˙ O2 kinetics of mild exercise are altered by RER V ´ AND JAMES J. HOFFMANN PAUL A. MOLE Human Performance Laboratory, Department of Exercise Science, University of California, Davis, Davis, California 95616

˙ O2 kinetics of Mole´, Paul A., and James J. Hoffmann. V mild exercise are altered by RER. J. Appl. Physiol. 87(6): 2097–2106, 1999.—We propose that variations in fat and carbohydrate (CHO) oxidation by working muscle alter O2 ˙ O2) kinetics. This hypothesis provides two predicuptake (V tions: 1) the kinetics should comprise two exponential components, one fast and the other slow, and 2) their contribution should change with variations in fat and CHO oxidation, as predicted by steady-state respiratory exchange ratio (RER). The purpose of this study was to test these predictions ˙ O2 kinetic model: V˙ O2(t) 5 aR 1 by evaluating the V aF51 2 exp[(t 2 TD)/2tF]6 1 aC51 2 exp[(t 2 TD)/2tC]6 for short-term, mild leg cycling in 38 women and 44 men, ˙ O2(t) describes the time course, aR is resting V˙ O2, t is where V time after onset of exercise, TD is time delay, aF and tF are asymptote and time constant, respectively, for the fast (fat) oxidative term, and aC and tC are the corresponding parameters for the slow (CHO) oxidative term. We found that 1) this ˙ O2 kinetics biexponential model accurately described the V over a wide range of RERs, 2) the contribution of the fast (aF, fat) component was inversely related to RER, whereas the slow (aC, CHO) component was positively related to RER, and 3) this assignment of the fast and slow terms accurately predicted steady-state respiratory quotient and CO2 output. Therefore, the kinetic model can quantify the dynamics of fat and CHO oxidation over the first 5–10 min of mild exercise in young adult men and women. humans; cycle ergometry; time delay; time constants; metabolism; energetics; fat and carbohydrate oxidation

is the standard paradigm for determining the energy transfer rate and oxidative fuel utilization during exercise in humans, but only ˙ CO2) under steady-state conditions where CO2 output (V is unaffected by acid-base disturbances. The requirement for steady state is a serious limitation, because most conditions of exercise are non-steady state. Consequently, there is an urgent need to extend the indirect calorimetric paradigm to determine the kinetics of muscle and whole body energetics and oxidation of fat and carbohydrate (CHO). The kinetic paradigm would provide opportunities to study 1) the contribution of fat and CHO to oxidative metabolism during the early transition to steady state, 2) the cellular mechanisms controlling oxidative metabolism, 3) the systemic controls involving alveolar gas exchange as well as convective O2 transport and conductive O2 diffusion, and 4) ˙ CO2 and acidthe interrelationship between oxidative V ˙ CO2. base effects on measured V INDIRECT CALORIMETRY

The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. http://www.jap.org

The complexity of physiological and metabolic processes makes developing a kinetic paradigm a daunting task. However, there are several principles of modeling that can help develop, as a first step, a simplified dynamic indirect calorimetric paradigm, which is based ˙ O2) kinetics. First, the shape of the on O2 uptake (V ˙ dynamic VO2 response provides information on pulmonary, circulatory, and metabolic processes. Second, mathematically decomposing the response into its component parts potentially can represent a specific process or some combination of processes. Critical to this task is the selection of a mathematical model, which contains parameters that reflect the dynamic characteristics of the process under study. At that point, each process can be identified by performing experiments that test the parameters of the model against theory. Previous authors (4–6, 43, 45) have initiated decom˙ O2 kinetics. Whipp and colleagues (45) position of V ˙ O2 kinetics during light to introduced the concept that V moderate exercise below the performer’s lactate threshold involves three phases. Phase 1 is the rapid rise in ˙ O2 after the onset of exercise, which reaches pulmonary V a transient plateau or inflection point in 15–20 s. They proposed that this phase involves the rapid delivery of O2-poor blood for alveolar exchange and called it the ˙ O2 rises ‘‘cardiodynamic phase.’’ After phase 1, the V exponentially in phase 2 and reaches an apparent ˙ O2 of steady state (phase 3), reflecting an increase in V metabolically active tissues involved in exercise, primarily working muscles. Accordingly, phases 2 and 3 represent oxidative dynamics of working muscles. Results from simulation experiments by Barstow and coworkers (4, 5) are consistent with this assignment of circulatory (phase 1) and metabolic components (phases ˙ O2 for light to 2 and 3) to the dynamics of pulmonary V moderately intense exercise. It is important to notice ˙ O2 kinetic response has informathat the shape of the V tion on the time course of cardiorespiratory (phase 1) and oxidative (phases 2 and 3) processes. Therefore, an ˙ O2 accurate mathematical model of phase 2 and 3 V kinetics may provide analytic information on the dynamics of oxidative metabolism of working muscles. ˙ O2 That the exercise-induced increase in pulmonary V primarily represents the oxidation of working muscles is well established experimentally (1–3, 21, 23, 30, 33, 36, 39), thereby confirming the theoretical work (4, 5). For example, recently Odland et al. (30) showed that ˙ O2 represents ,80% of whole body V˙ O2 with leg V virtually the same substrate oxidation [leg respiratory quotient (RQ) 5 0.91 vs. respiratory exchange ratio (RER) 5 0.92] during minutes 4–7 of cycling at 40% of ˙ O2 (V˙ O2 max). Therefore, it is reasonable to maximal V propose that the oxidation of fat and CHO in working ˙ O2 in phases 2 and 3. Furthermuscles contributes to V more, as developed below, we proposed that the initial

8750-7587/99 $5.00 Copyright r 1999 the American Physiological Society

2097

2098

˙ O2 KINETICS AND RER V

changes in the flux of these fuels are reflected in the ˙ O2. Briefly, we propose that the dynamics of phase 2 V flux of these fuels is initiated as a feedforward process driven by contraction-induced increases in calcium, cAMP, and epinephrine that simultaneously activate glycogenolysis, triglyceride (TG) lipolysis, and b-oxidation. Subsequently, feedback controls modulate their fluxes as dictated by alterations in the redox and phosphorylation potentials and intermediates, including pyruvate, citrate, and the ratio of acetyl CoA to CoA. Therefore, we propose that the contractioninduced changes in key enzymes controlling glycogen, intracellular TG, and fatty acid fluxes initially dictate the dynamic of redox and phosphorylation states, which in turn control and drive the dynamic of oxidative ˙ O2 of working muscle. phosphorylation and phase 2 V As recently reviewed by Tschakovsky and Hughson (40), if muscle O2 utilization, particularly for light exercise, is limited by intrinsic metabolic inertia and not by O2 transport inertia, then the primary determi˙ O2 kinetics will include enzyme nants of phase 2 V activation associated with phosphorylation and redox states in healthy individuals under normoxic conditions. Research from our laboratory supports this view (28). Our findings on myoglobin desaturation during plantar flexion exercise at various intensities suggest that intracellular PO2 is not limiting, and therefore neither convective nor conductive O2 transport is typically controlling oxidative dynamics in working skel˙ O2 controlled by etal muscle. How, then, is phase 2 V metabolic inertia at the early unsteady state of exercise? Briefly, muscle contraction will rapidly increase cellular Ca21 with activation of various ATPases and dehydrogenases. Also, the activity of various AMP protein kinases increases, which in turn activates glycogenolysis, intracellular TG lipolysis, and b-oxidation (38). Therefore, glycogen and intramuscular TG should be the primary fuels utilized during the early transient, non-steady state (10, 12, 15–17, 20, 31, 42, ˙ O2 by 48). Carbon flux from these sources is linked to V the formation of reducing equivalents, FADH2 and NADH, and their utilization by the electron transport chain (ETC). When fat oxidation dominates, such as when persons perform light to moderate exercise on a fat diet, we propose that TG lipolysis and b-oxidation are rapidly activated at exercise onset. Consequently, b-oxidation will be the dominant supplier of FADH2 and NADH to ETC and the formation of acetyl CoA. In this case, acetyl CoA from fat oxidation will dominate the early rapid increase in tricarboxylic acid (TCA) ˙ O2 in phase 2. cycle carbon flux and the rise of V Furthermore, TG oxidation would be expected to have feedback control over glycogenolysis (via citrate) and would attenuate activation of the pyruvate dehydrogenase (PDH) complex via acetyl CoA and NADH (7). More specifically, increases in the NADH-to-NAD and acetyl CoA-to-CoA ratios from TG oxidation could feed back to attenuate the activation of the PDH complex. Studies (9, 15) have shown that lactate rapidly accumulates after the onset of exercise, leading to an increased cytosolic NAD-to-NADH ratio. Brooks and co-workers

(8) recently proposed that lactate can be transported into the mitochondria and converted to pyruvate and NADH by matrix lactate dehydrogenase. We propose that this would rapidly enhance the mitochondrial NADH-to-NAD ratio and downregulate PDH, particularly when fat oxidation is dominant. Under these conditions in which fat (intracellular TG) is the dominant oxidative substrate, PDH complex activity must be insufficient initially in supplying acetyl CoA to the TCA cycle. Thus pyruvate oxidation would increase more slowly than intramuscular TG lipolysis and b-oxidation at the onset of exercise. We think this slow CHO ˙ O2 drift,’’ as oxidation is not part of the so-called ‘‘V described by Mole´ and Coulson (29), since it appears as an exponential dynamic as opposed to a linear drift, which often develops during prolonged exercise at moderate to high intensities. ˙ O2 distinctly tracks the dynamics of fat and CHO If V ˙ O2 in phase 2 would have oxidation as proposed, then V at least two components that change at different speeds. Moreover, the relative contribution of these substrates ˙ O2 kinetics, since their ATP-to-O2 could affect the V ratios differ because of differences in NADH and FADH2 production. This can be expressed by Eq. 1 ATP/O2(t) 5 6.5fCHO 1 5.64fFat

(1)

where fCHO and fFat are the time-dependent fractions for CHO and fat oxidation, respectively. The coefficients in Eq. 1 were derived by assuming that glycogen and mixed intramuscular TG are oxidized during the first minutes of exercise up to the initial attainment of steady state. If the rate of ATP utilization is constant during constant-load muscular work, then the rate of O2 utilization would vary according to the fractional oxidation of CHO and fat. MODEL OF O2 UPTAKE KINETICS

˙ O2 in phase 2 We propose that breath-by-breath V increases as a biexponential and quantitatively represents the dynamics of intramuscular TG and glycogen oxidation during the transition from rest to the initial steady state of exercise. Furthermore, the dynamics of ˙ O2 and the oxidation of these substrates are tightly V coupled, and local controls are the dominant factors determining their kinetics. The controls could operate, as reviewed above, such that intramuscular TG oxidation increases more quickly than CHO oxidation with a ˙ O2 kinetics step increase in work rate. In this case, V during phase 2 should be expressed by two components, a fast component representing fat oxidation and a slow component reflecting the time course for CHO oxidation, as given by Eq. 2, which is modified from that of Barstow and Mole´ (6) ˙ O2 (t) 5 aR 1 aF 51 2 exp31t 2 TD2/ 2 tF46 V 1 aC 51 2 exp [(t 2 TD2 / 2 tC46

(2)

˙ O2 describes the time course of muscle oxidawhere V ˙ O2, tion during phases 2 and 3, aR is the initial resting V t is the time starting from the onset of exercise, TD is

˙ O2 KINETICS AND RER V ˙ O2 equals aR, aF and aC the time delay where phase 2 V are the fast (fat) and slow (CHO) asymptotes represent˙ O2 at phase 3 due to ing the steady-state increase in V fat and CHO oxidation for exercise, and tF and tC are the time constants defining the speed of fat and CHO oxidation, respectively. Given this assignment, steadystate RQ can be estimated from the asymptotic parameters of Eq. 2 by calculating the fractional contribution of CHO (fCHO, Eq. 3) or fat (fFat, Eq. 4) RQ 5 0.707 1 0.293 f CHO

(3)

RQ 5 1.00 2 0.293 f Fat

(4)

where fCHO 5 aC/(aF 1 aC ) and fFat 5 1.00 2 fCHO. With ˙ CO2 can be calculated by using the defRQ, oxidative V ˙ CO2 to V˙ O2. inition of RQ as the ratio of oxidative V For mild to moderate intensities of exercise below ˙ TH ), steadylactate threshold or ventilatory threshold (V state RER will be equivalent to RQ (30). Then individuals with different RQs (RERs for mild exercise) should have different fCHO, fFat, aF, aC, and time constants. PROBLEM OF STUDY

In the present study we tested these predictions by ˙ O2 kinetic model (Eq. evaluating the parameters of the V 2) obtained for mild leg-cycling exercise in 82 subjects (38 women and 44 men) who had widely different steady-state RERs. The following hypotheses were tested. Hypothesis 1. Equation 2 will accurately describe ˙ O2 kinetics during mild exercise, as phase 2 and 3 V ˙ CO2 5 evidenced by the following equality: predicted V ˙ CO2. measured steady-state V Hypothesis 2. The asymptote of the fast term (aF ) and its fractional contribution [aF/(aF 1 aC )] will be inversely related to steady-state RER, since we propose that aF is the oxidation of fat. Hypothesis 3. Conversely, the asymptote of the slow term (aC ) and its fractional contribution [aC/(aF 1 aC )] will be positively related to steady-state RER, since we propose that aC is the oxidation of CHO. Hypothesis 4. The fast time constant (tF ) will be negatively, whereas the slow time constant (tC ) will be positively, related to RER. Hypothesis 5. The TD will not be related to RER.

2099

capacity and involved exercising for 5–10 min at 20–65 W ˙ O2 max). Because a wide range of RERs was desired, (28–40% V the subjects were studied 2–4 h postprandially or 10–14 h postabsorptively. Each subject was instructed to maintain his/her habitual dietary pattern, particularly on the day before and the day of testing. To assess compliance, each subject measured food intake by weight and volume on the day of testing and also over a 3-day period including 2 week days immediately before testing and 1 weekend day. Characterization of subjects. On arrival at the laboratory, the subject’s body composition was determined by underwater weighing or by air plethysmography by use of the BOD POD (Life Measurements, Concord, CA), as previously described (26). Residual lung volume was determined by the method of Wilmore (46). Percent fat mass was determined from body density with the Siri equation (37). Next, the ˙ O2 and leg-cycling power relationship between steady-state V was assessed in the following way. After seat height was adjusted and recorded and the preferred cycling cadence was determined on the electrically braked cycle ergometer (ErgoLine 800S, SensorMedics), the subject rested for 15 min while the flowmeter and gas analyzers of the metabolic cart (model 2900, SensorMedics) were calibrated with a 3-liter syringe and standard O2 and CO2 gases, respectively. On-line breathby-breath determinations of pulmonary ventilation (BTPS ), ˙ O2, V˙ CO2, and RER (i.e., V˙ CO2/V˙ O2) were made continuously V during consecutive 5-min periods of sitting rest and three intensities of exercise at 30, 60, and 90 W for women and 40, ˙ O2 values were averaged 80, and 120 W for men. The V between minutes 4 and 5 of exercise. These data and those obtained in subsequent experiments were regressed against cycling power by least-squares regression analysis (SigmaPlot, version 2.0, Jandel) and were used to define the relation˙ O2 and leg-cycling power output of Fig. 1. ship between V

METHODS AND PROCEDURES

Subjects. Eighty-two young adults (44 men and 38 women) volunteered to participate in the study after being informed of the requirements, procedures, and risks. Written consent was given in accordance with the requirements of the University’s Institutional Review Board for Human Subjects Experimentation. All subjects were physically active, but none were involved in intense training for athletic competition. Experimental protocol. The first session involved character˙ TH, izing each subject with respect to body composition, V ˙ O2 max, and steady-state V˙ O2 as a function of cycle power V output. Subsequently, subjects participated in three to eight ses˙ O2 kinetics sions at 2- to 7-day intervals to characterize their V for mild cycling exercise. The intensity of these mild exercise bouts varied between subjects depending on their aerobic

˙ O2) as related to leg-cycling power: comparison of Fig. 1. O2 uptake (V ˙ O2 5 (451 6 26.4) 1 (9.36 6 women and men. Regression for women: V 0.259)power, standard error of estimate (SEE) 5 162 ml/min, R2 5 ˙ O2 5 (485 6 33.1) 1 (11.20 6 0.194)power, 0.90; regression for men: V SEE 5 309 ml/min, R2 5 0.92. These coefficients are different from those reported in RESULTS, because they were derived by regression of each group’s data.

˙ O2 KINETICS AND RER V

2100

The subject rested for ,10 min while the metabolic cart ˙ O2 max was was recalibrated. Then the following test of V undertaken. Continuous breath-by-breath measurements were made throughout exercise to volitional exhaustion by using a ramp protocol. Exercise started with a power of 25 W increasing 25 W/min for the women or with a power of 30 W ˙ TH was estimated using increasing 30 W/min for the men. V ˙ CO2 and V˙ O2 and the V-slope method as 15-s averages of V ˙ O2 max was defined provided by the SensorMedics software. V ˙ O2 that did not change .110 ml/min or 1.6 as the highest V ml · min21 · kg21 (standard error for repeated measurements, unpublished observations) with increments in power output and was accompanied by a RER $ 1.10. All tests satisfied these criteria. ˙ O2 kinetics for mild cycling exercise. One week after characterV ization, the subject began a series of three to eight sessions at 2- to 7-day intervals. Data used for analyses came from an initial, single bout of submaximal cycle ergometer exercise performed each session. Breath-by-breath determinations of gas exchange were made at rest on the bicycle ergometer for 5 min and during the mild ‘‘warm-up’’ bout for 5–10 min ˙ O2 max). at 20–65 W (28–40% V ˙ O2 kinetics. The breathNonlinear regression analysis of V ˙ O2 data for each bout were smoothed with a by-breath V four-breath rolling average, interpolated to one value per second, and time aligned to the start of exercise. Next, the responses of three to eight experiments for the subject were ˙ O2 kinetics excluded phase 1. The end averaged. Analysis of V ˙ O2 was determined as of phase 1 or the start of phase 2 for V the breath before to the sudden, progressive fall in RER. This coincides with the inflection point or the end of the initial ˙ O2. The data after phase 1 (e.g., phases 2 and 3) plateau of V were then fit by a double-exponential model (Eq. 2) with a single TD by using standard nonlinear regression techniques provided by SigmaStat (version 1, Jandel Scientific). Because the initial estimates can influence the derived parameters of the model, they were standardized as follows. ˙ O2 was assigned to aR. The starting The average resting V values for TD, tF, and tC were 15, 15, and 60 s, respectively. ˙ O2 and the averaged V˙ O2 over The difference between resting V the last 30 s of exercise was halved, and this value was assigned to the initial estimate for aF and aC. Finally, tF and tC were constrained, such that tF , tC and tC , 500 s. The tolerance for convergence was initially set at 0.001 and modified, if necessary, with further trials to obtain a solution. Statistics. Values in RESULTS and Tables 1–4 are means 6 SD. Group data illustrated in Figs. 2–5 are means 6 SE. A difference was accepted as statistically significant for P # 0.05 at a power $0.8. Significant gender differences were tested using the independent t-test. The tests for differences between time constants were determined with the dependent t-test. One-way ANOVA was employed to evaluate the effect of RER on the parameters of Eq. 2, with Bonferroni’s post hoc test used when appropriate. RESULTS

Table 1 presents several characteristics of the volunteers. These results indicate that the participants’

Table 2. Steady-state metabolic response to mild leg exercise Group

n

Power, W

˙ O2 , V ml/min

˙ CO2 , V ml/min

RER, ˙ CO2 /V˙ O2 V

Women Men Combined

38 44 82

38.7 6 9.5 44.3 6 17.2 41.3 6 13.8

890 6 305 1,030 6 261 955 6 292

772 6 258 892 6 224 828 6 249

0.87 6 0.06 0.87 6 0.06 0.87 6 0.06

˙ O2 , O2 uptake; V˙ CO2 , CO2 output; RER, Values are means 6 SD. V ˙ O2 respiratory exchange ratio. This level of exercise represents a V ˙ O2 max for women and men, respectively. that is 40% and 28% of V

˙ O2 max, and V˙ TH were in the expected relative fatness, V range for nonobese, physically active young adults. ˙ O2 and Figure 1 shows the relationship between the V leg-cycling power output for the men and women. Comparison of the intercepts, derived by linear regression analysis of each individual’s relationship, showed no significant difference between the men and women [500 6 226 vs. 427 6 164 (SD) ml/min, respectively]. In contrast, the women responded with a significantly lower slope than the men (9.98 6 1.74 vs. 10.97 6 1.70 ml O2 · min21 · W21, P 5 0.02). These slopes are equivalent to 3.37 and 3.71 J metabolic energy/J mechanical work when the mean RER values of Table 2 are employed. The inverses of these values are estimates of the efficiency and indicate that the efficiency of leg cycling was 29.7 and 26.9% for the women and men, respectively. ˙ O2, V˙ CO2, and RER for mild exercise (20–65 Power, V W) are given in Table 2 for the women and men. The small differences between genders are attributed, in part, to the differences in efficiency and exercise intensity. On average, the exercise employed represented a ˙ O2 equivalent to 40 and 28% V˙ O2 max for the women and V men, respectively. In both groups the intensity was well ˙ TH of the participants (Table 1). Thus it is below the V reasonable to assume that steady-state RER closely corresponds to RQ for mild exercise used to evaluate ˙ O2 kinetics (30). In this case, the average RER of 0.87 V observed for both groups (Table 2) indicates that CHO was the dominant fuel (,56% on average) for oxidative metabolism. This is consistent with the analysis of the habitual diets for the groups (1,984 6 489 kcal, 14% protein, 59% CHO, and 27% fat for women and 2,952 6 774 kcal, 16% protein, 53% CHO, and 31% fat for men). However, there were large individual differences in diets. Furthermore, it will be shown that the subjects responded to mild exercise with a wide range of RERs. RER was not related to cycle power output (r 5 20.01) but was significantly associated with percent kilocalories from CHO in the diet (r 5 0.483, P , 0.01). ˙ O2 kinetic parameters obtained from nonlinear The V regression analysis of the model equation (Eq. 2) are

Table 1. Characteristics of women and men Group

n

Age, yr

Height, cm

Body Mass, kg

Fat Mass, %

˙ O2 max , V ml · min21 · kg21

˙ TH , V ˙ O2 max %V

Women Men

38 44

22.8 6 3.0 23.9 6 5.6

164.2 6 6.1 179.8 6 7.2

56.8 6 8.0 78.3 6 12.5

21.5 6 6.1 14.3 6 5.6

39.0 6 8.1 47.3 6 8.3

55.7 6 9.4 51.8 6 11.5

˙ O2 max , maximal O2 uptake; VTH , ventilatory threshold determined by V-slope method. Values are means 6 SD. V

˙ O2 KINETICS AND RER V

2101

˙ O2 kinetic model Table 3. Parameters for the V Group

aR , ml/min

aF , ml/min

aC , ml/min

TD, s

tF , s

tC , s

Women Men Combined

255 6 68 319 6 92 284 6 86

330 6 178 359 6 178 343 6 178

307 6 175 356 6 160 330 6 169

13 6 7.0 12 6 5.9 12 6 6.5

14 6 9.2 16 6 10.4 15 6 9.8

78 6 96.9 74 6 99.9 76 6 97.7

˙ O2 ; aF and tF , asymptote and time constant for fast (fat) oxidative term; aC and tC , Values are means 6 SD. TD, time delay; aR , resting V asymptote and time constant for slow (carbohydrate) oxidative term. tF vs. tC are significantly different, P , 0.0001. presented in Table 3. No significant differences (P . 0.05) were found for the women and men. Comparison of asymptotes (aF and aC ) and time constants (tF and tC ) by dependent (paired) t-test for the combined group indicated that only the latter was significant (P , 0.0001). The parameters of the model were grouped by RER in 0.05-unit increments (except for values ,0.8, which were lumped together) and analyzed by one-way ANOVA. As shown in Table 4, the TD was unchanged over the range of RER groupings. Similarly, tF was invariant with respect to RER 5 0.78–0.97. In contrast, tC was inversely related to RER (r 5 20.97). That is, for RER 5 0.78 the slow time constant was 163 6 59 s, but at RER 5 0.97 the time constant of 14 6 7.9 s was significantly faster (Table 4). At RER , 0.92, tF and tC were different (P , 0.05). These findings are illustrated in Fig. 2 by the relationships between the time constants and RER. Figure 2 shows that tC converges toward tF as RER approaches unity, making it impossible to discern a second-order dynamic response at RER $ 0.92. The fast (aF ) and the slow (aC ) steady-state parameters of the kinetic model were significantly related to RER (Table 4, Fig. 3; P , 0.0001). More specifically, the absolute rate of the fast component was inversely related to RER and approached zero at RER 5 1.00. The slow component showed the opposite (positive) relationship with RER. Because RER approximates RQ at steady state for mild exercise, as studied here, these relationships with RER imply that the fast and slow components represent fat and CHO oxidation, respectively. An initial test of this assignment is provided by the expected relationship between RQ and the fractional oxidation of fat (fFat ) at steady state (Eq. 4): RQ 5 1.00 2 0.293fFat, where fFat 5 [aF/(aF 1 aC )]. This prediction is illustrated by the theoretical dashed line of Fig. 4. Also shown are values we obtained for fFat for the RER groupings. Notice how closely they approximate the theorectical fFat at RER 5 0.78, 0.82, and 0.88.

At RER $ 0.92, there was a clear discrepancy between the estimated and theoretical fFat. This occurred where the time constants approached each other (Fig. 2), making it technically difficult to discern two unique ˙ O2 kinetics for mild cycling exercise. components for V Equation 4 provides a method for calculating RQ and ˙ CO2 from fFat. The estimated oxidative and oxidative V ˙ CO2 at steady state as well as the differmeasured V ences between them for the RER groupings are presented in Fig. 5, A and B, respectively. The discrepancies for the mean RERs of 0.92 and 0.97 were small (225 6 5.8 and 241 6 4.7 ml/min, respectively) but statistically significant (P , 0.05). However, the pre˙ CO2 were strongly related (Fig. dicted and measured V 6A; r 5 0.99). The mean difference was 211 6 3.0 (SE) ˙ CO2 evaluated. ml/min (Fig. 6B) over the domain of V DISCUSSION

The present study showed that 1) our double˙ O2 kinetexponential model adequately described the V ics for the subjects with RERs , 0.92, 2) aF was inversely correlated, whereas aC was positively correlated with steady-state RER, and 3) TD and tF were independent of RER, whereas tC was inversely related to RER. These findings suggest that the relative contributions of fat and CHO to oxidative metabolism determine the relative contributions of the fast and slow ˙ O2 kinetics and the speed of the slow components of V component. The possibility that these relationships reflect cause and effect is supported by the accurate ˙ CO2 obtained with estimates of steady-state RQ and V our kinetic model and indirect calorimetric paradigm. Therefore, it would appear that the fast component ˙ O2 represents fat oxidation while the slow compoof V ˙ O2 is due to CHO oxidation. nent of V ˙ O2 increased in Our findings show that steady-state V proportion to leg-cycling power output, indicating that aerobic metabolism behaves as a linear system, as has been consistently reported for exercise intensities be-

˙ O2 kinetic model Table 4. Grouping by steady-state RER for the parameters of the V Grouping by RER

n

Power, W

aR , ml/min

aF , ml/min

aC , ml/min

TD, s

tF , s

tC , s

0.78 6 0.016 0.82 6 0.015 0.88 6 0.014 0.92 6 0.014 0.97 6 0.014

9 23 23 22 5

39 44 37 44 39

308 6 88 289 6 83 264 6 95 297 6 89 260 6 44

557 6 159 445 6 203 255 6 73 272 6 119 211 6 49

206 6 104 297 6 160 336 6 135 393 6 124 397 6 94

12 6 4.0 12 6 7.5 12 6 5.4 13 6 7.5 15 6 6.0

21 6 7.1 14 6 9.2 17 6 9.7 13 6 9.7 10 6 9.0

163 6 59.0 93 6 69.0 67 6 80.3 48 6 58.5 14 6 7.9

Values are means 6 SD. TD and tF were not different, but tC was significantly lower (P , 0.05), when respiratory exchange ratio (RER) changed from 0.78 to 0.97. Below RER of 0.92, tF and tC were different (P , 0.05).

2102

˙ O2 KINETICS AND RER V

˙ O2 kinetic model as related to steady-state Fig. 2. Time constants of V respiratory exchange ratio (RER) for mild exercise. Values are group means 6 SE. Time constants for fast (fat, tF ) and slow [carbohydrate (CHO), tC] oxidative terms are significantly different (P , 0.05) at RER , 0.92.

˙ TH (6, 14, 44). Furthermore, V˙ O2 kinetics of mild low V ˙ TH were strongly associated with RER, exercise below V ˙ O2 kinetics in as seen in Table 4. For RER , 0.92, V phases 2 and 3 responded as a second-order linear system, inasmuch as respiration could be described by a fast and a slower exponential term. At RER $ 0.92, the system was statistically discerned only as a firstorder linear system, i.e., with only a fast exponential

Fig. 3. Fast (aF ) and slow (aC ) asymptotes of kinetic model as related to steady-state RER for mild exercise. Values are group means 6 SE.

Fig. 4. Fractional contribution of aF as related to steady-state RER for mild exercise. Values are group means 6 SE and individual values.

term. This apparent change in the system’s order came about because the slow time constant was inversely related to RER. That is, as RER increased, the slow time constant became faster, thereby converging toward the invariant fast time constant. This effect of

Fig. 5. Comparison of measured (Meas) and predicted (Pred) CO2 ˙ CO2, A) and differences (B) as related to steady-state production (V RER for mild leg cycling. Values are group means 6 SE.

˙ O2 KINETICS AND RER V

2103

difficult to discern two dynamic components at high RERs. Nevertheless, it is expected that detection will improve at exercise intensities higher than that for very mild exercise employed in this study because of ˙ O2. greater signal-to-noise ratio for breath-by-breath V We have presented only correlational evidence in support of our model. Preliminary results from research in progress show that dietary-induced metabolic ˙ O2 kinetics as predicted by our adaptations alter V model. That is, metabolic adaptations to a fat diet increase the contribution of the fast (fat oxidation) component and reduce the contribution of the slow (CHO oxidation) term. The converse occurs when adaptations are induced by a CHO diet. Figure 7 is an example of these preliminary results in one subject. ˙ O2 for the subject on the fat diet Notice the slow rise in V (Fig. 7A) that appears to disappear on the CHO diet (Fig. 7B). In both cases, there is excellent agreement between the predicted RQ and measured steady-state RER. These results are encouraging, but we must reserve judgment until a more complete assessment of

˙ CO2: individual Fig. 6. Comparison of measured and predicted V responses (A) and differences (B). Coefficient of determination (R2 ) was 0.99 (A), and difference (mean 6 SE) was 211 6 3.0 ml/min (B); 95% confidence interval is shown relative to mean difference (B). RER probably is an important reason why we (6) and ˙ O2 kinetics with a others (14, 44) could only describe V monoexponential model for exercise intensities below ˙ TH. Our experimental evidence helps explain this V apparent transition in the system’s order over the domain of RER. ˙ O2 at the It is thought that the rapid rise in muscle V onset of exercise represents the oxidation of CHO. However, our finding of an inverse relationship be˙ O2 component and RER suggests that tween the fast V this component reflects fat oxidation, not CHO oxidation. This is supported by the finding that our model ˙ CO2 accurately predicts steady-state RQ and oxidative V when the fast term is assigned to fat oxidation. Therefore, the transition from a double to a single exponential should occur as RER approaches unity, because the contribution of fat would progressively decrease and disappear at an RER of unity, leaving only CHO as the sole oxidative substrate. Unfortunately, it was difficult to identify the two components at RER $ 0.92. This practical limitation of the model probably is due to the inherent breath-by-breath noise and to the variability in the time constants in individuals with different muscle fiber type composition and recruitment pat˙ O2 kinetics independent of terns, which may affect V RER (24). These factors, combined with the acceleration of the slow time constant, make it technically

˙ O2 Fig. 7. Representative effect of fat (A) and CHO (B) diets on V kinetics during mild leg-cycling exercise at 39 W in a man. For fat ˙ O2 5 481 1 607exp[(t 2 diet, solution to model Eq. 2 was as follows: V 15.6)/217.5] 1 234exp[(t 2 15.6)/2197.8], R2 5 0.92, measured RER 5 0.79 vs. predicted respiratory quotient 5 0.79. For CHO diet, ˙ O2 5 239 1 456exp[(t 2 solution to model Eq. 2 was as follows: V 4.6)/232.5] 1 479exp[(t 2 4.6)/242.4], R2 5 0.98, measured RER 5 0.87 vs. predicted respiratory quotient 5 0.86.

2104

˙ O2 KINETICS AND RER V

dietary effects can be made and other experimental tests of the model are undertaken and reported. Other considerations suggest that the dynamics of muscle fat oxidation are fast and represent oxidation of intramuscular TGs, whereas the slow component represents primarily glycogen oxidation during the transition to the initial steady state (5–10 min). At the initiation of all intensities of exercise, glycogenolysis is rapidly activated. Wahren et al. (41) showed that, during the first 2 min of forearm exercise, net glucose is released, indicating that glycogen is the primary CHO substrate utilized during this early period. Connett et al. (10) found in contracting dog gracilis muscle that the initial burst in the glycolytic rate is independent of ˙ O2. Lactate rapidly accumulated, muscle PO2 and V indicating a mismatch between pyruvate production and oxidation. They suggested that aerobic glycolysis functions to reduce the cytosolic redox state via lactate accumulation to support mitochondrial fat oxidation. Consistent with this suggestion are the results of computer simulation studies on work-work transitions in the pyruvate-perfused heart. Garfinkel et al. (15) showed that pyruvate is mainly removed by lactate accumulation in the first 30 s of a step increase in work. Thus their findings are consistent with the view that intramuscular TG oxidation rapidly increases to support aerobic production of ATP with a step increase in work. However, studies on net utilization of muscle TG have reported conflicting results (34). Some studies (22, 38) have shown muscle TG concentration to be unaltered by exercise, whereas others (12, 18, 19, 35) have found exercise to decrease intracellular TG concentration. Intracellular TGs do turn over (35). Therefore, these discrepancies suggest that the balance between TG synthesis and utilization can be variable under various conditions of exercise. Differences between total fat oxidation by indirect calorimetry and plasma free fatty acid oxidation suggest that intracellular TG can account for ,50% of the fat oxidized at steady state during prolonged exercise (13, 16, 17, 20, 32). It is not known what contribution intramuscular TG makes to oxidative metabolism during the transient period of exercise. However, skeletal muscle has the enzymatic pathway for quickly increasing lipolysis of ˙ O2 intramuscular TG to support the rapid increase in V at the onset of exercise. Oscai and co-workers (31) studied an intramuscular TG lipase and showed that it is rapidly activated by protein phosphorylation mediated by cAMP-dependent protein kinase. Thus it is reasonable to assume that the activity of this lipase rapidly increases with the activation of phosphorylase at the onset of exercise. Presumably, b-oxidation also increases in concert with activation of lipolysis, inasmuch as the expected rise in AMP protein kinase would phosphorylate and inactivate acetyl CoA carboxylase, which is responsible for synthesis of malonyl CoA (47). Consequently, a rapid fall in malonyl CoA would disinhibit carnitine acyl transferase I, which is the ratelimiting enzyme in b-oxidation. Although more research on the dynamics of intracellular TG lipolysis is

needed, these speculations suggest that the machinery is present and could activate intracellular TG oxidation rapidly to support the fast increase in muscle oxidation. Consistent with this idea is indirect evidence which shows that the RER decreases at the start of phase 2 ˙ O2 (24). Estimated ‘‘apparent’’ CO2 retention from the V RER decline is markedly greater than can be accounted for by the alkalinization produced from net H1 removal and phosphocreatine utilization by the creatine kinase reaction (unpublished observations). Therefore, this suggests that fat oxidation likely is increasing faster ˙ O2. We than CHO oxidation during the early rise in V have reviewed evidence elsewhere (26) which suggests that intramuscular TG is the immediate source of fatty acid oxidation and is rapidly oxidized. For example, Zierler (48) found that production of 14CO2 lags the uptake of 14C-labeled plasma fatty acids in humans at rest when the arteriovenous RQ across the muscle was low, indicating dominance of intracellular fat oxidation. Paul and Issekutz (32) observed a similar delay in plasma fatty acid oxidation at the onset of exercise in the dog, even though the specific activity of 14C-labeled plasma fatty acids was at steady state. This probably occurred because the intramuscular TG pool was not labeled adequately. Consequently, this resulted in a ˙ CO2 from unlabeled TG. Also, Israpid increase in V sekutz and Paul (20) showed that, after prolonged infusion of [14C]palmitate for 190 min to label muscle TGs, 14CO2 production rapidly increased during the onset of exercise, even though the specific activity of plasma [14C]palmitate had fallen to a nominal level 5 min into exercise after infusion was stopped. All these findings suggest that intramuscular TG is the immediate fuel for fatty acid oxidation at the onset of exercise. Although it is unclear what specific factors influence the time constants, it is reasonable to assume that the fast time constant provides information related to the speed of fat oxidation. A comparison of our fast time constant with those reported for plasma [14C]palmitate oxidation demonstrates an important discrepancy between our data and the dynamics of plasma fatty acid oxidation. In this regard, plasma fatty acid oxidation probably does not contribute directly to the fast component during the initial phase of exercise, because its dynamic is very slow. Our fast time constant averaged 15 s (Table 3) and varied between 21 and 10 s for RERs of 0.78 and 0.97, respectively. In contrast, Havel et al. (16) reported that the time constant for plasma [14C]palmitate oxidation during walking at 3–4 miles/h in the fed state was 93–106 s and increased to 125–172 s in the fasted condition in six men. Notice that fasting slowed the kinetics, which is consistent with our find˙ O2 kinetics are slow when RER is low. For ing that V more vigorous exercise, Friedberg et al. (13) reported a time constant of 126 s for plasma [14C]palmitate oxida˙ O2 compotion. Clearly, the time constant for the fast V nent in our study is an order of magnitude faster than those reported for plasma fatty acid oxidation. There˙ O2 compofore, either our model is wrong or the fast V nent reflects the oxidation of intramuscular TGs.

˙ O2 KINETICS AND RER V ˙ O2 during the restIn summary, the dynamics of V work transition are influenced by RER. Statistically ˙ O2 kinetics as a double exponential and modeling V assigning the fast and slow terms to fat and CHO oxidation, respectively, were shown to accurately pre˙ CO2 over a wide dict steady-state RQ and oxidative V range of RERs for mild exercise. These results are consistent with available data and our theory, which ˙ O2 involves a proposes that the rapid rise in muscle V concerted interaction of modulators of the redox and phosphorylation states that feed back to modify the contraction-induced activation of intramuscular TG lipolysis, b-oxidation, and glycogen utilization. If future experiments confirm the validation of this model and paradigm, then we will have a new quantitative method for assessing the gross dynamics of energy transfer and fat and CHO oxidation, which can be used to investigate various aspects of exercise metabolism and respiratory control in vivo. We thank the volunteers who donated a considerable amount of time and effort in exercise and assessing habitual diets. We acknowledge the assistance of students who helped with data collection: Soleiman Osman, Ngu Pham, Winson Chan, Matt Sheehy, Jeff Trimmer, Art Schoenstadt, Lance Smith, and Christie Horvath. Address for reprint requests and other correspondence: P. A. Mole´, Dept. of Exercise Science, University of California, Davis, Davis, California 95616 (E-mail: [email protected]). Received 13 October 1998; accepted in final form 11 August 1999. REFERENCES 1. Ahlborg, G., and P. Felig. Substrate utilization during prolonged exercise preceded by ingestion of glucose. Am. J. Physiol. 233 (Endocrinol. Metab. Gastrointest. Physiol. 2): E188–E194, 1977. 2. Ahlborg, G., P. Felig, L. Hagenfeldt, R. Hendler, and J. Wahren. Substrate turnover during prolonged exercise in man: splanchic and leg metabolism of glucose, free fatty acids, and amino acids. J. Clin. Invest. 53: 1080–1090, 1974. 3. Barstow, T. J., S. Buchthal, S. Zanconato, and D. M. Cooper. Muscle energetics and pulmonary oxygen uptake kinetics during moderate exercise. J. Appl. Physiol. 77: 1742–1749, 1994. 4. Barstow, T. J., N. Lamarra, and B. J. Whipp. Modulation of muscle and pulmonary O2 uptakes by circulatory dynamics during exercise. J. Appl. Physiol. 68: 979–989, 1990. 5. Barstow, T. J., and P. A. Mole´. Simulation of pulmonary oxygen uptake during exercise in humans. J. Appl. Physiol. 63: 2253– 2261, 1987. 6. Barstow, T. J., and P. A. Mole´. Linear and nonlinear characteristics of oxygen uptake kinetics during heavy exercise. J. Appl. Physiol. 71: 2099–2106, 1991. 7. Behal, R. H., D. B. Buxton, J. G. Robertson, and M. S. Olson. Regulation of the pyruvate dehydrogenase multienzyme complex. Annu. Rev. Nutr. 13: 497–520, 1993. 8. Brooks, G. A., H. Dubouchaud, M. Brown, J. P. Sicurello, and C. Eric Butz. Role of mitochondrial lactate dehydrogenase and lactate oxidation in the intracellular lactate shuttle. Proc. Natl. Acad. Sci. USA 96: 1129–1134, 1999. 9. Chance, B., J. S. Leigh, J. Kent, K. McCully, S. Nioka, B. J. Clark, T. M. Maris, and T. Graham. Multiple controls of oxidative metabolism in living tissues as studied by phosphorus magnetic resonance. Proc. Natl. Acad. Sci. USA 83: 9458–9462, 1986. 10. Connett, R. J., T. E. J. Gayeski, and C. R. Honig. Energy sources in fully aerobic rest-work transitions: a new role for glycolysis. Am. J. Physiol. 248 (Heart Circ. Physiol. 17): H922– H929, 1985. 11. Erecinska, M., and D. F. Wilson. Regulation of cellular energy metabolism. Membr. Biol. 70: 1–14, 1982.

2105

12. Esse´n-Gustavsson, B., and P. A. Tesch. Glycogen and triglyceride utilization in relation to muscle metabolic characteristics in men performing heavy-resistance exercise. Eur. J. Appl. Physiol. 61: 5–10, 1990. 13. Friedberg, S. J., W. R. Harlan, Jr., D. L. Trout, and E. H. Estes, Jr. The effect of exercise on the concentration and turnover of plasma nonesterified fatty acids. J. Clin. Invest. 39: 215–220, 1960. 14. Gaesser, G. A., and D. C. Poole. The slow component of oxygen uptake kinetics in humans. In: Exercise and Sport Sciences Reviews, edited by J. O. Holloszy. Baltimore, MD: Williams & Wilkins, 1996, vol. 24, p. 35–70. 15. Garfinkel, D., M. C. Kohn, and M. J. Achs. Computer simulation of metabolism in pyruvate-perfused rat heart. V. Physiology implications. Am. J. Physiol. 237 (Regulatory Integrative Comp. Physiol. 6): R181–R186, 1979. 16. Havel, R. J., L. G. Ekelund, and A. Holmgren. Kinetic analysis of the oxidation of palmitate-1-14C in man during prolonged heavy muscular exercise. J. Lipid Res. 8: 366–373, 1967. 17. Havel, R. J., A. Naimark, and C. F. Borchgrevink. Turnover rate and oxidation of free fatty acids of blood plasma in man during exercise: studies during continuous infusion of palmitate1-14C. J. Clin. Invest. 42: 1054–1063, 1963. 18. Hopp, J. F., and W. K. Palmer. Effect of electrical stimulation on intracellular triacylglycerol in isolated skeletal muscle. J. Appl. Physiol. 68: 348–354, 1990. 19. Hurley, J. F., P. M. Nemeth, W. H. Martin III, J. M. Hagberg, G. P. Dalsky, and J. O. Holloszy. Muscle triglyceride utilization during exercise: effect of training. J. Appl. Physiol. 60: 562–567, 1986. 20. Issekutz, B., and P. Paul. Intramuscular energy source in exercising normal and pancreatectomized dogs. Am. J. Physiol. 215: 197–204, 1968. 21. Jorfeldt, L., and J. Wahren. Leg blood flow during exercise in man. Clin. Sci. (Colch.) 41: 459–473, 1971. 22. Kiens, B., B. Essen-Gustavson, N. J. Christensen, and B. Saltin. Skeletal muscle substrate utilization during submaximal exercise in man: effect of endurance training. J. Physiol. (Lond.) 62: 469: 459–478, 1993. 23. Klausen, K., N. H. Secher, J. P. Clausen, O. Hartling, and J. Trap-Jensen. Central and regional circulatory adaptations to one-leg training. J. Appl. Physiol. 52: 976–983, 1982. 24. Kushmerick, M. J., R. A. Meyer, and T. R. Brown. Regulation of oxygen consumption in fast- and slow-twitch muscle. Am. J. Physiol. 263 (Cell Physiol. 32): C598–C606, 1992. 25. Linnarsson, D. Dynamics of pulmonary gas exchange and heart rate changes at start and end of exercise. Acta Physiol. Scand. Suppl. 415: 1–68, 1974. 26. McCrory, M. A., T. D. Gomez, E. M. Bernauer, and P. A. Mole´. Evaluation of a new air displacement plethysmograph for measuring human body composition. Med. Sci. Sports Exerc. 27: 1686–1691, 1995. 27. Mole´, P. A. Exercise metabolism. In: Exercise Medicine, edited by A. A. Bove and D. T. Lowenthal. New York: Academic, 1983, p. 43–88. 28. Mole´, P. A., Y. Chung, T. K. Tran, N. Sailasuta, R. Hurd, and T. Jue. Myoglobin desaturation with exercise intensity in human gastrocnemius muscle. Am. J. Physiol. 277 (Regulatory Integrative Comp. Physiol. 46): R173–R180, 1999. 29. Mole´, P. A., and R. L. Coulson. Energetics of myocardial function. Med. Sci. Sports Exerc. 17: 538–545, 1985. 30. Odland, L. M., G. J. F. Heigenhauser, D. Wong, M. G. Hollidge-Horvat, and L. L. Spriet. Effects of increased fat availability on fat-carbohydrate interaction during prolonged exercise in men. Am. J. Physiol. 274 (Regulatory Integrative Comp. Physiol. 43): R894–R902, 1998. 31. Oscai, L. B., J. Gorski, W. C. Miller, and W. K. Palmer. Role of the alkaline TG lipase in regulating intramuscular TG content. Med. Sci. Sports Exerc. 20: 539–544, 1988. 32. Paul, P., and B. Issekutz. Role of extramuscular energy sources in the metabolism of exercising dogs. J. Appl. Physiol. 22: 615–622, 1967.

2106

˙ O2 KINETICS AND RER V

33. Poole, D. C., W. Schaffartzik, D. R. Knight, T. Derion, B. Kennedy, H. J. Guy, R. Prediletto, and P. D. Wagner. Contribution of exercising legs to the slow component of oxygen uptake kinetics in humans. J. Appl. Physiol. 71: 1245–1253, 1991. 34. Rico-Sanz, J., J. V. Hajnal, E. L. Thomas, S. Mierisova´, M. Ala-Korpela, and J. D. Bell. Intracellular and extracellular skeletal muscle triglyeride metabolism during alternating intensity exercise in humans. J. Physiol. (Lond.) 510: 615–622, 1998. 35. Romijn, J. A., E. F. Coyle, L. S. Sidossis, A. Gastaldelli, J. F. Horowitz, E. Endert, and R. R. Wolfe. Regulation of endogenous fat and carbohydrate metabolism in relation to exercise intensity and duration. Am. J. Physiol. 265 (Endocrinol. Metab. 28): E380–E391, 1993. 36. Rowell, L. B. Human cardiovascular adjustments to exercise and thermal stress. Physiol. Rev. 54: 75–159, 1974. 37. Siri, W. E. Body composition from fluid spaces and density: analysis of methods. In: Techniques for Measuring Body Composition, edited by J. Brozek and A. Henschel. Washington, DC: Natl. Acad. Sci./Natl. Res. Council, 1961, p. 223–224. 38. Starling, R. D., T. A. Trappe, A. C. Parcell, C. G. Kerr, W. J. Fink, and D. L. Costill. Effect of diet on muscle triglyceride and endurance performance. J. Appl. Physiol. 82: 1185–1189, 1997. 39. Sullivan, M. J., P. K. Binkley, D. V. Unverferth, and C. V. Leier. Hemodynamic and metabolic responses of the exercising lower limb of humans. J. Lab. Clin. Med. 110: 145–152, 1987.

40. Tschakovsky, M. E., and R. L. Hughson. Interaction of factors determining oxygen uptake at the onset of exercise. J. Appl. Physiol. 86: 1101–1113, 1999. 41. Wahren, J., G. Ahlborg, P. Felig, and L. Jorfeldt. Glucose metabolism during exercise in man. In: Muscle Metabolism During Exercise, edited by B. Pernow and B. Saltin. New York: Plenum, 1971, vol. 11, p. 189–203. 42. Wahren, J., P. Felig, G. Ahlborg, and L. Jorfeldt. Glucose metabolism during leg exercise in man. J. Clin. Invest. 50: 2715–2725, 1971. 43. Wasserman, K., B. J. Whipp, and J. Castagna. Cardiodynamic hyperpnea: hyperpnea secondary to cardiac output increase. J. Appl. Physiol. 36: 457–464, 1974. 44. Whipp, B. J., and M. Mahler. Dynamics of pulmonary gas exchange during exercise. In: Pulmonary Gas Exchange, edited by J. B. West. New York: Academic, 1980, vol. II, p. 33–96. 45. Whipp, B. J., S. A. Ward, N. Lamarra, J. A. Davis, and K. Wasserman. Parameters of ventilatory and gas exchange dynamics during exercise. J. Appl. Physiol. 52: 1506–1513, 1982. 46. Wilmore, J. H. A simplified method for determination of residual lung volume. J. Appl. Physiol. 27: 96–100, 1969. 47. Winder, W. W. Malonyl CoA as a metabolic regulator. In: Biochemistry of Exercise, edited by R. J. Maughan and S. M. Shirreffs. Champaign, IL: Human Kinetics, 1996, vol. IX, p. 173–184. 48. Zierler, K. L. Fatty acids as substrates for heart and skeletal muscle. Am. Heart Assoc. Monogr. 54: 35–39, 1977.