreased with increasing heart rate and increased with decreasing heart rate as expected. Various correction formulas were tested for their ability to account for heart rate effects on QT interval in these three species. The researchers found little difference between formulas, though regression analysis provided the best correction.
A typical Realization of the process with linear recovery of AldosteroneIJERA Editor
Hypercortisolism as a sign of hypothamamus-pituitary-adrenocortical (HPA) axis overactivity and sleep EEG
changes are frequently observed in depression. Closely related to the
HPA axis is the renin-angiotensin-aldosterone system (RAAS) as 1. adrenocorticotropic hormone (ACTH) is a
common stimulus for cortisol and aldosterone, 2. cortisol release is suppressed by mineralocorticoid receptor
(MR) agonists 3. angiotensin II (ATII) releases CRH and vasopressin from the hypothalamus. The first passage
time and the bounds of the survival functions for the application are also obtained
CARDIOLOGY - A propensity score-adjusted retrospective comparison of early an...MedConnect India
Feasibility and efficacy of mitral repair in the elderly remain controversial. This study aims to compare outcomes of mitral repair and replacement in octogenarians.
A typical Realization of the process with linear recovery of AldosteroneIJERA Editor
Hypercortisolism as a sign of hypothamamus-pituitary-adrenocortical (HPA) axis overactivity and sleep EEG
changes are frequently observed in depression. Closely related to the
HPA axis is the renin-angiotensin-aldosterone system (RAAS) as 1. adrenocorticotropic hormone (ACTH) is a
common stimulus for cortisol and aldosterone, 2. cortisol release is suppressed by mineralocorticoid receptor
(MR) agonists 3. angiotensin II (ATII) releases CRH and vasopressin from the hypothalamus. The first passage
time and the bounds of the survival functions for the application are also obtained
CARDIOLOGY - A propensity score-adjusted retrospective comparison of early an...MedConnect India
Feasibility and efficacy of mitral repair in the elderly remain controversial. This study aims to compare outcomes of mitral repair and replacement in octogenarians.
Marketo Roadshow - Scaling Marketing for High Growthtbdenney
Baxter Denney, VP Online Marketing and Operations at New Relic, shares how their team has scaled to support an ever-complex sales motion during a time of hyper-growth.
Decálogo presentadp al concurso de Airplanet, realizado por el equipo de los Earthcleaners, de 5º de Primaria del CEIP Enrique Tierno Galván de San Sebastián de los Reyes.
Marketo Roadshow - Scaling Marketing for High Growthtbdenney
Baxter Denney, VP Online Marketing and Operations at New Relic, shares how their team has scaled to support an ever-complex sales motion during a time of hyper-growth.
Decálogo presentadp al concurso de Airplanet, realizado por el equipo de los Earthcleaners, de 5º de Primaria del CEIP Enrique Tierno Galván de San Sebastián de los Reyes.
2. 202 JPM Vol. 32, No. 4
December 1994:201-207
tion potentials and increase refractory periods. In order
to assess the actions of various class III antiarrhythmics
in various species, we have made use of various mea-
sures of the QT interval as an indicator of drug effects
on ventricular action potential duration in various labo-
ratory species, i.e., rat, guinea pig, rabbit, and primates,
and for cross-species comparison we sought the most
precise measure of QT. Interpretation of such QT data is
complicated by the fact that some class III drugs lower
heart rate and therefore require correction, and so we
have made a systematic analysis of the effect of varia-
tions in heart rate on measures of QT interval in these
species. In the following study, anesthetized rats, guinea
pigs, rabbits, and primates were subjected to various
procedures chosen for their ability to slow or increase
heart rate.
Materials and Methods
Approval for experiments with rats, rabbits, and
guinea pigs was obtained from the Animal Care Com-
mittee of the University of British Columbia. Standard
laboratory animals were used for this part of the study.
Experiments on primates were performed at the Depart-
ment of Obstetrics and Gynaecology, National Univer-
sity of Singapore, Singapore.
General Surgery
Male Sprague-Dawley rats weighing between 250-
300 g, male Hartley guinea pigs weighing between
500-800 g, and male New Zealand white rabbits weigh-
ing between 1.5-2.5 kg were used in the studies. Two
different anesthetic regimens were used; pentobarbitone
was used to anesthetize both rats (60 mg/kg i.p.) and
rabbits (30 mg/kg i.v.), while guinea pigs were anesthe-
tized with urethane (1 g/kg i.p.). Additional anesthetic
was given when necessary. The anesthetic used in these
species were those conventionally used in our labora-
tory.
Primates were anesthetised with either halothane
(0.5-1.0%) or pentobarbitone (10-20 mg/kg i.v.) after
being initially tranquilized with ketamine (50 mg/kg
i.m.). The pentobarbitone anesthesia was used to induce
high heart rates whereas halothane anesthesia was used
for lower heart rates. The primates were baboons (Papio
anubis) or macaque monkeys (Macaque fascicularis).
Blood pressure was recorded from a transcutaneous
cannula placed in the femoral artery.
In the small species the left carotid artery was cannu-
lated for recording blood pressure on a Grass polygraph
(Model 79D) while the right external jugular vein was
cannulated for drug administration. The ECG was
recorded using a Lead II type of configuration along the
anatomical axis of the heart as determined by palpation.
(For a complete description, see Penz et al., 1992).
ECGs were recorded on Grass polygraph chart paper at
a standard chart speed of 100 mm/sec and on a Honey-
well E for M storage oscilloscope. Measurements of
intervals were made on the Grass polygraph chart re-
corder and from the memory trace of the monitor. Both
measurements were compared and did not differ signifi-
cantly in terms of ECG intervals.
Heart rate was altered by various procedures includ-
ing vagal stimulation, direct atrial stimulation and ad-
ministration of a bolus of cold saline or treatment with
various drugs and anaesthetics.
Vagal Stimulation
Vagal stimulation was accomplished by isolating
both the right branch of the vagal nerve as well as the
accompanying ascending cervical sympathetic nerves.
The nerves were separated by blunt dissection and cut
at the level of the submandibular gland. The cardiac end
of the vagus nerve was stimulated with bipolar elec-
trodes using square wave stimulation at twice threshold
current (it), pulse width of 0.5 msec, and at suitable
frequencies (25-60 Hz) to obtain a heart rate which
varied between 60 beats/min and sinus rhythm. In some
animals the resting heart rate was reduced by metoprolol
(1.0 mg/kg i.v.) administration. It was assumed that
vagal stimulation of the right branch of the nerve did not
result in the significant release of acetylcholine in the
ventricles. During vagal stimulation an attempt was
made to ensure that atrioventricular (AV) conduction
was still intact and that the heart rate was not of an AV
nodal origin.
Atrial Stimulation
For atrial stimulation studies, a specially constructed
stimulating electrode was prepared using polyethylene
(PE) tubing. Two teflon-coated silver wires were in-
serted into either PE50 (rats and guinea pigs) or PE90
(rabbits) tubing. In order to ensure that electrodes were
positioned equidistant apart, a second PE tubing was
inserted between the silver wires. For rats and guinea
pigs the second tubing was PE10, while for rabbits it
was PE50. All tubing was then filled with polyethyl-
ene glycol (PEG40o) to prevent blood from entering
the electrode and to electrically insulate the wires.
The Teflon coating of the wires was removed from the
stimulating end of the electrode to expose bare metal to
allow for pacing of the atria once the electrode was
threaded through the right jugular vein into the right
atria. The suitability of the electrode position was deter-
mined by ensuring that a minimum threshold current at
8 Hz and 1 msec gave a tachycardia with an ECG
recording close to that obtained with the animal in sinus
3. E. HAYES ET AL. 203
RELATIONSHIP BETWEEN QaT AND RR INTERVALS
rhythm. Tachycardia was induced over the range of
5.5-9.0 Hz.
Cold Saline and Drug Administration
In addition to electrical pacing, changes in heart rate
were also induced by the administration of ice-cold
saline and drugs including the beta adrenoceptor agonist
adrenaline or isopropylnoradrenaline (0.1-10.0 lxg/kg
i.v.). Ice-cold saline, at various volumes (1.0--5.0 ml),
was injected into the jugular vein, and the resulting
changes in heart rate monitored.
Q-T Interval Determination
In view of the difficulty that was sometimes experi-
enced in determining when the T wave returned to the
isoelectric line, an alternative measure of QT (QaT) was
used for species other than rat. QaT is measured to the
peak of the T wave, which could be more clearly
defined in this study. Calculations of the QaT interval
were made directly from the surface of paper charts. In
cases where a clear positive or negative T wave was
seen, the QT interval was the time between the negative
peak of the Q wave and the peak (negative or positive)
of the T wave. In some circumstances when heart rate
changed, there were alterations in the shape of the T
wave. When this occurred, an attempt was made to use
the part of T-wave configuration that best related to the
T wave seen in the normal condition, that is, sinus rate.
In the cases where changing heart rate produced a
change in the configuration of the T wave, a note was
made of the change in such configuration. The use of
QaT was suggested originally by Lepeschkin (1955),
and its value substantiated by Beck and Marriott (1959)
and more recently by Nierenberg and Ransil (1979).
Chernoff (1972) has discussed QaT in relation to QT in
detail.
In the case of rat, it is difficult to detect a T wave that
corresponds exactly with the T wave seen in other
species (Beinfeld and Lehr, 1968; Driscoll, 1981). In
this species therefore, T-wave calculations were made
on the basis of the repolarization wave that followed the
QRS complex. In previous studies we have considered
the difficulties associated with measurement of QT and
as a result measure a surrogate of QT ("QT"). This
measurement is taken from the Q wave to the first major
inflection point on the repolaxization phase. Exact de-
tails of this measurement is illustrated in Figure 1 of
Penz et al. (1992). The RR interval was the mean of 7
beats.
In individual animals, QaT estimates were made over
a range of heart rates, and the regression between QaT
and heart rate was plotted graphically. If the same
regression line occurred between different individuals of
the same species, the individual regressions were accu-
mulated to give an overall regression for the species.
These regressions were used to determine a regression
formula for each species. The effects of using such
derived formulae to correct QaT intervals to a common
RR of 250 msec were compared with the effects of the
other common correction factors.
The utility of the different correction factors was
tested by correcting all rates to a basic 240 beats/min
and examining the distribution of QaTc intervals for
closeness to normality (Gaussian) in terms of heterosce-
dascity by means of D'Agostino's test and calculation of
kurtosis and skewness. In addition, the coefficient of
variation (mean/standard deviation) was calculated as an
index of precision.
Results
Figures I(A-C) show the relationships between un-
corrected QaT as the dependent variable and heart rate
(RR) as the independent variable in individual rabbits,
guinea pigs, and rats, regardless of the technique used to
alter heart rate. As can be seen in Figure 1, the relation-
ships between heart rate and QaT were similar for the
different individuals of the same species, except for
rabbit 2 (Figure l-A), which was statistically signifi-
cantly different from the other animals (p < 0.05 for
difference for the parameters of its regression line from
the group). In the case of rats (Figure l-C), QaT did not
change with rate. However, with both rabbits and guinea
pigs, there was a clear positive regression between RR
and QaT.
In order to determine whether the method of chang-
ing heart rate had an effect on QaT intervals indepen-
dent of rate, data for all members of a particular species
were accumulated and plotted in Figures 2 and 3. In
Figure 2 the data points obtained by the different tech-
niques for changing heart rate show that the relationship
between RR and QaT for the rabbit (a) and the guinea
pig (b) were the same regardless of the procedures used
to change heart rate. In the case of primates (Figure 3),
values for monkeys and baboons initially were plotted
individually, but in view of the fact that the data were
not statistically different, they were grouped together for
calculation of the regression lines of best fit.
An attempt was made to determine the line of best fit
for the data using linear, square root, and polynomial
(power 2) functions, and the resulting best fit lines are
indicated in Figures 2 and 3. In the case of rabbit data,
the regression coefficients were 0.97, 0.90, and 0.97,
respectively, for the three types of regression. In the
case of guinea pig data, the corresponding coefficients
were 0.84, 0.84, and 0.85. For the accumulated pri-
mates, the values were 0.75, 0.75, and 0.75.
In view of the fact that a good regression relationship
between RR and QaT was obtained for the guinea pig,
5. E. HAYES ET AL. 205
RELATIONSHIP BETWEEN QaT AND RR INTERVALS
mIm
I-
o
11111
2a Rabbit . ....-
E
100
Q
=
0
2b Guinea Pig
f
AS & DS CS VS AS CS VS
I I I I I I I I I 0 I I I I
100 200 300 400 500 600 700 800 900 100 200 300 400
R-R Interval (mmac) R-R Interval (msac)
I
500
Figure 2. Regression between heart rate (as RR interval in msec) and QaT interval duration for rabbits (a) and guinea pigs (b). Data in
Figure l for rabbits and guinea pigs are replotted and the lines of best fit shown for the linear relationship QaT = aL + bE * RR + a (solid
line), square-root relationship QaT = bs * RR 1/2 (dotted line), and the polynomial relationship QaT = ae + be * RR + Ca * RR 2 (dashed
line), where QaT is the QaT interval in msec and RR in msec. No regression lines or data are shown for the rat because in this species
there was no relationship between QaT and RR (Figure lc). As discussed in the text, rabbit 2 was excluded from Figure 2a since it was
a statistically significant outlier and therefore can be excluded on statistical grounds according to the well-accepted Chauvenet's criteria.
The range of RR values over which various techniques were used to change heart rate are shown in the figure and denoted by VS for
vagal stimulation, AS for direct atrial stimulation, CS for cold saline, and DS for drug stimulation.
rabbit, and primates--whereas there was no such rela-
tionship for the rat--an attempt was made to determine
the most useful correction formulas for normalizing QaT
for changes in heart rate in guinea pigs, rabbits, and
primates (Table 1). The underlying premise for this
aspect of the study was that the best correction of QaT
would result in data that was normally (Gaussian) dis-
tributed, had the least heteroscedascity (by skewness
and kurtosis), and had the lowest coefficient of variation
(mean/standard deviation). The results of such correc-
tions are summarized in Table 2. This table compares
E
>=
@
p-
,=
o
3 Primate
2OO
0 I t
0 100 2OO
4-
4-
4- HI" ............
4-
4-
4-
I I I I I I I
300 4O0 500 60O 700 8OO 900
R-R Interval (msec)
Figure 3. Regression relationship between heart rate (as RR
interval in msec) and QaT interval duration for primates. Values
are accumulated data for both baboons and monkeys because no
significant difference between the two species was found. As in
Figure 2a and b, the lines of best fit are shown for linear (solid
line), square root (dotted line), and polynomial (dashed line)
formulas using the same equations as those found in the Figure 2
legend.
the correction using the above regression formulas with
a number of standard correction formulas that have been
used in human and other species. Examination of Table
2 shows that the best correction utilized the linear re-
gression equations found in Table 1 for the rabbit and
guinea pig, but not those for the rat because there was
no regression between QaT and heart rate in rats.
Discussion
Variations of QT interval with heart rate in humans
and other species has been noted many times. A com-
plete explanation, in terms of the underlying ionic
currents, is not available, but current hypotheses center
on the various potassium channels responsible for repo-
larization (Hume and Uehara, 1985; Carmeliet, 1993). It
is recognized that, at shorter diastolic intervals, a portion
of the repolarizing channels, such as those responsible
Table 1. Fitting Parameters for the Equations Used in Calculation
of Lines of Best Fit According to the Linear and Nonlinear
Formulas
Square
Linear root Polynomial
Species bL aL bs ap bp ce
G Pig 0.21 58.9 7.0 41.4 0.38 -0.00034
Rabbits 0.16 59.0 6.3 41.3 0.28 -0.00014
Primate 0.27 98.7 10.3 40.6 0.49 -0.00020
The equations used for the above fitting parameters were Linear:
QaT = aL + bL * RR; Square root: QaT = bs * VR'R; Polynomial: QaT =
ap + bp * RR + ce * RR2.
The fitting parameters in the above table are indicated by a, b, and c
with the appropriate subscript for the different functions.
6. 206 JPM Vol. 32, No. 4
December 1994:201-207
Table 2. StatisticalAnalysisof Effectsof Various CorrectionFormulason the QaT Interval
QaT Linear SqRt Poly Bazett Hodges Driscoll
Rabbit
Mean 108.9 99.0 102.0 102.4 102.7 113.7 6.5
SD 34.3 8.9 8.2 7.9 7.6 17.8 0.5
C.V.(%) 49.0 8.9 8.0 7.7 7.4 16.0 7.4
K -0.66 2.6 -0.23 -0.09 0.23 0.92 0.23
S 0.90 -0.26 -0.23 -0.08 0.23 0.92 0.23
P for Fit <0.01 <0.01 >0.20 <0.20 0.1--0.2 <0.01 0.1-0.2
Guinea Pig
Mean 101.3 112.4 113.8 116.8 115.2 118.6 7.3
SD 13.1 7.1 6.9 7.0 8.2 8.2 0.52
C.V.(%) 13 6.3 6.1 6.0 7.1 6.9 7.1
K 0.83 -0.35 -0.40 -0.50 -0.37 0.54 -0.37
S 0.32 -0.47 -0.38 -0.44 -0.25 0.10 -0.25
Fit 0.02-.05 >0.20 >0.20 >0.20 >0.20 >0.20 >0.20
Primate
Mean 223.8 163.7 165.7 149.0 164.4 202.6 10.4
SD 44.2 29.7 29.8 29.4 22.0 38.5 1.39
C.V.(%) 19 18 18 20 13 19 13
K 0.44 -0.27 -0.35 -0.56 0.16 0.30 0.16
S -0.14 -0.13 -0.06 0.08 -0.27 -0.08 -0.27
Fit >0.20 >0.20 >0.20 >0.20 >0.20 >0.20 >0.20
The above table shows the results of the following formulas used to correct the QaT interval for changes in
heart rate in rabbits, guinea pigs, and primates. For the linear correction the equation used was QaTc = QaT -
bE * (RR-250), while for the square root (SqRt) correction the equation used was QaTc = QaT + bs * (2~/~-0 -
~). Finally, the equation QaTc = QaT - be * (RR-250) - cp * ((250)2-(RR)2) was used as the polynomial
(Poly) correction. For each corrected mean, the standard deviation (SD) and coefficient of variation (CV) were
calculated. As well, the shape of the distribution of values for each species or kurtosis (K) and heteroscedasti-
city (S) can be seen, and the level of significance for the fit (P for fit) completes the table.
for iK~, would still remain activated from the previous
action potential and thereby shorten the action potential
(Hauswirth et al., 1972; Boyett and Fedida, 1984). In
addition, over time, changes in rate can be expected to
change the intracellular concentration of sodium and
calcium ions (Nierenberg and Ransil, 1979) as well as
extracellular potassium (Kunze, 1977), and, thus, influ-
ence repolarization. Such changes will, less directly,
result in corresponding changes in the repolarization
currents. The importance of such processes varies with
species, and, thus, rate dependent changes in QT could
be expected to vary with species. The lack of iK~ and
predominant ito in the rat (Josephson et al., 1984) may
account completely for the lack of any major effect on
rate on "QT" in this species.
It is necessary to recognize that changes in QT are
relatively poor indices in how repolarization might be
influenced by heart rate. The QT interval reflects, prob-
ably in a manner proportional to the number of each
type of cell present in the ventricle, repolarization pro-
cesses in the different types of cells found in ventricles.
Thus the repolarization process and its sensitivity to rate
could probably differ between ventricular and Purkinje
cells, between endocardial and epicardial cells, and be-
tween Spike and Dome cells, etc. Unfortunately the
possible differential effects of heart rate changes on the
shape and configuration of the QT interval has not been
systematically investigated. Regardless of such compli-
cations, the QT interval probably still represents a useful
approximation of action potential duration in the ventri-
cle and therefore is worth measuring.
If the QT is to be measured, there is a need to correct
for any concomitant changes in heart rate. This problem
has been studied many times for humans and even for
common laboratory species but without any consensus
as to the most useful and appropriate correction factor. It
is, however, apparent that there cannot be any one
correction formula that applies equally well to all spe-
cies.
In the case of the rat, there appeared to be little
variation in QT, as measured in this study, and heart
rate. Other investigators have previously investigated
this problem in rats and found some variability in QT
with rate (Beinfeld and Lehr, 1956).
In the guinea pig, a positive relationship between
heart rate and QaT was apparent and best corrected by
formulas derived from various regression equations.
Corrections according to the procedures of Bazett, Dri-
scoll, and Hodges were also beneficial in correcting
QaT for heart rate in this species.
In the rabbit, the various regression-corrected QaT
distributions are much better than the uncorrected data.
Corrections proposed by Bazett and Driscoll were of
some value, whereas the correction proposed by Hodges
7. E. HAYES ET AL. 207
RELATIONSHIP BETWEEN QaT AND RR INTERVALS
appeared to be of lesser use. In primates, the regression
corrected QaT distributions were not a great improve-
ment over the uncorrected data.
In conclusion, it appears that no one correction for-
mula has marked advantage over other correction for-
mulas for any species.
The BCHCRF and the BCYI-ISF are acknowledged for funding parts of
the above studies. MKP was a recipient of the B.C. Medical Services
Foundation Pre-Doctoral Scholarship.
References
Adaikan G, Beatch GN, Lee TL, Ratnam SS, Walker MJA (1992)
Antiarrhythmic actions of Tedisamil: Studies in rats and primates.
Cardiovasc Drugs Therapy 6:345-352.
Ashman R (1942) The normal duration of the Q-T interval. Am Heart J
23:522-534.
Bazett HC (1920) An analysis of the time relations of electrocardiograms.
Heart 7:353-370.
Beck G, Marriott H (1959) The electrocardiogram in hyperthyroidism.
Am J Cardiol 3:411-413.
Beinfield WH, Lehr D (1956) Advantages of ventral position in recording
electrocardiogram of the rat. J Appl Physiol 9:153-156.
Beinfield WH, Lehr D (1968) QRS-T variations in the rat electrocardio-
gram. Am J Physiol 214(1):197-204.
Boyett MR, Fedida D (1984) Changes in the electrical activity of dog
cardiac Purkinje fibers at high heart rates. J Physiol (Lond) 350:361-
391.
Browne KF, Prystowski E, Heger JJ, Zipes DP (1983) Modulation of the
Q-T interval by the autonomic nervous system. PACE 6(part
III):1050-1056.
Carlsson L, Abrahamsson C, Andersson B, Duker G, Schiller-Linhardt G
(1993) Proarrhythmic effects of the class III agent almokalant: impor-
tance of infusion rate, QT dispersion, and early after depolarizations.
Cardiovasc Res 27:2186-2193.
Carmeleit E. (1993) Mechanisms and control of repoladzation. Eur Heart
J 14(suppl H):3-13.
Chemoff HM (1972) The significance of the QT and QaT intervals. Conn
Med 36:236-360.
Driscoll P (1981) The normal rat electrocardiogram. In The Rat Electro-
cardiogram in Pharmacology and Toxicology. Eds., R. Budden, DK
Detweiler, and G Zbinden. New York: Pergamon Press; pp. 1-14.
Fredericia LS (1920) Die systolendauer im elektrokardiogram bei norma-
len menscben and bei herzkranken. Arch Med Scand 53:469-489.
Hauswirth IR, Noble D, Tsien RW (1972) The dependence of plateau
currents in cardiac Purkinje fibres on the interval between action
potentials. J Physiol (Lond) 222:27-49.
Hodges M, Salerno D, Erlien D (1983) Bazett's QT correction reviewed:
evidence that a linear QT correction for heart rate is better. J Am Coil
Cardiol 1:694.
Hume JR, Uehara A (1985) Ionic basis of the different action potential
configurations of single guinea-pig atrial and ventricular myocytes. J
Physiol (Lond) 368:525-554.
Josephson IR, Sanchez-Chapula J, Brown AM (1984) Early outward
current in rat single ventricular cells. Circ Res 54:157-162.
Kisch B (1953) The heart rate and the electrocardiogram of small ani-
mals. Exp Med Surg 11:117-130.
Kunze DL (1977) Rate-dependent changes in extracellular potassium in
the rabbit atrium. Circ Res 41:122-127.
Lepeschkin E (1955) The U wave of the electrocardiogram. Arch Intern
Med 9:600-617.
Nierenberg DW, Ransil BJ (1979) Q-aTe interval as a clinical indicator of
hypercalcemia. Am J Cardiol 44:243-248.
Penz WP, Pugsley MK, Hsieh MZ, Walker MJA (1992) A new ECG
measure (RSh) for detecting sodium channel blockade in vivo in rats.
J Pharmacol Meth 27(1):51-58.
Simonson E, Cady LD, Woodbury M (1962) The normal Q-T interval.
Am Heart J 63(6):747-753.
Taran LM, Szilagyi N (1947) The duration of the electrical systole (Q-T)
in acute rheumatic carditis in children. Am Heart J 33:14--24.
Van de Water A, Verheyen J, Xhonneux R, Reneman RS (1989) An
improved method to correct the QT interval of the electrocardiogram
for changes in heart rate. J Pharmacol Toxicol Meth 22:207-217.