European Journal of

Eur J Appl Physiol (1981) 46:135-147 Applied

Physiology
and Occupational PhysioLogy
9 Springer-Verlag 1981

Lactate After Exercise in Man:

II. Mathematical Model

P. Zouloumian and H. Freund

Centre d'Etudes Bioclimatiques du CNRS,
21, rue Becquerel, F-67087 Strasbourg Cedex, France

Summary. A mathematical model of lactate kinetics after exercise has been

constructed from the application of the mass conservation law and the
following assumptions:
1. The total lactate distribution space is composed of two compartments,
i.e., (M) the previously working muscles and (S) the remaining lactate
space;
2. The rates of lactate release and utilization in (M) and (S) are
proportional to the lactate contents of these compartments;
3. The post-exercise lactate production rates in (M) and (S) are
constants;
4. Arterial lactate concentration can represent the average lactate
concentration in (S).
Consideration of experimental facts reported in the literature shows
these assumptions to be reasonable. The relationships obtained express the
compatibility of parameters and time functions concerning lactate concen-
trations, as well as rates of production, uptake, release, and utilization. They
open the way to various applications, especially those involving numerical
fits to observed time courses of lactate concentrations.
Key words: Two-compartment model - Lactate kinetics, uptake, release,
utilization

Introduction

Constructing a model based on general laws concerning physical phenomena, as

well as putting it to the test through its predictions, constitute a closed-loop
procedure which can progressively allow an improved description of real-
ity.

Offprint requests to: Dr. P. Zouloumian (address see above)

0301-5548/81/0046/0135/$ 02.60
136 P. Zouloumianand H. Freund

Here we develop a mathematical model that, as will become apparent in the

present series of papers, can satisfactorily account for the observed lactate
evolution kinetics in the body.
The expression of the mass conservation law in a volume V where a time
varying lactate concentration L prevails involves rates of production (PR),
metabolic removal ( M R R ) , excretion ( R E X ) , and exchanges [uptake (L U) and
release (LR)] with other compartments:
dL
(PR) - ( R E X ) - ( M R R ) + (LU) - ( L R ) = V . - - (A)
dt
The spatial inhomogeneity of the lactate distribution during or after
muscular exercise is complex, as evidenced by the concentration differences
observed between arterial blood and muscle [3, 15, 21, 22], arterial and femoral
venous blood [1, 11, 15, 20], arterial and brachial venous blood [5, 11, 26], as
well as femoral venous blood and muscle [15]. The rigorous estimation of the
amounts of lactate and its kinetics for the whole body is thus extremely difficult,
since it would require the application of the relation (A) in several
spaces.
As already shown [9, 10], the very accurate hi-exponential fits obtained to
lactate concentrations in arterial blood, which perfuses tissues belonging to the
lactate distribution space, suggest a simplified approach of the problem
considering only two compartments. A first step was taken in this direction by
Freund and Gendry [9], but the need to consider lactate utilization in the
previously working muscles was unfortunately not taken into account in their
model.
The present work is designed to give a more general but still usable model to
predict in the previously working muscles (M) and in the remaining lactate
distribution space (S) time courses of lactate concentrations, rates of lactate
uptake, release, utilization and production, as well as the amounts of lactate
displaced and transformed during recovery.
The values of having such a model include:
a) The considered variables can either represent inputs or outputs of the
investigated system; thus, the mathematical relations binding them will serve as
support allowing confrontation of abstractions and/or observed phenomena.
b) The interpretation of various experimental situations can be undertaken
by means of a well-defined common tool; e.g., it should be possible to compare
active and passive recovery phases, to appraise the influence of the exercises
preceding them (type, duration, power), to study the combined effects of
climatic conditions (exposure to heat, cold, humidity, ionization, hypoxia,
hyperoxia, hyperbaria, etc.), as well as to investigate situations concerning
various muscular groups in man or in animals, in normal or pathologic
cases.
c) The investigation of any phase marked by lactate accumulation, such as
periods of muscular exercise, could be attempted either through direct reasoning
or computer simulation.
d) Finally, a suitable model of lactate movement could be inserted into wider
physiologic schema.
Model of Lactate Exchanges, Utilization, and Production 137

LACTATE PRODUCTION
f
A \
(PR) M = ~ c l (PR)s = 1 dl
(LR) N = a12 . VM . L M ( t )
d
v

LACTATE EXCHANGES
(M) ~(~MS(t) : (LR)N - LU)M~ (s)
LM ( t ) , VM Ls ( t ) , Vs

(LU) M = a21 . Vs Ls ( t )

~mM(t) : ~mS(t) =
c2.VM .kMlt) d2 . Vs . Ls (t)
\ /
v
LACTATE UTILIZATION
Fig. 1. Structure of the model

Definition of the Model

Fundamental assumptions (Fig. 1).

(AS1): The total lactate distribution space (TLS) is composed of two
compartments, (M) of the previously working muscles, and (S) of the remaining
lactate space, having respectively the constant volumes VM and Vs in which the
time-dependent concentrations LM(t) and Ls(t ) prevail.
(AS2): The rates of lactate release and utilization in (M) and (S) are
proportional to the lactate contents of these compartments; i.e., VM 9 LM(t) and
Vs" Ls(t). The coefficients of proportionality c2, d2, a12, and a21 are positive
constants.
(AS3): The post exercise lactate production rates (PR)M = c 1 and (PR)s -- dl
are positive constants;
(AS4): Arterial lactate concentration can represent that average in (S); i.e.,
Ls(t ) ~-La(t ).
In applications [29], fits to arterial concentrations [9, 10] will supply the
analytical expression of La(t ).

Discussion of the Assumptions

The following discussion intends to show that the assumptions are reasonable.
The success of the model predictions will also be a major justification of
them.
138 P. Zouloumian and H. Freund

L Assumptions (AS1) and (AS4)

1. Why Two Compartments? As reported by Karlsson [21], time courses of

muscular and arterial lactate concentrations after exercise show dissimilar forms.
This indicates that at least two compartments must be distinguished in the total
lactate distribution space (TLS). The subdivision of (TLS) into more than two
compartments, for instance one for each organ (passive/active skeletal muscles,
liver, kidney, heart, lung, etc.) undoubtedly would be more realistic, but would
lead to complicated models either requiring unavailable experimental infor-
mation or too many speculative assumptions. The present study will show [29]
that the predictions of even the simple two compartment model (the minimal
solution to the problem) are already subject to stringent experimental
tests.

2. Compartment (M). Some authors utilize local muscular concentrations for

overall lactate disappearance calculations. Thus, they assume either muscular
homogeneity, or at least extended representativeness of muscle samples [15, 21,
28].
Since muscle groups can work unequally, homogeneity with regard to lactate
concentration is not likely. However, a mean concentration Lv(t) can always be
defined in the compartment (M) which groups the worked muscles.

3. Compartment (S). Likewise, the other tissue spaces and the blood space (BS)
perfusing them can be grouped in a single compartment (S) where the average
concentration Ls(t) prevails.

a) The Mean Concentration in (BS). Mitchell and Cournand [25] and Holmgren
[17] showed no significant lactate concentration differences across the lungs.
Since the mixed venous blood concentration results from various local venous
concentrations weighted with local blood flows, either mixed venous or arterial
concentrations should represent the average concentration in (BS) with little
error.
Related precedents can be found in the assumptions of the representa-
tiveness of blood lactate concentrations for extracellular fluids [21] and that of
the homogeneity of the total body water [27]

b) Lactate Concentrations in (BS) and in (S). Since blood supplies all organs,
(BS) links (M) and (S), but the question of an equilibrium between (BS) and
tissue water spaces of (S) must be examined. This was discussed by Rowell et al.
[27], who finally assumed that the total lactate produced is rapidly and evenly
distributed throughout the total body water space. For the estimation of lactate
contents in the body after exercise, Karlsson [21] assumed that the previously
involved muscles concern a given fraction of the total body water, and that:
arterial concentrations represent the remainder composed of (BS) and tissue
water spaces. (AS4) follows this line of reasoning b y assuming that arterial
concentrations are representative of (S). If incomplete equilibrium exists
between blood and tissue water spaces, La(t) would likely overestimate the true
Model of Lactate Exchanges, Utilization, and Production 139

mean, Ls(t). Special investigations would be necessary to show by how

much.
The fact that only two exponentials describe the data, instead of at least three
as would be the case for an inhomogeneous compartment (S), is further evidence
for the mutual validity of (AS1) and (AS4).

II. Assumption (AS2)

Since exponential functions of time describe many biologic processes, the

instantaneous disappearance flow of a given substance is often assumed to be
proportional to its concentration. Concerning lactate specifically, this law was
chosen for its simplicity in a previous model [9]; but several observations support
(AS2).

1. Local and Overall Rates of Lactate Utilization 1. Proportionality between

arterial concentrations La(t ) and lactate removal rates (MRR) in several sites of
the body was mentioned by Rowell et al. [27], who observed it in the human
hepatic splanchnic tissues during exercise.
Likewise, linear relationships were found for the whole body between lactate
turnover rates (or MRR in steady states) and arterial concentrations by Depocas
et al. [6], Minaire et al. [24], Eldridge [8], and Issekutz et al. [18]. On the other
hand, Eldridge et al. [7] reported for similar data a curvilinear relationship which
nevertheless can be closely approximated by a linear dependence over the usual
range of concentrations.
Thus2:

(MRR) = a 9La(t) + ao. (B)

Fits of equation (B) to measurements obtained under various conditions

both in isolated organs [27] and in the whole body [6-8, 18, 24] clearly support
(AS2). The a0 denotes values that have been reported as positive or negative; the
value a0 = 0 is nevertheless consistent with the data within the experimental
uncertainties. The slope coefficients (a) which represent the efficiencies in
removing lactate are positive, and a given value of (a) can describe a given state
of rest [6-8, 18, 24], work [6, 8, 18, 24, 27], or cold exposure [24]. Extending this
observation to the estimation of local and overall (MRR) from arterial blood in
given states of recovery appears to be reasonable, even in situations involving
variations of blood flow, pH, or other factors.
In addition, applying (AS2) to LM(t) instead of La(t) will lead to utilization
rates ~bmM(t) involving two decreasing time exponentials (see the solution).

1 Hermansen and Stensvold [13], Hermansen et al. [14], Belcastro and Bonen [2], McGrail et al.
[23], as well as Bonen et al. [4] report lactate removal rates that agree neither with the detailed
studies of the time evolution of arterial lactate concentrations [9, 10] nor, as the authors [2, 23]
acknowledged, with the relation (A)
2 The amount of lactate removed by excretion is probably very small [14]; thus
(MRR) + (REX) = (MRR)
140 P. Zouloumian and H. Freund

These predictions may perhaps be related to the bi-exponential time evolutions

that Jorfeldt [19] observed for 14CO2-release from active forearm muscle after
injection of labeled lactate. All the experimental results cited above [6-8, 18,
19, 24, 27] may indicate that (AS2) is a general law concerning any lactate
concentrations La(t), LM(t), Ls(t), etc.

2. Muscular Lactate Release and Uptake. According to (AS2) the predicted rates
of muscular lactate release and uptake as well as their net result can be
expressed, respectively:

(LR)M = cq2 " VM" LM(t) , (O

(LU)M = a21" Vs" Ls(t) , (D)

and

~MS(O = a12" VM" LM(O - a21" Vs" Ls(t) . (E)

Simultaneous existence of ( L R ) M and (LU)M, as well as the analytical form

of (LU)M were already observed by Jorfeldt [19]. However, it is necessary to see
if the expression of q~MS(t) can estimate the true exchanges between (M) and (S)
even though blood flow, pH, and other factors do not explicitely appear in
it.
Among others, Graham et al. [12] stated that blood flow is probably a major
limiting factor in lactate movement from active muscle, but careful re-exam-
ination of their results shows that the correlation coefficient between the net
lactate release and blood flow has the small value R = 0.5. Thus, the major
fraction of the variance of the net lactate release can be accounted for by
variables other than blood flow.
Hirche et al. [16], Graham et al. [12], Hermansen and Vaage [15], and
Jorfeldt et al. [20] reported time varying mean rates of lactate release over wide
ranges of blood flow, oxygen uptake, and muscular and arterial lactate
concentrations (up to 1,200; 3,700; 1,700; and 1,100% of the rest values of these
variables, respectively). Since these results do not directly establish the validity
of the expression for ~bMs(t ), some trials were made to express numerically the
observed lactate releases as functions of muscular and blood concentrations:
a) Fits according to the classic laws of simple and facilitated diffusion show ~in
most cases [12, 15, 16] correlation coefficients of 0.82 < R < 0.99. Data of
Jorfeldt et al. [20] indicate the existence of muscular release against significant
muscle - blood lactate gradients. Fits to them give R-~ 0.22.
b) Fits according to ~MS(t) ~ b o + bl " LM(t) - b2 " La(t) provide 0.9
< R < 0.999 in all cases [12, 15, 16, 20] and often appear to be more suitable
than those obtained with the classic diffusion laws. The data of Jorfeldt et al. [20]
lead to b 0 = 0, b 1 < 0 and b 2 > 0; the obtained fit does not agree with the
positive a12 that (AS2) prescribes, and denotes the need of an additional
constant (b0) in the expression of CMS(t). However, this fit agrees with the
leveling off which appears in observed releases. On the other hand, the
Model of Lactate Exchanges, Utilization, and Production 141

expressions for lactate releases observed by the other authors [12, 15, 16] are
compatible with b0 ~ 0 and positive bl and b 2. This clearly supports the proposed
expression of q~MS(t) given by equation (E), provided that (AS4) [i.e.,
Ls(t) = La(t)] can be used [see also discussion of (AS1) and (AS4)].
These results must be taken with caution, since they were in most cases
obtained on mean values concerning exercise [12, 16, 20], Hermansen and
Vaage [15] being to our knowledge the only ones who described time courses of
lactate releases by the muscles during recovery in man.
Also in some cases the necessary data for our calculations was found in the
literature in forms that did not facilitate its exploitation.
Applications of the model will supply more discussion. A suitable expression
of q~MS(t) must agree with observed arterial-venous differences and lead to
plausible blood flows.

IlL Assumption (AS3)

There are two indications that serve to validate (AS3):

Since (PR) = (MRR) in steady states and according to relation (B), the
constant lactate concentrations observed at rest after recovery result from
constant lactate production rates.
As already shown [9, 10], very accurate bi-exponential fits can describe the
arterial concentrations over the entire recovery period. Since the observed noise
is always small even though important changes in respiration, heart rate and
catecholamine levels (as could possibly be observed for instance while muscular
biopsies are taken) may occur during this period, large variations of lactate
production during recovery may almost certainly be excluded.
To our knowledge, no other demonstrated facts can confirm or invalidate
(AS3).

Equation and Solution

1. By applying (AS1), (AS2), and (AS3) it is possible to express:

the lactate production rates during recovery:

(PR)M = Q , (1)

(nR)s = da ; (2)

the utilization rates of lactate in (M) and (S), respectively:

q'mM(t) = c 2 . LM(t) , (3)

q~ms(t) = d2 " Vs" Ls(t), (4)
142 P. Zouloumian and H. Freund

the net exchange rate between (M) and (S), defined as the result of simultaneous
muscular lactate uptake ( L U ) M and release ( L R ) M (see previous relations C , D,
and E)

@ms(t) = a~2 " VM " Lm(t) -- a21' Vs" Ls(t) = q(t) . [Lv(t) - L~(t)] . (5)

The time integraB of all these fluxes will describe the amounts of lactate
displaced and transformed during recovery.
T h e rates of variation in lactate contents of (M) and (S), according to relation
(A):
dLM(t)
I(M dt - c, + a21" Vs" Ls(t) - (a12 + c2)" VM" L M ( t ) , (6)

dLS(O
Vs'-- - d, + a12" VM" L u ( t ) - (a2i + d2)" Vs" L s ( t ) . (7)
dt

Thus, using the Laplace transforms:

cI l
(t7 + a,2 + c2)-~FLM(t) a2a VSM ~ L s ( t ) LM(O) + VM p (8)

dt 1
--a12" VMS "~5~LM(t ) + (p -I- a21 + d2)" ~CPLs(t)= Ls(O) + G" P (9)

T h e characteristic equation of the system is:

r 2 + (a12 + c2 + a21 -{- d2)" r + (a21 "C2 + a 1 2 " d2 + C2" d2) = 0 . (10)

Hence, its discfiminant:

A = [(a21 + d2) - (a12 + c2)] 2 + 4 a12 "a21.

Positive alz and a21 always lead to3:

a > o. (12)

Therefore, the equation 10 always has two real distinctive roots:

r1 = --Yl = [-- (a12 + c2 + a21 + d2) -- V ~ ] / 2 , (13)

and

r2 = --?'2 -----[-- (a~2 + c2 + a21 + d2) + V ~ ] / 2 9 (14)

3 The cases of possible negative a~ and/or a21, which violate (AS2), will not be examined in this
work
Model of Lactate Exchanges, Utilization, and Production 143

Thus:

}11 > }12 > 0 , (15)

Yl "}- }12 = a12 -'}-c2 -}- a21 -}- d2 > 0 , (16)

7i " }12 = a21 9 c2 + a12 d2 + c2 " d2 > 0 ,

" (17)

and
rl < r2 < 0 . (18)
T h e system always has non-oscillatory d a m p e d m o d e s . It is stable, and its
general solution can be e x p r e s s e d by m e a n s of combinations of two e x p o n e n t i a l
functions of time:

LM(t) = LM(~176-cgl " e -e' " t - ~ 2 " e -v~'t , (19)

Ls(t) = Ls(~176- ~ 1 " e -r' t - ~ 2 " e -7~t 9 (20)

Solving the system gives:

Cl
LM(O) " p2 + VM + a2~ " VsM " Ls(O)

]
+ (a21 + d2)' LM(O) "p + [a21(c 1 + d l ) + Cl"d2]/VM
~LM(O = , (21)
p (p - r~) 9 (p - r2)

Ls(O) " p2 +
[< Vs + el2" VMS"LM(O)

+ (al 2 + C2)" Ls(O)] "p + [a12(c 1 + d l ) + c2" dl]/gM

s = (22)
p (p - r l ) ' (p - r2)

Thus, the t i m e functions

a21(Cl -{- d l ) + Cl " d2 1 /

LM(O = -I- - - I }11 " LM(O) -- Cl/VM
}11 " }12 " VM Yl - - Y2
-- (221. VSM " Ls(O) - (a21 -1- d2)LM(O )

+ a21(Cl +_d i) + q " d2 I . e -r't 1

Y2" LM(O) - Q / V M
}11 " VM J Yl -- }12

- a21" VSM "Ls(O) - (a21 + d2)LM(O)

+ a21(Cl q- d l ) + Cl " de /
g : ~VM [ " e-r~t' (23)
144 P. Zouloumian and H. Freund

a12(Cl 4- d,) + C2 ' dl 1

Ls(t) = + - - " 7t " Ls(O) - da/Vs
}'' " }'5 " V S }'1 - - }'2

-- 0 { , 2 - VMS" LM(O) -- ( e l 2 4- c2)ts(O )

6~12(Cl 4- y l ) _ ~ C2" d, l . e-r,t 1

+ 72 ' Ls(O) - d l / V s
}'1 9 Vs J }'1 - }'2

- a12" v . s - L . ( O ) - (a12 + c s ) L s ( O )

+ a12(q + dO + c2- d, t j e -y~t

"
(24)

from which:

a,2- c~ 1 ----- Vsm(a21 4- 62 -- }'1) " 9 1 , (25)

(a12 4- C 2 - - } ' 1 ) " c~1 = a21 " VSM" 91 , (26)

a 1 2 ' (~2 = VsM(a21 4- d5 - }'2) " 9 5 , (27)

( a 1 2 4- c 2 - } ' 2 ) 9 6~2 = a 2 1 " VSM" 92 , (28)

and, by letting

~, = ai2 " V M L M ( ~ ) - a21" V s Ls(o~) = q ' M s ( ~ ) , (29)

cl = # + c2" VM" L M ( ~ ) , (30)

dl = -f~ + d2" Vs" L s ( ~ ) 9 (31)

2. Reducing the relations obtained allows to write the following corre-

spondences:
a) Between c2 and d2 [i.e., the specific efficiencies of (M) and (S) in lactate
utilization]:

~(& - c2) = v M . LM(oo) 9 (r, - c2). (72 - c5)

+ Vs" Ls(~1769 (Yl - d2)" (Y2 - d2), (32)

from which the parameter c 2 can be written as a function of d2 and #

C2 = [}', 4- }'5 - - /Z/VM" LM(~176q- V~c2]/2 = f ( d 2 , / z ) , (33)

Model of Lactate Exchanges, Utilization, and Production 145

where:

A c 2 = [}~/VM" L M ( m ) -- Yl -- 72] 2 -- 4 "71 "72

+ 4. [ u . d 2 - Vs. Ls(~). (7, - d2). ( 7 2 - d2)]/VM'LM(~). (34)

b) The expressions of the mutual lactate transfer coefficients as functions of

c2 and d2:

a12 = (71 - c2)" (72 - c2)/(d2 - c2), (35)

a21 ~--- -(71 - d2). (r2 - d2)/(d2 - c2), (36)

with:

d2 =/= C2 9 (37)

As shown, a12 and a21 allow writing the expressions of muscular lactate
release (LR)M and uptake (LU)M, as well as their net result qOMs(t).

c) The expressions for lactate productions:

C1 ~-" [,~ -}- V M " LM(~) "f(d2, #), (38)

dl= -[A + d2" Vs" Ls(m). (31)

d) The relations between the coefficients of the functions LM(t) and

Ls(t):

71 - - d2
(~1 = -- " VSM " 9 1 , (39)
71 -- C2

~2 -- d2
C~2 = -- " VSM" 92, (40)
72 -- C2

from which:

LM(O) = LM(OO)--~ -- ~2- (41)

Conclusion
The two-compartment model is completely solved within the scope of the
present hypotheses. If sufficient experimental information is supplied, the
obtained relationships provide the way to various applications which can serve as
tests of the model. For instance, if the parameters of bi-exponential fits to either
LM(t) or Ls(t) were known, numerical applications would allow exploration of
the properties of the model for given values of VM, Vs, d2, and #.
146 P. Zouloumian and H. Freund

Since the basic assumptions are reasonable, realistic predictions can be

expected. That this is indeed the case will be seen in the following report
[29].

Acknowledgements. We gratefully acknowledge the important contribution made to this work by

A. Heitz, J. Marbach, and C. Ott.
We also extend our gratitude to P. Kehayoff, S. Oyono, A. Pap6, R. Bucher, J. Becht,
F. Halbwachs, and E. Lampert for help in various stages of the project, and to Professor B. Metz for
his continual interest and encouragement.

List of Abbreviations and Symbols

a12, a21 Coefficients denoting efficiency in lactate transfer from (M) to (S) and (S) to (M),
respectively (rain-1)
a~ a 0 Coefficients of linear fits to observed lactate removal rates
bo, bl, b2 Coefficients of linear fits to observed lactate releases
(BS) Blood space
C2, d2 Coefficients denoting efficiency in lactate utilization by (M) and (S) (min-1)
Amplitudes of the exponential terms of LM(t) (mmol' 1-1)
~1,~2 Amplitudes of the exponential terms of Ls(t) (mmol- 1-1)
L Lactate concentration (mmol. 1-1)
Lo(O Lactate concentration in arterial blood at time t obtained by fits to experimental
data (mmol 91-1)
LM(O, Ls(t) Lactate concentrations in (M) and (S) at time t (mmol. 1-1)
(LU), (LR) Lactate uptake and lactate release rates (mmol 91-1)
(LU)M, (Ln)u Simultaneous muscular lactate uptake and lactate release rates (mmol- min -1)
L~(t) Lactate concentration in blood leaving (M) at time t (mmol. 1-1)
(M), (S) Worked muscle space and remaining lactate space
(MRR) Metabolic removal rate of lactate (mmol - min -1)
(nn) Lactate production rate (mmol. min -1)
(PR)M = q; Lactate production rates in (M) and (S), respectively (mmol. min -1)
(PR)s = dl
q (t) Blood flow peffusing (M) at time t (1. min -1)
(REX) Rate of lactate excretion (mmol 9min -1)
rl, r2 Roots of the characteristic equation (min-1)
t Time after the end of exercise (rain)
(TLS) Total lactate distribution space
V Volume of a compartment (1)
VM, Vs Volumes of (M) and (S) (1)
VMS VM to Vs ratio
VSM Vs to VM ratio
71, 72 Theoretical velocity constants of the time functions (min 1)
q'mM(t), ~s(O Lactate utilization rates in (M) and (S) (mmol. min<)
~MS(t) Net muscular release rate of lactate (mmol 9min -1)
# Net muscular release rate of lactate at t---~ o0 (mmol 9min -1)

References

1. Ahlborg G, Felig P (1977) Substrate utilization during prolonged exercise preceded by ingestion
of glucose. Am J Physiol 233:E188-E194
2. Belcastro AN, Bonen A (1975) Lactic acid removal rates during controlled and uncontrolled
recovery exercise. J Appl Physiol 39:932-936
3. Bergstr6m J, Guarnieri G, Hultman E (1971) Carbohydrate metabolism and electrolyte changes
in human muscle tissue during heavy work. J App1 Physiol 30:122-125
4. Bonen A, Campbell CJ, Kirby RL, Belcastro AN (1979) A multiple regression model for blood
lactate removal in man. Pflfigers Arch 380:205-210
Model of Lactate Exchanges, Utilization, and Production 147

5. De Coster A, Denolin H, Messin R, Degre S, Vandermotten P (1969) Role of the metabolites in

the acid-base balance during exercise. In: Poortmans JR (ed) Biochemistry of exercise. Karger,
Basel New York, pp 15-34
6. Depocas B, Minaire Y, Chatonnet J (1969) Rates of formation and oxidation of lactic acid in
dogs at rest and during moderate exercise. Canad J Physiol Pharmacol 47:603-610
7. Eldridge FL, T'so L, Chang H (1974) Relationship between turnover rate and blood
concentration of lactate in normal dogs. J Appl Physiol 37:316-320
8. Eldridge FL (1975) Relationship between turnover rate and blood concentration of lactate in
exercising dogs. J Appl Physiol 39:231-234
9. Freund H, Gendry P (1978) Lactate kinetics after short strenuous exercise in man. Eur J Appl
Physiol 39:123-135
10. Freund H, Zouloumian P (1981) Lactate after exercise in man: I. Evolution kinetics in arterial
blood. Eur J Appl Physiol 46:121-133
11. Freyschuss U, Strandell T (1967) Limb circulation during arm and leg exercise in supine
position. J Appl Physiol 23:163-170
12. Graham TE, Sinclair DG, Chap ler CK (1976) Metabolic intermediates and lactate diffusion in
active dog skeletal muscle. Am J Physiol 231:766-771
13. Hermansen L, Stensvold I (1972) Production and removal of lactate during exercise in man.
Acta Physiol Scand 86:191-201
14. Hermansen L, Maehlum S, Pruett EDR, Vaage O, Waldum H, Wessel-Aas T (1975) Lactate
removal at rest and during exercise. In: Howald H, Poortmans JR (eds) Metabolic adaptations
to physical exercise. Birkh~iuser, Basel, pp 101-105
15. Hermansen L, Vaage O (1977) Lactate disappearance and glycogen synthesis in human muscle
after maximal exercise. Am J Physiol 233:E422-E429
16. Hirche H, Wacker U, Langohr HD (1971) Lactic acid formation in the working gastrocnemius of
the dog. Int Z Angew Physiol 30:52-64
17. Holmgren A (1959) Arterio-venous lactic acid differences in man at rest and during muscular
work. Scand J Clin Lab Invest 11:150-153
18. Issekutz B Jr, Shaw WAS, Issekutz AC (1976) Lactate metabolism in resting and exercising
dogs. J Appl Physiol 40:312-319
19. Jorfeldt L (1970) Metabolism of L(+) - Lactate in human skeletal muscle during exercise. Acta
Physiol Scand 78: [Suppl] 338
20. Jorfeldt L, Dannfeldt AJ, Karlsson J (1978) Lactate release in relation to tissue lactate in human
skeletal muscle during exercise. J Appl Physiol 44:350-352
21. Karlsson J (1971) Lactate and phosphagen concentrations in working muscle of man. Acta
Physiol Scand 81: [Suppl] 358
22. Knuttgen HG, Saltin B (1972) Muscle metabolites and oxygen uptake in short-term submaximal
exercise in man. J Appl Physiol 32:690-694
23. McGrail JC, Bonen A, Belcastro AN (1978) Dependence of lactate removal on muscle
metabolism in man. Eur J Appl Physiol 39:89-97
24. Minaire Y, Pernod A, Jomain MJ, Mottaz M (1971) Lactate turnover and oxidation in normal
and adrenal-demedullated dogs during cold exposure. Can J Physiol Pharmacol
49:1063-1070
25. Mitchell AM, Cournand A (1955) The fate of circulating lactic acid in the human lung. J Clin
Invest 34:471-476
26. Olsen C, Strange Petersen E (1973) The lactate/pyruvate ratio in muscular work and following
injection of lactate in man. Pflfigers Arch 342:359-365
27. Rowell LB, Kraning II KK, Evans TO, Kennedy JW, Blackmon JR, Kusumi F (1966)
Splanchnic removal of lactate and pyruvate during prolonged exercise in man. J Appl Physiol
21:1773-1783
28. Sahlin K, Harris RC, Nylind B, Hultman E (1976) Lactate content and pH in muscle samples
obtained after dynamic exercise. Pflfigers Arch 367:143-149
29. Zouloumian P, Freund H (1981) Lactate after exercise in man: III. Properties of the
compartment model. Eur J Appl Physiol 46:149-160

Accepted January 5, 1981