Prediction of Stroke Volume During Upper and Lower Body Exercise in Men and Women

Transcription

1 713 Prediction of Stroke Volume During Upper and Lower Body Exercise in Men and Women Yagesh N. Bhambhani, PhD ABSTRACT. Bhambhani YN. Prediction of stroke volume during upper and lower body exercise in men and women. Arch Phys Med Rehabil 1995;76: In this study, regression equations were derived to predict cardiac stroke volume (SV, ml/beat), measured by carbon dioxide rebreathing, from oxygen pulse (Oz pulse, ml/beat) measurements in healthy men (n = 25) and women (n = 12) during upper and lower body exercise at the ventilatory threshold. The equations for upper body exercise were as follows: men, Y = 10.21X - 1.0, SE = 13.0; r = 0.85; women, Y = 12.70X - 4.8, SE = 15.4, r = The equations for lower body exercise were as follows: men, Y = 5.22X , SE = 17.9, r = 0.76; women, Y = 7.41X , SE = 13.5, r = No significant differences (p >.05) were observed between the exercise modes for the slopes and intercepts of the regression equations in both genders. The relationships between 02 pulse and SV were indirectly validated by using raw data from previous studies that had used the direct Fick method to determine cardiac output (Q) for each exercise mode and gender. Comparisons between the values of SV reported in several studies that used direct and indirect methods to determine Q and those predicted from the current equations indicated an error that was usually within 10% of the reported values. These observations suggest that the equations derived for predicting SV from Oz pulse measurements at the ventilatory threshold are quite accurate and can be generalized to a variety of techniques currently used to determine Q by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation The acute cardiovascular and metabolic responses during dynamic upper and lower body exercise have been well documented.~'2 It is generally accepted that during incremental exercise, cardiac output (Q), which is determined by the product of heart rate (HR) and stroke volume (SV), increases linearly until the peak oxygen uptake (peak ~ro2) is attained. While HR increases linearly with oxygen uptake (~'O2) during dynamic exercise, the SV tends to reach a plateau at approximately 40% to 50% of the peak "VO2. Hence, the linear increase in Q observed at higher exercise intensifies appears to be caused solely by an increase in HR. 2 Although the trend in these variables is quite similar during these two exercise modes, it is well established that the magnitude of the peak responses for each of these variables is significantly lower during upper body compared with lower body exercise. 1.3 Moreover, the slope of the relationship between the absolute VO2 and HR during these two modes of incremental exercise is significantly lower during upper body exercise. 4'5 As well, studies that have examined gender differences for both these exercise modes, have demonstrated that the peak value for each of these variables is significantly lower in women than in men. 6'7 Routine assessment of cardiorespiratory fitness in the laboratory usually includes measurement of the peak VO2 and From the Department of Occupational Therapy, Faculty of Rehabilitation Medicine, University of Alberta, Edmonton, Canada. Funded in part by a grant from the Small Faculties Committee, University of Alberta, Edmonton. Submitted for publication July 19, Accepted in revised form January 19, No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organization with which the authors are associated. Reprint requests to Dr. Y. Bhambhani, Room 3-73, Corbett Hall, Faculty of Rehabilitation Medicine, University of Alberta, Edmonton, Canada, T6G 2G by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation /95/ /0 peak HR. The measurement of Q and SV are usually not undertaken because of the specialized equipment necessary for their indirect assessment from carbon dioxide (CO2) rebreathing, 8 electrical impedance, 9 or radionuclide angiography.l As well, unfamiliar maneuvers such as hyperventilation or breath holding for short intervals during exercise that are necessary to undertake these measurements during the former two procedures may not be feasible in certain patient populations. It is important therefore that a simple alternative method be available to evaluate SV during exercise so that the cardiovascular responses can be documented more comprehensively. Recently Bhambhani and colleagues ~ established regression equations for predicting SV from oxygen pulse (O2 pulse, ml/beat; defined as the ratio between absolute ~ro2 and HR) measurements in untrained and trained men during lower body exercise and recommended their use for estimating SV when actual measurements from the indirect techniques was not practical. Wheelchair-dependent individuals rely primarily on their upper extremities for ambulation. As well, many occupational and recreational activities are performed with the upper extremities. ~ Because the relationship between the absolute ~ro2 and HR differs significantly between upper and lower body exercise, 4'5 it is likely that the regression equation established by Bhambhani and colleagues ~I to predict SV from O2 pulse measurements during lower body exercise may not be applicable to upper body exercise. As well, the fact that the cardiovascular and metabolic responses during both these exercise modes differ significantly between men and women 6'7 suggests that gender-specific regression equations need to be established for this prediction. Hence, the purpose of this study was to develop and compare regression equations for predicting SV from 02 pulse measurements during submaximal upper (arm cranking) and lower (leg cycling) body exercise in healthy men and women.

2 714 STROKE VOLUME PREDICTION DURING EXERCISE, Bhambhani METHODS Subjects Written informed consent was obtained from 25 male and 12 female volunteers for this study. All the subjects were recreationally active and were currently not involved in regular exercise training programs of the upper and lower body. The testing procedures undertaken were approved by a duly constituted human ethics review committee at this institution. Each subject completed four testing sessions over a 2- week period, the details of which are presented next. Determination of Peak Oxygen Uptake and Ventilatory Threshold In the first two sessions, the subjects completed an incremental arm cranking or leg. cycling test in random order so as to determine their peak VO2 for each mode, according to the protocol previously described by Bhambhani and coworkers. 3 Both tests were performed on mechanically braked ergometers a'b and initiated at zero workload for 2 minutes. During arm cranking, the power output was increased by 12.5W (0.5kg at 50rpm) every 2 minutes, whereas during leg cycling, the power output was increased by 30W (0.5kg at 60rpm) every 2 minutes. The tests were continued until volitional fatigue, ie, until the subjects were unable to maintain the required cadence specified for each exercise mode. Test-retest reliability of the peak VO2 during upper body exercise in novice subjects is ~2 During the tests, respiratory gas exchange measurements were continuously recorded every 30 seconds using an automated metabolic measurement cart. c This instrument was calibrated before each test with 16% oxygen and 4% carbon dioxide (balance nitrogen) according to manufacturer's specifications, and the calibration was verified for its accuracy on completion of the test. The HR (CM5 lead) was monitored by interfacing an electrocardiogram d with the metabolic cart using an analog input board. For each subject, the respiratory gas exchange data were plotted against power output. These plots were examined by an experienced evaluator to identify the ventilatory threshold using the criteria established by Davis and colleagues~3: s.ystematic increase in the ventilatory equivalent for oxygen (VE/ VO2 ratio), without a concomitant increase in the ventilatory equivalent for carbon dioxide (VE/~/CO2 ratio). The power output at which the ventilatory threshold occurred was used in the subsequent submaximal tests to evaluate Q and SV. This intensity was selected as a reference point because it has been established j4 that SV reaches a plateau at the ventilatory threshold and is not significantly altered at higher intensities. Measurement of Cardiac Output In the next two sessions, the subjects completed a 10- minute arm cranking or leg cycling test in random order at a power output, which elicited the ventilatory threshold. The test was initiated with a 2-minute zero load period, and thereafter, the desired resistance was placed on the ergometer. The Collier ~5 technique was used to determine Q during the 7th and 10th minutes of the tests. The Advanced Exercise Testing Program e available with the metabolic cart was used for these measurements. This software incorporates the following assumptions outlined in Jones16: (1) an arterial ph of 7.4; (2) arterial and venous saturation levels of 95% and 100%, respectively during rebreathing; and (3) conversion of the arterial and venous CO2 pressures into concentrations using the equation: loge CCO2 = [0.396 loge PCO2] In this study, Hb concentrations were not measured so as to make the test noninvasive. Because all the subjects were considered normal, values of 15.8g/100/mL and 13.9g/100/mL of blood were assumed for men and women, respectively 2 and entered into the above equation to obtain estimates of the CO2 concentrations. After verifying that the subject was in a steady state (ie, no significant change in the cardiorespiratory responses during the previous 1 minute of exercise), the rebreathing maneuver to determine Q was initiated. The subject hyperventilated from a 5L anaesthesia bag containing 11% to 14% CO2 (balance oxygen) until an equilibrium was attained between the gas in the bag.and lungs. The gas concentration selected was based on the VO2 and end-tidal CO2 criteria available in Jones.~6 The end-tidal CO2 tension was considered to be reflective of arterial PCO2, whereas the bag CO2 was assumed to be indicative of venous PCO2. A "downstream" correction factor was applied to increase the validity of the latter assumption. 16 The criterion used by the computer program for identifying PCO2 equilibrium between the bag and lungs was a change of less than lmmhg pressure over a 5- second interval. In cases where the computer was unable to detect this equilibrium point during the rebreathing maneuver, the value was extrapolated from the line joining the points for expired PCO2 at 6 to 10 seconds of rebreathing to that at 20 seconds. This value is reported to be within 2mmHg of the equilibrium value. ~6 Test-retest reliability coefficients of this procedure from this laboratory (unpublished observations) during upper and lower body exercise at the ventilatory threshold were 0.87 and 0.88, respectively. From the values of Q obtained, SV was calculated as the ratio between Q and HR, whereas (a-v)o2 difference was computed from the Fick equation as the ratio between VO2 and Q. Because the reliability of the SV measurements was 0.94 (p <.01), and no significant difference (p >.05) was observed between the mean of the two trials, the average value of the two maneuvers performed during the test was used for analysis. The values of Q and SV obtained were divided by the subject's body surface area (BSA) to obtain the cardiac index (CI) and stroke volume index (SVI), respectively. ~7 The 02 pulse value obtained during the 30 seconds before the initiation of each rebreathing maneuver was averaged and used for examining the interrelationships among these variables. Statistical Analysis Subject characteristics were compared using a Student t test for independent samples. Analysis of variance with repeated measures was used to examine differences in the mean values between exercise modes and gender for pertinent physiological variables. Significant F ratios were analyzed on a post hoc basis using the Scheffe multiple comparison procedure. Pearson's product-moment correlations were used to examine the relationships among the physiological

3 STROKE VOLUME PREDICTION DURING EXERCISE, Bhambhani 715 Table 1: Peak Physiological Responses During Upper and Lower Body Exercise in Men and Women (means _+ sd) Variable Gender Upper Lower Oxygen uptake (L/min) M* 2.49 _+.51 F 1.48 _.25 Oxygen uptake (ml/kg/min) M* _ 6.6 F 25.1 _+ 4.9 Heart rate (beats/min) M 170 _ F 167 _ 16.2 Oxygen pulse (ml/beat) M* 14.7 _ 3.0 F Ventilation volume (L/min) M* F Ventilatory equiv, for oxygen M 46.4 _+ 8.9 F 47.2 _ 8.4 Respiratory exchange ratio M 1.20 _.09 F _ ' 2.16 _+.3U 43.9 _+ 6.3* ' 183 +_ 10.2* 183 _+ 12.0' 19.8 _+ 3.3* ' _ 25.4* ' 43.1 _ _ _ _+.07 * Indicates significant gender difference for that variable during upper and lower body exercise. * Indicates significant difference between lower and upper body exercise for that gender. variables of interest. Linear regression analysis was used to derive the equations for predicting SV from O2 pulse measurements. Multiple analysis of variance was used to compare the slopes and intercepts of the regression equations. Results were considered to be significant at the.05 level of confidence. All statistical analyses were performed using SPSS f programs. RESULTS Characteristics of Subjects The mean _+ SD for the age, height, body mass, and BSA of the men was years, _ 6.6cm, kg, and m 2, respectively. The corresponding values for the women were years, cm, kg, and 1.63 _+.10m 2, respectively. With the exception of age, the mean values for males were significantly greater (p <.05) than that of women for the remaining variables. Peak Physiological Responses The peak values of the physiological responses monitored during upper and lower body exercise in men and women are summarized in table 1. The VO2 (absolute and relative), HR, O2 pulse, and VE were significantly higher during lower compared with upper body exercise in both genders, whereas no significant differences were observed between exercise modes for the VE/VO2 ratio and respiratory exchange ratio (RER.). Gender comparisons for these variables indicated that VO2 (absolute and relative), 02 pulse, and VE were significantly higher in men compared with women during both the exercise modes, whereas HR, VE/VO2 ratio, and PER were not significantly different. Responses at the Ventilatory Threshold The physiological responses monitored at the ventilatory threshold are presented in table 2. In both genders, ~/O2 (absolute, relative, and % peak), HR, 02 pulse, Q, CI, SV, SVI, and (a-v)o2 differences were significantly higher during Table 2: Physiological Responses During Upper and Lower Body Exercise at the Ventilatory Threshold in Men and Women (means sd) Variable* Gender Upper Lower Oxygen uptake (L/min) M* 1.40 _ _.49 F _ _+.31 Oxygen uptake (ml/kg/min) M* F 14.4 _ _+ 6.0 Oxygen uptake (% peak) M 57.6 _ F Heart rate (beats/min) M 130 _ _+ 16.3' F 131 _ _ 15.6 Oxygen pulse (ml/beat) M* _ F Cardiac output (L/rain) M* 14.2 _ F Cardiac index (L/m 2) M 7.1 _ _ 1.4 F Stroke volume (ml/beat) M* F Stroke volume index (ml/m ~-) M 55.9 _ _ 13.6 F (a-v)o2 diff (ml/100ml) M* 10.0 _ _ F * Indicates significant difference between upper and lower body exercise for all the variables in both genders. * Indicates significant gender difference for that variable during upper and lower body exercise. * Indicates significant gender difference for that variable during that exercise mode. lower compared with upper body exercise. In men, the Oz pulse and SV were higher by 32% and 19%, respectively, whereas in women the values were higher by 33% and 22%, respectively. Gender comparisons showed that with the exception of HR, CI, and SVI during the two exercise modes, the remaining variables were significantly higher in men compared with women. The HR was significantly higher in women compared with men during lower but not upper body exercise. During upper body exercise, the 02 pulse and SV were higher in men by 40% and 29%, respectively, whereas during lower body exercise, they were higher by 39% and 27%, respectively. Prediction of Stroke Volume The relationships between 02 pulse and SV, SVI, and (av)o2 differences at the ventilatory threshold are reported in table 3. Significant correlations were observed between 02 pulse and SV as well as SVI, but not with (a-v)o2 difference during each mode in both genders. The regression equations Table 3: Relationship Between the Oxygen Pulse and Stroke Volume, Stroke Volume Index, and Arteriovenous Oxygen Difference During Upper and Lower Body Exercise in Men and Women* Mode Gender SV SVI (a-v)o2 diff Male Upper Female Lower Male Female * All the correlation coefficients were significant, except those observed between the 02 pulse and (a-v)o2 difference. Arch Phys Med Rehabi! Vol 76, August 1995

4 716 STROKE VOLUME PREDICTION DURING EXERCISE, Bhambhani A: Males o9..../ 250 The scatterplots and 95% confidence intervals for these equations are illustrated in figures A and B for men and women, respectively. It is evident that during both modes of exercise 200 in each gender, majority of the data points were close to or within the 95% confidence interval. There were no significant differences between exercise modes for the slopes and intercepts of these regression equations in either gender. 150 DISCUSSION o >. o [.o ~ 120 0"",. " 0 o 0 = lower body, r = 0.76, p<.o1 = upper body, r = 0.85, p<.o1 I r I I I Oxygen Pulse, mk/beat B: Females..../"...'" /"..,-Q''"".... "5 100 Peak Oxygen Uptake and Ventilatory Threshold Sawka I in an extensive review of literature pertainin~ to upper body exercise, reported that the ratio for the peak VO2 50 between upper and lower body exercise ranged from 36% to 89% in healthy untrained subjects, with a mean of 73%. In the current study, the mean value obtained for this variable was very similar to that reported by Sawka ~ in both genders. 0 The ventilatory threshold occurred at approximately 58% and 68% of peak ~zo2 during upper and lower body exercise, respectively, in both genders, which is consistent with previous reports for untrained subjects. 3'~3J8 The fact that there 160 was a significant difference between upper and lower body exercise for this variable also concurs with the findings of Davis and colleagues. TM In general, the subjects evaluated in this study could be considered to be representative of the average population with regard to the ratio between upper 120 and lower body cardiorespiratory fitness. d ~ 8o o > o 40 co e~/./" o-o./ / / 0 = lower body, r = 0.78, p<.01 = upper body, r = 0.78, p<.o1 0 I I i Oxygen Pulse, ml4beat Scatterplots and regression coefficients for predicting SV from oxygen pulse measurements during upper and lower body exercise in men (n = 25) and women (n = 12). Dotted lines indicate 95 % confidence intervals for each exercise mode. for predicting SV (Y in the equation) from 02 pulse (X in the equation) during upper and lower body exercise in men were as follows: Y = 10.21X - 1.0, SE = 13.0, r = 0.85 (upper) Y = 5.22X , SE = 17.9, r = 0.76 (lower) The corresponding equations for women were as follows: Y = 12.70X - 4.8, SE = 15.4, r = 0.78 (upper) Y = 7.41X , SE = 13.5, r = 0.78 (lower) Validity of the Predictive Equations In an earlier study, Bhambhani and colleagues ~ developed regression equations to predict SV from 02 pulse measurements in untrained and trained men during cycle exercise. The validity of those equations was indirectly established by analyzing raw data from previous studies 17J9 that used the direct Fick method to determine Q (from which SV was calculated). In the current investigation, a similar approach was used to examine the validity of the equations derived for predicting SV from 02 pulse measurements during upper and lower body exercise in men and women. The raw data from three studies 19-2~ were subjected to linear regression analysis, and the results are presented in table 4. It is evident that with the exception of the results of females derived from Stenberg and colleagues, 2~ moderate to high relationships were observed between the 02 pulse and SV during the two exercise modes in both genders. The low association between these two variables in women obtained from the data of Stenberg and coworkers 21 was most likely caused by the small number of subjects (n = 4) who appeared to be fairly homogeneous in their responses during the two exercise modes. However, the female data of Astrand and colleagues j9 obtained during lower body exercise on 11 subjects support the observations that the 02 pulse is significantly associated with SV in this gender. In none of the studies cited earlier was there a significant relationship between the 02 pulse and mixed (a-v)o2 difference, although mathematically, these two variables are directly related to each other { 02 pulse = SV (a-v)o2 difference}. This observation concurs with the earlier report jl in untrained and trained subjects during lower body exercise. Overall, the results of studies that have

5 STROKE VOLUME PREDICTION DURING EXERCISE, Bhambhani 717 Table 4: Validity of the Equations for Predicting Stroke Volume From Oxygen Pulse Measurements During Submaximal Upper and Lower Body Exercise 02 Gender Mode Authors* Pulse SV r Equation Male (N = 6, N = 6) Upper Stenberg et al. 2~ Bevegard et al Female (N = 4) Upper Stenberg et al. 2~ Male (N = 6, N = 6) Lower Stenberg et all -~ Bevegard et all Female (N = 4, N = 11) Lower Stenberg et al. 2~ Astrand et al Y = 9.24X , SE = 16.7 Y = 11.34X , SE = 8.3 No equation derived because of the low correlation Y = 8.36X , SE = 20.3 Y = 5.31X , SE = 11.7 Y = -3.35X , SE = 3.7 Y = 10.43X , SE = 8.4 * All three studies cited used the direct Fick technique to determine cardiac output. Stroke volume was calculated as the ratio between cardiac output and heart rate. Linear regression analysis was used to predict stroke volume from oxygen pulse. See text for details. used the direct Fick technique to determine Q appear to support the author's observations that O2 pulse can be used to predict SV during upper and lower body exercise in both genders and therefore validate the findings obtained from the indirect technique. Generalizability of the Predictive Equations To evaluate the generalizability of the equations derived for predicting SV at the ventilatory threshold, the mean O2 pulse values reported at comparable metabolic rates from several studies were substituted in the mode and gender-specific equations, and the predicted SV (SVp~) was calculated. The difference between the SV reported (SV~p) in the study cited and SVp,~ by the author's equations was computed and expressed as a percentage. In addition to the studies used for the validation procedure described earlier, those that had used indirect methods for evaluating Q, such as, l~'22-24 electrical Table 5: Generalizability of Regression Equation for Predicting Stroke Volume From Oxygen Pulse Measurements During Upper and Lower Body Exercise at the Ventilatory Threshold in Men VO2 O2 Pulse SV~p SVpre (L/ (ml! (ml/ (ml/ % Researchers/Method min) beat) beat) beat) diff Upper body Present study Bevegard et al) Clausen et all Cummins/Gladden Lower body Present study Astrand et al. ' Bhambhani et al. tl Berry et al Brandao et al. m Radionuclide Ang. Clausen et al Abbreviations: SVr~p, stroke volume reported by the researchers; SVp~, stroke volume predicted by equations derived in present study; % diff, calculated as ([SVpre -- SV,.cp] + SVrep) 100. impedance 25 and radionuclide angiography ~ also were included. This was performed to examine whether the SVpr~ by these equations was applicable to other methods of determining Q, or whether it was specific only to the method used in this study. The results of some of these comparisons during upper and lower body exercise in both genders are presented in tables 5 and 6. It is evident that in each case, the SVpr~ was within one standard error of estimate of the regression equation derived in this study. With the exception of the direct Fick data of Astrand and colleagues ~9 in women, the predictive error of the equations derived in the current study range from approximately -9% to 9%. The 16.4% discrepancy observed for the data of Astrand and colleagues j9 could be owing to the fact that the error in estimating Q from the technique when compared with the direct Fick method is approximately 12% on the average. 26 However, it should be noted that the discrepancies observed for the other studies that used the direct Fick method 2 '27 were not quite as large. Hence, the overall evidence suggests that the regression equations derived in this study for predicting SV from 02 pulse measurements can be generalized to other studies, despite the differences in techniques used to evaluate Q and therefore SV. In summary, the results of this study indicated that O2 Table 6: Generalizability of Regression Equation for Predicting Stroke Volume From Oxygen Pulse Measurements During Upper and Lower Body Exercise at the Ventilatory Threshold in Women VO2 02 Pulse SVr~p SVpre Researchers/ (L! (ml! (ml/ (ml/ % Method rain) beat) beat) beat) diff Upper Body Present study Miles et aly Impedance Lower Body Present study Astrand et all Miles et al SV~cp, stroke volume reported by the researchers; SVp~, stroke volume predicted by equations derived in present study; % diff, calculated as [(SVpr~ -- SVrep) -~- SVrep] X 100.

6 718 STROKE VOLUME PREDICTION DURING EXERCISE, Bhambhani pulse can be used to predict SV at the ventilatory threshold during upper and lower body exercise in healthy men and women. The slopes and intercepts of the regression equations developed were not significantly different between exercise modes in either gender. The error in the SVpre from these equations is usually within 10%, regardless of whether direct or indirect techniques are used to determine Q. Hence, it is recommended that these equations be used to predict SV when actual methods of estimation are not feasible. References 1. Sawka MN. Physiology of upper body exercise. Exer Sports Sci Rev 1986; 14: Astrand PO, Rodahl K. Textbook of work physiology: physiological bases of exercise. 3rd ed. Toronto: McGraw-Hill Book Company, Bhambhani Y, Eriksson P, Gomes P. Transfer effects of endurance training with the arms and legs. Med Sci Sports Exer 1991;23: Pendergast DR. Cardiovascular, respiratory, and metabolic responses to upper body exercise. Med Sci Sports Exer 1989;21:S Vokac Z, Bell H, Bautz-Holter E, Rodahl K. Oxygen uptake/heart rate relationship in leg and arm exercise, sitting and standing. J Appl Physiol 1975;39: Falkel JE, Sawka MN, Levine L, Pimental NA, Pandolf KB. Upperbody exercise performance comparison between men and women. Ergonomics 1986;29: Washburn R, Seals DR. Peak oxygen uptake during arm cranking for men and women. J Appl Physiol 1984;56: Wilmore JH, Farrell PA, Norton AC, Cote III RW, Coyle EF, Ewy GA, Temkin LP, Billing JE. An automated, indirect assessment of cardiac output during rest and exercise. J Appl Physiol 1982; 52: Kubicek WG, Karnegis JN, Patterson RP, Witose DA, Matson RH. Development and evaluation of an impedance cardiac output system. Aero Med 1966;37: Brandao MUP, Wajngarten M, Rondon E, Giorgi MCP, Hironaka F, Negrao CE. Left ventricular function during dynamic exercise in untrained and moderately trained subjects. J Appl Physiol 1993;75: Bhambhani Y, Norris S, Bell G. Prediction of stroke volume from oxygen pulse measurements in untrained and trained men. Can J Appl Physiol 1994; 19: Bar-Or O, Zwiren LD. Maximal oxygen consumption during arm exercise-reliability and validity. J Appl Physiol 1975;38: Davis JA, Frank MH, Whipp BJ, Wasserman K. Anaerobic threshold alterations caused by endurance training in middle aged men. J Appl Physiol 1979;46: Clausell N, Ludwig E, Narro F, Ribeiro JP. Response to left ventricular diastolic filling to graded exercise relative to lactate threshold. Eur J Appl Physiol 1993;67: Collier CR. Determination of mixed venous COz tensions by rebreathing. J Appl Physiol 1956;9: Jones NL. Clinical Exercise Testing. 3rd ed. Philadelphia: Saunders, Hossack KF, Bruce RA, Green B, Kusumi F, DeRouen TA, Trimble S. Maximal cardiac output during upright exercise: approximate normal standards and variations with coronary heart disease. Am J Cardiol 1980;46: Davis JA, Vodak P, Wilmore JH, Kurtz P. Anaerobic threshold and maximal aerobic power for three modes of exercise. J Appl Physiol 1976; 41: Astrand PO, Cuddy ET, Saltin B, Stenberg J. Cardiac output during submaximal and maximal work. J Appl Physiol 1964; 19: Bevegard S, Freychuss U, Strandell T. Circulatory adaptation to ann and leg exercise in supine and sitting position. J Appl Physiol 1966; 21: Stenberg J, Astrand PO, Ekblom B, Royce J, Saltin B. Hemodynamic response to work with different muscle groups, sitting and supine. J Appl Physiol 1967;22: Berry MJ, Zehnder TJ, Berry CB, Davis SE, Anderson SK. Cardiovascular responses in black and white males. J Appl Physiol 1993;74: Cummins T, Gladden B. Responses to submaximal and maximal ann cycling above, at, and below heart level. Med Sci Sports Excr 1983; 15: Miles DS, Critz JB, Knowlton RG. Cardiovascular, metabolic, and ventilatory responses of women to equivalent cycle ergometer and treadmill exercise. Med Sci Sports Exer 1980; 12: Miles DS, Sawka MN, Wilde SW, Doerr BM, Frey MAB, Glaser RM. Estimation of cardiac output by electric impedance during ann exercise in women. J Appl Physiol 1981;51: Marks C, Katch V, Beekman R, Rosenthal A. Validity and reliability of cardiac output by CO,_ rebreathing. Sports Med 1985;2: Clausen JP, Klausen K, Rasmussen B, Trap-Jensen J. Central and peripheral circulatory changes after training of the arms or legs. Am J Physiol 1973;225: Suppliers a. Monark Model 881; High Level Fitness Equipment, Avenue, Edmonton, Alberta, Canada. b. Monark Model 818E; High Level Fitness Equipment, Avenue, Edmonton, Alberta, Canada. c. Model MMC Horizon; Sensormedics Corporation, Savi Ranch Parkway, Yorba Linda, CA d. Model 1500B; Hewlett Packard, Goreway Drive, Missisauga, Ontario, Canada, L4V 1M8. e. Advanced Exercise Testing Program; Sensormedics Corporation, Savi Ranch Parkway, Yorba Linda, CA f. SPSS for Windows; SPSS Inc., 444 North Michigan Avenue, Chicago, IL