An Improved Method for Adjusting the
QT Interval for Heart Rate
(the Framingham Heart Study)
Alex Sagie, MD, Martin G. Larson, ScD, Robert J. Goldberg, PhD,
James R. Bengtson, MD, and Daniel Levy, MD
Several form&s have been proposed to adjust
the QT Interval for heart rate, the most commonly
used bslng the QT correctlon formula (QT, =
QT/m) propwed In 1920 by Raxett. The QT,
formula was derived from observations In only 39
young subjects. Recently, ths adequacy of Ba-
&ts formula has been questtoned. To evaluate
ths heart rate QT assocMon, the QT Interval was
measured on the hdtlal baseline electrocardiogram
of S,OlS subjects (2,229 men and 2,779 women)
from the Framlngham Heart Study with a mean
age of 44 years (range 26 to 62). Persons wlth
coronary artery disease were excluded. A linear
regmdon modei was developed for correcting QT
according to RR cycle length. The large sample al-
lowed for subdlvlslon of the populatlon into sex-
speclflc deciles of RR Intervals and for comparison
of QT, Raxetts QT, and linear corrected QT
(QT&. The mean RR interval was 0.81 second
(range 0.5 to 1.47) heart rate 74 bsats/mln (range
41 to 120), and mean QT was 0.35 second (range
0.24 to 0.49) In men and 0.36 second (range 0.26
to 0.46) In women. The linear regression model
yleldedacorrectlonformula(fora&erenceRR
Interval of 1 second): QTu: = QT + 0.154 (1 - RR)
that applies for men and women. This equatlon
corrects QT more reliably than the Razetts formu-
la, which overcorrects the QT Interval at fast heart
rates and undercorrects lt at low heart rates. Low-
er and upper limits of normal QT values In relation
to RR were generated. A simple linear equation
was developed that Is mere accurate than Baxetts
correction at dlfferent cycle lengths and more con-
venient for cllnkal practke. This formula allevl-
ates the need to apply sscondary corrections to
9axetts fomuda. Addltlonal Investlgatlon Is war-
ranted to determined whether QTLC Improves the
ldentlftcation of subjects at high risk for mallg-
nant arrhythmias or sudden death.
(Am J Cardlol1992;70:797-601)
From the Framingham Heart Study, Framingham, the Divisions of
Epidemiology and Preventive Medicine of Boston University School of
Medicine, Boston, the Department of Medicine, University of Massa-
chusetts Medical School, Worcester, and the Divisions of Cardiology
and Clinical Epidemiology, Beth Israel Hospital, Boston, Massachu-
setts; and the National Heart, Lung, and Blood Institute, Bethesda,
Maryland. Dr. Sagie was a visiting Research Fellow from the Beiin
Medical Center, Tel Aviv, Israel. Manuscript received May 1, 1992;
revised manuscript received and accepted May 21,1992.
Address for reprints: Daniel Levy, MD, Framingham Heart Study,
5 Thurber Street, Framingham, Massachusetts 01701.
P
rolongation of the QT interval is associated with
increased risk for malignant ventricular arrhyth-
mias and sudden cardiac death in congenital long
QT syndromes, electrolyte disturbances, after use of
several antiarrhythmic drugs and after myocardial in-
farction.-* The duration of the QT interval has been
shown to depend on the length of the precedmg cardiac
cycle. It has been over 70 years since a formula for cor-
recting QT for cycle length (QTJ was introduced by
Bazett.13 However, the adequacy of thii approach (di-
viding the measured QT by the square root of the mea-
sured RR) to adjust correctly for differences in heart
rate has been questioned. i4-*0 Although many formulas
have been proposed to adjust the QT interval for heart
rate, only a few have been derived or validated in a
large population-based cohort.14-22 In the present study
we examined the initial baseline electrocardiogram in
over 5,000 persons originally enrolled in the Framing-
ham Heart Study. This report examines the relation of
QT (and QT,) to heart rate and presents results of a
linear regression equation (QTLc) that better corrects
QT for heart rate.
METHODS
The original cohort of the Framingham Heart Study
consisted of 5,209 subjects who were between the ages
of 28 and 62 years on entering the study >40 years ago.
The initial electrocardiograms of these subjects were
used to measure the duration of the QT interval. From
1985 to 1986 all initial electrocardiograms were mea-
sured manually by 2 trained readers with the aid of cali-
pers and magnifying lens. Measurements were obtained
from a 1Zlead resting electrocardiogram, with a paper
speed of 25 mm/s. The average of 2 to 3 QT intervals
from the electrocardiographic lead with the longest QT
interval was analyzed. The QT interval was measured
from the beginning of the QRS complex to the end of
the T wave where its terminal limb joined the baseline.
Each week a blinded duplicate review was performed on
a random sample of all electrocardiograms read during
that week. There was a high degree of inter- and in-
traobserver agreement. Any disagreements found were
subsequently adjudicated and corrected. Additionally, a
sample of 480 baseline electrocardiograms were blindly
and independently reread 5 times. A high degree of cor-
relation was obtained between the multiply read sam-
ples and the singly read baseline measurements with re-
gard to the mean RR, and average QT interval (p
<O.OOl). Subjects with coronary heart disease (n = 82)
at the initial examination were excluded from consider-
ation because coronary disease may have affected the
ADJUSTING QT INTERVAL FOR HEART RATE 797
 duces 2 regression lines: QT = Be + B1 RR + B2 l l

TABLE I Characteristics of Study Population

male, where male is a 0 to 1 indicator for male sex.
Men Women Total Linearity of the QT-RR relation was demonstrated
Age (years) 44 k 8.6 44 r 8.5 44 k 8.6 by grouping subjects by sex-specific decile values of RR
(range) (29-62) (28-62) (28-62) intervals and by plotting the mean QT versus mean RR
Heart rate (beats/min) 74 -t 12 77 f 12 76 f 12 for each decile. The fitted linear regression lines were
(range) (41-120) (43-120) (41-120) overlaid for comparison. To approximate the typical
RR interval (set) 0.84 + 0.14 0.80 2 0.12 0.81 + 0.13
(range) (0.50-1.47) (0.50-1.38) (0.50-1.47) 95% range of QT values that might be observed, we
QT interval (set) 0.35 k 0.03 0.36 + 0.03 0.35 lr 0.03 constructed 95% prediction intervals for persons at spec-
(range) (0.24-049) (0.26-0.48) (0.24-0.49) ified RR values. A table of lower and upper limits is
Values are means 2 standard deviation. presented.
An RR interval of 1.OOwas selected as the reference
point for the linearly corrected QT. Since Bazetts for-
duration of the QT interval. In addition, subjects taking mula QT, = QT/(e) is widely used, we applied
antiarrhythmic drugs or tricyclic antidepressants that both methods to the data for each subject and plotted
might have affected the QT interval (n = 2) were also mean values after grouping subjects by sex-specific RR
excluded. An additional 21 subjects were excluded be- deciles as mentioned before. This indicated whether
cause of extreme values for the RR interval (<0.5 or each method corrected QT evenly across the range of
>1.5 seconds), because there were not enough data for RR values.
analyses at the extremes. Finally, subjects with missing To see whether other nonlinear expressions for the
baseline electrocardiograms were also excluded (n = QT-RR relation were better than the linear model, we
84). After these exclusions, a total of 5,018 persons also fitted several alternatives motivated by a prior in-
(2,239 men and 2,779 women) formed the study vestigation by Puddu et al. I7 Goodness-of-fit of the pre-
sample. dicted values to the observed data was assessedvisually
Data analysis: Least-squares multiple regression23 by overlaying several cures, and quantitatively by lack
was used for modeling the QT-RR relation in men and of fit statistics and root-mean-squared errors (RMSE).
women. Analyses were performed on a SUN Spare Sta-
tion 1 running version 6.07 SAS software.24 At first, RESULTS
separate models were fitted for each sex, but their simi- Subjects Characteristics of the sample are summa-
larities allowed us to combine men and women for sub- rized in Table I. The mean age of 44 years was the
sequent analyses. The influence of RR, sex and age same for men and women. Women had shorter mean
were examined through stepwise selection applied to RR, faster heart rate and longer QT intervals than men.
RR interval, sex, age, RR2 (to accommodate a curvilin- QT-RR ngrauions: Separate QT-RR regressions
ear QT-RR relation), and sex - RR, sex - age, RR . age for men and women produced similar results (fitted
and sex . RR - age to accommodate higher-order influ- QT = 0.228 + 0.147 - RR, RMSE = 0.0220 and R
ences. The fmal choice was the linear model that pro- squared = 0.467 for men; fitted QT = 0.227 + 0.167 .

0.42 T

0.40
0

v Men: actual
- -Men: Linear
0 Women: actual
-Women: Linear

0.32

0.55 0.55 0.75 0.55 0.95 1.05 1.15

RR Interval (set)

798 THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 70 SEPTEMBER 15, 1992

 TABLE II Mean QT, QTc and QTLC at Different Deciles of RR TABLE Ill Mean QT, QTc and QTL~ at Different Deciles of RR
Interval in Men Interval in Women
RR Deciles Heart Rate No. of Mean Mean Mean RRDeciles Heart Rate No. of Mean Mean Mean
bed (beats/min) Men QT QT, QTLC bed (beats/min) Women QT QTc QTLC
0.50-0.67 go-120 246 0.316 0.402 0.375 0.50-0.64 94-120 303 0.323 0.417 0.385
0.68-0.72 83-88 228 0.333 0.398 0.379 0.65-0.69 87-92 315 0.335 0.409 0.386
0.73-0.76 79-82 238 0.337 0.391 0.377 0.70-0.72 83-86 225 0.340 0.403 0.384
0.77-0.80 75-78 276 0.345 0.389 0.378 0.73-0.76 79-82 333 0.349 0.404 0.388
0.81-0.83 72-74 171 0.350 0.387 0.378 0.77-0.79 76-78 235 0.354 0.401 0.388
0.84-0.86 70-71 220 0.353 0.384 0.377 0.80-0.82 73-75 285 0.360 0.400 0.390
0.87-0.90 67-69 219 0.356 0.379 0.374 0.83-0.85 71-72 261 0.366 0.399 0.390
0.91-0.95 63-66 206 0.363 0.377 0.374 0.86-0.89 67-70 277 0.370 0.395 0.389
0.96-1.02 59-63 219 0.373 0.376 0.375 0.90-0.96 63-67 301 0.380 0.394 0.391
1.03-1.47 41-58 216 0.391 0.370 0.373 0.97-1.38 43-62 244 0.391 0.384 0.385

RR, RMSE = 0.0224 and R squared = 0.445 for wom-
en). Because of these similarities men and women were
pooled for further analyses.
Stepwise regression was used to investigate the si-
multaneous effects on QT from RR, sex and age and
their higher-order interactions, as well as RR2 to allow
for curvature. The single most important variable was
the duration of the RR cycle, which explained 42.2% of
the QT variance; this was followed by sex which ex-
plained an additional 3.6% of the variance. Although
age, RR2, and the interaction term RR . age also were
statistically significant (each at p <O.OOS), their clinical
and biologic importance were minimal, because none of
these individually accounted for even 1% of the QT
variance; all together they accounted for only 1.4% after
adjustment for RR and sex. Therefore, we selected the
2-variable linear model: Fitted QT = 0.234 + 0.154 .
RR - 0.012 - male. This model yielded an RMSE of
0.0224 and R2 of 0.458, and was the simplest model
that was adequate to predict QT in our combined sam-
ple of men and women.
The basic linearity and goodness-of-fit for the QT-
RR relation is demonstrated in Figure 1, where each
data point represents mean QT versus RR decks. This
plot establishes that the fitted lines adhere closely to the
group means with no evidence of gross departure from
linearity at any RR value either in men or women.
QT eorre&ionr hear versus Bazettr The linearly
corrected QT (QT&, has the same expression for men
and women, namely: QTLC = QT + 0.154 . (1.00 -
RR) where QT and RR are measured values and 0.154
is the estimated regression slope. A subject with a heart
rate >60 beats/mm (RR <l) has a QTLC >QT, where-
as one with a heart rate <60 beats/min (RR >l) has a
QTLC <QT (i.e., the linear correction is upward when
the RR interval is shorter than 1 second and downward
when it is longer than 1 second).
Mean QT, and QTLC were computed for each RR
decile (Figure 2). The mean QTLC values are virtually
constant across the range of RR values. In contrast the
QTC values decline with increasing RR decks. Thus,
the linear correction for QT yields a valid correction;

0.42
\ \
\ , , Women QT c

MenQT;L-,
-. 1---
--. \ \
\
\ --- \

I I I I I

0.75 0.65 0.95 1.05 1.15

RR Interval (set)

ADJUSTING QT INTERVAL FOR HEART RATE 799

 TABLE IV Mean Predicted QT Values (and lower and upper 95% limits) at Different RR
Cycle Lengths
QT for Men (set) QT for Women (set)

RR Heart Rate Mean Lower Upper Mean Lower Upper

b?c) (beats/min) Value Limit Limit Value Limit Limit

0.50 120 0.299 0.255 0.343 0.311 0.267 0.354

0.55 109 0.307 0.263 0.351 0.318 0.274 0.362
0.60 100 0.314 0.270 0.358 0.326 0.282 0.370
0.65 92 0.322 0.278 0.366 0.334 0.290 0.378
0.70 86 0.330 0.286 0.374 0.341 0.297 0.385
0.75 80 0.337 0.293 0.381 0.349 0.305 0.393
0.80 75 0.345 0.301 0.389 0.357 0.313 0.401
0.85 71 0.353 0.309 0.397 0.364 0.321 0.408
0.90 67 0.361 0.317 0.404 0.372 0.328 0.416
0.95 63 0.368 0.324 0.412 0.380 0.336 0.424
1.00 60 0.376 0.332 0.420 0.388 0.344 0.432
1.05 57 0.384 0.340 0.428 0.395 0.351 0.439
1.10 55 0.391 0.347 0.435 0.403 0.359 0.447
1.15 52 0.399 0.355 0.443 0.411 0.367 0.455
1.20 50 0.407 0.363 0.451 0.418 0.374 0.462
1.25 48 0.414 0.370 0.458 0.426 0.382 0.470
1.30 46 0.422 0.378 0.466 0.434 0.390 0.478
1.35 44 0.430 0.386 0.474 0.441 0.397 0.486
1.40 43 0.438 0.394 0.482 0.449 0.405 0.493
1.45 41 0.445 0.401 0.489 0.457 0.413 0.501
1.50 40 0.453 0.409 0.497 0.465 0.421 0.509

the group mean QTLC values were between 0.313 and
0.379 for men, and between 0.384 and 0.391 for wom-
en. Small sex differences in QTLC were observed with
mean values of 0.376 in men and 0.388 in women. Be-
cause of the large sample size, limits for the central 95%
of individual QTLC values are approximated by the ex-
pression QTLC (mean) f 1.96 . RMSE. These 95% lim-
its for QTLC are 0.332 to 0.420 for a man and 0.344 to
0.432 for a woman. A subject whose QTLC is outside
these sex-specific limits is considered to have an unusu-
ally short (or long) QT interval. Lower and upper limits
of the QTC distribution are not constant over the ob-
served range of RR, but must be determined anew at
each RR value. The impact of different decks of RR
intervals on QT, QTC and QTLC in men and women is
presented in Tables II and III.
Bazetts formula, QTC = QT/fl, also adjusts up
ward when RR is <l, and downward when RR is > 1
second. But, it yields substantial overadjustment at
short RR intervals and underadjustment at long RR
intervals (Figure 2). Bazetts correction is flawed,
whereas the linear adjustment QTlc produces a common
reference value for all members of the same sex.
Another way to use the linear model for cliical pur-
poses is to construct 95% confidence limits for an indi-
vidual QT observation at different RR cycle lengths.
These data are listed in Table IV. A subject whose mea-
sured QT falls outside the appropriate knits has an un-
usually short (or long) duration QT interval. This table
alleviates the need to compute QTLc, although it may
require linear interpolation when a measured RR lies
between tabulated values.
DISCUSSION
Assessment of a heart rate-adjusted QT interval is
considered to be clinically important, since prolonga-
tion of this interval is associated with increased inci-
dence of malignant ventricular arrhythmias and sudden
death.-I2 To use heart rate-corrected QT to predict
life-threatening arrhythmias and sudden death, it would
be appropriate to use a QT correction formula that is
developed and validated in a large population-based co-
hort. Using Framingham Heart Study subjects, we de-
veloped a linear regression equation (QTLc) that more
accurately corrects QT for heart rate than the tradition-
ally used Bazetts (QT,) formula. We found that QTc
undercorrects QT at slow heart rates and overcorrects it
at fast heart rates, whereas QTLC reliably corrects QT
across a wide range of RR cycle lengths. The large
number of subjects studied also enabled us to subdivide
the sample into deciles of RR intervals and to examine
the impact of different cycle lengths on QT, QTC and
QTLc.
Several formulas have been proposed to adjust the
QT interval for heart rate.13-22 The most frequently
used is Bazetts square root formula which was intro-
duced more than 70 years ago.13 However, the adequa-
cy of this nonlinear formula, obtained from data on 39
young men, has been questioned because the QTC over-
corrects the measured QT interval at fast heart rate and
undercorrects it at a low heart rate.16-22 Our data con-
firm this deficiency of Bazetts correction. Puddu et all7
examined 10 mathematically different QT prediction
formulas in 881 middle-aged men and provided similar
evidence that the QT is overcorrected by Bazetts for-
mula at high heart rates. They observed that the cubed
root Fridericias formula20 fitted the data better than
Bazetts or other formulas. In the present study we test-
ed the regression models used by Puddu and found that
most of them (except Bazetts formula) adequately and
equally fitted the QT-RR data. Our results generally
agree with those of Puddu,l but we included women as

800 THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 70 SEPTEMBER 15, 1992

well as men, and we did not make arbitrary distinctions
between several models that fit equally well.
Since the adequacy of the linear regression model
was not inferior to the others, it was chosen as the most
suitable simple equation that could be used conveniently
in clinical practice. With use of this formula, there is no
need for further adjustment using a secondary correct-
ing nomogram as one does with Bazetts formula. Alter-
natively, a table of lower and upper limits of QT inter-
val for different RR cycles (Table IV) can be used by
clinicians for references purposes. This approach obvi-
ates the need for using any QT correction formula (QTc
or QTLc). The reliability of this model results from its
derivation in a large population-based sample. The
same formula applies for men and women (owing to the
almost identical linear regression slope).
One limitation of the present study is that QT mea-
surements were collected from resting electrccardio-
grams. We did not have the option to test whether our
linear model applies at extreme values of heart rate.
Also, the regression model accounts for <50?6 of the
variance of the QT interval. This indicates that the QT
interval is related not only to heart rate, but also to sev-
eral other factors such as changes in autonomic neural
tone and metabolic factors.25-28 It is clear that individu-
al variability is large.
Controversy persists regarding the significance of
QT prolongation as a risk factor for sudden cardiac
death in healthy subjects29-31 and in patients after myo-
cardial infarction.8-12 Each of these earlier studies used
Bazetts formula to adjust QT for heart rate and ap-
plied the same value for QT,: prolongation of 440 ms as
the upper limit of normal. Additional research is war-
ranted to determine if linear correction of the QT inter-
val improves our ability to detect patients at risk for
malignant ventricular arrhythmias and sudden death.
REFERENCES
1. Mass AJ. Prolonged QT interval syndromes. JAMA 1986;256:2985-2987.
2. Surawicz B, Knobel SB. Long QT: good, bad or indifferent? JAm Co11 Cardiol
1984;4:398-413.
3. Bhandar AK, Sceinman M. The long QT syndrome. Mod Concepts Cardicwsc
Dis 1985;54:45-50.
4. Ahnve S, Lundman T, Shoalch-var M. The relation between QT interval and
ventricular arrhythmias in acute myocardial infarction. Acta MedScmd 1978;20
417-19.
5. ,Puddu PE, Jouve R, Torresan J, Jouve A. QT interval and primary ventricular
fibrillation in acute myocardial infarction. Am Heart J 1981;101:118-119.
6. Ahnve S, Gilpin E, Madsen EB, Froelicher V, Henning H, Ross J Jr. Prognos-
tic importance of Qtc interval at discharge after acute myocardial infarction: a
multicenter study of 865 patients. Am Heart J 1984;108:395-400.
7. Haynes RE, Hallstrom A, Cobb LA. Repolarization abnormalities in survivors
of out of hospital ventricular fibrillation. Circulation 1978;57:654-658.
8. Puddu PE, Bourassa MG. Prediction of sudden death from QTc interval
prolongation in patients with chronic ischemic heart disease. J Electrocardiol
1986;19:203-212.
9. Pohjola-Sintonen S, Siltanen P, Haapakoski J. Usefulness of QTc interval on
the discharge electrocardiogram for predicting survival after myocardial infarc-
tion. Am J Cardiol 1986;57:1066-1068.
10. Schwartz PJ, Wolf S. QT interval prolongation as predictor of sudden death
in patient with myocardiil infarction. Circulation 1978;57:1074-1077.
11. Juul-Moller S. Corrected QT interval during one year follow-up after an
acute myocardial infarction. Eur Heart J 1986;7:299-304.
12. Fioretti P, Tijssen GJP, Azar Lazzeroni E, Brewer RW, Katen HJ, Lubsen J,
Hugenholtz PG. Prognostic value of predischarge 12 lead electrocardiogram after
myocardial infarction compare with other routine clinical variable. Br Heart J
1987;57:306-312.
13. Bazett HC. An analysis of the time-relations of electrocardiogram. Heart
1920;7:353-370.
14. Samara JSM, Samara RJ, Bilitch M, Katz D, Song SL. An exponential
formula for heart rate dependence of QT interval during cardiac pacing in hu-
mans: reevaluation of Bazetts formula. Am J Cardiol 1984;54:103-108.
15. Kovacs SJ Jr. The duration of the QT interval as a function of heart rate: a
derivation based on physical principles and a comparison to measured values. Am
Heart J 1985:110:872-878.
16. Ahnve S. Correction of the QT interval for heart rate: review of different
formulas and the use of Bar&ts formula in myocardial infarction. Am Heart J
1985;109:568-574.
17. Puddu PE, Jouve R, Mariotti S, Giampaoil S, Lanti M, Real A, Menotti A.
Evaluation of 10 QT prediction formulas in 881 middle-aged men from the seven
countries study: emphasis on the cubic root Fridericias equation. J Electrocardiol
1988;21:219-229.
19. Van De Water A, Verheyen J Xhonneux, Reneman RS. An improved method
to correct the QT interval of the electrocardiogram for changes in heart rate. J
Pharmocol Methods 1989;22:207-217.
19. Manion CV, Whit&t TL, Wilson MF. Applicability of the correcting QT
interval for heart rate (letter). Am Hearf J 1980;99:678.
20. Fridericia LS. Die systlendauer in elektrokardiogramm bei normalen men-
then und bei Herzkranken. Acta Med Scami 1920;53:469-486.
21. Ashman R. Normal duration of QT interval. Am Heart J 1942;23:522-553.
22. Adams W. The normal duration of the electrocardiographic ventricular
complex. J Clin Irwest 1936;15:335-642.
23. Kleinbaum DB, Kupper LL, Muller KE. Applied regression analysis and
other multivariable methods. 2nd ed. Boston: PWS Kent Publishing, 1988:
109-l 10.
24. SAS Institute, Inc., SAS/STAT Users Guide, Version 6.4th ed. ~012. The
REG procedure Gary, NC: SAS Institute, Inc, 19891351-1456.
25. Brown KZ, Zipes DP, Heger JJ, Prystowsky EN. Influence of the autonomic
nervous system on the QT interval in man. Am J Cardiol 1982;50:1099-1103.
26. Fananapazir L, Bennett DH, Faragher EB. Contribution of heart rate to QT
interval shortening during exercise. Eur Heart J 1983;4:265-271.
27. Staniforth DH. The QT interval and cycle length: the influence of atropine,
hyoscine and exercise. Br J C/in Phormocol 1983;16:615-621.
28. Davidowski TA, Wolf S. The QT interval during reflex cardiovascular adap
tation. Circulation 1984;69:22-25.
29. Goldberg RJ, Bengtson J, Chen Z, Anderson KM, Lo&i E, Levy D. Dura-
tion of the QT interval and total cardiovascular mortality in healthy persons (The
Framingham Study experience). Am J Cardiol 1991;67:55-58.
30. Algra A, Tijssen GP, Roeland RTC, Pool J, Lubsen J. QTc prolongation
measured by standard 12 lead electrocardiography is an independent risk factor
for sudden death due to cardiac arrest. Circularion 1991;83:1888-1894.
31. Schouten EG, Dekker JM, Meppelink P, Kok FJ, Vandenbroucke JP, Pool J.
QT interval prolongation predicts cardiovascular mortality in an apparently
healthy population. Circulation 1991;84:1516-1523.
ADJUSTING QT INTERVAL FOR HEART RATE 801