Sexual Dimorphism in the Electrocardiographic Dynamics of Human Ventricular Repolarization
Characterization in True Time Domain
Background—Previous characterizations of sex differences in ST-T waveform voltages have largely focused on amplitudes at selected time points during repolarization, subject to potential distortions from variations in heart rate (HR) or reliance on a JT-normalized time scale.
Methods and Results—Using digitized 12-lead ECGs from 553 normal adults (426 males) with HRs confined to 60±1, 70±1, or 80±1 bpm, we derived X, Y, and Z lead voltages and then generated, for each HR category by sex, summary (population mean) resultant spatial vector amplitudes (ST-TXYZ) and instantaneous slopes (dV/dtXYZ) at successive 4-ms intervals following the J point. Within each HR category, there was an early intersex divergence of ST-TXYZ trajectories (95% CIs nonoverlapping), with men exhibiting 2- to 3-fold greater dV/dtXYZ values during the ST segment and achieving greater maximum TXYZ and dV/dtXYZ values than women; descending TXYZ limbs were relatively more concordant between sexes but still steeper in men. The early sex differences in repolarization dynamics persisted in multiple regression analyses that took into account age and a morphometric index of left ventricular mass. In men, absolute values of extrema of TXYZ and dV/dtXYZ varied inversely with HR.
Conclusions—At physiological resting HRs, the spatial ST-T vector voltage time trajectory is steeper in men than in women, beginning virtually from the J point. In addition to its mechanistic implications, the demonstration of marked sensitivity of ST-TXYZ and especially dV/dtXYZ to sex raises the possibility that these time-based, ECG-derived parameters might be informative in pathophysiological studies of ventricular repolarization.
Recent awareness of an increased susceptibility of women to development of torsade de pointes ventricular tachycardia1 has highlighted the need to understand better the nature and extent of sex modulation of normal human ventricular repolarization. The existence of quantitative sex differences in the ECG manifestations of myocardial repolarization has been appreciated since the early 20th century, when Bazett2 reported that women have a longer average rate-corrected QT interval (QTc) than men. Such a sex disparity has since been confirmed repeatedly and found to apply to various ECG indices of repolarization duration.3 4 5 6 7 8
Over the years, investigators have also described differences between the sexes with respect to selected ST- and T-wave amplitudes, ie, typically higher mean values in men.3 4 9 10 11 These studies, however, did not correct for variations in heart rate (HR), a deficiency that other researchers attempted to remedy by reliance on a JT-normalized time scale.12 13 14 Thus, to date, a systematic comparison of the ST-T voltage trajectory and its dynamics, as a true function of time, in men and women has not been accomplished.
In the present study, we circumvented rate-related or temporal distortions inherent in previous methodologies by utilizing ECG data obtained at uniform HRs. This enabled us, for the first time, to compare by sex successive (4 ms) ST-T spatial vector voltages and first derivatives of those voltages in true time domain.
We used a database of ECGs recorded on a MAC-12 System (Marquette Electronics Inc) in normal adults, as described previously.5 8 In brief, subjects contributing to the database had normal histories and physical examinations and were not taking any cardioactive medication. None had ECG evidence of ventricular hypertrophy, intraventricular conduction disturbance, or myocardial infarction. The study subset was limited to individuals in the database whose resting HR by ECG fell into 1 of 3 narrow ranges: 60±1, 70±1, and 80±1 bpm, hereafter designated as HR 60, HR 70, and HR 80, respectively.
The 12-lead ECG signals were digitized and recorded at 250 Hz (4-ms sampling intervals). HRs, as well as markings of P onset, QRS onset and offset (ie, J point), and T offset, based on the median beat, were calculated by the MAC-12 software (12SL program) as previously described and validated15 ; however, QRS onset and offset were the only predetermined fiducial points we used.
Of 618 subjects (482 men and 136 women) in the ECG database who met the inclusion criteria described above, 34 (29 men and 5 women) had to be excluded because of incomplete ECG data files. An additional 31 subjects (27 men and 4 women) were excluded because of the absence of height or weight data needed for estimation of body surface area (described below). ECGs for the remaining 553 subjects formed the basis of the present study.
Study Population Characteristics
Of the 553 study subjects, 426 (77%) were men and 127 (23%) were women. These individuals hailed from across the United States and had the following racial profile: 416 (75.2%) were white, 40 (7.2%) were black, 40 (7.2%) were Hispanic, 13 (2.4%) were Oriental, 12 (2.2%) were Native American, and 32 (5.8%) were of unknown race. Mean age of the men was 34.0±12.1 years (range 15 to 74 years), and mean age of the women was 40.9±16.5 years (range 17 to 81 years).
Computed ECG Parameters
Digitized and MAC-12–preprocessed 12-lead ECG signals were stored and processed on a personal computer.8 For purposes of the present study, we confined our analysis primarily to the ST-T waveform, ie, ECG voltages recorded during the JT interval, at successive 4-ms intervals, with t=0 ms defined at the J point. Various ECG-derived parameters were calculated for distinct population subsets within the database, stratified by HR category and sex. Our analysis was based on ST-T vector voltages in the 3 orthogonal Frank X, Y, and Z leads (ST-Tx, ST-Ty, and ST-Tz, respectively), which were mathematically synthesized (at each successive sample time) from linear combinations of standard ECG leads V1 through V6 and leads I and II.16
Instantaneous Voltages of Reconstructed Spatial ST-T Vectors as a Function of Time
For each subject, the voltage magnitude of the reconstructed instantaneous spatial ST-T vector was calculated as the 3D resultant, ST-Txyz, of the 3 orthogonal components ST-Tx, ST-Ty, and ST-Tz. Thus, at any given sample time ti:
For a population of n subjects, we could then calculate at each sample time a summary (population mean) instantaneous resultant spatial ST-T voltage as follows: where ST-Txyzk(ti) is the value of ST-Txyz in the k-th subject at sample time ti. The ST-TXYZ waveform, so generated over successive sampling sample times during repolarization, consisted of contiguous portions STXYZ and TXYZ, followed by UXYZ. Analogous to calculation of ST-TXYZ, we also calculated at each sample time summary (population mean) instantaneous orthogonal lead ST-T voltages, ST-TX, ST-TY, and ST-TZ, as means of the respective individual ST-Tx, ST-Ty, and ST-Tz values.
Instantaneous Slopes (dV/dt) of ST-T Waveform
For any given sequence of ST-T(ti) voltages, the raw instantaneous slope (change in voltage, or dV/dt) of the ST-T(t) curve at t=ti is approximated by [ST-T(ti)−ST-T(ti−1)]/[ti−ti−1], provided ti−ti−1 is sufficiently small (4 ms in our study). Owing to the jagged nature of the resulting dV/dt of ST-T waveform, we used a 7-point moving average of this raw dV/dt (centered at each ti) to generate a smoothed dV/dt of ST-T waveform,8 beginning at 16 ms after the J point. Throughout the remainder of this article, dV/dt is assumed to refer to smoothed dV/dt.
Using the dV/dt values for each ST-Txyz curve, designated dV/dtxyz, it was then possible to calculate for the n-sized population of interest, at any sample time ti (beginning 16 ms after the J point), a summary (population mean) instantaneous slope of the spatial ST-T voltage vector magnitude waveform, as follows:
A more mathematically precise notation for this parameter is d(VXYZ)/dt, but we have opted for dV/dtXYZ as a simpler linear notation, with the understanding that the spatial subscript (XYZ) actually applies to the ST-T voltage (V) rather than time (t).
Specific parameters were extracted from the summary (not individual) ST-TXYZ and dV/dtXYZ curves. Three extrema were derived: Max TXYZ, the peak value of TXYZ, and Max (or Min) dV/dtXYZ, the peak positive (or peak negative) value of dV/dtXYZ. Two duration parameters were extracted: J-to-Max TXYZ, the time from the J point to the peak of TXYZ, and J-to-End TXYZ, the time from the J point to the nadir between TXYZ and UXYZ.
Summary values of a given parameter with corresponding 95% CIs were plotted on a graph to depict quantitative differences in summary waveforms between the sexes. Inasmuch as age is known to affect various ECG parameters of depolarization and repolarization,4 6 9 10 14 we repeated our sex comparisons within HR categories dichotomized by a 40-year age cut point. ST-TXYZ and dV/dtXYZ were also recalculated as parameters normalized for each subject’s height (in meters) to the 2.7 power and, alternatively, body surface area estimated17 from height and weight to the 1.5 power. These morphometric indices (especially the former) have been shown to largely adjust for sex differences in echocardiographically measured left ventricular mass.18
Intersex and intrasex comparisons of mean age, height2.7, and QRS duration for the 3 HR categories were performed initially with ANOVA, followed where appropriate by Student’s unpaired t test. Within each HR category, multiple regression analysis was used to assess the potential influence of sex, age (as a continuous variable), and height2.7 on both ST-TXYZ and dV/dtXYZ at 40, 80, and 120 ms after the J point and on the extrema Max TXYZ, Max dV/dtXYZ, and Min dV/dtXYZ. Analogous multiple regression analyses were performed to analyze the potential influence of HR on these extrema of TXYZ and dV/dtXYZ for each sex. Statistical significance in these analyses was defined as P<0.05.
Table 1⇓ provides age, morphometric, and QRS duration data for study subjects by HR category. Within each HR category, men were younger, with greater mean height2.7 and QRS duration. For each sex, there were no significant differences across HR categories with respect to any of the 3 parameters except for age in men, at HR 60 versus HR 80 (P<0.03).
ST-TXYZ and dV/dtXYZ in Men vs Women
Figure 1⇓ shows, for each HR category, a comparison by sex of summary ST-TXYZ voltages (top panels) and summary dV/dtXYZ values (bottom panels) as a function of time (referenced to the J point). At each HR, the ST-TXYZ mean voltages, from the J point through the entire ascending limb of TXYZ, achieved greater values and peaked 28 to 32 ms sooner in men. The 95% CIs for these mean voltages, at each 4-ms sample time, were nonoverlapping between the sexes from the J point (or just beyond) through their respective TXYZ peak amplitudes. Although the J-to-End TXYZ interval was also shorter in men (by 20 to 36 ms), mean spatial vector voltages were relatively more concordant between sexes over the course of the descending limb of TXYZ after transient overlap of 95% CIs, with mean voltages declining more rapidly in men. The UXYZ peak also occurred earlier (by 32 to 52 ms) in men than in women.
As evident from the dV/dtXYZ curves in Figure 1⇑, for each HR category, early values of dV/dtXYZ, as well as absolute maximum and minimum values, were consistently greater in men. At each HR, the 95% CIs for mean instantaneous slopes of the spatial vector voltage time function were nonoverlapping between men and women, from the first plotted value (at 16 ms after the J point) through the timing of Max dV/dtXYZ in men. After the initial crossing of the male and female dV/dtXYZ curves (near Max dV/dtXYZ in women), the 95% CIs were again nonoverlapping through the timing of Min dV/dtXYZ in men. Subsequently, in their terminal portions, these dV/dtXYZ curve pairs exhibited somewhat greater concordance. Max dV/dtXYZ was attained 28 to 40 ms and Min dV/dtXYZ 24 to 36 ms sooner in men than in women.
Even at the 40- and 80-ms points of the ST segment, the male to female ratios for summary parameters ST-TXYZ and dV/dtXYZ ranged approximately from 1.2 to 2.5 and from 2.3 to 3.6, respectively, at each HR, with nonoverlap of even 99% CIs between the sexes for ST-TXYZ and dV/dtXYZ values (Table 2⇓). In all 3 HR categories, the sex ratios for these ST-related parameters (especially dV/dtXYZ) typically exceeded the female to male ratios (range 1.1 to 1.2) for either of the repolarization duration indices, J-to-Max TXYZ or J-to-End TXYZ.
ST-T Voltages in ECG-Derived Orthogonal Leads
Figure 2⇓ shows summary ST-T voltages at successive 4-ms intervals in each of the 3 mathematically synthesized orthogonal component leads that were used to reconstruct successive ST-TXYZ vectors, at HR 70. Although mean ST-T voltages until the T-wave peak were of greater absolute magnitude in men for each of the 3 orthogonal leads (but most apparent in leads X and Z), the algebraic signs of these components from the J point onward were similar in both sexes (ie, positive in X, positive in Y, and negative in Z), yielding a spatial orientation of each instantaneous resultant vector that was grossly comparable (directed leftward, inferiorly, and anteriorly) in men and women. Similar results were obtained at HR 60 and HR 80 (data not shown).
Adjustment for Age and Body Size
The sexual dimorphism in ST-TXYZ voltages and dV/dtXYZ observed at HR 70 (Figure 1⇑) was maintained when ECGs were dichotomized with a 40-year age cut point (Figure 3⇓; more evident in the younger age group) or when the analysis was performed with ST-T voltages normalized by height2.7 (Figure 4⇓). Qualitatively similar results were obtained at HR 60 and HR 80 (data not shown). For all 3 HR categories, the described sex differences also persisted when morphometric normalization was performed with body surface area1.5 rather than height2.7 (data not shown).
Within HR groups, multiple regression analysis was performed to assess the possible predictive effects of sex, age, and height2.7, specifically on early spatial vector amplitudes and slopes, at 40, 80, and 120 ms after the J point. Sex was found to be a highly significant independent predictor of the 3 early ST-TXYZ parameters (P<0.0001 at HR 60 and HR 70, P<0.004 at HR 80) and the 3 early dV/dt XYZ parameters (P<0.0001 at HR 60 and HR 70, P<0.001 at HR 80). Age was also a significant independent (inversely related) predictor of these various parameters at HR 60 and HR 70 (P<0.0001) but not at HR 80. Height2.7, however, was virtually never an independent predictor of any of the 6 parameters (exception mainly at 40 ms).
When extrema of TXYZ and dV/dtXYZ were similarly analyzed by multiple regression analysis, within HR categories, sex was independently predictive of Max TXYZ at all 3 HRs (P<0.001), Max dV/dtXYZ at HR 60 and HR 80 (P<0.02), and Min dV/dtXYZ at HR 60 and HR 70 (P<0.05). For all 3 HRs, age, but not height2.7, was a significant (inverse) predictor of Max TXYZ (P<0.02) and absolute values of Max (P<0.03) and Min (P<0.01) dV/dtXYZ.
Effect of HR on ST-TXYZ and dV/dtXYZ
Superimposed HR plots of ST-TXYZ and dV/dtXYZ for men and women are shown in Figure 5⇓. When extrema of TXYZ and dV/dtXYZ were compared by HR, within each sex, multiple regression analysis (incorporating age and height2.7) revealed a statistically significant HR effect in men such that the absolute values for Max TXYZ, Max dV/dtXYZ, and Min dV/dtXYZ varied inversely with HR (P<0.0001, HR 70 and HR 80 versus HR 60 for Max TXYZ; P<0.02, HR 80 versus HR 60 for Max dV/dtXYZ; and P<0.0001, HR 80 versus HR 60, and P<0.01, HR 70 versus HR 60, for Min dV/dtXYZ). There were no statistically significant differences among values of these parameters in women, however, when compared by HR. In both sexes, the multiple regression analysis also showed age, but not height2.7, to be a significant (inversely related) predictor of the absolute values of Max TXYZ, Max dV/dtXYZ, and Min dV/dtXYZ (P<0.0001 for men, P<0.02 for women).
The present study is the first to compare at common HRs in men and women successive mean ECG-derived spatial vectorcardiographic voltages (ST-TXYZ) and concomitant mean instantaneous slopes (dV/dtXYZ) at high temporal resolution during ventricular repolarization. At each of the 3 physiological resting rates studied (HR 60, HR 70, and HR 80), our findings attest to clear quantitative sex differences in the ECG dynamics of repolarization, with more brisk dynamics of ST-T–wave generation seen in men. This sex difference largely persisted after adjustment for age and body size.
Conventional3 4 9 10 and orthogonal11 lead scalar ECG studies have demonstrated in men greater peak T-wave amplitudes (especially anteriorly), select ST-segment voltages, and precordial mean ST angles.19 Limiting these prior investigations, however, was the fact that measured ST-T parameters were averaged within study groups regardless of intersubject variations in HR. In contrast, Green et al14 used a time-normalization procedure on digitized body-surface potential maps in normal subjects and reported “slightly” larger average ST potentials and greater average early and mid T-wave potentials in men than in women. Analogous observations were made in previous scalar orthogonal lead ECG studies in which intersex time-normalized repolarization voltages were compared at successive 1/8 time divisions of the JT interval.12 13 More recently, Macfarlane20 found men to have greater average ST angles, as measured between the J point and 3/8 of the JT interval (P.W. Macfarlane, PhD, personal communication, 1999), in selected conventional ECG leads. The time-normalization approach, however, is acknowledged to introduce some distortions of its own, potentially affecting intergroup comparisons.14
The present study design was intended to overcome limitations of the earlier methodologies by focusing on ECGs at common HRs, thereby permitting direct and comprehensive comparisons of derived 3D vectorcardiographic voltages as a true function of time (at 4-ms intervals) during ventricular repolarization. We demonstrated greater and more rapidly attained absolute values of maximal instantaneous slopes of the ascending and descending limbs of the spatial T vector waveform in men, extending our findings in lead V5.8 Importantly, the present study enabled us to trace systematically the temporal origin of these more brisk male dynamics of T-wave generation back to the ST segment (Figure 1⇑), where dV/dtXYZ values were already some 2- to 3-fold greater in men (Table 2⇑). Thus, at each study HR, a steeper ST-TXYZ voltage time trajectory was found in men relative to women, essentially from the ECG inception of ventricular repolarization. Moreover, sex differences in early ST-T amplitudes and slopes persisted in multiple regression analyses that included age and height2.7.
We found expected sex differences in repolarization duration5 to be of relatively smaller magnitude than the sex disparity in early dV/dtXYZ values. Whereas J-to-Max TXYZ and J-to-End TXYZ were shorter (by ≈20 to 40 ms) in men, the female to male ratios were just slightly >1 for both parameters.
The ST-T waveform is primarily a manifestation of successive, instantaneous net transmural ventricular voltage gradients during repolarization, with superimposed effects of regional, ie, “geographic,” differences in electrical recovery properties (especially within the more rapidly repolarizing epicardial layer).21 The general concordance in orthogonal lead distribution of early ST-T voltages in men and women (Figure 2⇑), however, argues against a gross sex disparity in geographic sequence of repolarization as the basis for the observed sexual dimorphism in ST-TXYZ dynamics.
More tenable is the hypothesis that during repolarization, there are sex-related quantitative differences in the action potential voltage time course of 1 or more ventricular transmural cell types that may have an effect on the morphology of the ST-T waveform. This conjecture is motivated by experimental data implying that the aggregate ventricular repolarization gradient reflected in the ST-T waveform until the peak of the T wave is driven largely by the voltage time dynamics of phases 2 and 3 of action potentials in the epicardium relative to those of deeper myocardial cell layers.22
Sex steroids (especially androgens) appear likely mediators23 of the peripubertal shortening of QTc6 and the lower propensity to JTc prolongation24 and torsade de pointes1 exhibited by men versus women. By extension, these hormones, which are capable of affecting repolarizing currents,25 26 also could have a modulating influence on the transmural voltage gradients that contribute to the genesis and dynamics of the ST-T waveform. This hypothesis is supported by a recent report describing differences in precordial lead T-wave amplitudes and average ST angles under altered androgenic states.27
Influence of HR and Age
Our findings, at least in men, of a decrease in Max TXYZ as resting HR increases (Figure 5⇑) confirm definitively earlier scalar ECG impressions.3 Such behavior, implying the need for “rate correction” of maximum T-wave amplitude measurements, is consistent with observations in the perfused left ventricular wedge preparation.28 We further documented a similar inverse relationship in men between absolute value of Max (and Min) dV/dtXYZ and HR.
In both men and women, age was found to be an independent (inverse) predictor of the absolute values of the extrema of TXYZ and dV/dt XYZ (at all 3 HRs), as well as the values of these parameters during the ST segment (at HR 60 and HR 70). Although of unclear mechanistic origin, these observations at controlled HRs extend traditional ECG analyses showing age-related declines in ST-T parameters, especially in men, during adulthood.4 9 10
Conceivably, differences in resting HR among the study population may have reflected, in part, variations in other physiological parameters, eg, sympathetic and parasympathetic tone. Such an HR selection artifact could have had a confounding influence on the quantitative comparison of repolarization dynamics among the HR categories we studied but would not be expected to negate the highly consistent sex difference we documented within each HR group. We recognize that the onset of detectable ventricular repolarization actually precedes the J point (our operational “time zero”) by an average of ≈5 to 12 ms.29 However, slight interindividual differences in true t0 would not explain the consistent early upward displacement of the dV/dtXYZ curves and steeper STXYZ trajectory in men relative to women. The relative paucity of women in our study population (<1:3 female:male ratio) may have limited our ability to precisely define the relationship between HR and Max TXYZ, Max dV/dtXYZ, and Min dV/dtXYZ in women (Figure 5⇑). Expanded studies are needed to address this issue and to explore ST-T dynamics in nonwhite populations, older subjects, and groups with defined levels of physical activity.
The present findings add to accumulating evidence for the existence of fundamental differences between the sexes in the physiology of ventricular repolarization. Our observations of the marked sex sensitivity of instantaneous ST-T spatial vector amplitudes, and particularly slopes, raise the possibility that these time-based parameters may convey information regarding a variety of processes capable of modulating repolarization.30
The authors thank Gregor Altvater, BSEE, for reviewing the data calculations; Dan Montgomery, BS, for assistance with statistical analysis and preparation of the figures; and Michael R. Rosen, MD, and Janice M. Jenkins, PhD, for their critical reviews of the manuscript.
Reprint requests to Michael H. Lehmann, MD, University of Michigan Medical Center, Women’s Hospital Room L3119, Box 0273, 1500 E Medical Center Dr, Ann Arbor, MI 48109-0273.
- Received January 26, 2001.
- Revision received April 4, 2001.
- Accepted April 11, 2001.
- Copyright © 2001 by American Heart Association
Makkar RR, Fromm BS, Steinman RT, et al. Female gender as a risk factor for torsade de pointes associated with cardiovascular drugs. JAMA. 1993;270:2590–2597.
Bazett H. An analysis of the time relations of electrocardiogram. Heart. 1920;7:353–370.
Ashman R, Hull E. Essentials of Electrocardiography. New York, NY: Macmillan Co; 1941:145–167.
Macfarlane PW, Lawrie TDV. The normal electrocardiogram and vectorcardiogram. In: MacFarlane PW, Lawrie TDV, eds. Comprehensive Electrocardiology: Theory and Practice in Health and Disease. New York, NY: Pergamon Press; 1989;1:407–445.
Merri M, Benhorin J, Alberti M, et al. Electrocardiographic quantitation of ventricular repolarization. Circulation. 1989;80:1301–1308.
Rautaharju PM, Zhou SH, Wong S, et al. Sex differences in the evolution of the electrocardiographic QT interval with age. Can J Cardiol. 1992;8:690–695.
Lepeschkin E, Surawicz B. The duration of the Q-U interval and its components in electrocardiograms of normal persons. Am Heart J. 1953;46:9–20.
Yang H, Elko P, Fromm BS, et al. Maximal ascending and descending slopes of the T wave in men and women. J Electrocardiol. 1997;30:267–276. (Correction. 2001;34:183)
Simonson E. Differentiation between normal and abnormal in electrocardiography. St Louis, Mo: Mosby Co; 1961:99–101.
Gambill CL, Wilkins ML, Haisty WK, et al. T wave amplitudes in normal populations. J Electrocardiol. 1995;28:191–197.
Sotobata I, Richman H, Simonson E. Sex differences in the vectorcardiogram. Circulation. 1968;37:438–448.
Nemati M, Doyle JT, McCaughan D, et al. The orthogonal electrocardiogram in normal women: implications of sex differences in diagnostic electrocardiography. Am Heart J. 1978;95:12–21.
Ishikawa K. Correlation coefficients for electrocardiographic and constitutional variables. Am Heart J. 1976;92:152–161.
Green LS, Lux RL, Haws CW, et al. Effects of age, sex, and body habitus on QRS and ST-T potential maps of 1100 normal subjects. Circulation. 1985;71:244–253.
12SL ECG Analysis Program Physician’s Guide. Milwaukee, Wis: GE Marquette Medical Systems Inc; 1999.
Edenbrandt L, Pahlm O. Vectorcardiogram synthesized from a 12-lead ECG: superiority of the inverse Dower matrix. J Electrocardiol. 1988;21:361–367.
Mosteller R. Simplified calculation of body-surface area. N Engl J Med. 1987;317:1098. Letter.
de Simone G, Daniels SR, Devereux RB, et al. Left ventricular mass and body size in normotensive children and adults: assessment of allometric relations and impact of overweight. J Am Coll Cardiol. 1992;20:1251–1260.
Bidoggia H, Maciel JP, Capalozza N, et al. Sex-dependent electrocardiographic pattern of cardiac repolarization. Am Heart J. 2000;140:430–436.
Macfarlane PW, Lawrie TDV, eds. Comprehensive Electrocardiology: Theory and Practice in Health and Disease. New York, NY: Pergamon Press; 1989;3:1459.
Surawicz B. ST-T abnormalities. In: MacFarlane PW, Lawrie TDV, eds. Comprehensive Electrocardiology: Theory and Practice in Health and Disease. New York, NY: Pergamon Press; 1989;3:511–563.
Antzelevitch C, Yan GX, Shimizu W, et al. Electrical heterogeneity, the ECG and cardiac arrhythmias. In: Zipes DP, Jalife J, eds. Cardiac Electrophysiology: From Cell to Bedside. 3rd ed. Philadelphia, Pa; WB Saunders Co; 2000:222–238.
Drici MD, Burklow TR, Haridasse V, et al. Sex hormones prolong the QT interval and downregulate potassium channel expression in the rabbit heart. Circulation. 1996;94:1471–1474.
Lehmann MH, Hardy S, Archibald D, et al. JTc prolongation with d,l-sotalol in women versus men. Am J Cardiol. 1999;83:354–359.
Liu XK, Katchman A, Drici MD, et al. Gender difference in the cycle length-dependent QT and potassium currents in rabbits. J Pharmacol Exp Ther. 1998;285:672–679.
Hara M, Danilo P Jr, Rosen MR. Effects of gonadal steroids on ventricular repolarization and on the response to E4031. J Pharmacol Exp Ther. 1998;285:1068–1072.
Bidoggia H, Maciel JP, Capalozza N, et al. Sex differences on the electrocardiographic pattern of repolarization: possible role of testosterone. Am Heart J. 2000;140:678–683.
Shimizu W, Antzelevitch C. Sodium channel block with mexiletine is effective in reducing dispersion of repolarization and preventing torsade des pointes in LQT2 and LQT3 models of the long-QT syndrome. Circulation. 1997;96:2038–2047.
Spach MS, Barr RC, Benson W, et al. Body surface low-level potentials during ventricular repolarization with analysis of the ST segment: variability in normals. Circulation. 1979;59:822–836.
Sasaki A, Takimiya A, Arai T, et al. Abnormalities of T waves in effort angina pectoris patients at rest evaluated by spatial velocity electrocardiogram. Jpn Heart J. 1996;37:879–889.