Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
International Journal of Obesity (1997) 21, 1104±1110 ß 1997 Stockton Press All rights reserved 0307±0565/97 $12.00 Effect of obesity and regional adiposity on the QTc interval in women J-J Park1 and PD Swan2 1 Department of Kinesiology, University of Maryland, MD, USA and 2 Department of Exercise Science and Physical Education, Arizona State University, Tempe, AZ, USA OBJECTIVES: To determine whether differences in body fat composition and body fat distribution patterns are associated with a prolongation of the corrected QT interval for heart rate (QTc) on the electrocardiogram (EKG) during rest and exercise. DESIGN: Cross-sectional evaluation of the QTc interval in three groups of premenopausal women during rest and two _ 2 max). _ 2 max and VO exercise conditions (50% VO SUBJECTS: Thirty-one healthy women with a mean age of 35 y (27±44 y) were classi®ed as either obese (n 22; percent body fat [%BF] > 30%) or nonobese (NO; n 9; %BF 27%) by hydrostatic weighing. Obese subjects matched for age and %BF were grouped by waist to hip ratio (WHR) into two groups: upper body obesity (UBO; n 11; WHR 0.85) and lower body obesity (LBO; n 11; WHR 0.75). MEASUREMENTS: RR and QT intervals were measured in a double-blind design with the aid of calipers and magnifying lens for seven consecutive beats in lead II from a 12-lead EKG at a paper speed of 25 mm/sec. Five consecutive cardiac cycles excluding the longest and shortest RR and QT intervals were averaged and calculated for QTc interval using Bazett's formula. RESULTS: Mean QTc intervals were signi®cantly different (P < 0.001) across the groups for each condition. For all conditions, UBO had the longest QTc interval as compared to LBO and NO respectively (i.e., Rest: 0.426; 0.413; 0.399 sec1/2; Mid50%: 0.447; 0.426; 0.409 sec1/2; Max: 0.390; 0.374; 0.357 sec1/2). CONCLUSIONS: The QTc interval is positively associated with UBO even at the same level of body fat in moderately obese women. It is clear that abdominal obesity may be one of the risk factors for a prolonged QTc interval in premenopausal women. Keywords: body fat distribution; electrocardiogram; exercise; waist±hip ratio Introduction Obesity has been implicated as a risk factor for sudden cardiac death as well as cardiovascular morbidity and mortality for several decades.1±4 In the Framingham Heart Study, obesity was found to be a strong predictor of sudden cardiac death.1,3 A primary mechanism of sudden cardiac death in obesity has been shown to be malignant ventricular arrhythmias (MVA), which may relate to electrocardiogram (EKG) abnormalities.2,5,6 Prolongation of the QT interval on the EKG, represents delayed repolarization of the ventricular myocardium and is considered a precursor of MVA and sudden cardiac death in obesity.5,7±10 Prolongation of the QT interval is the most important electrocardiographic abnormality found in obesity.11 Frank et al12 found that 28.3% of their obese outpatients had a mild prolonged QT interval and 4% of them had a markedly prolonged QT interval. A prolonged QT interval Correspondence: Pamela D Swan, Ph. D., FACSM, Department of Exercise Science and Physical Education, Mail Code 870701, Arizona State University, Tempe, AZ, 85287-0701, USA. E-mail: [email protected] Received 16 January 1997; revised 2 July 1997; accepted 21 July 1997 has also been found in obese patients on very low calorie diets (VLCD) or after surgery for obesity.6,13 From these results, however, it is impossible to distinguish whether obesity alone or other factors associated with obesity (for example, VLCD, metabolic abnormalities etc.) increases the risk for prolongation of the QT interval.7,10 Since the duration of the QT interval is altered by the length of the preceding cardiac cycle, an adjustment for differences in heart rate (QTc) has been developed.9,14 The length of the QTc interval was found to be exponentially related to the risk of developing MVA in patients with QT prolongation. In healthy obese subjects, El-Gamal et al15 reported that prolongation of the QTc interval was signi®cantly associated with relative body mass and fatness, whereas Peiris et al,16 found no signi®cant association. It has become increasingly evident that obesity is a heterogeneous disorder. Body fat location, speci®cally abdominal or visceral obesity, is a higher predictor of cardiovascular and metabolic disease risk, than the amount of body fat alone.17±19 In addition, abdominal fat deposition has been suggested as an independent risk factor for prolongation of the QTc interval.16 Peiris et al16 found that increasing intra-abdominal deposition of fat was closely associated with prolon- QTc interval and regional adiposity in women J-J Park and PD Swan gation of the QTc interval independent of obesity and other cardiovascular risk factors. The data obtained from these studies suggest that the risk for prolongation of the QTc interval in obesity, may be in¯uenced by body fat distribution pattern (that is, abdominal vs gluteal/femoral). No previous studies have examined the relationship between prolongation of the QTc interval and obesity subtypes. In addition, most studies of prolonged QTc syndromes have observed QTc intervals only during rest. Nonetheless, exercise is commonly recommended for obese individuals. The effect of progressive exercise on the QTc interval is unknown. Therefore, the purpose of this study was to determine differences in the QTc intervals between three groups of women with distinctly different body fat composition and body fat distribution patterns during rest and exercise. It was hypothesized that the obese women would have a prolongation of the QTc interval, as compared to nonobese women, and that upper body obese (UBO) women would show a larger prolongation of the QTc interval than lower body obese (LBO) women during both rest and exercise. Methods metabolic risk in women with a WHR of > 0.85 and average risk in women with a WHR of 0.75. By using these `cutoff' values, it was possible to distinguish distinctly different body fat distribution subgroups. Anthropometry Subjects were measured to the nearest 0.25 inch in height and 0.25 lb in weight. These measures were converted to metric units for BMI determination. WHR was established by standardized procedures.23 Measurement sites were ascertained with the subject assuming a standing position and then these points were marked on the subject's skin for a subsequent supine measurement. Standardized waist circumferences were determined at the narrowest point between the ribs and hips from a posterior view. If it was dif®cult to determine this point, the level of the umbilicus was measured.23 Standardized hip circumferences were determined at the widest protrusion of the buttocks from a side view of the subject. All circumferences were measured with the subject supine and wearing only undergarments. Measurements were performed twice to the nearest 0.1 cm using a Gulick II tape measure which is nonstretchable and tension controlled. WHR was calculated as waist circumference in centimeters divided by hip circumference in centimeters. Subjects Obese and non obese (NO) premenopausal women, aged 25±46 y, were recruited from the Metropolitan Phoenix Area, surrounding Arizona State University, Tempe, AZ. All women were sedentary (2 h per week of planned exercise or physical activity) with no evidence of current (within the last 6 months) participation in diet reduction programs. All subjects were asked to complete a personal health and medical history questionnaire, which served as a screening tool and provided basic descriptive data.20 Subjects were screened by questionnaire for any metabolic abnormalities or for any medication use which has a cardiac effect. They were deemed apparently healthy, as de®ned by the American College of Sports Medicine (ACSM) guidelines.21 After being screened, each subject was informed of the procedures and risks of the study, and signed an informed consent, according to procedures established by the University Human Subjects19 Research Review Board. Each screened subject was classi®ed as either obese (percent body fat [%BF] > 30%) or nonobese (%BF 27%) based on hydrostatic weighing. According to waist to hip circumference ratio (WHR), obese subjects matched for age, body mass index (BMI) and %BF were divided into an upper body obesity group (UBO; WHR 0.85; n11) and a lower body obesity group (LBO; WHO 0.75; n 11). Nonobese subjects were a control group (NO; n 9). WHR criteria were based on Bray's22 nomogram that indicates high Body composition The simpli®ed oxygen dilution technique of Wilmore et al24 was used to measure residual lung volume using a Beckman O2 analyzer (Model OM-11, Fullerton, CA) and a Beckman CO2 analyzer (Model LB2) with ¯ow control (Model R-1). Prior to testing, O2and CO2 analyzers were calibrated to register the correct percentage of calibrated gas (certi®ed to the nearest 0.01%) and ¯ow control was adjusted to 500 ml/min. Measurements of underwater weight consisted of 5±10 trials for each subject. Within each trial, 3±5 readings from the computer were obtained. Mean underwater weight was calculated from the highest values with the least variability from each trial. This underwater weight was used to determine body density following the equation of Buskirk.25 Percent body fat (% BF) was estimated using the formula of Lohman et al26 developed for women. Exercise testing Pre-exercise testing. The protocol and sequence of the tests were explained to all subjects. Each subject was prepared for standard 12-lead electrocardiogram (EKG) and rested in a darkened quiet room for 10 min. Resting heart rate, blood pressure and 12-lead EKG were taken in the supine, sitting and standing positions. 1105 QTc interval and regional adiposity in women J-J Park and PD Swan 1106 Exercise testing. Each subject was instructed to walk on the treadmill to determine a comfortable walking speed for a 3 min warm-up. According to individual walking speed, a walking protocol, with a range of walking speeds of 2.5±3.7 mph, was selected. Once the speed was selected, treadmill grade was increased by 2% every 2 min. Arterial blood pressure was measured by mercury sphygmomanometry every 2 min. Subjects were verbally encouraged to walk, until they reached their volitional exhaustion. Metabolic measures were made using indirect calorimetry. An AeroSport TEEM 100, portable indirect calorimeter (AeroSport, Ann Arbor, MI), was calibrated for volume, using a 3 liter calibrated syringe, prior to use. Gas analyzers were calibrated using certi®ed calibraÇ O2 max was tion gases prior to exercise testing. V Ç attained when VO2 did not increase (> 150 ml/min) despite increasing exercise work rate or if the respiratory exchange ratio (R) was greater than 1.12 and heart rate was within 10 beats of age predicted maximum.21 Electrocardiography A standard 12-lead EKG was measured to monitor the changes of the adjusted QT intervals for heart rate (QTc interval) both at rest and during exercise using a Quinton 4000 stress test monitor (Quinton Instrument Co., Seattle, WA). For a standard 12-lead EKG recording, electrode sites were prepared for a modi®ed 10-electrode placement.21 Skin on the chest was abraded with ®ne-grain emery paper to remove epidermis and oil, and cleaned with an alcohol gauze pad.21 A resting 12-lead EKG was recorded in a supine position at a paper speed of 25 mm/sec. The RR and QT intervals were measured from seven cardiac cycles from a recording of lead II of the resting EKG. During exercise, the EKG was continuously monitored and recorded every min. Measurements of exercise EKG parameters were made during maximal exercise (Max) and at the time period that represented 50% of the maximal aerobic capacity (Mid50%). All EKGs were coded and evaluated blindly, such that the subject group was not known to the investigator. The same well trained non-physician researcher read all the EKGs. The QT interval was measured following the method described by Browne et al27 from the earliest onset of the QRS complex to the terminal portion of the T wave, where it met the baseline. The RR interval from the preceding cardiac cycle was measured from the peaks of the R waves to correct the QT interval for heart rate (QTc). The QT and the preceding RR intervals were measured with the aid of calipers and magnifying lens for seven consecutive beats in lead II.11 To calculate QTc interval, the mean value of ®ve consecutive beats, excluding the longest and the shortest RR and QT intervals, were used following the equation from Bazett:28 QT QTc p RR Statistical analysis A 3 6 3 study design with one between-subject variable (treatment group) and one within-subject variable (condition) was used. A two-way factorial ANOVA with repeated measurement was used to analyze differences between means. Signi®cant F ratios were then followed with Newman-Keuls test. The Statistical Package for the Social Sciences (SPSS)29 was used for all statistical analysis. Results The physical characteristics of subjects are shown in Table 1. The mean age and height were similar in all three groups; UBO, LBO and NO. Body weight, BMI and %BF were matched between the two obese groups, but were signi®cantly higher (P < 0.05) than the NO group. WHR was signi®cantly higher ( < 0.05) in UBO than LBO or NO groups. Maximal aerobic Ç O2max) was signi®power relative to body weight (V cantly higher (P < 0.05) in NO than the obese groups. Mean RR intervals were not signi®cantly different across groups for any condition; Rest, Mid50% and Max. In all groups, the duration of the RR intervals were signi®cantly shortened (data not shown) across conditions paralleling the heart rate increase with progressive exercise. The group mean adjusted QT intervals for heart rate (QTc intervals) were signi®cantly different (P < 0.05) Table 1 The physical characteristics of subjects Variable Age (y) BMI (kgm72) % BF WHR _ 2max (mlkg71min71) VO UBO (n 11) LBO (n 11) NO (n 9) 34.64 6.01 30.46 5.45 38.15 5.62 0.87 0.04* 27.19 5.48 35.73 5.29 30.52 5.51 39.86 5.57 0.75 0.03 26.37 4.15 33.67 4.47 20.56 1.38* 24.59 4.87* 0.73 0.03 33.73 4.78* Values are mean s.d. *Signi®cantly different from other groups, P < 0.001. UBO upper body obesity; LBO lower body obesity; NO nonobese; BMI body mass index; % BF percentage body fat; WHR waist-hip ratio. QTc interval and regional adiposity in women J-J Park and PD Swan Table 2 The QTc intervals (sec1/2) across groups within condition Condition UBO (n 11) LBO (n 11) NO (n 9) Rest Mid50% Max 0.426 0.009* 0.447 0.014* 0.390 0.014* 0.413 0.115* 0.426 0.123* 0.374 0.018* 0.399 0.020* 0.409 0.023* 0.357 0.019* Values are mean s.d. *Signi®cantly different from other groups, P < 0.05 (UBO > LBO > NO). UBO upper body obesity; LBO lower body obesity; _ 2max. _ 2max; Max VO NO nonobese; Mid50% 50% VO for each condition (Table 2). Within each condition, UBO had the longest QTc interval, NO had the shortest and LBO had intermediate values. Figure 1 shows the changes in QTc intervals during rest and exercise conditions. As seen from Figure 1, in the obese groups, mean QTc intervals were signi®cantly (P < 0.05) longer during Mid50%, but shorter during Max. However, in the NO group, the change in the QTc interval between Rest and Mid50% did not achieve signi®cance. A Pearson's correlation matrix29 to examine the relationship between QTc interval and anthropometric conditions is summarized in Table 3. There was a signi®cant correlation (P < 0.001) between WHR and QTc interval at all conditions. Body weight, BMI and %BF were also signi®cantly correlated with QTc interval at all conditions. A partial correlation between WHR and QTc after adjusting for %BF also indicated a signi®cant relationship between WHR and QTc for all conditions (Table 4). Table 3 Pearson's Correlation of QTc interval and selected body composition parameters Variable Rest QTc Mid50% QTc Max QTc Weight (kg) BMI (kgm72) % BF WHR 0.4757* 0.4717* 0.4975* 0.5248** 0.4168* 0.4367* 0.3789* 0.6111** 0.5890** 0.6335** 0.5551** 0.6101** Values are correlation coef®cients. *Signi®cant correlation, P < 0.05; ** Signi®cant correlation, P < 0.001. _ 2max; BMI body mass _ 2max; Max VO Mid50% 50% VO index; % BF percentage body fat; WHR waist-hip ratio. Table 4 Partial Correlation of QTc interval and WHR after adjusting for % BF Variable Rest QTc Mid50% QTc Max QTc WHR 0.4154 P 0.02 0.5441 P 0.002 0.5158 P 0.004 Values are correlation coef®cients. WHR waist-hip ratio; % BF percentage _ 2max. _ 2max; Max VO Mid50% 50% VO body fat; Discussion As anticipated, both obese groups (UBO and LBO) had signi®cantly longer QTc intervals than NO at rest. However, the duration of the QTc intervals of all subjects were normal considering that the generally accepted upper limit of normal for QTc interval is 0.44 sec1=2 .11,29 There were signi®cant relationships between the QTc interval, body weight, %BF and Figure 1 Changes of the QTc interval across condition. *Signi®cantly different, P < 0.05, Rest vs Mid50% and Max. **Signi®cantly different, P < 0.05, Mid50% vs Max. ***Signi®cantly different, P < 0.05, Max vs Rest and Mid50%. UBO Upper body obesity. _ 2max _ 2max. Max VO LBO Lower body obesity. NO Nonobese. Mid50% 50% VO 1107 QTc interval and regional adiposity in women J-J Park and PD Swan 1108 BMI. These ®ndings con®rm previous studies of the QTc interval in obesity.12,16 Frank et al12 found a low but statistically signi®cant relationship (r 0.22, P < 0.001) between percent overweight and QTc interval in obese patients without cardioactive medications. In addition, El-Gamal et al15 recently found signi®cant correlations between QTc intervals and both BMI (r 0.31, P 0.0002) and %BF (r 0.25, P 0.003) in healthy obese people. In contrast to these studies, Peiris et al16 reported no association between obesity and QTc interval in healthy premenopausal women. However, the subjects in Peiris's study were all signi®cantly obese and very similar in %BF (Mean s.d. 42.0 0.9%). Perhaps the homogeneity in body fat between subjects may have masked any correlations with QTc in this population. Results from the current study, clearly indicate that excess fatness, in healthy premenopausal women, is associated with a longer QTc interval at rest, compared to lean women. Body fat distribution patterning also had a signi®cant effect on the length of the QTc intervals at rest. The QTc interval was signi®cantly greater in UBO than LBO even though age, body weight, BMI and %BF were matched between groups. Signi®cant correlations (r 0.52; P < 0.001) were observed between the QTc interval and the WHR. In addition, the partial correlation between WHR and QTc after adjusting for %BF was also signi®cant at all conditions. These ®ndings support those of Peiris et al16 who found a signi®cant linear association (r 0.54, P < 0.05) between the QTc interval and intra-abdominal deposition of fat as measured by computerized tomography (CT) in healthy premenopausal obese women. These data indicate that increasing abdominal adiposity is accompanied by prolonged QTc interval in obese people even after adjusting for fatness. Thus, although obesity is associated with a prolonged QTc interval, UBO may be a higher risk for potential EKG abnormalities. During exercise, the associations between the QTc interval and body fat distribution patterns were maintained. For each condition, the obese groups (UBO and LBO) had longer QTc intervals than NO and the UBO had longer values than LBO. The results indicated no interaction between exercise condition and group on the QTc interval. However, there were signi®cant effects of exercise condition on the QTc interval. For all groups, the RR and QT intervals progressively shortened during exercise in response to the increase in heart rate that accompanies increased physical work. According to the ACSM21 description of normal electrocardiographic response to exercise, these intervals should normally shorten with an increase in exercise intensity. However, the intervals did not shorten proportionally. At Mid50%, the proportional decrease in RR interval was greater than that for the QT interval. Thus, the QTc interval which is calculated from the ratio of QT to RR, actually increased slightly compared to resting values. The reason for this surprising increase in the QTc interval may be due to an uncoupling between heart rate and the oxygen requirements at this intensity. Work intensity at Mid50% was based on 50% of Ç O2max) and not 50% of maximal aerobic power (V maximal heart rate. Thus, the heart rate response at this work rate represented approximately 75% of the maximal heart rate. Any uncoupling or unproportional shortening of the heart rate in comparison to the repolarization wave (i.e., QT interval) would alter the QTc interval. At Mid exercise, the greater proportional change in heart rate as compared to the QT interval shortening, resulted in a slight increase in QTc interval. This ®nding suggests that the QTc interval is responsive to the absolute heart rate at any submaximal intensity of work. The precise mechanisms of the prolonged QTc interval in obesity are unknown. However, obesity and UBO have been documented to be independent risk factors for coronary heart disease (CHD), which may induce prolongation of the QTc interval.7 Possible mechanisms of the prolonged QTc interval in obesity include cardiac autonomic nervous system dysfunction, electrolyte abnormalities, left ventricular hypertrophy (LVH), hyperinsulinemia and CHD.5,7,12,30±33 Obesity and UBO have been associated with autonomic nervous system dysfunction34,35 which may result in prolongation of the QTc interval.5,7,27,30,31 An imbalance in cardiac sympathetic enervation has been shown to prolong the QTc interval in men and women.30 Gao et al,34 found that women with UBO had signi®cantly higher cardiac sympathetic and parasympathetic activity than those with LBO. They compared autonomic function by power spectral analysis in women with subcutaneous vs visceral abdominal body fat as assessed by CT. Signi®cantly higher autonomic nerve frequencies were reported for the women with predominantly visceral body fat. Browne et al,27 who investigated prolonged QTc intervals in patients with ventricular arrhythmia during sleep, demonstrated that increased parasympathetic activity may result in prolongation of the QTc interval. Kahn et al,31 reported that cardiac autonomic neuropathy may result in sympathetic imbalance and QTc interval prolongation. Furthermore, El-Gamal et al15 suggested that an elevated cardiac output in obesity, due to an increased stroke volume and overall greater body mass, may contribute to alter autonomic nervous system balance. It is possible that some unknown dysfunction in autonomic nervous system balance could have in¯uenced the ®ndings of increased QTc interval in this study. Another possible mechanism contributing to prolongation of the QTc interval in obesity and UBO may be electrolyte abnormalities. High levels of cytosolic free calcium and low levels of free magnesium have been found in even healthy obese individuals.36 Phinney et al,32 found that chronic oral magnesium supplementation signi®cantly shortened the QTc intervals QTc interval and regional adiposity in women J-J Park and PD Swan in obese patients with dietary-induced (VLCD) prolonged QTc interval. In the current study, electrolyte balance was not measured. However, the dietary habits of the women were not signi®cantly different between the groups. It is possible, however, that the UBO group had electrolyte abnormalities that were not accounted for in this study. Electrolyte abnormalities are also known to cause increased heart muscle contractility, increased vasoconstriction and may predispose one to LVH.33 In nonhypertensive obese individuals, LVH of the eccentric type, which is an increase in myocardial mass combined with chamber dilation, is considered a primary risk factor for ventricular arrhythmias.2,12 In their investigation of sudden death in obesity, Messerli et al2 suggested that the hypertrophied myocardium in obesity may induce ventricular arrhythmias. Indeed, it is recommended that the QTc interval must be monitored if LVH is found.12 In the current study, no women had electrocardiographic evidence of LVH. However, since echocardiographic measures of LVH were not made, it is possible that the UBO group had some evidence of LVH. Results from this study suggest that UBO may be a higher risk factor for prolonged QTc interval than obesity alone. In addition, it is well known that UBO is associated with the group of risk factors called the metabolic syndrome.37 The increased portal free fatty acid (FFA), hyperinsulinemia and insulin resistance in UBO are closely aligned with CHD.38±40 Abraham et al 41 reported that increased plasma FFA levels in patients with acute myocardial infarction, had signi®cant effects on ventricular arrhythmias. Peiris et al16 also suggested that hyperinsulinemia and insulin resistance of abdominal obesity may induce cardiovascular disease and may play a role in prolonged QTc interval. Recently, Dekker et al 42 found that the QTc interval was signi®cantly longer in elderly men with glucose intolerance and newly diagnosed diabetes, compared to men with normal carbohydrate metabolism. The authors suggest that the observed association might be evidence that insulin-induced sympathetic activity, may precipitate myocardial repolarization abnormalities and/or that disturbed glucose metabolism may alter ion exchange at the myocardium, which may also manifest in QTc prolongation. Thus, any factor associated with the development of CHD such as diabetes or hyperinsulinemia would obviously effect cardiac function. The increased QTc intervals noted in this study may be evidence of an initial development of some undiagnosed factor associated with the metabolic syndrome. Further research into the association of UBO with diabetes, hyperinsulinemia and CHD is warranted. Exercise induced a shortening of the QTc interval in all the healthy women evaluated in this study. However, group differences in the duration of the QTc interval persisted throughout exercise condition. Despite the common practice of recommending exer- cise for overweight individuals, research about EKG abnormalities, during exercise in obesity, is rare. Sudden death in an obese patient with a prolonged QTc interval during exercise has been reported.31 Thus, it seems important to carefully monitor the QTc interval especially in UBO individuals during exercise. It remains to be shown, however, what value of the QTc interval during exercise would be deemed critical. In addition, it is unknown why the QTc interval is prolonged in UBO individuals and what effects exercise training would have on this EKG index. In conclusion, this study demonstrated that prolongation of the QTc interval is associated with increased body fatness and increased abdominal fat distribution even at the same level of body fatness. It is clear that abdominal obesity may be one of the risk factors for prolongation of the QTc interval in otherwise healthy premenopausal women. Acknowledgements The authors wish to thank Ms Mary Boynton, MS for her help and collaboration in all aspects of this project. Also we would like to acknowledge the Arizona State University Women's Studies program for partial funding of this project. References 1 Hubert HB, Feinleib M, McNamara PM, Castelli WP. Obesity as an independent risk factor for cardiovascular disease: A 26year follow-up of participants in the Framingham Heart Study. Circulation 1983; 67: 968±977. 2 Messerli FH, Nunez BD, Ventura HO, Snyder DW. Overweight and sudden death: Increased ventricular ectopy in cardiopathy of obesity. Arch Intern Med 1987; 147: 1725± 1728. 3 Kannel WB, Plehn JF, Cupples LA. Cardiac failure and sudden death in the Framingham Study. Am Heart J 1988; 115: 869±875. 4 Berenson GS, Srinivasan SR, Hunter SM, Nicklas TA. Risk factors in early life as predictors of adult heart disease: The Bogalusa Heart Study. Am J Med Sci 1989; 298: 141±151. 5 Vlay SC, Mallis GI, Brown EJ, Cohn PF. Documented sudden cardiac death in prolonged QT syndrome. Arch Intern Med 1984; 144: 833±835. 6 Drenick EJ, Fisler JS. Sudden cardiac arrest in morbidly obese surgical patients unexplained after autopsy. Am J Surg 1988; 155: 720±726. 7 Moss AJ. Prolonged QT interval syndrome. JAMA 1986; 256: 2985±2987. 8 Akhtar M. Clinical spectrum of ventricular tachycardia. Circulatioin 1990; 82: 1561±1573. 9 Sagie A, Larson MG, Goldberg RJ, Bengtson JR, Levy D. An improved method for adjusting the QT interval for heart rate (The Framingham Heart Study). Am J Cardiol 1992; 70: 797± 801. 10 Moss AJ. Measurement of the QT interval and the risk associated with QTc interval prolongation: A review. Am J Cardiol 1993; 72: 23B±25B. 11 Blumberg VS, Alexander J. Obesity and the heart. In: BjoÈrntrop P, Brodoff BN (eds). Obesity. Lippincott: Philadelphia, 1992, pp 517±531. 12 Frank S, Colliver JA, Frank A. The electrocardiogram in obesity: Statistical analysis of 1029 patients. J Am Coll Cardiol 1986; 7: 295±299. 1109 QTc interval and regional adiposity in women J-J Park and PD Swan 1110 13 Van Itallie TB, Yang MU. Cardiac dysfunction in obese dieters: A potentially lethal complication of rapid, massive weight loss. Am J Clin Nutr 1984; 39: 695±702. 14 Funck-Brentano C, Jaillon P. Rate-corrected QT interval: Techniques and limitations. Am J Cardiol 1993; 72: 17B±22B. 15 El-Gamal A, Gallagher D, Nawras A, Gandhi P, Gomez J, Allison DB, Steinberg JS, Shumacher D, Blank R, Heyms®eld SB. Effects of obesity on QT, RR, and QTc intervals. Am J Cardiol 1995; 75: 956±959. 16 Peiris AN, Thakur RK, Sothmann MS, Gustafson AB, Hennes MI, Wilson CR, Kissebah AH. Relationship of regional fat distribution and obesity to electrocardiographic parameters in healthy premenopausal women. South Med J 1991; 84: 961± 965. 17 BjoÈrntorp P. Classi®cation of obese patients and complications related to the distribution of surplus fat. Am J Clin Nutr 1987; 45: 1120±1125. 18 Haffner SM, Fong D, Hazuda HP, Pugh JA, Patterson JK. Hyperinsulinemia, upper body adiposity and cardiovascular risk factors in non-diabetics. Metabolism 1988; 37: 333±345. 19 Peiris AN, Sothmann MS, Hoffmann RG, Hennes MI, Wilson CR, Gustafson AB and Kissebah, AH. Adiposity, fat distribution, and cardiovascular risk. Ann Intern Med 1989; 110: 867± 872. 20 Allison D, Heshka S. Toward an empirically derived typology of obese persons: Derivation in a nonclinical sample. Int J Eating Disorders 1993; 13: 93±108. 21 American College of Sports Medicine. ACSM's Guidelines for Exercise Training and Prescription. Williams and Wilkins: Baltimore, 1995. 22 Bray G. Obesity: Basic considerations and clinical approaches. Disease-a-Month 1989; 35: 460±540. 23 Lohman TG, Roche AF, Martorell R. Anthropometric standardization reference manual. Human Kinetics: Champaign, 1988. 24 Wilmore JH, Vodak PA, Parr RB. Further simpli®cations of a method for determination of residual lung volume. Med Sci Sports Exerc 1980; 12: 216±218. 25 Buskirk E. Underwater weighing and body density: A review of procedures. In: Brozek J, Henschel A (eds). Techniques for measuring body composition. National Academy of Sciences: Washington DC, 1961, pp 90±106. 26 Heyward VH, Stolarczyk LM. Applied Body Composition Assessment. Human Kinetics: Champaign, 1996. 27 Browne KF, Prystowsky E, Heger JJ, Chilson DA, Zipes DP. Prolongation of the QT interval in man during sleep. Am J Cardiol 1983; 52: 55±59. 28 Bazett HC. An analysis of the time relations of electrocardiograms. Heart 1920; 7: 353±367. 29 SPSS Inc. SPSS for Windows Release 6.1. Chicago 1995. 30 Schwartz PJ, Wolf S. QT interval prolongation as predictor of sudden death in patients with myocardial infarction. Circulation 1978; 57: 1074±1077. 31 Kahn JK, Sisson JC, Vinik AI. QT interval prolongation and sudden cardiac death in diabetic autonomic neuropathy. J Clin Endocrinol Metab 1987; 64: 751±754. 32 Phinney SD, Newton JM, Brown RL, Tsuboi CN, Maybury K, Roesler J, Burnette-Ellis GJ. Chronic magnesium supplementation reverses the prolonged QT EKG interval associated with very low calorie dieting [Abstract]. Obes Res 1996; 4(Suppl. 1): 54S. 33 Resnick LM. Cellular ions in hypertension, insulin resistance, obesity, and diabetes: a unifying theme. J Am Soc Nephrol 1992; 3: S78±S85. 34 Gao YY, Lovejoy JC, Sparti A, Bray GA, Keys LK, Partiongton C. Autonomic activity assessed by heart rate spectral analysis varies with fat distribution in obese women. Obes Res 1996; 4: 55±63. 35 Valensi P, Phan-Thi BN, Lormeau B, Paries J, Attali JR. Cardiac autonomic function in obese patients. Int J Obes 1995; 19: 113±118. 36 Resnick L. Calcium metabolism in hypertension and allied metabolic disorders. Diabetes Care 1991; 14: 505±520. 37 Bouchard C, DespreÂs J, Mauriege P. Genetic and nongenetic determinants of regional fat distribution. Endocr Rev 1993; 14: 72±93. 38 BjoÈrntorp P. Hypertension and other complications in human obesity. J Clin Hypertens 1986; 2: 163±165. 39 Larsson B, SvaÈrdsudd K, Welin L. Abdominal adipose tissue distribution, obesity, and risk of cardiovascular disease and death: 13-year follow-up of participants in the study of men born in 1913. BMJ 1984; 288: 1401±1404. 40 Lapidus L, Bengtsson C, Larsson B. Distribution of adipose tissue and risk of cardiovascular disease and death: A 12-year follow-up of participants in the population study of women in Gothenburg, Sweden. BMJ 1984; 289: 1257±1261. 41 Abraham R, Riemersma RA, Wood D, Elton R, Oliver MF. Adipose fatty acid composition and the risk of serious ventricular arrhythmias in acute myocardial infarction. Am J Cardiol 1989; 63: 269±272. 42 Dekker JM, Feskens EJM, Schouten EG, Klootwijk P, Pool J, Kromhout D. QTc duration is associated with levels of insulin and glucose tolerance. Diabetes 1996; 45: 376±380.