Journal Logo

Basic Sciences: Original Investigations

Lactate distribution in the blood during progressive exercise


Author Information
Medicine & Science in Sports & Exercise: May 1997 - Volume 29 - Issue 5 - p 654-660
  • Free


It is well known that red blood cells (RBCs) have transport systems which move lactate across the RBC membrane (16). Since the rate of lactate production by RBCs is relatively slow, these transport pathways may be most important during exercise when lactate efflux into the blood from exercising muscles can be quite rapid(6,17). Because most of the lactate transport across the RBC membrane is handled by a monocarboxylate carrier which displays saturation kinetics, it is possible that equilibration of RBC lactate concentration ([La]) with plasma [La] could be limited during exercise.

Buono and Yeager (2) were among the first to investigate the distribution of lactate between the plasma and RBCs during exercise. They examined the plasma to RBC [La] gradient during progressive incremental exercise on a cycle ergometer using increment durations of 2 min. Although they observed no [La] gradient at rest or at exercise intensities less than 50% of maximal oxygen uptake (˙VO2max), they did report an increasing gradient at exercise intensities above 50% ˙VO2max. In contrast, Foxdal et al. (9) observed a [La] gradient from plasma to RBCs at rest. During progressive incremental exercise with 5-min increment durations, Foxdal et al. (9) found that the [La] gradient from plasma to RBCs increased with increasing exercise intensity. However, the ratio of plasma [La] to RBC [La] was unchanged from rest through the highest exercise intensity.

Comparison of the results of Buono and Yeager (2) and Foxdal et al. (9) is difficult since the two groups employed different increment durations in their progressive exercise tests. It is a reasonable assumption that progressive incremental exercise protocols with longer increment durations provide more time to achieve lactate equilibrium between the plasma and RBC than protocols with shorter increment durations. If the increment duration is less than the time required for lactate equilibration between the plasma and RBCs, plasma [La] might increase proportionally more than the RBC [La].

Possible differences in the distribution of lactate with differing exercise durations during incremental load tests may have practical application to the detection of the lactate threshold (LT). Certain methods of LT detection would be influenced more by a changing distribution than others. For example, when a fixed plasma [La] is used to detect LT, higher [La] in the plasma relative to RBCs might cause an underestimation of the LT. Such an underestimation may be exacerbated by shorter increment durations. Therefore, the purpose of this investigation was to examine the effect of increment durations of 1 min and 4 min during progressive incremental exercise tests on: 1) the distribution of lactate between plasma and RBCs, and 2) LT detection by three conventional methods using whole blood [La] or plasma [La].


Subjects. Eight male volunteers gave their informed consent to perform two progressive incremental exercise tests on a Monark (Monark Crescent AB, Varberg, Sweden) cycle ergometer. Physical and physiological characteristics of the subjects (mean ± SE) were: age, 22.5 ± 0.6 yr; height, 172.0 ± 2.3 cm; weight, 76.0 ± 3.1 kg; and peak cycle ergometer O2 uptake (˙VO2peak), 42.8 ± 2.0 mL·kg-1·min-1. Although each regularly engaged in some form of exercise, none was specifically trained in any aerobic event.

Progressive incremental exercise tests. The two protocols were separated by 1 wk and performed in random order. Both protocols began with unloaded pedaling at 60 rpm for 4 min. Thereafter, work rate was increased by 30 W at 1-min or 4-min intervals, depending upon the testing protocol.

Inspired minute ventilation (˙VI), expired oxygen fraction(FEO2), and expired carbon dioxide fraction(FECO2) were measured continuously throughout each test and averaged over 30-s intervals. Subjects inspired room air through a Hans Rudolph (Kansas City, MO) valve that directed the expired air into a Plexiglass mixing chamber. ˙VI was measured by a calibrated Parkinson-Cowan (Ventura, CA) gas meter that was fitted with a rotary potentiometer. Mixed expired air was continuously drawn from the mixing chamber through a small tube containing a drying agent (Aquasorb, Mallinckrodt, Paris) and through gas analyzers that measured FEO2 (Ametek S-3A/I) and FECO2 (Beckman Instruments LB-2, Fullerton, CA). The gas analyzers were calibrated with gases from tanks of known concentrations prior to each test. Output from the gas meter and analyzers was directed to an automated gas exchange system(Rayfield, Inc., Waitsfield, VT) which calculated ˙VO2, carbon dioxide output (˙VCO2), and respiratory exchange ratio (R) over 30 s intervals using standard equations. ˙VO2peak was verified by demonstration of at least three of the following criteria: 1) a plateau or decrease in ˙VO2 with increasing work rate; 2) an R of > 1.15; 3) attainment of age predicted maximal heart rate; and 4) an immediate post-exercise whole blood [La] of 8 mM or greater.

Blood sampling and biochemical analysis

At the beginning of each testing session, a 19-gauge catheter (Baxter Healthcare Corp., Deerfield, IL) was inserted into a prominent forearm vein. Arterialization (blood PO2 ≥ 70 mm Hg) of blood samples drawn from this vein was achieved by warming the subject's hand in a heated box that circulated air with a temperature of ≈60°C and by wrapping the upper arm in a heating pad (medium setting). The catheter was flushed with a small amount of sterile saline solution containing 100 USP U·mL-1 of heparin to maintain patency between samples. A total of 6 mL of blood was drawn into two heparinized 3-mL syringes at rest and during the last 30 s of each work rate. The first syringe of each pair was used for whole blood measurements. From this syringe, approximately 1 mL of blood was placed in a spot plate from which 200 μL were immediately transferred to each of two test tubes containing 2 mL of cold perchloric acid. The perchloric acid extracts were frozen (-80°C) for subsequent analysis of lactate. Also from the first syringe approximately 0.5 mL of blood was placed in a test tube from which hematocrit (Hct) samples were drawn. Hct was determined in duplicate by centrifugation for 5 min in an IEC MB Centrifuge. A 4% correction factor was applied for trapped plasma (2,4). The syringe, with the remaining blood sample, was capped and then placed in ice water for the short time before it was analyzed for blood gases and whole blood pH (IL 1304 Blood Gas, pH Analyzer). Subsequently, water content was determined.

To minimize transport of lactate between the plasma and RBCs, the second 3-cc syringe of each sample pair was immediately capped and submerged in a dry ice-ethanol slurry for 10 s and then placed in ice water. Although equilibration of lactate across the RBC membrane occurs quickly at 37°C(4,13,14), equilibration is greatly slowed at 0°C, taking as long as 100 h (13). Our own preliminary experiments showed that 10 s in the dry ice-ethanol slurry was sufficient to drop the temperature to 4°C without hemolysis. At completion of the progressive exercise tests, syringes were centrifuged (2000 g, 15 min, 4°C) with plungers down. This allowed easy separation of the plasma from RBCs. Two-hundred μL of the separated plasma was placed in each of two test tubes containing cold perchloric acid for later lactate analysis. The separated plasma was also analyzed for water content. After removal of the plasma from the second syringe, the remaining packed RBCs were lysed by freezing three times in a dry ice-ethanol slurry; each freeze was interrupted by 10-15 min of thawing. Syringes of lysed RBCs were then centrifuged (2000 g, 15 min, 4°C). The pH of the supernatant was considered to represent RBC pH. Water content of whole blood and plasma was determined by drying samples weighing approximately 200-400 mg at 100°C for 24 h.

Whole blood and plasma [La] were determined by a spectrophotometric modification of the enzymatic technique described by Gutmann and Wahlefeld(11). RBC [La] was calculated from whole blood and plasma lactate concentrations as follows(2,4,5,8,12):Equation

where Cc is the [La] in cell water, Cb is [La] in whole blood water, Cp is the [La] in plasma water, and Hct is the corrected hematocrit.

Methods of LT detection. LT was determined by the following methods: 1) Visual inspection of [La] versus time in the progressive exercise test (Visual); 2) computerized linear regression analysis of a two-segment model plot of log [La] versus log ˙VO2 (Log-Log); and 3) a fixed[La] of 2 mM. A computer program (NLIN, Thomas Clanton, Worthington Concepts, Columbus, OH) was used to fit two regression lines to the Log-Log data and determine their intersection. The ˙VO2 that corresponded with the point of intersection of the two regression lines was considered LT. For the fixed 2 mM method, the time at which the [La] reached 2 mM was determined from a plot of [La] versus time in the progressive exercise test. This time was then inserted into an equation relating exercise time to ˙VO2; the resulting ˙VO2 at the time at which 2 mM [La] was achieved was considered LT. Each method was applied to both whole blood [La] and plasma[La] for both the 1-min and 4-min protocols.

Statistical analysis. Repeated measures ANOVA comparisons of means were used to determine significant differences. Pairwisepost-hoc contrasts were used to determine where significant differences occurred. The 0.05 level was used for statistical significance. The analyses were done with SuperANOVA (Abacus Concepts, Inc., Berkeley, CA), a statistical package for the MacIntosh computer.


It has been shown (15,19) that LT occurs at higher work rates with shorter increment durations in progressive exercise tests. Similar results were found in the present study as illustrated inFigure 1 which shows the blood lactate response to the two exercise protocols for a typical subject. Therefore, to compare the plasma to RBC [La] gradient and RBC:plasma [La] ratio between the 1-min and 4-min protocols of this investigation, data were normalized to each individual's LT for each increment duration. The work rate at which whole blood [La] was closest to, but not exceeding 2 mM, represented LT in each protocol. All eight subjects completed at least two work rates below LT and three work rates above LT during both the 1-min and 4-min protocols. Therefore, lactate distribution patterns were compared at these six work rates during incremental load tests of 1-min and 4-min increment duration.

Figure 2 illustrates changes in whole blood [La], plasma [La], and RBC [La] with increasing work rate for both the 1-min and 4-min increment durations. No significant differences were observed between protocols for whole blood [La] or plasma [La] when the results were normalized to LT. RBC [La] did not significantly differ between protocols until work rates were 90 W above LT. At these work rates, RBC [La] was significantly lower in the 1-min protocol as compared with the 4-min protocol. Within both the 1-min and 4-min protocols, whole blood [La], plasma [La], and RBC [La] at LT were not significantly different from values at the work rates below LT. However, at each work rate above LT, there was a significant increase in whole blood [La], plasma [La], and RBC [La] over that of the preceding work rate during the 1-min protocol. The same was true for the 4-min protocol, except that plasma [La] and RBC [La] did not become significantly increased over previous work rates until work rates 60 W above LT.

There was a significant plasma to RBC [La] gradient at rest (0.74 ± 0.08 mM), and this gradient did not change significantly with increasing work rate up to the LT (Fig. 3). However, with each increase in work rate above LT there was a significant increase in the [La] gradient from plasma to RBCs; the gradient rose to 3.1 ± 0.6 mM at work rates that were 90 W greater than LT.

The RBC:plasma [La] ratio averaged 0.60 ± 0.04 at rest and did not change significantly with increasing work rate either below or above LT(Fig. 3). Neither the plasma to RBC [La] gradient nor the RBC:plasma [La] ratio was significantly different between the two protocols of incremental exercise tests when the work rates were normalized to LT(Fig. 3).

Figure 4 shows the plasma pH and RBC pH responses to progressive exercise in the two different protocols. There was a significant difference between plasma pH and RBC pH at all times. However, no statistical difference was noted in the pH values between the two protocols. Plasma pH at one work rate below LT (-30 W) was not significantly different from the plasma pH at the LT work rate. Plasma pH declined significantly with each work rate above LT. No statistical difference was noted for RBC pH between any of the work rates below LT and LT. Similarly, RBC pH at a work rate of 30 W greater than LT was not significantly different from the values observed at LT. However, RBC pH values at work rates of 60 and 90 W above LT were significantly lower than LT. The plasma and RBC pH values were used to calculate the distribution of hydrogen ions in the blood during the two protocols of progressive exercise. The plasma:RBC [H+] ratio averaged 0.62 ± 0.01 at rest and did not change with increasing work rate either below or above the LT in either the 1-min or 4-min exercise protocols as shown in Figure 5.

Figure 6 summarizes the results of LT determination by three commonly used methods: Visual, Log-Log, and Fixed 2 mM [La]. These three methods were applied to whole blood [La] data and plasma [La] data for both the 1-min and 4-min progressive exercise protocols. When whole blood [La] data were analyzed, there was no significant difference between LTs in the two exercise protocols as determined by the Visual and Log-Log methods. However, when LT was determined by the Fixed 2 mM [La] method, the LT was significantly lower in the 4-min protocol (˙VO2 = 23.9 ± 1.5 mL·kg-1·min-1) than in the 1-min protocol(˙VO2 = 27.2 ± 1.8 mL·kg-1·min-1).

When plasma [La] data were analyzed, there was no significant difference between LTs for the 1-min and 4-min protocols when LT was determined via any of the three methods. However, for both protocols, the LT, as determined by the fixed 2 mM plasma [La] method, was significantly lower than the LT determined by the Visual and Log-Log methods on either plasma [La] data or whole blood [La] data. There was no significant difference in LT via the Visual and Log-Log methods as determined on either plasma or whole blood [La] data.


In this investigation, a [La] gradient was observed between plasma and RBCs at rest and at all levels of progressive incremental exercise. As a result of the [La] gradient at rest, the ratio of RBC:plasma [La] averaged 0.60 ± 0.04. This finding contradicts the report of equal [La] in plasma and RBCs at rest by Buono and Yeager (2), but agrees with most studies of lactate distribution in the blood(3,9,12,14). The RBC:plasma [La] ratio in blood at rest was mirrored by a plasma:RBC [H+] ratio of 0.62± 0.01. Such gradients between plasma and RBCs for lactate and other ions appear to be largely because of a Donnan equilibrium(1,7,10,12,13).

Despite the increasing plasma to RBC [La] gradient at work rates above the LT, there was no significant change in the RBC:plasma [La] ratio in either of the exercise protocols used in this study. This result agrees with of Foxdal et al. (9) who reported an unchanging ratio of approximately 0.44 for a 5-min increment duration. The lack of a change in the ratio with increasing work rate in the 1-min increment duration protocol was surprising. Overall, these observations suggest that there is adequate time for equilibration of lactate between plasma and RBCs even in progressive exercise protocols using an increment duration as short as 1 min. In other words, the relative or proportional increase in both plasma and RBC [La] was the same, thus maintaining a constant ratio between RBC and plasma [La]s. This idea is reinforced by the finding of a constant plasma:RBC [H+] ratio at all work rates as well (Fig. 5).

The major finding of this study was that there were no significant differences between progressive exercise protocols with either 1-min or 4-min increment durations in terms of the [La] gradient between plasma and RBCs or the RBC:plasma [La] ratio. Two important factors should be considered in interpreting the present results. First, exchange of lactate between plasma and RBCs must be minimized in all blood samples to assess the distribution of lactate in the blood at the time of blood collection. The half-time for equilibration of RBC lactate with plasma lactate has been reported to be between 50 and 120 s at 37°C (13,14). Arrest of lactate exchange can be accomplished by immediate, rapid centrifugation(14) or by rapid cooling followed by centrifugation as in the current study. In practical terms, blood sampling as typically performed allows time for considerable exchange of lactate between plasma and RBCs.

A second important consideration is the fact that blood was sampled from a forearm vein during leg exercise in the present study. During cycle ergometry, lactate is most likely released in greatest quantities from the contracting leg muscles. Therefore, some time for uptake of lactate into RBCs from the plasma has elapsed before the blood sample is drawn from the forearm vein. Even if the RBC membrane lactate transport pathways are relatively slow in taking lactate into the RBCs from the plasma, equilibration may be mostly complete in this circulation time. With enough time allowed for equilibration of lactate across the RBC membrane, RBC and plasma [La] should increase at the same relative rate, thus maintaining the resting RBC:plasma [La] ratio during times of increasing lactate entry into the blood. Since the [La] in the plasma at rest is approximately twice that of the RBCs, a five-fold increase in both the plasma and RBC [La] would result in a five-fold increase in the [La] gradient. However, the ratio between the RBC and plasma [La]s would remain unchanged at approximately 0.6.

There have been few reports of RBC pH changes during exercise. In the present study, RBC pH was always lower than plasma pH. However, the declines in plasma pH above LT were accompained by similar declines in RBC pH as illustrated by Figure 4. The unchanging plasma:RBC[H+] ratio with increasing work rate (Fig. 5) indicates that the relative increases in [H+] in plasma and RBCs were the same. However, since the [H+] inside the RBCs was greater to begin with, the absolute increase in RBC [H+] was 1.2 times greater than that of the plasma with increasing work rates above the LT.

Given that the plasma and RBC [La]s change by different amounts during progressive incremental exercise, a practical question is raised regarding the possible effect on LT determination via different methods. In the present study, the Visual and Log-Log methods of LT detection did not yield different estimates (in units of ˙VO2) of the LT regardless of whether whole blood or plasma [La] data were analyzed or whether the exercise protocol used increment durations of 1-min or 4-min (Fig. 6). However, in the 4-min protocol, the LT, as determined by the Fixed 2 mM [La] method on whole blood [La], was significantly lower than LT as determined via the same method in the 1-min protocol and LT as determined via Visual and Log-Log methods in either the 1-min or 4-min exercise protocols. In terms of˙VO2, this difference in LT amounted to about 3 mL·kg-1·min-1 or 7% of ˙VO2peak.

The present data suggest that the most serious problem in LT estimation may arise when the Fixed 2 mM [La] method is applied to plasma [La] data. AsFigure 6 illustrates, LT as determined by this method using either the 1-min or 4-min exercise protocol seriously underestimates the LT in comparison with any other combination of method, increment duration, and blood sample type. In terms of ˙VO2, this underestimation of LT amounted to about 10 mL·kg-1·-1 or 22% of˙VO2peak. In general, this result supports previous studies(8,18) that report that plasma [La] is greater than whole blood [La] at any given ˙VO2 during progressive exercise. Therefore, a given [La] (e.g., 2 mM) will be achieved at a lower˙VO2 or work rate when plasma [La] data are analyzed rather than whole blood [La] data. This difference must be considered when comparisons are made on LT estimates that are based on different types of blood sampling.

In summary, lactate is unequally distributed between the plasma and RBCs at rest. With increasing exercise intensity and concomitant increasing blood[La], the [La] gradient from plasma to RBCs increases. Despite this increase in plasma to RBC [La] gradient, lactate appears to be equilibrated between the plasma and RBCs under the exercise and blood sampling conditions of the present study. This equilibration is evidenced by unchanging ratios of RBC:plasma [La] and plasma:RBC [H+] from rest to low intensity (below LT) exercise to high intensity (above LT) exercise. Therefore, under the conditions of this study, there is no evidence of insufficient transport of lactate from plasma to RBCs.

A practical consequence of the greater absolute increase in plasma [La] with increasing exercise intensity is the fact that LT, as determined by the Fixed 2 mM [La] method, is underestimated in comparison with other methods examined in this investigation. Accordingly, care must be taken in comparing LT determinations from one study to another depending on the LT methods used.

Figure 1-Comparison of LT expressed in Watts between progressive exercise tests of 1-min and 4-min increment duration for a typical subject. LT is represented as the work rate at which whole blood [La] was closest to, but not exceeding, 2 mM. Note the occurrence of LT at a lower work rate during the 4-min protocol than during the 1-min protocol.
Figure 2-Changes in whole blood [La] (a), and changes in plasma and RBC [La] (b) during progressive exercise tests of 1-min and 4-min increment duration. * represents a significant difference between [La] at that work rate and the [La] at the LT work rate; † represents a significant difference between the 1-min and 4-min increment durations(:
P < 0.05).
Figure 3-Changes in plasma to RBC [La] gradient (a), and changes in RBC:plasma [La] ratios (b) during progressive exercise tests with 1-min and 4-min increment durations. * represents a significant increase over value at previous work rate (:
P < 0.05).
Figure 4-Changes in plasma and RBC pH during progressive exercise tests of 1-min and 4-min increment durations. Values are plotted at work rates relative to the work rate at LT. * represents a significant difference between the pH value at that work rate and the value at LT(:
P < 0.05).
Figure 5-Plasma:RBC [H+] ratio values at various work rates relative to LT during progressive exercise tests of 1-min and 4-min increment durations. No significant change was observed with either protocol. Standard errors are within the size of the symbols.
Figure 6-Effect of using whole blood [La] (a) or plasma [La] (b) on LT as determined via three conventional methods during progressive exercise tests of 1-min and 4-min increment duration. * represents a significantly lower estimation of LT using whole blood [La] with the 2 mM method and the 4-min protocol; ** represents a significantly lower estimation of LT with plasma [La] using the 2 mM method than with any other method of LT detection. (:
P < 0.05).


1. Bromberg, P. A., J. Theodore, E. D. Robin, and W. N. Jensen. Anion and hydrogen ion distribution in human blood. J. Lab. Clin. Med. 66:464-475, 1965.
2. Buono, M. J. and J. E. Yeager. Intraerythrocycte and plasma lactate concentration during exercise in humans. Eur. J. Appl. Physiol. 55:326-329, 1986.
3. Consolazio, C. F., R. E. Johnson, and L. J. Pecora.Physiological Measurements of Metabolic Functions in Man. New York: McGraw-Hill, 1963, pp. 125-126.
4. Daniel, S. S., H. O. Morishima, L. S. James, and K. Adamsons. Lactate and pyruvate gradients between red blood cells and plasma during acute asphyxia. J. Appl. Physiol. 19:1100-1104, 1964.
5. Decker, D. G. and J. D. Rosenbaum. The distribution of lactic acid in human blood. Am. J. Physiol. 138:7-11, 1942.
6. Deuticke, B. Monocarboxylate transport in erythrocytes.J. Membr. Biol. 70:89-103, 1982.
7. Fitzsimons, E. J. and J. Sendroy. Distribution of electrolytes in human blood. J. Biol. Chem. 236:1595-1601, 1961.
8. Foxdal, P., A. Sjodin, B. Ostman, and B. Sjodin. The effect of different blood sampling sites and analyses on the relationship between exercise intensity and 4.0 mmol·L-1 blood lactate concentration. Eur. J. Appl. Physiol. 63:52-54, 1991.
9. Foxdal, P., B. Sjodin, H. Rudstam, C. Ostman, B. Ostman, and G. Hedenstierna. Lactate concentration differences in plasma, whole blood, capillary finger blood and erythrocytes during submaximal graded exercise in humans. Eur. J. Physiol. 61:218-222, 1990.
10. Funder, J. and J. O. Wieth. Chloride and hydrogen ion distribution between human red cells and plasma. Acta Physiol. Scand. 68:234-245, 1966.
11. Gutmann, I., and A. W. Wahlefeld. Determination with lactate dehydrogenase and NAD. In: Methods of Enzymatic Analysis, 2nd Ed. U. Bergmeyer (Ed.). New York: Academic Press Inc., 1974, pp. 1464-1468.
12. Harris, R. T. and G. A. Dudley. Exercise alters the distribution of ammonia and lactate in blood. J. Appl. Physiol. 66:313-317, 1989.
13. Johnson, R. E., H. T. Edwards, D. B. Dill, and J. W. Wilson. Blood as a physicochemical system: the distribution of lactate.J. Biol. Chem. 157:461-473, 1945.
14. Juel, C., J. Bangsbo, T. Graham, and B. Saltin. Lactate and potassium fluxes from human skeletal muscle during and after intense, dynamic, knee extensor exercise. Acta Physiol. Scand. 140:147-159, 1990.
15. Kim, S. W., N. Ichimaru, M. Kakimura, and M. Ishii. Effect of work load durations in progressive exercise relationships between lactate and anaerobic thresholds. Ann. Physiol. Anthropol. 7:151-157, 1988.
16. Poole, R. C. and A. P. Halestrap. Transport of lactate and other monocarboxylates across mammalian plasma membranes. Am. J. Physiol. 264:C761-C782, 1993.
17. Roth, D. A. The sarcolemmal lactate transporter: transmembrane determinants of lactate flux. Med. Sci. Sports Exerc. 23:925-934, 1991.
18. Williams, J. R., N. Armstrong, and B. J. Kirby. The influence of the site of sampling and assay medium upon the measurement and interpretation of blood lactate responses to exercise. J. Sports Sci. 10:95-107, 1992.
19. Yoshida, T. Effect of exercise duration during incremental exercise on the determination of anaerobic threshold and the onset of blood lactate accumulation. Eur. J. Appl. Physiol. 53:196-199, 1984.


©1997The American College of Sports Medicine