Effect of endurance training on possible determinants of irO, during heavy exercise RICHARD CASABURI, THOMAS W. STORER, ISSACHAR BEN-DOV, AND KARLMAN WASSERMAN Division of Respiratory Physiology and Medicine, Harbor- UCLA Medical Center, Torrance 90509; and Department of Physical Therapy, University of Southern California, Downey, California 90840 CASABURI,RICHARD,THOMAS W. STORERJSSACHAR BENDov, AND KARLMAN WASSERMAN. Effect of endurance training on possible determinants of VO, during heavy exercise. J. Appl. Physiol. 62( 1): 199-207, 1987.--When moderate exercise begins, 0, uptake (VO,) reaches a steady state within 3 min. However, with heavy exercise, VO, continues to rise beyond 3 min (VO, drift). We sought to identify factors contributing to VO, drift. Ten young subjects performed cycle ergometer tests of 15 min duration for each of four constant work rates, corresponding to 90% of the anaerobic threshold (AT) and 25, 50, and 75% of the difference between maximum VO, tvo, ,,,) \ and AT for that subject. Time courses of \jo2, minute ventilation (VE), and rectal temperature were recorded. Blood lactate, norepinephrine, and epinephrine were measured at the end of exercise. Eight weeks of cycle ergometer endurance training improved average Vo2 In8x by 15%. Subjects then performed four tests identical to pretraining studies. For the above AT tests, training reduced TO, drift substantially; reduction in each of the possible mediators we measured was also demonstrated. The training-induced decrease in VO, drift was well correlated with decreases in end exercise lactate and less well correlated with the drift in VE seen at above AT work rates. The traininginduced reduction in VO, drift was not significantly correlated with attenuation of rectal temperature rise or decrease in endexercise level of the catecholamines. Thus the slow rise in $70~ during heavy exercise seems linked to lactate, though a component dictated by the work of breathing cannot be ruled out. lactate; epinephrine; temperature norepinephrine; work of breathing; body EXERCISE BEGINS, or increases in intensity, Oz uptake (vo2) increases. If the work rate is moderate, the increase will have a half time of -30 s and a new steady state will be achieved within -3 min (7, 33). If the work rate is heavy, however, there will be a delay in reaching the steady state or it may not be achieved at all before exhaustion ensues (20, 32, 33). In this report, we shall designate the delayed rise in VO, as Vo2 drift (though some may find this nomenclature unsatisfactorily vague) and define its magnitude as the difference between Vo2 at end exercise and at 3 min after onset of exercise. More important than definitional issues, however, is that the mediator(s) of To2 drift have not been identified with certainty. We can easily identify four factors which increase in WHEN 0161-7567/87 $1.50 Copyright magnitude as heavy exercise proceeds; all can be predieted to contribute to an increasing Oz demand. The challenge is to separate the major from the minor contributors. 1) Body temperature rises progressively during exercise (1222, 25). By the Q10effect, metabolic rate is raised (30). Hagberg et al. (14) have asserted that body temperature rise, as assessedby changes in rectal temperature, is the predominant mediator of the VO, drift. 2) Both epin.ephrine and norepinephrine rise progressively when exercise exceeds roughly 40% of maximum 0, uptake . wo 2 max)(2, 10). Both of these catecholamines are calorigenic, and thus are potential contributors to VO, increase (8, 27, 28). 3) Ventilation has been observed to drift upward during heavy exercise (17, 20) in a manner qualitatively similar to Vo2. As ventilation increases, so must the work of breathing which contributes to additional VO, (‘26). 4) Serum lactate rises during heavy exercise. Margaria et al. (21) introduced the concept that lactate acid metabolism was linked to Vo2. Clearly, that portion of lactate catabolism which occurs while exercise proceeds and which results in gluconeogenesis will result in increased VO, because gluconeogenesis is an energy requiring process (15). The present study is designed to identify the predominant mediator(s) underlying \io2 drift. It has been observed that endurance training can lower the amount of To2 drift seen at a given heavy work rate (35). We reasoned that the level of any major mediator must also be reduced by endurance training. Furthermore, it may be expected that the reduction in any substantial stimulus and the consequent reduction in VO, drift brought about by endurance training should be well correlated among work rates and among subjects. We thus studied the responses of a group of subjects to a range of work rates before and after 8 wk of endurance training. METHODS The responses of 10 young healthy volunteers form the basis of this report. An eleventh subject was enrolled but was unable to complete the training regimen because of orthopedic problems. All participants were specifically free of cardiac or pulmonary disease and were nonsmokers. These six women and four men were all undergraduate or graduate students at local universities and were 0 1987 the American Physiological Society 199 200 DETERMINANTS OF iTo DURING not aware of the hypotheses being tested. None had engaged in physical training in the previous 6 mo, though several had participated in competitive sports in the past. Their physical characteristics are listed in Table 1. All gave informed consent for their participation in this study. Subjects underwent cycle ergometer exercise testing on three separate occasions before and after undergoing an exercise training program. Each was tested at the same time of day before and after training, though the time of day varied among subjects. Subjects limited themselves to a light meal and consumed no caffeinated beverages before exercise testing. On the first day of testing, the subject performed an incremental exercise test, consisting of 4 min of unloaded cycling, followed by constant rate of increase in work intensity [ramp protocol (31)]) until exhaustion supervened. The rate of increase of work rate was selected as either 20 or 25 W/ min, depending on our assessment of the subject’s relative fitness. In pretraining studies, exhaustion occurred 8-12 min into the ramp protocol, which has been found to allow good discrimination of both anaerobic threshold (AT) and VO, max from gas exchange measurements (3). Both pretraining and posttraining incremental exercise test responses were reviewed independently by at least two of the authors for determination of AT and Voz max, by previously described criteria (3) e From the pretraining incremental exercise tests, a series of four work rates were computed for each subject, the lowest of which would constitute moderate work (work not engendering sustained blood lactate increase) and the highest of which would be very severe exercise for the subject. These work rates were calculated as 1) 90% AT, 2) AT + 25%~& 3) AT + 50%A, and 4) AT + 75%A, where AT is the work rate corresponding to the pretraining anaerobic threshold and A is the difference between the work rates corresponding to the anaerobic threshold and VO, max for that subject. Studies on the second and third day of exercise consisted of 4 min of unloaded cycling followed by 15 min of exercise against one of the four chosen work rates (unless exhaustion occurred before 15 min had elapsed). Two tests were performed on each day, with a minimum of 1 h separating tests. On 1 day the sequence of work intensity was 1 followed by 3, on the other day 2 was followed by 4. Which of these day’s studies was erformed first by a TABLE Age, NO. Yr 1 2 3 4 5 6 7 8 9 10 heart 25 22 22 25 26 23 22 23 23 22 rate is expressed Ht, cm 174 170 163 169 168 185 170 178 175 183 as first Training Heart Rate, beats/min Gender M F F F M M F M F F 4 wk/second 174/186 156/168 180/186 180/188 1711177 156,‘171 177/180 17yl.79 168/174 4 wk. AT, anaerobic EXERCISE given subject was randomly assigned. An important feature of the experimental design was that pre- and posttraining tests were at the same work rate and were of identical duration. Specifically, work rates were not readjusted based on any improvement in exercise tolerance produced by training. Furthermore, in those pretraining tests terminated by exhaustion, the time to exhaustion was noted, and in the posttraining test at that work rate, exercise was stopped at the same time. During these exercise studies, subjects exercised on an electrically braked cycle ergometer (Godart), the work rate profile was computer-generated (Hewlett-Packard, system 1000). Subjects wore a noseclip and respired through a mouthpiece. Both inspired and expired volume were continuously monitored by a turbine volume transducer (Alpha Technologies). The dead space of the mouthpiece-volume transducer assembly was 90 ml. Gas was withdrawn from a point just distal to the mouthpiece at a rate of 1 ml/s by a mass spectrometer (Perkin-Elmer MGA 1100); signals proportional to the fractional concentration of 02, COZ, and N2 were generated. Rectal temperature was assessed by a flexible thermocouple probe (Yellow Springs 702A). Calibration of the temperature probe was confirmed daily using two constant temperature water baths. Analog signals from these devices were transmitted to a 12-channel chart recorder (Beckman Dynograph) for signal conditioning and display. These signals also underwent analog-to-digital conversion 50 times/s by a digital minicomputer (HewlettPackard system 1000). This computer is programmed to utilize these signals to calculate a range of cardiopulmonary variables on a breath-by-breath basis. The details of the calculational algorithms have been previously published (1). Of relevance to the present study, ventilation (VE) is expressed BTPS. To2 is expressed STPD and is corrected breath by breath for variation in lung gas stores [alveolar gas exchange (l)]. These calculations are available on-line on the chart recorder and are also stored on disk and magnetic tape for later analysis. At the end of each constant work rate exercise test, the subject dismounted the cycle ergometer and sat in a chair while a 5 ml blood sample was drawn over approximately 30 s from a superficial antecubital vein. To facilitate venipuncture, light tourniquet was applied, but was left in place less than 20 s in most cases. The time 1. Physical characteristics and responsesto training of subjects Subj Training HEAVY threshold. Pretraining/Posttraining kg AT, l/min vo 2 max, l/min 99194 57/57 62/62 69/69 74166 77176 59158 92188 70170 5184 1.80/1.93 0.92/1.34 1.oy1.37 1.55/1.92 1.17/2.24 1.44/2.02 0.90/1.35 1.80/2.21 1.09/1.79 1.7411.74 3.4y4.00 2.04/2.40 1.90/2.47 2.24/2.48 2.68,‘3.04 3.23/3.45 1.70,‘2.08 3.5113.77 2.42/2.76 2.68/2.94 m DETERMINANTS OF To, elapsing between the end of exercise and the beginning of blood drawing averaged 49 t 12 (SD) s. We attempted to match the pre- and posttraining time before blood drawing and, as a result, the average difference between end exercise and blood drawing for 40 corresponding pairs of studies was only -6.7 t 8.2 s, assuring comparability between pre- and posttraining measurements. A l-ml aliquot from each end exercise blood sample was pipetted and mixed with an equal volume of iced perchlorate. The supernatant was subsequently obtained using a refrigerated centrifuge and then frozen until assayed for lactate concentration by an enzymatic technique. The remainder of the blood sample was placed in an iced tube containing EDTA. Plasma was separated by refrigerated centrifugation and then frozen; subsequent assay for epinephrine and norepinephrine was by radioimmunoassay (23). I3ecause of mishandling of blood samples, two norepinephrine and four epinephrine assays could not be performed (on a total of 80 blood samples). We chose to approximate end-exercise blood levels of lactate and the catecholamines from superficial antecubital vein samples drawn some 50 s after the end of exercise because of the large number of exercise tests in this study. However, several factors mitigate this approximation. In the first few minutes after the end of constant work rate exercise, both blood lactate and catecholamine levels remain constant, or rise slightly (9, 29). Though differences between arterial and deep venous lactate concentrations exist during incremental. exercise (36), these differences will be less during constant-load exercise, in which arterial lactate reaches a near-steady level. Also, since blood samples from corresponding preand posttraining studies were drawn in exactly the same manner, full comparability can be expected. Small differences have been detecte lactate levels of women performing he pending on whether the exercise was performed in the luteal or follicular phase of the menstrual cycle (11, 16) We attempted to perform pre- and posttraining studies in the same phase of the menstrual cycle in each woman (based on dates of menses onset); blood samples dr at the time of exercise were assayed for progesterone to confirm menstrual phase. The endurance training program consisted of exercise on stationary cycle ergometers for 5 days/wk and 45 min/session over an 8-wk period. One of two exercise leaders supervised all sessions. We designed our training regimen after that described by Davis et al. (6). For the first 4 wk, the subjects exercised at a target heart rate which corresponded to the end-exercise heart rate from the pretraining study whose work rate was designed to be halfway between AT and Vo2 max(too& rate 3). The average heart rate for the 10 subjects was 174 beats/min. The target heart rate was increased in the final 4 wk to that seen in the test corresponding to 75% of the difference between AT and vo2 max (work rate 4); the average heart rate was 181 beats/min (see Table 1). Heart rate was frequently checked by the exercise leader by palpation and the work rate adjusted to elicit the target heart rate. The paired t test or analysis of variance was used to DURING HEAVY EXERCISE 201 assess pre- and posttraining differences. Correlation coefficients were calculated to determine relationships among responding variables. Significance was accepted if 1” < 0.05. RESULT'S The improvement in indexes of exercise performance produced by 8 wk of endurance training is documented in Table 1. For the PO subjects, on the average, the anaerobic threshold increased by 0.45 l/min (38%) and the vo2 maxincreased by 0.36 l/min (15%). Appreciable changes in body weight were not seen except in three subjects (I., 5, and 8) who pursued weight reduction diets during the training period (Table 1) oAll subjects reported that the four constant work rate tests, which were conducted at identical work rates pre- and posttraining, were subjectively less stressful in the posttraining period. In four female subjects, progesterone assay demonstrated that the pre- and posttraining studies were in the same phase of her menstrual cycle. Due to a misestimation of the menstrual cycle, subjects 9 and POperformed the pretraining studies in the luteal phase and the posttraining studies in the follicular phase. Data from previous studies (11, 16) suggest that this misestimation would tend to produce a mild underestimation of the decrease in end-exercise lactate produced by endurance training in these two subjects. Each subject performed four constant work rate tests before and after training. In pretraining studies, 8 of 10 subjects could not tolerate the highest work rate for 15 min (the average duration was 9.8 zt 3.4 min). The peak Vo2 reached in the highest work rate pretraining study was quite near the VoB maxdetermined from the incremental exercise study in most subjects. (The constant work rate test, in fact, produced a peak VOW which averaged 1.8% d- 4.3% higher for the 10 subjects.) All subjects could tolerate 15 min of exercise at the three lower work rates. Figure 1 shows the time course of Vo2 for four identical tests before and after training for subject 6. In the pretraining studies, the steady state is achieved within 3 min for the lowest work rate, but is increasingly delayed at the higher work rates. In comparision, the posttraining responses show less delay in achieving the steady state at the higher work rates, and end-exercise Voz is lower. We chose to quantitate the VQ~ drift as the difference between the end-exercise and 3-min VO,. Though other investigators have chosen to fit a second slower exponential to the TjoZ time course (20) or to calculate the difference between 6- and 3-min vo2 (32), we felt our chosen approach facilitated the comparisions to be made in this study. Figure 2 presents the average values for VQ~ drift for each work rate before and after training. Endurance training clearly produced substantial reductions in Voz drift for the work rates above the pretraining AT. In a similar fashion, Fig. 3 reports the average responses of each of the postulated determinants of VO, drift before and after endurance training for each of the four work rates studied. Examining Fig. 3, end-exercise lactate for the lowest work rate observed averaged 1.36 meq/l pretraining, suggesting that for at least some of DETERMINANTS 202 OF ire, DURING HEAVY EXERCISE FIG. 1. Effect of endurance training on time course of 00,. Each panel compares time course of O2 uptake (\jo2) in response to an identical work rate test before (solid Lines) and after (hatched lines) endurance training for subject 8. A-D: responses to work rates of 47,152,197, and 241 W, respectively. These data are expressed as a 9-s moving average to deemphasize breath-tobreath variation in 00,. 0 6 TIME lb 0 l-5 (mid 5 TIME -PRE 3OOr / / a 0 WORK 0 I POST @ RATE FIG. 2. Effect of endurance training on O2 uptake (vo2) drift (difference between vo2 at 3 min and at end exercise) at 4 work rates. Values plotted are average for 10 subjects before (solid Line) and after (dashed line) training. Vertical bars, t SE. the subjects the chosen work rate may have been slightly above the anaerobic threshold. The reduction following training to an average of 0.88 meq/l achieved statistical significance (P < 0.01). At higher work rates, end-exercise lactate increased progressively. Endurance training produced a marked reduction in end-exercise lactate at each of these work rates. Rectal temperature rose over the course of the exercise test at each work rate. The rate of increase of rectal temperature was greater at higher work rates. The explanation for the smaller rectal temperature change at the IO IS (mid highest work rate (Fig. 3) relates to the shorter duration of this exercise test in -most subjects. Figure 4 demonstrates another feature of the rectal temperature time course. In contrast to VOW, where most of the drift occurs early in the exercise study and an apparent steady state is often achieved, rectal temperature change occurs progressively throughout the exercise study, not suggestive of a cause and effect relationship between rectal temperature and Voz drift (see DISCUSSION). Endurance training tended to reduce the magnitude of the rectal temperature drift for all four work rates studied, although statistical significance was achieved only for the two higher work rates (P < 0.05). In the pretraining studies both end-exercise norepinephrine and epinephrine increased dramatically as work rate increased. Figure 3 shows that endurance training radically reduced end-exercise levels of norepinephrine at the higher work rates; there was an 81% reduction at the highest work rate. This trend was even more profound for epinephrine; there was a 91% decrease at the highest work rate. At the lowest (below AT) work rate, ventilation reached a steady state within 3 min. For higher work rates, ventilation continued to rise; the steady state was either delayed or did not occur at all in those studies curtailed by exhaustion. We calculated ventilation drift as the difference between ventilation at 3 min and end exercise; these values are plotted in Fig. 3. Ventilation drift was reduced by endurance training at the three above AT work rates. At the highest work rate the DETERMINANTS OF Voz 0.6~ PRE G 0 Y PRE s if POST Oa4 POST z! 0.2t; ii a O- PRE = E 0.4 \ CI, POST 0 0 WORK 0 @ RATE WORK 0 0 0 WORK RATE RATE @ FIG. 3. Effect of endurance training on possible mediators of O2 uptake (VOW) drift at 4 work rates. Values plotted are average of responses of 10 subjects before (solid lines) and after (dashed lines) endurance training. Vertical bars, t SE. Panels depict end-exercise lactate, rectal temperature rise during exercise, end-exercise norepinephrine, end-exercise epinephrine, and VE drift (difference between VE at 3 min and end exercise). 0.6 F Y 1 -0.2 ! 0 HEAVY 203 EXERCISE the decrease of a given mediator is well correlated with the decrease in Vo2 drift. Table 2 lists the calculated correlations, for each of these possible mediators, between the size of the decrease in the candidate mediator and the size of the decrease inVoz drift for the 40 preand posttraining pairs of studies of identical work rate and duration. Figure 5A presents the data which produced the most striking correlation. Plotted is the relation between end-exercise blood lactate and the size of the Vo2 drift. Lines connect the pretraining to the posttraining responses for each of the 40 pairs of studies in the 10 subjects. The correlation between the size of the decrement in end-exercise lactate and decrement in Voz drift is 0.64 (P < 0.001). The discontinuous lines in Fig. 5A represent those studies which terminated in exhaustion in the pretraining state. In these studies, it may be argued (see DISCUSSION), that the rate of lactate production outstrips the rate of lactate catabolism and blood lactate continues to rise; in this circumstance blood lactate level may be a particularly poor reflection of lactate catabolic rate (and thus may be a poor reflection of the Vo2 cost of lactate catabolism). If the studies involving exhaustive exercise are removed, the correlation between changes in lactate and Vo2 drift rises to 0.81. In contrast, Figure 5B demonstrates the poor relation between changes in rectal temperature and V,Z drift brought about by endurance training. The correlation between the changes in these two variables is not significant (r = 0.15). Also, for the data presented in Fig. 5, C and D, the correlation between changes in epinephrine and in norepinephrine and VO~ drift produced by training failed to reach statistical significance (r = 0.13 and 0.25, respectively). Finally a significant correlation was detected between the change in ventilation drift and in VO, drift resulting from training (Fig. 5E) (r = 0.51, P < O.Ol), though this correlation was somewhat less impressive than that between lactate and 30, drift changes. DISCUSSION 0 W u a DURING II 5 II IO I 15 TIME (mid FIG. 4. Rise in rectal temperature during exercise in subject 6 in reponse to 4 work rates in the pretraining studies. Note that rectal temperature increases progressively throughout each exercise bout. Highest work rate test (224 W) was terminated by exhaustion after about 6 min of exercise (see text). training-induced reduction was 67%. Thus all of the possible mediators of Voz drift we considered were altered by endurance training in a direction consistent with a lower 00~. However, important information relevant to the cause of VO, drift is whether Following the onset of exercise of mild to moderate intensity, the time course of rise of Vo2 is principally dictated by the intramuscular processes underlying aerobic energy production and by changes in body O2 stores (33). Furthermore, the steady-state requirement for a given level of cycle ergometer exercise can be predicted based on knowledge of the efficiency of the energy yield of substrate combustion, which. translates to approximately 10.1 ml/min increase in VO, for each watt incre- 2. Correlation between decrease in drift and decrease in possible mediators of TABLE r End-exercise lactate Rectal temperature rise End-exercise norepinephrine End-exercise epinephrine Ventilation drift Observation vo2, O2 uptake. r = 0.81. 0.64* 0.15 0.25 0.13 0.51 VO, Vo2 drift P <O.OOl NS NS NS co.01 was made in 40 pairs of pre- and posttraining studies. * If studies terminated by exhaustion are eliminated, 204 DETERMINANTS OF 600 DIJRING (mEq/L) 1 EXERCISE T 0.2 r 0.4 A RECTAL 600 HEAVY 1 0 LACTATE voz 0.6 TEMPERATURE 0.8 1.0 (‘Cl t FIG. 5. Relation between reduction in O2 uptake (qo2) drift and reduction in 5 possible mediators of Vo2 drift brought about by endurance training. Lines connect pre- and posttraining responses to identical exercise tests (head of arrow points to posttraining responses). In all panels, ordinate is VO, drift. Plotted on abscissas are end-exercise lactate (dashed lines are for studies curtailed by exhaustion pretraining; see text), rectal temperature rise, end-exercise norepinephrine, end-exercise epinephrine, and minute ventilation (VE) drift. I 2 3 NOREPINEPHRINE 4 (ng /ml 5 1 EPINEPHRINE (rig/ml) 600 / 9~ DRIFT ( LBmln) ment of work rate. For heavy work rates, the situation is more complex and the Voz increase continues beyond 3 min (20, 32, 33). There has been disagreement about the principal mechanism accounting for this continued rise. The es- sense of the problem is that -several processes which might potentially account for TO, drift also increase as heavy exercise proceeds. In the past, investigators have attempted to predict the 0, cost of the observed increase in the level of a given postulated mediator. This may be DETERMINANTS OF ‘iTo, DURING hazardous, as the prediction is only valid if the estimate of the O2 cost is known to be relevant to the situation of heavy exercise. We decided to take an approach to this problem not dependent on prior estimates of the O2 cost of increases in a given mediator. Rather, we studied an intervention in which the Vo2 at a given level of heavy exercise was altered, and observed which of the possible mediators demonstrated parallel alterations. Our findings shed light on the role of four possible mediators. Body temperature. Hagberg et al. concluded that the slow increase in VO, seen during heavy exercise is related (both directly and indirectly) to the increase in body temperature (14). This conclusion was based on the observation of a rectal temperature rise during heavy exercise in comparison with the size of the slow component of Vo2 increase in a group of normal subjects. However, assignment of a precise VO, cost of a given increase in rectal temperature may be subject to error, as Qlo values have not been firmly established for humans. Furthermore, the calculation of temperature-induced O2 cost is made more difficult by the nonuniformity of temperature changes during exercise: temperature of the exercising muscle changes to a greater extent (24) and measures of core thoracic temperature (e.g., esophageal temperature) changes at a slightly faster rate (22). Thus the increase in metabolic rate produced by the Q10 effect ought to be calculated as the integral over all body tissues of varying metabolic rate, mass and temperature change. The use of rectal temperature change in the current study to assess a thermal effect is based on our impression that it represents, as well as any other single estimator, the temperature of the mass of the body’s tissues. Furthermore, our conclusions are based on relative changes rather than absolute values; relative changes are likely to be similar at different measurement sites. In Figs. 3 and 4 we confirm the work of others (e.g., 24), that the size of the thermal stress increases with work rate. We were also able to demonstrate that endurance training decreases the size of the temperature increase (l2), though this achieved significance only at the higher two work rates. However, Fig. 5B clearly shows that the size of the rectal temperature change did not correlate with the size of the Voz drift. Further evidence casting doubt on the candidacy of body temperature change is the dissimilarity of the time courses of these two variables. As seen in Fig. 4, rectal temperature rises rather slowly at first and then increases more steeply, in accord with previous descriptions that a gradual change occurs until about 20 min into the exercise (22, 25). In contrast, Fig. 1 shows that the drift in VO, is most prominent in the first few minutes of heavy exercise, with a tendency to approach a steady state late in the test (except for the highest work rate pretraining study). Catecholamines. Evidence has been obtained in resting subjects that the calorigenic effect of catecholamines may stem in part from the stimulation of lipolysis and glycogenolysis (28). We are not aware of any studies where the metabolic stimulatory effects of epinephrine and norepinephrine have been assessed during exercise. Both Sjostrom et al. (27) and Fellows et al. (8) have recently reported that infusions of epinephrine, which achieved HEAVY EXERCISE 205 serum levels similar to those seen on average in our subjects in the highest work rate pretraining study, increased resting VO, by roughly 30 ml/min. Extrapolation of these resting values to the situation in exercise is likely inappropriate. Study of the response to norepinephrine infusion may have limited relevance, since norepinephrine exerts its effects primarily as a neurotransmitter; a steep gradient exists between the synaptic cleft and blood during endogenous sympathetic stimulation but not during infusion. The data obtained here confirm the trend toward higher serum levels of both epinephrine and norepinephrine with higher work intensities (2, lo), though little elevation in either hormone occurred at the below-AT work level. We further have confirmed the ability of endurance training to substantially decrease both epinephrine and norepinephrine levels at identical work rates (34). We suspect that, in part, perceived emotional stress may have contributed to the markedly high catecholamine levels in the pretraining responses of a few of our subjects (10). However, as can be seen in Fig. 5, dramatic changes in both epinephrine and norepinephrine levels produced by training did not dictate proportional changes in Vo2 drift. This finding seems inconsistent with a substantial role of either catecholamine as a mediator of Vo2 drift. Ventilation. Estimates for the 02 cost of breathing in normal subjects vary widely (26). The data obtained here demonstrate the ability of endurance training to lower the ventilatory requirement for heavy exercise. Figure 5C demonstrates that there was a significant correlation between the size of the ventilatory decrement and the size of the decrease in the Vo2 drift. Thus a contribution to Voz drift from the 02 cost of ventilation seems possible, though a quantitative estimate of the contribution cannot be obtained from these data. An additional issue of some import relates to the mechanism by which TjE drifts upward during heavy exercise. Evidence can be cited for lactic acidosis (19), body temperature (l4), and catecholamine (28) mediation. A companion report (4) discusses the relevance of the results of this study to this controversy. Lactate. The blood level of lactate is only indirectly related to the 02 cost of lactic acid metabolism. The blood level is the resultant of both the rates of lactate production and catabolism. When glycolysis results in lactate formation, ATP is produced without O2 being consumed. Therefore, the production of lactate during heavy exercise is (mildly) O2 sparing for the exercising muscle. However, as exercise proceeds, lactate catabolism ensues principally in the exercising and nonexercising muscle and in the liver (5). Lactate which is metabolized gluconeogenically requires a net expenditure of ATP (and thus has an 0, cost), but the apportionment between gluconeogenic and oxidative fates is not known with precision (5, 15). Attempts to use tracer molecules to dissect out the fine detail of lactate catabolism have been stymied by the complexity of evaluating a substance which originates in the intracellular space (18). Nevertheless, it seems reasonable to suppose that, for work rates at which a near-steady level of lactate can be 206 DETERMINANTS OF Vo, DURING HEAVY EXERCISE findings are sufficient to conclusively estabish that lactate metabolism is, in itself, a major contributor to O2 uptake during heavy exercise. In fact, it has been asserted that reasonable assumptions about lactate catabolic fate lead to calculations (13) which suggest that lactate catabolism cannot be responsible for the entirety of the VO, drift. What may be asserted, on the basis of these data, is that Vo2 drift seems linked to lactate; it is possible that an as yet unidentified O2 requiring process is somehow closely coupled to blood lactate level. Further research will be necessary to clarify this issue. In summary, the present study suggests that the \joz during heavy exercise is linked to the blood lactate level the exercise engenders. A component dictated by the work of breathing seems plausible as well. Body temperature and catecholamine levels do not seem to be changed by endurance training in proportion to changes in the Voz drift, casting doubt on the possibility that they play a substantial role in determining Vo2 during heavy exercise. LACTATE (mEq/L) FIG. 6. Effect of endurance training on relationship between endexercise lactate and O2 uptake (vo2) drift for exercise studies not involving exhaustion. SoLid Line connects pretraining studies; dashed line connects posttraining studies. Points are average (& SE) of responses of 10 subjects to same relative work rates (see text). achieved (i.e., catabolic rate can adjust to the increased production rate), the rate of catabolism is at least monotonically related to the blood level of lactate. For work rates associated with steadily rising lactate levels (those producing exhaustion), the catabolic rate of lactate will have failed to keep pace with the production rate and the blood level will thus poorly reflect the catabolic rate. We have confirmed that endurance training can substantially decrease blood lactate levels during heavy exercise (34) (Fig. 3). As Fig. 5A shows, decrease in VOW drift was well correlated with the lactate decrease, especially when those studies which produced exhaustion in the pretraining state are removed from consideration (as argued above). An additional question that might be asked in evaluating the relation between lactate and Voz drift is whether the quantitative relationship between blood lactate and Vo2 drift is altered by endurance training. Figure 6 shows that the average responses of the 10 subjects to the pretraining work intensities (eliminating the highest work rate, which produced exhaustion in most subjects) and to the four posttraining work intensities. It is clear that training produces a downward shift; a lower Voz drift is associated with a given level of blood lactate. Though any interpretation of this shift must be highly speculative, we can conceive of three plausible explanations: 1) Vop drift is substantially influenced by a mediator besides lactate, and the level of this second mediator is also decreased by endurance training; 2) training produces a systematic decrease in the rate of lactate catabolism (and in the O2 cost of lactate catabolism at a given level of blood lactate); or 3) training produces a shift from a gluconeogenic to an oxidative fate for lactate catabolism. We hasten to emphasize that we do not feel that these The authors thank Bruce Beekley and Quenton Sims for acting as training supervisors, Dr. Marianne Frank for assistance in subject recruitment, and Dr. James Davis for his critical review of the manuscript. We also thank our subjects for their cheerful participation in this research. R. Casaburi is a Trudeau Scholar of the American Lung Association. This investigation was supported by National Heart, Lung, and Blood Institute Grant HL-11907. Received 10 March 1986; accepted in final form 11 August 1986. REFERENCES 1. BEAVER, W. L., N. LAMARRA, AND K. WASSERMAN. Breath-bybreath measurement of true alveolar gas exchange. J. Appl. Physiol. 51: 1662-1675, 1981. 2. BLOOM, S. R., R. H. JOHNSON, D. M. PARK, M. J. RENNIE, AND W. R. SULAIMAN. Differences in the metabolic and hormonal response to exercise between racing cyclists and untrained individuals. J. Physiol. Lond. 258: 1-18, 1976. 3. BUCHFUHRER, M. J., J. E. HANSEN, T. E. ROBINSON, D. Y. SUE, K. WASSERMAN, AND B. J. WHIPP. Optimizing the exercise protocol for cardiopulmonary assessment. J. AppZ. Physiol. 55: 15581564,1983. Endurance train4. CASABURI, R., T. STORER, AND K. WASSERMAN. ing reduces ventilatory demand during heavy exercise (Abstract). Am. Rev. Respir. Dis. 133: 45, 1986. Dia5. COHEN, R. D., AND H. F. WOODS. Lactic acidosis revisited. betes 32: 181-191, 1983. 6. DAVIS, J. A., M. H. FRANK, B. J. WHIPP, AND K. WASSERMAN. Anaerobic threshold alterations caused by endurance training in middle-aged men. J. AppZ. Physiol. 46: 1039-1046, 1979. 7. DIAMOND, L. B., R. CASABURI, K. WASSERMAN, AND B. J. WHIPP. Kinetics of gas exchange and ventilation in transitions from rest or prior exercise. J. AppZ. Physiol. 43: 704-708, 1977. 8. FELLOWS, I. W., T. BENNETT, AND I. A. MACDONALD. The effect of adrenaline upon cardiovascular and metabolic functions in man. CZin. Sci. Lond. 69: 215-222, 1985. 9. FREUND, H., AND P. ZOULOUMIAN. Lactate after exercise in man. I. Evolution kinetics in arterial blood. Eur. J. AppZ. Physiol. Occup. Physiol. 46: 121-133, 1981. and MetaboZic Adaptation to Exercise. New 10. GALBO, H. Hormonal York: Thieme-Stratton, 1983. F., L. STRINDBERG, AND I. WAHLBERG. Female work 11. GAMBERALE, capacity during the menstrual cycle. Physiological and psychological reactions. Stand. J. Work Environ. Health 1: 120-127, 1975. 12. GISOLFI, C., AND S. ROBINSON. Relations between physical training, acclimitization, and heat tolerance. J. AppZ. Physiol. 26: 530534,1969. 13. HAGBERG, J. M., J. P. MULLIN, AND F. J. NAGLE. Effect of work DETERMINANTS 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. OF Vo2 intensity and duration on recovery Oz. J. Appl. Physiol. 48: 540544,198O. HAGBERG, J. M., J. P. MULLIN, AND F. J. NAGLE. Oxygen consumption during constant-load exercise. J. Appl. Physiol. 45: 38l384,1978. JONES, N. L. Hydrogen ion balance during exercise. CZin. Sci. Land. 59: 85-91,198O. JURKOWSKI, J. E. H., N. L. JONES, C. J. TOEWS, AND J. R. SUTTON. Effects of menstrual cycle on blood lactate, O2 delivery, and performance during exercise. J. AppZ. Physiol. 51: 1493-1499, 1981. KATCH, F. I., R. N. GIRANDOLA, AND F. M. HENRY. The influence of the estimated oxygen cost of ventilation on oxygen deficit and recovery oxygen intake for moderately heavy bicycle ergometer exercise. Med. Sci. Sports 4: 71-76, 1972. KATZ, J., F. OKAJIMA, M. CHENOWETH, AND A. DUNN. The determination of lactate turnover in vivo with 3H and 14C-labelled lactate. Biochem. J. 194: 513-524, 1981. KOYAL, S. N., B. J. WHIPP, D. HUNTSMAN, G. A. BRAY, AND K. WASSERMAN. Ventilatory responses to the metabolic acidosis of treadmill and cycle ergometry. J. AppZ. Physiol. 40: 864-867, 1976. LINNARSSON, D. Dynamics of pulmonary gas exchange and heart rate changes at start and end of exercise. Actu Physiol. Stand. Suppl. 415: l-68, 1974. MARGARIA, R., H. T. EDWARDS, AND D. B. DILL. The possible mechanisms of contracting and paying the oxygen debt and the role of lactic acid in muscular contraction. Am. J. Physiol. 106: 689-715,1933. NIELSEN, B., AND M. NIELSEN. Body temperature during work at different environmental temperatures. Acta Physiol. Stand. 56: 120-129,1962. RAUM, W. J. Methods of catecholamine measurement including radioimmunoassay. Am. J. Physiol. 247 (Endocrinol. Metab. 6): E4E12,1984. SALTIN, B., A. P. GAGGE, AND J. A. J. STOLWIJK. Muscle temperature during submaximal exercise in man. J. AppZ. Physiol. 25: View publication stats DURING HEAVY EXERCISE 207 679-688,1968. 25. SALTIN, B., AND L. HERMANSEN. Esophageal, rectal and muscle temperature during exercise. J. AppZ. Physiol. 21: 1757-1762, 1966. 26. SHEPHARD, R. J. The oxygen cost of breathing during vigorous exercise. Q. J. Exp. Physiol. Cog. Med. Sci. 51: 336-350, 1966. 27. SJOSTROM, L., Y. SCHULTZ, F. GUDINCHET, L. HEGNELL, P. G. PITTET, AND S. E. JEQUIER. Epinephrine sensitivity with respect to metabolic rate and other variables in women. Am. J. Physiol. 245 (Endocrinol. Metub. 4): E431-E442, 1983. 28. SVEDMYR, N. Studies on the mechanism for the calorigenic effect of adrenaline in man. Actu Physiol. Stand. 68: 84-95, 1966. 29. WATSON, R. D. S., C. A. HAMILTON, D. H. JONES, J. L. REID, T. J. STALLARD, AND W. A. LITTLER. Sequential changes in plasma noradrenaline during bicycle exercise. CZin. Sci. Lond. 58: 37-43, 1980. 30. WEIBEL, E. The Pathway for Oxygen. Cambridge: Harvard Univ. Press, 1984. 31. WHIPP, B. J., J. A. DAVIS, F. TORRES, AND K. WASSERMAN. A test to determine the parameters of aerobic function during exercise. J. AppZ. Physiol. 50: 217-221, 1981. 32. WHIPP, B. J., AND K. WASSERMAN. Oxygen uptake kinetics for various intensities of constant-load work. J. AppZ. Physiol. 33: 35l356,1972. 33. WHIPP, B. J., AND M. MAHLER. Dynamics of pulmonary gas exchange during exercise. In: Pulmonary Gas Exchange. New York: Academic, 1980, vol. II, p. 33-96. 34. WINDER, W. W., R. C. HICKSON, J. M. HAGBERG, A. A. EHSANI, AND J. A. MCLANE. Training-induced changes in hormonal and metabolic responses to submaximal exercise. J. AppZ. Physiol. 46: 766-771,1979. 35. YOSHIDA, T., Y. SUDA, AND N. TAKEUCHI. Endurance training regimen based upon arterial blood lactate: effects on anaerobic threshold. Eur. J. AppZ. Physiol. Occup. Physiol. 49: 223-230, 1982. 36. YOSHIDA, T., N. TAKEUCHI, AND Y. SUDA. Arterial versus venous blood lactate increase in the forearm during incremental bicycle exercise. Eur. J. AppZ. Physiol. Occup. Physiol. 50: 87-93, 1982.