The following is the abstract of the article discussed in the subsequent letter:
and then decreased over the more dependent regions (zone 4). Thus lung volume has a critical role, and to state that “gravity is a minor determinant of pulmonary GLENNY, ROBB W., WAYNE J. E. LAMM, RICHARD K. ALblood flow distribution” without designating the lung volBERT,AND H. THUMA~ROBERTSON. Grwityisaminordeterminant of pulmonary blood flow distribution. J. Appl. Physiol. ume at which the measurements were made is very misof Glenny et al. (1) 71(Z): 620-629,1991.-Regional pulmonary blood flow in dogs leading. In fact, the measurements made below normal FRC, because we under zone 3 conditions was measuredin supine and prone were apparently posturesto evaluate the linear gravitational model of perfusion know that anesthesia reduces FRC (3); it is therefore not distribution. Flow to regions of lung that were I.9 cm3in volsurprising that the authors found little effect of gravity. ume wasdetermined by injection of radiolabeledmicrospheres A more accurate title for the paper would have been, in both postures. There was marked perfusion heterogeneity “Gravity is a minor determinant of pulmonary blood flow within isogravitational planes (coefficient of variation = distribution in zone 3 at low lung volume.” The results of 42.5%)aswell aswithin gravitational planes(coefficient of varithe study are not at all surprising when the conditions ation = 44.2 and 39.2% in supine and prone postures, respectively; P = 0.02). On average,vertical height explained only 5.8 are clarified. 3) We have never taken the position that perfusion is and 2.4% of the flow variability in the supine and prone postures, respectively. Whereas the gravitational model predicts uniform at a particular level in the lung. Indeed we demperfusion heterogeneity some that regional flows should be negatively correlated when mea- onstrated isogravitational sured in supine and prone postures, flows in the two postures 20 years ago (4, 6) and also proposed a mechanism for were positively correlated, with an r2 of 0.708* 0.050.Regional perfusion as a function of distance from the center of a lung explained 13.4 and 10.8% of the flow variability in the supine and prone postures,respectively. A linear combination of vertical height and radial distance from the centers of each lung provided a better-fitting modelbut still explained only 20.0 and 12.0%of the flow variability in the supineand prone postures, respectively. The entire lung wassearchedfor a region of contiguous lung pieces(22.8 cm3) with high flow. Such a region was found in the dorsal area of the lower lobesin six of sevenanimals, and flow to this region was independent of posture. Under zone 3 conditions, neither gravity nor radial location is the principal determinant of regional perfusion distribution in supine and prone dogs.
nongravitational REFERENCES
1. GLENNY, R. W., W. J. E. LAMM, R. K. ALBERT, AND H. T. ROBERT2.
3. 4.
5.
Gravity and pulmonary blood flow distribution To the Editor: The paper by Glenny et al. (1) is mislead-
ing in several respects. I) The title of the paper, “Gravity is a minor determinant of pulmonary blood flow distribution,” is misleading. The average reader would conclude that gravity generally is a minor determinant in the whole lung, for example, in the upright human lung during chest radiography or scintigraphy. However, the paper deals only with zone 3, which was shown in the original publication (5) to have little topographical inequality of blood flow (see Fig. 8A in Ref. 5). The striking increase in perfusion from nondependent to dependent lung regions in the whole lung is explained by the transition from zone 1 (if one exists) through zone 2 to zone 3. 2) The authors ignore the critical influence of lung volume. The importance of this was pointed out 23 years ago by Hughes et al. (2) when they showed that, whereas at total lung capacity there was a dramatic increase in blood flow down the upright human lung, at residual volume basal blood flow exceeded apical flow. In other words, at residual volume gravity had no influence. At functional residual capacity (FRC), blood flow increased down the lung to a point - 10 cm below the second rib
inhomogeneity (7).
6.
7.
SON. Gravity is a minor determinant of pulmonary blood flow distribution. J. Appl. PhysioZ. 71: 620-629, 1991. HUGHES, J. M. B., J. B. GLAZIER, J. E. MALONEY, AND J. B. WEST. Effect of lung volume on the distribution of pulmonary blood flow in man.Respir. Physiol. 4: 58-72, 1968. REHDER, K., A. D. SESSLER, AND H. M. MARSH. General anesthesia and the lung. Am. Reu. Respir. Dis. 112: 541-563, 1975. WARRELL, D. A., J. W. EVANS, R. 0. CLARKE, G. P. KINGABY, AND J. B. WEST. Patterns of filling in the pulmonary capillary bed. J. Appl. Physiol. 32: 346-356, 1972. WEST, J. B., C.T, DOLLERY, AND A. NAIMARK. Distribution of blood flow in isolated lung: relation to vascularand alveolarpressures. J. Appl. Physiol. 19: 713-724, 1964. WEST, J. B., J. E. MALONEY, AND B. L, CASTLE. Effect of stratified inequality of blood flow on gas exchange in liquid-filled lungs. J. Appl. Physiol. 32: 357-361, 1972. WEST, J. B., A. M. SCHNEIDER, AND M. M. MITCHELL. Recruitment in networks of pulmonary capillaries. J. Apple. Physiol. 39: 976-984, 1975.
John B. West
Department of Medicine University of Culifornia, Sun Diego La Jullu, California 92093-0623 REPLY To the Editor: Dr. West has raised some legitimate concerns about the relevance of our study (2), and we appreciate the opportunity t;o address them. Our study was designed to assess the relationship between gravity and regional perfusion variability by use of high-resolution measurements. As the volume of pieces into which an organ is sectioned decreases, the measured heterogeneity of perfusion increases (1, 3). Because we diced our lungs into relatively small pieces, we observed more perfusion heterogeneity than has previously been described. In quantitating the effect of gravity, we used
the correlation
coefficient
0161-7567/92 $2.00 Copyright 0 1992 the American Physiological
(r’) to measure the proportion
Society
2201
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
2202
LETTERS
TO THE
of perfusion heterogeneity explained by gravity. As the total amount of flow variability increases, the relative proportion explained by gravity must decrease. Hence for the same experimental methods, a greater proportional effect of gravity would be observed by using larger lung pieces or a smaller effect by dicing the lung into even smaller pieces. At finer scales of measurement, nearer the level of gas exchange, the perfusion heterogeneity should be even greater, with the relative contribution of gravity becoming less. When the volume resolution of our study is reduced to that of prior studies by averaging flows within isogravitational planes, our results agree with previous studies of Dr. West and colleagues (4,7,8). Hence our results do not conflict with their experiments but rather with their interpretation. We would therefore like to refocus attention on our most notable observation, that regional perfusion is similar regardless of posture. The mechanism invoked by Dr. West et al. (8) to explain perfusion distribution in the gravitational model is that flow increases “linearly with distance down the lung . . . (because) the degree of distention will depend on the transmural pressure difference . . . which increase(s) through the hydrostatic effect . . . [and hence] the resistance to flow will decrease down the lung.” This hydrostatic mechanism should have its greatest influence at low lung volumes when the vascular compliance is greatest (5). We found that, although regional hydrostatic pressures in a given region differ when animals are supine or prone, regional perfusion changed minimally (see Fig. 5, Ref. 2). This observation alone indicates that an alternate mechanism is needed to explain regional pulmonary perfusion distribution. It is our opinion that a fractal vascular model may provide new insights into this mechanism. Dr. West has raised three general issues that were not fully addressed in our paper. 1) Dr. West points out that studying perfusion under zone 1, 2, and 3 conditions would produce a greater effect of gravity. Although perfusion would be dependent on height up the lung, the perfusion gradient would no longer be due to gravity alone but also on alveolar and intravascular pressures. Dr. West references Fig. 8A in his original publication (8) to demonstrate the topographic equality of regional perfusion under zone 3 conditions. The companion figure (Fig. 8B) shows a lung entirely under zone 3 conditions, in which the regional perfusion has a strong linear relationship to height up the lung and is remarkably similar to Fig. 2A of our paper (2). The lack of perfusion heterogeneity in Dr. West’s study is not due to the absence of a gravitational effect but rather the relatively low spatial resolution attainable by his methods. Subsequent experiments of ours with lung pieces of the same volume as in our previous study (1.9 cm3) showed that the dependence of regional perfusion on height up the lung increases when the lung is in all three zones. However, to produce all three zones, we had to place dogs in a head up position, reduce their cardiac output, and make them hypotensive. In three dogs under these very abnormal conditions, height up the lung still explained only . 18, 25, and 30% of the variability in regional perfusion.
EDITOR
2) We agree with Dr. West that the functional residual capacities of our anesthetized animals were likely less than conscious functional residual capacity at the time of microsphere injection. We have since repeated our studies, injecting the microspheres during tidal ventilation, and have not observed perfusion distributions different from our original findings. \We also studied the effect of varying levels of positive end-expiratory pressure (PEEP) on the distribution of perfusion. Again using piece sizes of 1.9 cm3 in prone dogs, we observed an increase in the vertical perfusion gradient with increasing PEEP. In three animals at 15 cmH,O of PEEP, height explained 13, 20, and 48% of the variability in regional perfusion. Although Dr. West also suggests that the lack of a gravitational effect at our lower lung volumes may be attributable to “zone 4,” we intentionally excluded “zone 4” from our analyses to circumvent this argument (see p. 622, Ref. 2). 3) All of the studies Dr. West cites (to support his statement that he has previously demonstrated isogravitational perfusion heterogeneity) measured perfusion at the microscopic level and not at the macroscopic level of our study (6,9). The mechanism he proposed for isogravitational perfusion heterogeneity (10) was also restricted to the capillary bed and does not explain the marked heterogeneity of flow observed among larger regions. We agree that isogravitational perfusion is heterogeneous. We wish to emphasize that this heterogeneity is not random, because regional flow is spatially organized, with neighboring regions having similar flow (3). Neither the gravitational model nor Dr. West’s proposed mechanism for nongravitational inhomogeneity in the microcirculation can account for this local correlation of regional perfusion. REFERENCES J. B., AND J. H. G. M. VAN BEEK. Lightning and the heart: fractal behavior in cardiac function. Proc. IEEE 76:
1. BASSINGTHWAIGHTE, 693-699,1988. 2. GLENNY, R.
W., W. J. LAMM, R. K. ALBERT, AND H. T. ROBERTSON. Gravity is a minor determinant of pulmonary blood flow distribution. J. Appl. Physiol. 71: 620-629, 1991. Fractal properties of pulmo3. GLENNY, R. W., AND H. T. ROBERTSON. nary blood flow: characterization of spatial heterogeneity. J. Appl. Physiol. 69: 532-545, 4. HUGHES, J. M. B.,
1990.
J. B. GLAZIER, J. E. MALONEY, AND J. B. WEST. Effect of extra-alveolar vessels on the distribution of blood flow in the dog lung. J. Appl. Physiol. 25: 701-712, 1968. 5. SMITH, J. C., AND W. MITZNER. Analysis of pulmonary vascular interdependence in excised dog lobes. J. Appl. Physiol. 48: 450-467, 1980. D, A., J. W. EVANS, R. 0. CLARKE, G. P. KINGABY, AND 6. WARREL, J. B. WEST. Patterns of filling in the pulmonary capillary bed. J. Appl. Physiol. 32: 346-356, 1972. 7. WEST, J. B., AND C. T. DOLLERY. Distribution of blood flow and ventilation-perfusion in the lung, measured with radioactive CO,. J. Appl. Physiol. 15: 405-410, 1960. 8. WEST, J. B., C. T. DOLLERY, AND A. NAIMARK. Distribution of blood flow in isolated lung: relation to vascular and alveolar pressures. J. Appl. Physiol. 55: 1341-1348, 1964.
Robb W. Glenny Division of Pulmonary and Critical Care Medicine University of Washington Seattle, Washington 98195
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
LETTERS
TO
The following is the abstract of the article discussed in the subsequent letter: BINDER, NANCY D., DORINE DAY, FREDERICK C. BATTAGLIA, GIACOMO MESCHIA, AND JOHN W. SPARKS. Role of the circulation in measurement of lactate turnover rate. J. Appl.
Physiol. ‘70(4):2203-2205,1991.-Previous studieshave shown that venous lactate specific activity during arterial tracer lactate infusion differs from arterial lactate specific activity during systemic venous tracer lactate infusion. We performed paired experiments on chronically catheterized rabbits to compare left ventricular (LV) infusion with femoral venous (FV) infusion of L-[U-14C]lactate. Blood was sampledfrom both the femoral artery (FA) and right ventricle (RV) during both modesof infusion. The meanlactate specific activity measured for each combination (infusion site, sampling site) was (FV, FA) 4,380 t 452, (FV, RV) 4,370 t 471, (LV, FA) 4,364 t 239, and (LV, RV) 3,325 I!Z240 (SE) dpm/pmol. Lactate turnover calculated from the specific activity in the (LV, RV) modewas significantly higher than from the other three modes (P < 0.001). Models of lactate turnover are discusseddemonstrating that the (FV, FA) and analogousmodesof infusion sampling measure the turnover rate of lactate molecules that cycle through the circulation. This estimate of turnover is lessthan the turnover rate by the whole organism to the extent that someproduced lactate is metabolized locally without entering the general circulation. The turnover calculated by the (LV, RV) mode overestimates the turnover of circulating lactate and relates to whole body lactate turnover in a complex manner. A- Vprocedure does not overestimate lactate turnover To the Editor: We wish to compliment Binder et al. (1) on their recent paper. Binder et al. present models similar to those in our recent paper (4) and reach some similar conclusions. Moreover, Binder et al. agree with us that infusion of lactate tracer into the descending aorta with blood sampling from a superior great vein (i.e., the arteriovenous A-V mode of tracer infusion and blood sampling) as employed by Katz et al. (6) and Okajima et al. (7) will result in unequal tissue distribution and biased blood sampling, a fact that has been demonstrated experimentally (9). Moreover, Binder et al. conclude that for calculating lactate turnover arterial sampling will be superior to mixed venous sampling when significant lactate turnover occurs in the pulmonary circulation. Although not recognized by Binder et al., the A-V mode of tracer infusion and blood sampling will ignore coronary lactate metabolism (2, 4). Although the compartmental models developed by Binder et al. (1) (Figs. 4 and 5) are similar to our models, we believe that conceptual and analytical errors in the distributed model (Fig. 6) obscure the relationship between the A-V and V-A modes and, ironically, lead Binder et al. to neglect the role of distribution of circulation in causing a difference between turnover estimates from the two types of sampling sites. In the case where tracee is produced and consumed in separate compartments, as in Fig. 4 of Binder et al. (l), arterial specific activity is equal to tissue specific activity, and therefore turnover is correctly measured using arterial sampling. This is the case for glucose but not for lactate, as argued in Lehman and Stanley (5).
THE
EDITOR
2203
If tracee is both produced and consumed in a single compartment, as in Fig. 5 of Binder et al. (l), then arterial specific activity exceeds tissue specific activity, so R, < R, s (using their notation). Indeed, arterial specific activity kxceeds tissue specific activity by the infusion rate divided by the flow of tracee mass from arterial plasma, as was shown previously for lactate (5). We have used this relationship to estimate true tissue (i.e., whole body) turnover from R, and blood flow (4,5). Errors in turnover computed from venous specific activity can be estimated using the same model (Fig. 5, Ref. 1; Fig. 2B, Ref. 4). The result is that R, is always less than R o s. Venous specific activity exceeds tissue specific activity by a fraction of the difference between arterial and tissue specific activity. The fraction is the ratio of the shunt arterial-to-venous flow to the total flow of tracee out of arterial plasma (Ref. 5, Eq. 15; Ref. 4, Eq. 3; Ref. 3, Eq. 8). If all the blood flows through tissues exchanging lactate, this fraction is zero, and Rv = R,,. In no case is R, greater than R. s, as stated by Bindei et al. (1). The factor by which venous specific activity exceeds tissue specific activity is left out of the model in Fig. 6 by assumption. In this distributed model, the entire tissue pool lies in series between arterial and venous blood, without a shunt flow. Binder et al, (1) thereby neglect tissues that receive blood flow but do not exchange tracee in this model. It is the existence of these nonexchanging tissue pools in parallel with the exchanging tissues that causes venous samples to underestimate whole body turnover. Estimation of turnover is complicated by the fact that tissues with different lactate exchanging properties are distributed throughout the body. The model of Fig. 6 addresses the series distribution of lactate exchanging tissues between arterial and venous blood. Arterially infused tracer must fall in concentration with distance along the tissue bed. If we assume that tracee concentration remains constant, specific activity must also decrease along the bed. Therefore, estimates of turnover from samples along the bed must increase from the arterial to the venous end, as concluded by Binder et al. (1). However, the venous specific activity is identical to the whole tissue specific activity in this case of no shunt flow, so a venous sample correctly estimates whole body turnover, and intermediate samples underestimate turnover rate by the whole system. An error in the analysis of the distributed model leads to the incorrect assertion that Rv “bears no fixed relationship to the turnover rate by the whole system.” If the net loss of tracer concentration is to be proportional to its concentration in the blood (Eq. 15), and the fluxes and entry and exit rates of tracee per unit capillary length are to be constant, as stated in the modeling assumptions, then the proportionality constant K must surely be independent of distance along the capillary and not a function of x as it is given in Eq. 16. (If K is indeed a function of X, the step from Eq. 15 to Eq. 17 is certainly wrong.) Either way, the deduced Eqs. 20 and 21 are faulty. The fact that degree of turnover is distributed among tissues is not adequately addressed by modeling the series distribution of tissue with homogeneous lactate ex-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
2204
LETTERS
TO
change properties between arterial and venous blood. The issue is rather the parallel distribution of blood flow acruss tissues with different exchange properties. Analysis of a compartmental model that includes a shunt flow through nonexchanging tissues [like Fig. 5 of Binder et al. (l)] led us to the conclusion that either arterial or venous specific activity exceeds tissue specific activity, so either sample underestimates tissue turnover. Our compartmental analysis leads to two predictions testable from the data of Binder et al. (1): given arterial infusion, arterial specific activity should exceed venous specific activity by a constant; the constant should be the infusion rate divided by the flow of tracee out of the arterial compartment (Ref. 4, Eq. 1). We measure a slope of one for the regression line fit to venous specific activity as a function of arterial specific activity (arterial specific activity exceeds venous specific activity by a constant). With an infusion rate r = 192,000 dpm/min, a blood flow of 607 ml/min, and a lactate concentration of 0.275 pm&ml, we predict an A-V difference of 1,150 dpm/ pmol. The difference recorded by Binder et al. (1) is 4,380 - 3,325 = 1,055 dpm/pmol, an estimate well within the standard errors of the measurements. Binder et al. (1) set out to answer three basic questions: 1) Does the A-V mode of isotope infusion and blood sampling measure lactate turnover by the whole organism? 2) What aspects of lactate metabolism do the A-V and V-A modes measure? and 3) If the V-A procedure underestimates lactate turnover by the whole organism, could the A-V procedure overestimate lactate turnover? We believe that compartmental analysis has already answered all three. The answers are: 1) only if there is no shunt flow through nonexchanging tissues does the A-V mode accurately estimate lactate turnover; 2) both the A-V and V-A modes underestimate lactate turnover, by amounts depending on distribution of blood flow; and 3) lactate turnover is not overestimated by the A-V mode, both modes underestimate tissue turnover. In summary, we concur with must of the conclusions of Binder et al. (l), and applaud their efforts, However, their Eqs. 20 and 21 are faulty and lead to faulty conclusions. REFERENCES 1. BINDER, N. D., D. DAY, F. C. BATTAGLIA, G. MESCHIA, AND J. W. SPARKS. Role of the circulation in measurement of lactate turnover rate. J. Appl. Physiol. 70: 1469-1476, 1991. 2. GERTZ, E. W., J. A. WISNESKI, R. A. NEESE, J. D. BRISTOW, G. L. SEARLE, AND J. T. HANLON. Myocardial lactate metabolism: evidence for lactate release during net chemical extraction in man. Circulation 63: 1273-12’79, 1981. 3. LEHMAN, S. L. Measurement of lactate production by tracer techniques. 1Med. Sci. Sports Exercise 23: 935-938, 1991. 4. LEHMAN, S. L., AND G. A. BROOKS. Obtaining a representative blood sample in lactate tracer studies. Harm. Metub. Res. 22: 470-477, 1990. 5. LEHMAN, S. L., AND W. C. STANLEY. Measuring tracee turnover from tracer specific activity in the steady state. Am. J. Physiol. 255 (Endocrinol. Aletab. 18): E94-E98, 1988. 6. KATZ, J. F., F. OKAJIMA, M. CHENOWETH, AND A. DUNN. The determination of lactate turnover in vivo with 3H- and 14C-labelled lactate: the significance of sites of tracer administration and sampling. Biochem, J. 194: X3-524.1981.
THE
EDITOR
7. OKAJIMA, F., M. CHENOWETH, R. ROGNSTAD, A. DUNN, AND J. KATZ. Metabolism of 3H and 14C-labelled lactate in starved rats. Biochem. J. 194: 525-540, 1981. 8. STANLEY, W. C., AND S. L. LEHMAN. A model for measurement of lactate disappearance with isotopic tracers in man. Biochem. J. 256: 1035-1038, 1988. 9. WZSNESKI, J. A., G. A. BROOKS, R. A. NEESE, W. C. STANLEY, D. L. MORRIS, AND E. W. GERTZ. Tracer studies with aortic infusion result in improper tracer distribution. Harm. Met& Res. 22: 159-164, 1990.
Steven L. Lehman
and George A. Brooks
Department of Physical Education University of California Berkeley, California 94 720 REPLY
2% the Editor: We welcome the opportunity to respond to the letter by Lehman and Brooks, which focuses attention on one of our publications (1) and illustrates how different investigators hold vastly different opinions about the interpretation of turnover measurements in vivo. A resolution of these conflicting views can come only from an open discussion of the basic issues. It is possible to construct many different models to represent the path and fate of labeled and unlabeled lactate within the body. These models will be more or less useful to the extent, that their underlying assumptions agree with what is known about anatomy, blood flow, exchange of molecules across cellular boundaries, and regional tissue differences. We are pleased to note that a paper submitted by Lehman and Brooks (2) after the submission of our paper (1) acknowledges that the anatomy of the circulatory system is not a trivial matter. Unfortunately, experimental evidence bearing on the relevant points is scant; our study was an attempt to obtain some of the answers necessary to make useful assumptions for constructing models. Specifically, tracer was infused into the left ventricle so that all tissues supplied by the systemic circulation would be presented with blood with the same specific activity. Our data showed that the difference in sample specific activity between the “A-V” and “V-A” modes was not the result of an artifact from perfusing parts of the systemic circulation with different specific activities. We did not claim that arterial sampling will be preferable to venous sampling when there is significant turnover by the lungs. We used a method that isolated this part of total turnover in vivo and demonstrated that under the conditions of our study this part of total turnover Can be neglected. With appropriate placement of catheters it would be possible to isolate other regions of interest for in vivo studies. The essential difference between the model of in vivo turnover proposed by Lehman and Stanley (3) and our model is that their model assumes that one does not need to have any information about the metabolic exchange between cells and blood to prove that venous specific activity is identical to the specific activity. In essence, the model claims that end-capillary specific activity represents intracellular specific activity. Our position is radically different,. We maintain that rates of exchange between extracellular and intracellular fluid as well as the
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
LETTERS
TO THE EDITOR
2205
rate of production and consumption within cells are of arterial blood. This shows that some site(s) of net lactate crucial importance in determining whether there is spe- production is anatomically separated from some site(s) cific activity equilibration between end-capillary blood of net consumption within the body. Among the comand sites of metabolism. Even in an organ with homogepound models needed to describe this situation are those nous cells, cells at the arterial and venous ends of the in which R, > R,,. capillary will have intracellular specific activities differWe do not claik that our models are adequate to deent from the average tissue specific activity. Furtherscribe whole body lactate turnover. The purpose of all more, different cell types handle metabolic substrates the models presented in our study (1) is to demonstrate differently even within an anatomically defined tissue. that R, is not equal to Ros in models that are much Some cells are producers, some cells are consumers of a simpler than the whole orginism. The rationale for this given substrate, and some cells are both. Some cell popuapproach is that if Rv does not measure Ro,s in a set of lations exchange the substrate readily with extracellular simple models, there is no basis for the expectation that fluid and some do not. R v = R, ,s in a much more complex system. Lehman and Brooks take exception to our conclusion that a turnover rate calculated on the basis of venous REFERENCES specific activity (R,) bears no fixed relation to the turnover rate by the whole organism (R, s). This is puzzling, BINDER,N. D.,D. DAY, F.C. BATTAGLIAJLMESCHIA, AND J. W. because our study (1) presents a stiaightforward arguSPARKS. Role of the circulation in the measurement of lactate turnover rate. J. A&. Physiol. 70: 1469-1476, 1991. ment to support the conclusion. To summarize, we first LEHMAN, S. L., AND G. A. BROOKS. Obtaining a representative blood considered a hypothetical model in which production and sample in lactate tracer studies. Harm. Metab. Res. 22: 470-77,199O. consumption sites of tracee are anatomically separated. LEHMAN, S. L., ANDW. C, STANLEY.Measuring tracee turnover This model shows unambiguously that turnover rates from tracer specific activity in the steady state. Am. J. Physiol. 255 calculated by use of the specific activity in either the ve(Endocrinol. Metab. 18): E94-E98, 1988. REILLY, P. E., AND L. G. CHANDRASENA. Sheep lactate entry-rate nous effluent of the producing site or the mixed venous measurements: error due to sampling jugular blood. Am. J. Physiol. effluent overestimates whole body turnover (i.e., Rv > 233 (Endocrinol. Metub. Gastrointest. Physiol. 2): E138-E140, 1977. RoS). We then considered the opposite model in which prhduction and metabolic disposal go on within each cell Nancy D. Binder, Giacomo Meschia, John W. Sparks, of a homogenous intracellular pool. We analyzed this secDorine Day, and Frederick C. Battaglia ond model for the relatively simple situation in which Diuisiun of Perinatal Medicine bIood flow rate throughout the tissue capillaries is exDepartments of Pediatrics, Obstetrics-Gynecology, tremely high with respect to the rate of metabolic exand Physiology change and demonstrate that in this particular situation University of Colorado School of Medicine the turnover rate calculated by use of the specific activity Denver, Colorado 80262 in the venous effluent underestimates whole body turnover rate (i.e., Rv < R,,,). Thus, by shifting sites of production and utilization and blood flow rates, one can easily produce models in which the A-V mode of turnover The following is the abstract of the article discussed in calculation yields a result that is either higher or lower the subsequent letter: than the whole body turnover rate. Equations 20 and 21 (1) are not necessary for grasping the essence of our arMACRAE, HOLDENS-H., STEVEN C. DENNIS,ANDREWN. BOSCH,ANDTIMOTHYD. NoAKEs.E~~&w~ trainingonkzctate guments against assuming that there is a fixed relationship between Rv and R, sI These equations define how production and removal during progressive exercise in humans. J. determine whether turnover rates relate to’ tissue blood flow and tracee Appl. Physiol. 72(5): 2205-2207,1992.-To fluxes for the final and more detailed model (Fig. 6) of the reducedblood lactate concentrations [La] during submaxithe study. Lehman and Brooks claim that these equa- mal exercisein humans after endurancetraining result from a decreasedrate of lactate appearance(Ra) or an increasedrate tions are faulty because they find the method of derivaof lactate metabolic clearance (MCR), interrelationships tion objectionable but do not present what, according to among blood [La], lactate Ra, and lactate MCR were investithem, is the correct solution. gatedin eight untrained men during progressiveexercisebefore Lehman and Brooks also object to our conclusion that and after a 9-wk endurancetraining program. Radioisotope dimixed venous specific activity may, in some instances, lution measurementsof L- [U-14C]lactate revealed that the slower rise in blood [La] with increasing 0, uptake (VOW) after overestimate whole body turnover rate. Their objection is based on a division of the body into two groups of training wasdue to a reducedlactate .Raat the lower work rates P < 0.01]. At tissues: those that produce and consume lactate and [VO, < 2.27 l/min, ~60% maximum VO, (VO,,,); those that do not exchange lactate with the blood (2). power outputs closer to maximum, peak lactate Ra values before (215 t 28 pm01 min-’ . kg-l) and after training (244 t, 12 Their analysis is based on model II of Lehman and Stan- pmol mine1 kg-l) became similar. In contrast, submaximal ley (3) in which production and consumption of lactate (~75% voz,,) and peak lactate MCR values were higher after occur within a homogenous tissue compartment. The nec- than before training (40 t 3 vs. 31 t 4 ml min-‘a kg-‘, P < essary assumption that production and consumption are 0.05). Thus the lower blood [La] values during exercise after merged within the body is doubtful. For example, Reilly training in this study were caused by a diminished lactate Ra at and Chandrasena (4) showed that tracee lactate concen- low absolute and relative work rates and an elevated MCR at tration in jugular venous blood is ~20% higher than in higher absolute and all relative work rates during exercise. l
l
l
l
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
2206
LETTERS
Endurance
training
vs. lactate production
TO THE EDITOR
and removal
To the Editor: I read with interest the recent article by MacRae et al. As described in the introduction of the paper, the mechanism(s) responsible for the lower blood lactate concentrations after an endurance training program remains unclear. However, the conclusions reached by MacRae et al. that the rate of appearance of lactate is decreased after training at low relative [i.e., % maximum 0, uptake (% VO 2max)] intensities and the rate of disappearance is increased at high relative intensities need to be clarified. In the first paragraph of RESULTS, the authors describe changes in peak 0, uptake (vo2) that occurred during the 6-min, 40-W incremental test. Mean values for the pretraining data were 2.27 l/min attained at a power output of 220-260 W. hfter training, at a power output close to 300 W, peak VO, was 2.60 l/min. Even with the individual variations that may occur for the 0, cost of cycling at a given power output, one would expect mean values to be close to 3.0 l/min at the power outputs achieved during the pretraining test and -3.9 l/min during the posttraining evaluation. Can the authors explain these large discrepancies between the measured and expected 0, cost of cycling? Further inconsistencies continued for the presentation of the data. For example, MacRae et al. were concerned about the interpretation of changes in the rates of lactate appearance and disappearance at both absolute and relative metabolic rates. From the data presented in their Table 2, they concluded that the rate of appearance was decreased at relative intensities < 60% Tjoamax and the rate of disappearance was increased at intensities > 60% iT02max.The only problem with these conclusions concerns the fact that the mean absolute $70, values presented in their Table 2 do not represent 50 or 60% for the pretraining data and that the peak TO, values represented markedly different % VO, maxlevels. For example, according to the 00, m8xvalues presented in their Table 1 (i.e., 2.59 and 3.45 l/min pre- and posttraining, respectively), absolute VO, values of 1.45, 1.75, and 2.27 l/min (Table 2) correspond to 56,68, and 88% 00, max, respectively, for the pretraining data. The corresponding absolute values posttraining sof 1.73,2.07, and 2.60 l/min represent 50, 60, and 75% VO, max. In Fig. 1, I have shown a graphical representation of the rates of lactate appearance and disappearance at these different relative exercise intensities. Apparently, from this presentation of the data, the conclusion would be that the rate of appearance of lactate is “increased” after training at high relative intensities and that the rate of lactate disappearance is increased at all relative intensities > 50% VO, m8x. These are quite different from the conclusions presented by MacRae et al., and the discrepancies need to be clarified. One additional area of confusion arose in reading RESULTS. In their Table 2, metabolic clearance of lactate is presented as 26 and 24- ml min-’ kg-l at 1.45 and 1.75 l/min Vo2, respectively, for pretraining. However, for the same data presented in their Fig. 4A, mean values for metabolic clearance are close to 35 and 32 ml. min-’ kg-l at these respective metabolic rates. Why do the values in Table 2 and Fig. 4A not agree? l
l
I would greatly appreciate a response to these inconsistencies from the authors and hope that the problems outlined above can be clarified. T. M. Mclellan Defence and Civil Institute of Environmental North York, Ontario M3M 3B9 Canada
Medicine
REPLY
To the Editor: We would like to respond to the comments raised by McLellan regarding our article. In the first place, McLellan expresses concern about the discrepancies between the measured and expected 0, cost of cycling during the 6-min, 40-W incremental exercise test before and after training. We agree with McLellan that the measured 0, cost during the experiments is lower than would be expected. However, the 60, values that should be elicited for those work loads are based on the use of a Monark cycle ergometer (P. 0. Astrand and K. Rodahl. Textbook of Work Physiology: Physiological Bases of Exercise. New York: 300
-
G y 200 2*-5 z
-
E 3
M 100
1Q
/
-1
0
Post-Training
.
!
20
I
.
40
I
I
60 %b2max
80
.
1
100
300
2
h
z 200 -2 .a
m
E co E
Y
100 E
M
Pre-Training
__o__
Post-Training
l
0
- 20
I
I
40
1
I
60 %i02max
I
I
80
1
1
100
FIG, 1. Rates of lactate appearance (Ra, to,p) and disappearance (Rd; bottom). % TO, mBX,% maximal 0, uptake.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
LETTERS
TO
McGraw Hill, 1986, p. 366). We have observed that the VO, values at similar work rates on the Uniwork (Quinton electrically braked) ergometer are lower than those obtained on the Monark. No comparative data are available for the Tunturi ergometer. Because the 0, cost during the progressive exercise tests was lower than we had expected, we compared our data at equivalent metabolic rates (VO,'s) before and after training. By expressing our data in this way, any possible differences in 0, cost due to equipment or to changes in mechanical efficiency of cycling are eliminated. We stand by our conclusion that the rate of lactate appearance was decreased at relative exercise intensities 60% VO 2max after training. We made a transcription/editing error in that our pretraining . vo 2max was reported as 2.59 l/min in Table 1, when in actuality it was 2.89 llmin. The submaximal and peak exercise data reported in Table 2 are correct. The data showing rates of lactate metabolic clearance (MCR) presented in Table 2 should not be confused with the data shown in Fig. 4A, Lactate MCR in Fig. 4A was calculated for each subject by dividing lactate rates of disappearance by blood lactate concentration and then showing these mean values at absolute metabolic rates. The submaximal data in Table 2, excluding the peak exercise data, are calculated estimates of the values at rela-
THE
2207
EDITOR
tive metabolic rates. These data were obtained by fitting y = Ap + C curves to the data shown in Figs. 2B and 3B and then using these equations to calculate the relative lactate concentration and lactate rate of disappearance for each subject. The lactate MCR value for each subject was then calculated with the mean value for all subjects reported in Table 2. Discrepancies between the lactate MCR data in Fig. 4A and the submaximal data in Table 2 are best explained by imperfections inherent in curvefitting of data. We thank McLellan for his interest in this paper and hope that the questions raised by him have been clarified. Holden
MacRae
Natural Science Division Department of Sports Medicine Pepperdine University Malibu, California 90263
Steven C. Dennis, Andrew N. Bosch, and Timothy D. Noakes Liberty Life Chair of Exercise and Sport Science and Bioenergetics of Exercise Research Unit Department of Physiology University of Cape Town Medical School Observatory 7925 Republic of South Africa
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
and then decreased over the more dependent regions (zone 4). Thus lung volume has a critical role, and to state that “gravity is a minor determinant of pulmonary GLENNY, ROBB W., WAYNE J. E. LAMM, RICHARD K. ALblood flow distribution” without designating the lung volBERT,AND H. THUMA~ROBERTSON. Grwityisaminordeterminant of pulmonary blood flow distribution. J. Appl. Physiol. ume at which the measurements were made is very misof Glenny et al. (1) 71(Z): 620-629,1991.-Regional pulmonary blood flow in dogs leading. In fact, the measurements made below normal FRC, because we under zone 3 conditions was measuredin supine and prone were apparently posturesto evaluate the linear gravitational model of perfusion know that anesthesia reduces FRC (3); it is therefore not distribution. Flow to regions of lung that were I.9 cm3in volsurprising that the authors found little effect of gravity. ume wasdetermined by injection of radiolabeledmicrospheres A more accurate title for the paper would have been, in both postures. There was marked perfusion heterogeneity “Gravity is a minor determinant of pulmonary blood flow within isogravitational planes (coefficient of variation = distribution in zone 3 at low lung volume.” The results of 42.5%)aswell aswithin gravitational planes(coefficient of varithe study are not at all surprising when the conditions ation = 44.2 and 39.2% in supine and prone postures, respectively; P = 0.02). On average,vertical height explained only 5.8 are clarified. 3) We have never taken the position that perfusion is and 2.4% of the flow variability in the supine and prone postures, respectively. Whereas the gravitational model predicts uniform at a particular level in the lung. Indeed we demperfusion heterogeneity some that regional flows should be negatively correlated when mea- onstrated isogravitational sured in supine and prone postures, flows in the two postures 20 years ago (4, 6) and also proposed a mechanism for were positively correlated, with an r2 of 0.708* 0.050.Regional perfusion as a function of distance from the center of a lung explained 13.4 and 10.8% of the flow variability in the supine and prone postures,respectively. A linear combination of vertical height and radial distance from the centers of each lung provided a better-fitting modelbut still explained only 20.0 and 12.0%of the flow variability in the supineand prone postures, respectively. The entire lung wassearchedfor a region of contiguous lung pieces(22.8 cm3) with high flow. Such a region was found in the dorsal area of the lower lobesin six of sevenanimals, and flow to this region was independent of posture. Under zone 3 conditions, neither gravity nor radial location is the principal determinant of regional perfusion distribution in supine and prone dogs.
nongravitational REFERENCES
1. GLENNY, R. W., W. J. E. LAMM, R. K. ALBERT, AND H. T. ROBERT2.
3. 4.
5.
Gravity and pulmonary blood flow distribution To the Editor: The paper by Glenny et al. (1) is mislead-
ing in several respects. I) The title of the paper, “Gravity is a minor determinant of pulmonary blood flow distribution,” is misleading. The average reader would conclude that gravity generally is a minor determinant in the whole lung, for example, in the upright human lung during chest radiography or scintigraphy. However, the paper deals only with zone 3, which was shown in the original publication (5) to have little topographical inequality of blood flow (see Fig. 8A in Ref. 5). The striking increase in perfusion from nondependent to dependent lung regions in the whole lung is explained by the transition from zone 1 (if one exists) through zone 2 to zone 3. 2) The authors ignore the critical influence of lung volume. The importance of this was pointed out 23 years ago by Hughes et al. (2) when they showed that, whereas at total lung capacity there was a dramatic increase in blood flow down the upright human lung, at residual volume basal blood flow exceeded apical flow. In other words, at residual volume gravity had no influence. At functional residual capacity (FRC), blood flow increased down the lung to a point - 10 cm below the second rib
inhomogeneity (7).
6.
7.
SON. Gravity is a minor determinant of pulmonary blood flow distribution. J. Appl. PhysioZ. 71: 620-629, 1991. HUGHES, J. M. B., J. B. GLAZIER, J. E. MALONEY, AND J. B. WEST. Effect of lung volume on the distribution of pulmonary blood flow in man.Respir. Physiol. 4: 58-72, 1968. REHDER, K., A. D. SESSLER, AND H. M. MARSH. General anesthesia and the lung. Am. Reu. Respir. Dis. 112: 541-563, 1975. WARRELL, D. A., J. W. EVANS, R. 0. CLARKE, G. P. KINGABY, AND J. B. WEST. Patterns of filling in the pulmonary capillary bed. J. Appl. Physiol. 32: 346-356, 1972. WEST, J. B., C.T, DOLLERY, AND A. NAIMARK. Distribution of blood flow in isolated lung: relation to vascularand alveolarpressures. J. Appl. Physiol. 19: 713-724, 1964. WEST, J. B., J. E. MALONEY, AND B. L, CASTLE. Effect of stratified inequality of blood flow on gas exchange in liquid-filled lungs. J. Appl. Physiol. 32: 357-361, 1972. WEST, J. B., A. M. SCHNEIDER, AND M. M. MITCHELL. Recruitment in networks of pulmonary capillaries. J. Apple. Physiol. 39: 976-984, 1975.
John B. West
Department of Medicine University of Culifornia, Sun Diego La Jullu, California 92093-0623 REPLY To the Editor: Dr. West has raised some legitimate concerns about the relevance of our study (2), and we appreciate the opportunity t;o address them. Our study was designed to assess the relationship between gravity and regional perfusion variability by use of high-resolution measurements. As the volume of pieces into which an organ is sectioned decreases, the measured heterogeneity of perfusion increases (1, 3). Because we diced our lungs into relatively small pieces, we observed more perfusion heterogeneity than has previously been described. In quantitating the effect of gravity, we used
the correlation
coefficient
0161-7567/92 $2.00 Copyright 0 1992 the American Physiological
(r’) to measure the proportion
Society
2201
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
2202
LETTERS
TO THE
of perfusion heterogeneity explained by gravity. As the total amount of flow variability increases, the relative proportion explained by gravity must decrease. Hence for the same experimental methods, a greater proportional effect of gravity would be observed by using larger lung pieces or a smaller effect by dicing the lung into even smaller pieces. At finer scales of measurement, nearer the level of gas exchange, the perfusion heterogeneity should be even greater, with the relative contribution of gravity becoming less. When the volume resolution of our study is reduced to that of prior studies by averaging flows within isogravitational planes, our results agree with previous studies of Dr. West and colleagues (4,7,8). Hence our results do not conflict with their experiments but rather with their interpretation. We would therefore like to refocus attention on our most notable observation, that regional perfusion is similar regardless of posture. The mechanism invoked by Dr. West et al. (8) to explain perfusion distribution in the gravitational model is that flow increases “linearly with distance down the lung . . . (because) the degree of distention will depend on the transmural pressure difference . . . which increase(s) through the hydrostatic effect . . . [and hence] the resistance to flow will decrease down the lung.” This hydrostatic mechanism should have its greatest influence at low lung volumes when the vascular compliance is greatest (5). We found that, although regional hydrostatic pressures in a given region differ when animals are supine or prone, regional perfusion changed minimally (see Fig. 5, Ref. 2). This observation alone indicates that an alternate mechanism is needed to explain regional pulmonary perfusion distribution. It is our opinion that a fractal vascular model may provide new insights into this mechanism. Dr. West has raised three general issues that were not fully addressed in our paper. 1) Dr. West points out that studying perfusion under zone 1, 2, and 3 conditions would produce a greater effect of gravity. Although perfusion would be dependent on height up the lung, the perfusion gradient would no longer be due to gravity alone but also on alveolar and intravascular pressures. Dr. West references Fig. 8A in his original publication (8) to demonstrate the topographic equality of regional perfusion under zone 3 conditions. The companion figure (Fig. 8B) shows a lung entirely under zone 3 conditions, in which the regional perfusion has a strong linear relationship to height up the lung and is remarkably similar to Fig. 2A of our paper (2). The lack of perfusion heterogeneity in Dr. West’s study is not due to the absence of a gravitational effect but rather the relatively low spatial resolution attainable by his methods. Subsequent experiments of ours with lung pieces of the same volume as in our previous study (1.9 cm3) showed that the dependence of regional perfusion on height up the lung increases when the lung is in all three zones. However, to produce all three zones, we had to place dogs in a head up position, reduce their cardiac output, and make them hypotensive. In three dogs under these very abnormal conditions, height up the lung still explained only . 18, 25, and 30% of the variability in regional perfusion.
EDITOR
2) We agree with Dr. West that the functional residual capacities of our anesthetized animals were likely less than conscious functional residual capacity at the time of microsphere injection. We have since repeated our studies, injecting the microspheres during tidal ventilation, and have not observed perfusion distributions different from our original findings. \We also studied the effect of varying levels of positive end-expiratory pressure (PEEP) on the distribution of perfusion. Again using piece sizes of 1.9 cm3 in prone dogs, we observed an increase in the vertical perfusion gradient with increasing PEEP. In three animals at 15 cmH,O of PEEP, height explained 13, 20, and 48% of the variability in regional perfusion. Although Dr. West also suggests that the lack of a gravitational effect at our lower lung volumes may be attributable to “zone 4,” we intentionally excluded “zone 4” from our analyses to circumvent this argument (see p. 622, Ref. 2). 3) All of the studies Dr. West cites (to support his statement that he has previously demonstrated isogravitational perfusion heterogeneity) measured perfusion at the microscopic level and not at the macroscopic level of our study (6,9). The mechanism he proposed for isogravitational perfusion heterogeneity (10) was also restricted to the capillary bed and does not explain the marked heterogeneity of flow observed among larger regions. We agree that isogravitational perfusion is heterogeneous. We wish to emphasize that this heterogeneity is not random, because regional flow is spatially organized, with neighboring regions having similar flow (3). Neither the gravitational model nor Dr. West’s proposed mechanism for nongravitational inhomogeneity in the microcirculation can account for this local correlation of regional perfusion. REFERENCES J. B., AND J. H. G. M. VAN BEEK. Lightning and the heart: fractal behavior in cardiac function. Proc. IEEE 76:
1. BASSINGTHWAIGHTE, 693-699,1988. 2. GLENNY, R.
W., W. J. LAMM, R. K. ALBERT, AND H. T. ROBERTSON. Gravity is a minor determinant of pulmonary blood flow distribution. J. Appl. Physiol. 71: 620-629, 1991. Fractal properties of pulmo3. GLENNY, R. W., AND H. T. ROBERTSON. nary blood flow: characterization of spatial heterogeneity. J. Appl. Physiol. 69: 532-545, 4. HUGHES, J. M. B.,
1990.
J. B. GLAZIER, J. E. MALONEY, AND J. B. WEST. Effect of extra-alveolar vessels on the distribution of blood flow in the dog lung. J. Appl. Physiol. 25: 701-712, 1968. 5. SMITH, J. C., AND W. MITZNER. Analysis of pulmonary vascular interdependence in excised dog lobes. J. Appl. Physiol. 48: 450-467, 1980. D, A., J. W. EVANS, R. 0. CLARKE, G. P. KINGABY, AND 6. WARREL, J. B. WEST. Patterns of filling in the pulmonary capillary bed. J. Appl. Physiol. 32: 346-356, 1972. 7. WEST, J. B., AND C. T. DOLLERY. Distribution of blood flow and ventilation-perfusion in the lung, measured with radioactive CO,. J. Appl. Physiol. 15: 405-410, 1960. 8. WEST, J. B., C. T. DOLLERY, AND A. NAIMARK. Distribution of blood flow in isolated lung: relation to vascular and alveolar pressures. J. Appl. Physiol. 55: 1341-1348, 1964.
Robb W. Glenny Division of Pulmonary and Critical Care Medicine University of Washington Seattle, Washington 98195
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
LETTERS
TO
The following is the abstract of the article discussed in the subsequent letter: BINDER, NANCY D., DORINE DAY, FREDERICK C. BATTAGLIA, GIACOMO MESCHIA, AND JOHN W. SPARKS. Role of the circulation in measurement of lactate turnover rate. J. Appl.
Physiol. ‘70(4):2203-2205,1991.-Previous studieshave shown that venous lactate specific activity during arterial tracer lactate infusion differs from arterial lactate specific activity during systemic venous tracer lactate infusion. We performed paired experiments on chronically catheterized rabbits to compare left ventricular (LV) infusion with femoral venous (FV) infusion of L-[U-14C]lactate. Blood was sampledfrom both the femoral artery (FA) and right ventricle (RV) during both modesof infusion. The meanlactate specific activity measured for each combination (infusion site, sampling site) was (FV, FA) 4,380 t 452, (FV, RV) 4,370 t 471, (LV, FA) 4,364 t 239, and (LV, RV) 3,325 I!Z240 (SE) dpm/pmol. Lactate turnover calculated from the specific activity in the (LV, RV) modewas significantly higher than from the other three modes (P < 0.001). Models of lactate turnover are discusseddemonstrating that the (FV, FA) and analogousmodesof infusion sampling measure the turnover rate of lactate molecules that cycle through the circulation. This estimate of turnover is lessthan the turnover rate by the whole organism to the extent that someproduced lactate is metabolized locally without entering the general circulation. The turnover calculated by the (LV, RV) mode overestimates the turnover of circulating lactate and relates to whole body lactate turnover in a complex manner. A- Vprocedure does not overestimate lactate turnover To the Editor: We wish to compliment Binder et al. (1) on their recent paper. Binder et al. present models similar to those in our recent paper (4) and reach some similar conclusions. Moreover, Binder et al. agree with us that infusion of lactate tracer into the descending aorta with blood sampling from a superior great vein (i.e., the arteriovenous A-V mode of tracer infusion and blood sampling) as employed by Katz et al. (6) and Okajima et al. (7) will result in unequal tissue distribution and biased blood sampling, a fact that has been demonstrated experimentally (9). Moreover, Binder et al. conclude that for calculating lactate turnover arterial sampling will be superior to mixed venous sampling when significant lactate turnover occurs in the pulmonary circulation. Although not recognized by Binder et al., the A-V mode of tracer infusion and blood sampling will ignore coronary lactate metabolism (2, 4). Although the compartmental models developed by Binder et al. (1) (Figs. 4 and 5) are similar to our models, we believe that conceptual and analytical errors in the distributed model (Fig. 6) obscure the relationship between the A-V and V-A modes and, ironically, lead Binder et al. to neglect the role of distribution of circulation in causing a difference between turnover estimates from the two types of sampling sites. In the case where tracee is produced and consumed in separate compartments, as in Fig. 4 of Binder et al. (l), arterial specific activity is equal to tissue specific activity, and therefore turnover is correctly measured using arterial sampling. This is the case for glucose but not for lactate, as argued in Lehman and Stanley (5).
THE
EDITOR
2203
If tracee is both produced and consumed in a single compartment, as in Fig. 5 of Binder et al. (l), then arterial specific activity exceeds tissue specific activity, so R, < R, s (using their notation). Indeed, arterial specific activity kxceeds tissue specific activity by the infusion rate divided by the flow of tracee mass from arterial plasma, as was shown previously for lactate (5). We have used this relationship to estimate true tissue (i.e., whole body) turnover from R, and blood flow (4,5). Errors in turnover computed from venous specific activity can be estimated using the same model (Fig. 5, Ref. 1; Fig. 2B, Ref. 4). The result is that R, is always less than R o s. Venous specific activity exceeds tissue specific activity by a fraction of the difference between arterial and tissue specific activity. The fraction is the ratio of the shunt arterial-to-venous flow to the total flow of tracee out of arterial plasma (Ref. 5, Eq. 15; Ref. 4, Eq. 3; Ref. 3, Eq. 8). If all the blood flows through tissues exchanging lactate, this fraction is zero, and Rv = R,,. In no case is R, greater than R. s, as stated by Bindei et al. (1). The factor by which venous specific activity exceeds tissue specific activity is left out of the model in Fig. 6 by assumption. In this distributed model, the entire tissue pool lies in series between arterial and venous blood, without a shunt flow. Binder et al, (1) thereby neglect tissues that receive blood flow but do not exchange tracee in this model. It is the existence of these nonexchanging tissue pools in parallel with the exchanging tissues that causes venous samples to underestimate whole body turnover. Estimation of turnover is complicated by the fact that tissues with different lactate exchanging properties are distributed throughout the body. The model of Fig. 6 addresses the series distribution of lactate exchanging tissues between arterial and venous blood. Arterially infused tracer must fall in concentration with distance along the tissue bed. If we assume that tracee concentration remains constant, specific activity must also decrease along the bed. Therefore, estimates of turnover from samples along the bed must increase from the arterial to the venous end, as concluded by Binder et al. (1). However, the venous specific activity is identical to the whole tissue specific activity in this case of no shunt flow, so a venous sample correctly estimates whole body turnover, and intermediate samples underestimate turnover rate by the whole system. An error in the analysis of the distributed model leads to the incorrect assertion that Rv “bears no fixed relationship to the turnover rate by the whole system.” If the net loss of tracer concentration is to be proportional to its concentration in the blood (Eq. 15), and the fluxes and entry and exit rates of tracee per unit capillary length are to be constant, as stated in the modeling assumptions, then the proportionality constant K must surely be independent of distance along the capillary and not a function of x as it is given in Eq. 16. (If K is indeed a function of X, the step from Eq. 15 to Eq. 17 is certainly wrong.) Either way, the deduced Eqs. 20 and 21 are faulty. The fact that degree of turnover is distributed among tissues is not adequately addressed by modeling the series distribution of tissue with homogeneous lactate ex-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
2204
LETTERS
TO
change properties between arterial and venous blood. The issue is rather the parallel distribution of blood flow acruss tissues with different exchange properties. Analysis of a compartmental model that includes a shunt flow through nonexchanging tissues [like Fig. 5 of Binder et al. (l)] led us to the conclusion that either arterial or venous specific activity exceeds tissue specific activity, so either sample underestimates tissue turnover. Our compartmental analysis leads to two predictions testable from the data of Binder et al. (1): given arterial infusion, arterial specific activity should exceed venous specific activity by a constant; the constant should be the infusion rate divided by the flow of tracee out of the arterial compartment (Ref. 4, Eq. 1). We measure a slope of one for the regression line fit to venous specific activity as a function of arterial specific activity (arterial specific activity exceeds venous specific activity by a constant). With an infusion rate r = 192,000 dpm/min, a blood flow of 607 ml/min, and a lactate concentration of 0.275 pm&ml, we predict an A-V difference of 1,150 dpm/ pmol. The difference recorded by Binder et al. (1) is 4,380 - 3,325 = 1,055 dpm/pmol, an estimate well within the standard errors of the measurements. Binder et al. (1) set out to answer three basic questions: 1) Does the A-V mode of isotope infusion and blood sampling measure lactate turnover by the whole organism? 2) What aspects of lactate metabolism do the A-V and V-A modes measure? and 3) If the V-A procedure underestimates lactate turnover by the whole organism, could the A-V procedure overestimate lactate turnover? We believe that compartmental analysis has already answered all three. The answers are: 1) only if there is no shunt flow through nonexchanging tissues does the A-V mode accurately estimate lactate turnover; 2) both the A-V and V-A modes underestimate lactate turnover, by amounts depending on distribution of blood flow; and 3) lactate turnover is not overestimated by the A-V mode, both modes underestimate tissue turnover. In summary, we concur with must of the conclusions of Binder et al. (l), and applaud their efforts, However, their Eqs. 20 and 21 are faulty and lead to faulty conclusions. REFERENCES 1. BINDER, N. D., D. DAY, F. C. BATTAGLIA, G. MESCHIA, AND J. W. SPARKS. Role of the circulation in measurement of lactate turnover rate. J. Appl. Physiol. 70: 1469-1476, 1991. 2. GERTZ, E. W., J. A. WISNESKI, R. A. NEESE, J. D. BRISTOW, G. L. SEARLE, AND J. T. HANLON. Myocardial lactate metabolism: evidence for lactate release during net chemical extraction in man. Circulation 63: 1273-12’79, 1981. 3. LEHMAN, S. L. Measurement of lactate production by tracer techniques. 1Med. Sci. Sports Exercise 23: 935-938, 1991. 4. LEHMAN, S. L., AND G. A. BROOKS. Obtaining a representative blood sample in lactate tracer studies. Harm. Metub. Res. 22: 470-477, 1990. 5. LEHMAN, S. L., AND W. C. STANLEY. Measuring tracee turnover from tracer specific activity in the steady state. Am. J. Physiol. 255 (Endocrinol. Aletab. 18): E94-E98, 1988. 6. KATZ, J. F., F. OKAJIMA, M. CHENOWETH, AND A. DUNN. The determination of lactate turnover in vivo with 3H- and 14C-labelled lactate: the significance of sites of tracer administration and sampling. Biochem, J. 194: X3-524.1981.
THE
EDITOR
7. OKAJIMA, F., M. CHENOWETH, R. ROGNSTAD, A. DUNN, AND J. KATZ. Metabolism of 3H and 14C-labelled lactate in starved rats. Biochem. J. 194: 525-540, 1981. 8. STANLEY, W. C., AND S. L. LEHMAN. A model for measurement of lactate disappearance with isotopic tracers in man. Biochem. J. 256: 1035-1038, 1988. 9. WZSNESKI, J. A., G. A. BROOKS, R. A. NEESE, W. C. STANLEY, D. L. MORRIS, AND E. W. GERTZ. Tracer studies with aortic infusion result in improper tracer distribution. Harm. Met& Res. 22: 159-164, 1990.
Steven L. Lehman
and George A. Brooks
Department of Physical Education University of California Berkeley, California 94 720 REPLY
2% the Editor: We welcome the opportunity to respond to the letter by Lehman and Brooks, which focuses attention on one of our publications (1) and illustrates how different investigators hold vastly different opinions about the interpretation of turnover measurements in vivo. A resolution of these conflicting views can come only from an open discussion of the basic issues. It is possible to construct many different models to represent the path and fate of labeled and unlabeled lactate within the body. These models will be more or less useful to the extent, that their underlying assumptions agree with what is known about anatomy, blood flow, exchange of molecules across cellular boundaries, and regional tissue differences. We are pleased to note that a paper submitted by Lehman and Brooks (2) after the submission of our paper (1) acknowledges that the anatomy of the circulatory system is not a trivial matter. Unfortunately, experimental evidence bearing on the relevant points is scant; our study was an attempt to obtain some of the answers necessary to make useful assumptions for constructing models. Specifically, tracer was infused into the left ventricle so that all tissues supplied by the systemic circulation would be presented with blood with the same specific activity. Our data showed that the difference in sample specific activity between the “A-V” and “V-A” modes was not the result of an artifact from perfusing parts of the systemic circulation with different specific activities. We did not claim that arterial sampling will be preferable to venous sampling when there is significant turnover by the lungs. We used a method that isolated this part of total turnover in vivo and demonstrated that under the conditions of our study this part of total turnover Can be neglected. With appropriate placement of catheters it would be possible to isolate other regions of interest for in vivo studies. The essential difference between the model of in vivo turnover proposed by Lehman and Stanley (3) and our model is that their model assumes that one does not need to have any information about the metabolic exchange between cells and blood to prove that venous specific activity is identical to the specific activity. In essence, the model claims that end-capillary specific activity represents intracellular specific activity. Our position is radically different,. We maintain that rates of exchange between extracellular and intracellular fluid as well as the
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
LETTERS
TO THE EDITOR
2205
rate of production and consumption within cells are of arterial blood. This shows that some site(s) of net lactate crucial importance in determining whether there is spe- production is anatomically separated from some site(s) cific activity equilibration between end-capillary blood of net consumption within the body. Among the comand sites of metabolism. Even in an organ with homogepound models needed to describe this situation are those nous cells, cells at the arterial and venous ends of the in which R, > R,,. capillary will have intracellular specific activities differWe do not claik that our models are adequate to deent from the average tissue specific activity. Furtherscribe whole body lactate turnover. The purpose of all more, different cell types handle metabolic substrates the models presented in our study (1) is to demonstrate differently even within an anatomically defined tissue. that R, is not equal to Ros in models that are much Some cells are producers, some cells are consumers of a simpler than the whole orginism. The rationale for this given substrate, and some cells are both. Some cell popuapproach is that if Rv does not measure Ro,s in a set of lations exchange the substrate readily with extracellular simple models, there is no basis for the expectation that fluid and some do not. R v = R, ,s in a much more complex system. Lehman and Brooks take exception to our conclusion that a turnover rate calculated on the basis of venous REFERENCES specific activity (R,) bears no fixed relation to the turnover rate by the whole organism (R, s). This is puzzling, BINDER,N. D.,D. DAY, F.C. BATTAGLIAJLMESCHIA, AND J. W. because our study (1) presents a stiaightforward arguSPARKS. Role of the circulation in the measurement of lactate turnover rate. J. A&. Physiol. 70: 1469-1476, 1991. ment to support the conclusion. To summarize, we first LEHMAN, S. L., AND G. A. BROOKS. Obtaining a representative blood considered a hypothetical model in which production and sample in lactate tracer studies. Harm. Metab. Res. 22: 470-77,199O. consumption sites of tracee are anatomically separated. LEHMAN, S. L., ANDW. C, STANLEY.Measuring tracee turnover This model shows unambiguously that turnover rates from tracer specific activity in the steady state. Am. J. Physiol. 255 calculated by use of the specific activity in either the ve(Endocrinol. Metab. 18): E94-E98, 1988. REILLY, P. E., AND L. G. CHANDRASENA. Sheep lactate entry-rate nous effluent of the producing site or the mixed venous measurements: error due to sampling jugular blood. Am. J. Physiol. effluent overestimates whole body turnover (i.e., Rv > 233 (Endocrinol. Metub. Gastrointest. Physiol. 2): E138-E140, 1977. RoS). We then considered the opposite model in which prhduction and metabolic disposal go on within each cell Nancy D. Binder, Giacomo Meschia, John W. Sparks, of a homogenous intracellular pool. We analyzed this secDorine Day, and Frederick C. Battaglia ond model for the relatively simple situation in which Diuisiun of Perinatal Medicine bIood flow rate throughout the tissue capillaries is exDepartments of Pediatrics, Obstetrics-Gynecology, tremely high with respect to the rate of metabolic exand Physiology change and demonstrate that in this particular situation University of Colorado School of Medicine the turnover rate calculated by use of the specific activity Denver, Colorado 80262 in the venous effluent underestimates whole body turnover rate (i.e., Rv < R,,,). Thus, by shifting sites of production and utilization and blood flow rates, one can easily produce models in which the A-V mode of turnover The following is the abstract of the article discussed in calculation yields a result that is either higher or lower the subsequent letter: than the whole body turnover rate. Equations 20 and 21 (1) are not necessary for grasping the essence of our arMACRAE, HOLDENS-H., STEVEN C. DENNIS,ANDREWN. BOSCH,ANDTIMOTHYD. NoAKEs.E~~&w~ trainingonkzctate guments against assuming that there is a fixed relationship between Rv and R, sI These equations define how production and removal during progressive exercise in humans. J. determine whether turnover rates relate to’ tissue blood flow and tracee Appl. Physiol. 72(5): 2205-2207,1992.-To fluxes for the final and more detailed model (Fig. 6) of the reducedblood lactate concentrations [La] during submaxithe study. Lehman and Brooks claim that these equa- mal exercisein humans after endurancetraining result from a decreasedrate of lactate appearance(Ra) or an increasedrate tions are faulty because they find the method of derivaof lactate metabolic clearance (MCR), interrelationships tion objectionable but do not present what, according to among blood [La], lactate Ra, and lactate MCR were investithem, is the correct solution. gatedin eight untrained men during progressiveexercisebefore Lehman and Brooks also object to our conclusion that and after a 9-wk endurancetraining program. Radioisotope dimixed venous specific activity may, in some instances, lution measurementsof L- [U-14C]lactate revealed that the slower rise in blood [La] with increasing 0, uptake (VOW) after overestimate whole body turnover rate. Their objection is based on a division of the body into two groups of training wasdue to a reducedlactate .Raat the lower work rates P < 0.01]. At tissues: those that produce and consume lactate and [VO, < 2.27 l/min, ~60% maximum VO, (VO,,,); those that do not exchange lactate with the blood (2). power outputs closer to maximum, peak lactate Ra values before (215 t 28 pm01 min-’ . kg-l) and after training (244 t, 12 Their analysis is based on model II of Lehman and Stan- pmol mine1 kg-l) became similar. In contrast, submaximal ley (3) in which production and consumption of lactate (~75% voz,,) and peak lactate MCR values were higher after occur within a homogenous tissue compartment. The nec- than before training (40 t 3 vs. 31 t 4 ml min-‘a kg-‘, P < essary assumption that production and consumption are 0.05). Thus the lower blood [La] values during exercise after merged within the body is doubtful. For example, Reilly training in this study were caused by a diminished lactate Ra at and Chandrasena (4) showed that tracee lactate concen- low absolute and relative work rates and an elevated MCR at tration in jugular venous blood is ~20% higher than in higher absolute and all relative work rates during exercise. l
l
l
l
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
2206
LETTERS
Endurance
training
vs. lactate production
TO THE EDITOR
and removal
To the Editor: I read with interest the recent article by MacRae et al. As described in the introduction of the paper, the mechanism(s) responsible for the lower blood lactate concentrations after an endurance training program remains unclear. However, the conclusions reached by MacRae et al. that the rate of appearance of lactate is decreased after training at low relative [i.e., % maximum 0, uptake (% VO 2max)] intensities and the rate of disappearance is increased at high relative intensities need to be clarified. In the first paragraph of RESULTS, the authors describe changes in peak 0, uptake (vo2) that occurred during the 6-min, 40-W incremental test. Mean values for the pretraining data were 2.27 l/min attained at a power output of 220-260 W. hfter training, at a power output close to 300 W, peak VO, was 2.60 l/min. Even with the individual variations that may occur for the 0, cost of cycling at a given power output, one would expect mean values to be close to 3.0 l/min at the power outputs achieved during the pretraining test and -3.9 l/min during the posttraining evaluation. Can the authors explain these large discrepancies between the measured and expected 0, cost of cycling? Further inconsistencies continued for the presentation of the data. For example, MacRae et al. were concerned about the interpretation of changes in the rates of lactate appearance and disappearance at both absolute and relative metabolic rates. From the data presented in their Table 2, they concluded that the rate of appearance was decreased at relative intensities < 60% Tjoamax and the rate of disappearance was increased at intensities > 60% iT02max.The only problem with these conclusions concerns the fact that the mean absolute $70, values presented in their Table 2 do not represent 50 or 60% for the pretraining data and that the peak TO, values represented markedly different % VO, maxlevels. For example, according to the 00, m8xvalues presented in their Table 1 (i.e., 2.59 and 3.45 l/min pre- and posttraining, respectively), absolute VO, values of 1.45, 1.75, and 2.27 l/min (Table 2) correspond to 56,68, and 88% 00, max, respectively, for the pretraining data. The corresponding absolute values posttraining sof 1.73,2.07, and 2.60 l/min represent 50, 60, and 75% VO, max. In Fig. 1, I have shown a graphical representation of the rates of lactate appearance and disappearance at these different relative exercise intensities. Apparently, from this presentation of the data, the conclusion would be that the rate of appearance of lactate is “increased” after training at high relative intensities and that the rate of lactate disappearance is increased at all relative intensities > 50% VO, m8x. These are quite different from the conclusions presented by MacRae et al., and the discrepancies need to be clarified. One additional area of confusion arose in reading RESULTS. In their Table 2, metabolic clearance of lactate is presented as 26 and 24- ml min-’ kg-l at 1.45 and 1.75 l/min Vo2, respectively, for pretraining. However, for the same data presented in their Fig. 4A, mean values for metabolic clearance are close to 35 and 32 ml. min-’ kg-l at these respective metabolic rates. Why do the values in Table 2 and Fig. 4A not agree? l
l
I would greatly appreciate a response to these inconsistencies from the authors and hope that the problems outlined above can be clarified. T. M. Mclellan Defence and Civil Institute of Environmental North York, Ontario M3M 3B9 Canada
Medicine
REPLY
To the Editor: We would like to respond to the comments raised by McLellan regarding our article. In the first place, McLellan expresses concern about the discrepancies between the measured and expected 0, cost of cycling during the 6-min, 40-W incremental exercise test before and after training. We agree with McLellan that the measured 0, cost during the experiments is lower than would be expected. However, the 60, values that should be elicited for those work loads are based on the use of a Monark cycle ergometer (P. 0. Astrand and K. Rodahl. Textbook of Work Physiology: Physiological Bases of Exercise. New York: 300
-
G y 200 2*-5 z
-
E 3
M 100
1Q
/
-1
0
Post-Training
.
!
20
I
.
40
I
I
60 %b2max
80
.
1
100
300
2
h
z 200 -2 .a
m
E co E
Y
100 E
M
Pre-Training
__o__
Post-Training
l
0
- 20
I
I
40
1
I
60 %i02max
I
I
80
1
1
100
FIG, 1. Rates of lactate appearance (Ra, to,p) and disappearance (Rd; bottom). % TO, mBX,% maximal 0, uptake.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
LETTERS
TO
McGraw Hill, 1986, p. 366). We have observed that the VO, values at similar work rates on the Uniwork (Quinton electrically braked) ergometer are lower than those obtained on the Monark. No comparative data are available for the Tunturi ergometer. Because the 0, cost during the progressive exercise tests was lower than we had expected, we compared our data at equivalent metabolic rates (VO,'s) before and after training. By expressing our data in this way, any possible differences in 0, cost due to equipment or to changes in mechanical efficiency of cycling are eliminated. We stand by our conclusion that the rate of lactate appearance was decreased at relative exercise intensities 60% VO 2max after training. We made a transcription/editing error in that our pretraining . vo 2max was reported as 2.59 l/min in Table 1, when in actuality it was 2.89 llmin. The submaximal and peak exercise data reported in Table 2 are correct. The data showing rates of lactate metabolic clearance (MCR) presented in Table 2 should not be confused with the data shown in Fig. 4A, Lactate MCR in Fig. 4A was calculated for each subject by dividing lactate rates of disappearance by blood lactate concentration and then showing these mean values at absolute metabolic rates. The submaximal data in Table 2, excluding the peak exercise data, are calculated estimates of the values at rela-
THE
2207
EDITOR
tive metabolic rates. These data were obtained by fitting y = Ap + C curves to the data shown in Figs. 2B and 3B and then using these equations to calculate the relative lactate concentration and lactate rate of disappearance for each subject. The lactate MCR value for each subject was then calculated with the mean value for all subjects reported in Table 2. Discrepancies between the lactate MCR data in Fig. 4A and the submaximal data in Table 2 are best explained by imperfections inherent in curvefitting of data. We thank McLellan for his interest in this paper and hope that the questions raised by him have been clarified. Holden
MacRae
Natural Science Division Department of Sports Medicine Pepperdine University Malibu, California 90263
Steven C. Dennis, Andrew N. Bosch, and Timothy D. Noakes Liberty Life Chair of Exercise and Sport Science and Bioenergetics of Exercise Research Unit Department of Physiology University of Cape Town Medical School Observatory 7925 Republic of South Africa
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (148.088.067.084) on January 14, 2019.
