Influence of Blood Handling Techniques on Lactic Acid Concentrations
I
Abstract
P. A. Bishop, M May, I F. Smith, I Kime, J. Mayo and M. Murphy, Influence of Blood Handling Techniques on Lactic Acid Concentrations. mt j Sports Med, Vol 13,Nol,pp56—59, 1992. Accepted after revision: April 25, 1991
lytic agents, treatment with stabilizing elements for long-term storage, dilution with lysing agents, and separation of plasma by centrifugation. Although these procedures are frequently employed in different combinations, there is limited knowledge of the influences of blood handling techniques on the re-
sulting measured lactic acid concentrations. Yeh et a!. (5) point out in the introduction of their paper that the blood sampling site might influence the point of "anaerobic threshold". Likewise, the variations in blood handling techniques could
influence the lactate concentrations and subsequent judge-
Despite the popularity of measuring blood lactic acid concentrations, many of the common variations in technique have not been evaluated. The purposes of this study were to: 1) establish the relationship between plasma and blood lactate concentrations, 2) determine the interanalyzer reliability, and 3) assess the stability of lactate concentration in blood stored for up to one week. Blood was sampled from 26 volunteers before exercise, at 80 Yo of esti-
mated maximum heart rate, and 5 minutes after a treadmill run to exhaustion. Inter-machine reliability was tested between two Yellow Springs Instruments analyzers with buf-
fer treated with a lysing agent and between two without. Blood lactate levels at all three levels could be predicted from plasma with R2> .95. Correlations between duplicates on the same machine were greater than .96 for blood and .97 for plasma. In the worst cases, between duplicate differences and between machine differences were 2%. Lactate in stored blood was in some cases significantly different after 24 hours of storage. Moderate and high lactate concentrations in plasma were not significantly altered after 2 days of storage.
ments made.
The purposes of this study were: 1) To compare blood lactate levels of whole blood, blood treated with Triton X-l00 as a lysing agent, and plasma, at low, moderate, and high physiological concentrations, 2) to determine the effects of storage (with a preservative agent) on lactate concentration for whole blood and plasma at low, moderate and high physiological concentrations, 3) to determine differences between duplicate measures and between automated analyzers for blood and plasma at low, moderate, and high physiological concentrations.
Methods
Subjects for this study were male and female volunteers (18 M, 8 F) who provided written informed consent
in accordance with the standards of our institution. Subjects had a mean (± SD) age of 24 (5) years, were 177 (9) cm, weighed 70.1 (14.1) kg, with a V02 of 3.48 (0.99) 1/mm (49.5 ml/kg), and a maximal heart rate of 196 (10) bpm. Subjects re-
ported to the laboratory and remained standing for 20 Key words
Measurement of lactic acid, reliability, stability of lactic acid, blood handling, plasma lactate concentration
minutes. An 18 gauge catheter was inserted into an antecubital vein and approximately 14 ml of blood were withdrawn and transferred into two open 7 ml vacutainers each containing 14 mg potassium oxalate and 17.5 rug sodium floride. The blood was immediately placed on ice and subdivided into 0.5 ml and 1.5 ml aliquots for immediate and delayed analyses with Yellow Springs Instruments Inc Model 23L automated analyzers (YSI). The balance of the blood was spun at 3200 rpm for 5 mm
whereupon the serum was removed and respun for an addiIntroduction
Measuring blood lactic acid concentration is
tional five minutes to ensure serum was free of red blood cells. The serum was subsequently sub-divided in the same volumes as the blood.
common in physiological research and training athletes. Blood for lactate analysis can be handled a number of ways, in-
Subjects performed a continuous graded
cluding treatment with various anti-clotting and anti-glyco-
treadmill run at either 9.7 or 11.2 km/hr (6 or 7mph) with a 2% elevation increase every two minutes. When subjects achieved
mt. J. SportsMed. 13(1992)56—59 GeorgThieme Verlag Stuttgart New York
approximately 80% of estimated maximum heart rate (HR), they were stopped and blood immediately withdrawn. HR's
Downloaded by: University of Pennsylvania Libraries. Copyrighted material.
P. A. Bishop1, M. May2, IF. Smith1, J. Kime1 , Mayo1, M Murphy' Human Performance Laboratory, The University of Alabama Department of Chemistry, Ohio State University
ml. J. Sports Med. 13(1992) 57
Influence of Blood Handling Techniques on Lactic Acid Concentrations
were recorded on a Quinton 630A ECG. Blood was immediately removed and treated as previously described. Subjects were then connected to gas collection equipment (Sensormedics MMC-l). Subjects continued the treadmill run to maximal volitional fatigue. Four to six minutes following the termination of the treadmill run, blood was again removed and han-
Table 1 Mean (SD) lactate concentrations (mmol/l) by type (NON, TX), and by level (LOW, MODERATE, HIGH) for whole blood (B) and
dled as previously described. After the first two blood removals, the catheter was kept patent with heparinized saline which was discarded prior to blood collection.
B P
1.30 (0.64) 1.76 (0.89)
B P
1.19 (0.67) 1.73 (0.89)
B P
1.25 (0.56) 1.63 (0.711*
B P
1.40 (0.63)*
B P
1.44 (0.71)* 1.77 (0.87)
Machine
LOW Initial
.05. That is, p was adjusted by dividing by the number of
1.51 (0.90) 1.75 (0.93) MODERATE Initial
2.66 (1.22) 3.08 (1.17)
2.18 (0.98) 3.04 (1.29)
P
24 Hour P
1.92 (1.12)* 3.06 (1.82)
B P
2.14 (1.29) 3.04 (1.76)
B P
2.35 (1.16)* 3.32 (1.56)
2.59 (1.43) 3.18 (1.77) 2 Day
2.66 (1.42)* 3.16 (1.82) 4 Day 2.74 (1.34)
3.44 (1.61) 7 Day 2.57 (1.19) 3.20 (1.48)
2.42 (1.09)* 3.12 (1.451*
B P
HIGH Initial P
7.31 (2.20) 10.86 (3.31)
B P
7.26 (2.72) 10.62 (3.80)
B
8.09 (2.93)* 10.62 (3.80)
B
9.42 (2.99) 11.33 (3.39)
24 Hour
9.87 (345)* 11.29 (3.76) 2 Day
P
9.77 (3.39) * 11.39 (3.95)
4 day P
8.71 (2.62) * 10.99 (3.16)
B P
8.60 (2.81)* 10.36 (3.12)*
B
were grouped to maximize statistical power and multiple MANOVA's were adjusted to yield family-wise alpha levels of
1.47 (0.57)* 1.67 (0.72)
1.71 (0.72) *
7 Day
B
MANOVA analysis is particularly sensitive to missing values. Occasional samples were missed due to inability to obtain a full blood sample or loss of whole or partial samples in handling. Lost data represented less than one percent of all data collected. Consequently, to ensure maximal efficacy of data analysis, omnibus MANOVA's could not be run. Instead variables
1.51 (0.56) 1.71 (0.72) 4 Day
through the normal range of exercise lactate values.
Data were analyzed with MANOVA and mul-
1.62 (0.73) * 1.82 (1.04) 2 Day
B
tiple regression using the SAS GLM procedure. The
1.48 (0.73) 1.80 (0.89)
24 Hour
throughout the test day a 15 mmol/l lactate standard (YSI lithium lactate solution) was analyzed to verify linearity Sixteen blood and plasma samples were stored at room temperature (19 °C) and read with the automated analyzers approximately 24 hours after initial collection. Samples were stored at room temperature to emulate field collections in sports and other situations wherein immediate refrigeration is often not possible. Samples were then refrigerated at 5 °C and read the next day (24 hour), two days later (two day), and four days later (four day) and one week (seven day) after initial collection. Stored blood was preserved with 0.015 micrograms of citric acid and 0.015 micrograms of potassium citrate per 1 .5 ml of blood or plasma.
TX
NON
9.92 (2.70) 11.56 (3.14) 7 Day
9.42 (3.10) 10.89 (3.36)* _______
* Indicates a significant (p < .05) difference from initial value of lac-
MANOVA's.
tate concentration.
Results The 20 separate {5 points in time x 2 machine
comparisons x 2 treatments (blood and plasma)] repeated measures MANOVA's revealed that with only two exceptions,
neither duplicate measures nor between machine measures were significantly different (p > .05/20) for whole blood, lysed blood, or plasma. Only the analyses of initial whole blood with no Tritzon X-lOO in the buffer (NON) showed that
there was significant interaction between lactate concentration levels (moderate, high) and duplicate samples for blood analyzed on TX machines. Additionally, there was a sig-
nificant interaction between blood concentration levels and
also between NON machines for plasma at all lactate concentration levels. Follow-up tests indicated that differences existed in both instances at medium and high lactate levels; but, inspection of the mcans suggests that in the worst case (high
concentrations), initial lactate concentration differences amounted to only 0.17 mmol/1 (2%) for duplicates and only 0.21 mmol/l (2%) between machines. Correlations between duplicates on the same machines 3 & 4 (TX) were all greater than .96 for all lactate concentrations for whole blood at Day 0 and Day 7. For plasma, the lowest correlation between these
Downloaded by: University of Pennsylvania Libraries. Copyrighted material.
Shortly after the blood was drawn, it was analyzed with four YSI automated analyzers. Two of these analyzers were supplied with a buffer solution (YSI #2357) (NON) and two were supplied with buffer to which 0.25% Triton X100 (TX) was added as a lysing agent. Analyzers were read at approximately the same time, in alternating order. Analyzers were calibrated with a 5.0 mmol/liter lactate standard (YSI, lithium lactate solution) between each reading. All readings were made in duplicate yielding 8 analyses per blood sample. Plasma was read in the same way. All analyzers were zeroed before each reading. At the start of each day and periodically
plasma (P), at initial, 24 hour, 2, 4, and 7 day
mt. J. Sports Med. 13(1992)
P. A. Bishop, M May, I F Smith, J. Kime, J. Mayo, M Murphy
____
Table 2 Regression equations for estimating whole blood (B) lactate concentrations from plasma (P) lactate and initial hematocrit (Hct) B Status
P Status
P Coeff
Hct Coeff
NON
NON TX
0.678 0.787
0.0080
TX
NON TX
0.667 0.768
NON TX
NON TX
0.643 0.860
Intercept
R2
Sxy
—0.24 0.29
.91
.95
.20 .16
0.04 —0.18
.98 .97
.14 .15
0.63 —2.60
.96 .98
.43 .42
LOW Lactates TX
0.0004
MODERATE Lactates NON
0.0004
0.0070
0.0085 0.0574
measures was .98. Correlations were greater than .99 for many
of the plasma-plasma comparisons. Because of the lack of practical differences, for subsequent analyses, duplicate and between machine values were averaged (1. e. TX'd machines were combined, and NON machines were combined). TX'd blood mean lactate concentrations were significantly higher (p < .05) than NON blood mean lactate concentrations at all lactate levels. TX plasma lactate concentrations were also sig-
or TX) for YSI analyses. The larger R squares for the higher lactate concentrations is at least partially attributable to the larger range of lactate concentrations. The inclusion of initial hematocrit levels contributed little to the equation indicated by the relatively low coefficients and inspection of the partial regression plots. This is not surprising in that the range of hematocrit was limited. Cross-validation of these equations is needed.
nificantly higher than NON plasma concentrations at all levels. Actual mean lactates by sample and machine for all collection times are shown in Table I.
The changes in lactate concentration relative to the initial value are also shown in Table 1. At least one lactate
concentration for blood was significantly different by 24 hours. Plasma values were unchanged after 24 hours. Plasma values were unchanged after 2 days for moderate and high lac-
tate concentrations. The occurrence and direction of lactate concentration shift followed no clear pattern.
Whole blood and lysed blood lactate levels were predicted from plasma lactates and hematocrits. Linear regression equations, R squares, and standard errors of the estimate are displayed in Table 2.
Discussion
In the present study we compared lactate concentrations in whole blood, lysed blood, and plasma at three levels and also examined within machine and between machine reliability as well as the effects of storage with a preservative.
Previous studies have reported differences between lactate levels inside and outside the erythrocytes (3, 4). When red blood cells are lysed, the intercellular lactate raises the measured concentration per unit of sample (3, 4). When Triton X-1 00 was added to plasma, the apparent lactate level was 0.04 mmol/l higher for low and moderate lactate concentrations. Such a difference would generally not be detectable since the analyzer reads only to one decimal. At the high lactate levels, plasma with Triton X-l00 showed mean concentrations 0.47 mmol/l higher (p < .05) than untreated plasma. This 4% increase is not attributable to the analyzer, because the zero, 5 and 15 mmol/l calibrations were made with buffer containing Triton X- 100. Instead, the Triton X- 100 must react with some plasma constituent to produce a lactate or peroxide (6) analog. The use of citric acid and potassium citrate as a preservative of blood lactate concentration appears only partially effective. The stability of the lactate concentration appeared greater for plasma than for nonlysed or lysed blood. The mean lactate concentrations appeared to vary randomly, which was disconcerting. With this preservative it appears that only plasma is reliably stable but only for one day. Blood stability follows no clear pattern. Since TX blood is frequently
stored for varying time periods without any preservative, Both within machine and between machine re-
liability was high. In two instances, significant interactions were identified, but inspection of the actual means suggests that, in most applications, these differences would not be important. Significant differences were not observed in the other 18 comparisons.
further research is needed to examine this practice. In the present study, blood was stored at room temperature for 24 hours following the initial measurement to simulate field conditions. Better results might have been obtained if the blood had been collected on ice and then refrigerated or frozen. Buono (2) has reported that deproteinized blood can be stored up to 90 days
by freezing. We have not yet been successful in analyzing
Differences in lactate concentration between whole blood, lysed (TX'd) blood and plasma were significant
blood treated with perchloric acid with the YSI analyzers.
but predictable. The derived regression equations yielded consistently high R squares with reasonable low standard errors for the derivation sample. It appears that whole blood can be predicted from plasma lactates of the same type (either NON
when blood cannot be immediately analyzed. Incomplete hemolysis could be problematic. In another study we com-
TX or other lysing agents are often employed pared blood treated with TX in the same way as in the present
Downloaded by: University of Pennsylvania Libraries. Copyrighted material.
HIGH Lactates
mt. J. Sports Med. 13(1992) 59
Influence of Blood Handling Techniques on Lactic Acid Concentrations
It appears that if storage of blood for later lac-
tate measurement is required, plasma would be the best storage form, and storage should not exceed two days with our preservation technique. Assuming adequate cross-validation, the equations derived should allow interconversion of plasma, whole and lysed blood lactate values, suggesting that the technique for handling blood samples can be selected depending upon collection conditions.
In summary, both duplicate and inter-machine reliability is good at low, moderate, and high lactate levels. The
use of TX does influence lactate at all lactate concentrations, even in plasma. YSI whole blood lactates are predictable from YSI plasma values. And, moderate and high concentrations of lactate in plasma appear stable in acidified samples for up to two days, but this was not true for whole blood. Acknowledgement This study was supported in part by a grant from Yellow Springs Instruments INC.
References I
BishopP.A.,SmithJ.F.,KimeJ.C.,MayoJ.,TinY.H.:Comparison of a manual and an automated enzymatic technique for determining blood lactic acid concentrations. mt j Sports Med 1991 (In
2 Press). Buono M. J.: Freezing provides a viable method for the prolonged,
stable storage of blood lactate. Can J App! Spt Sd 11(2): 80—81, 1986.
Buono M. J., Yeager J. E.: Intraerythrocyte and plasma lactate concentrations during exercise in humans. Eur JAppi Physiol 55: 326—329,1986.
Daniel S. S., Morishima H. 0., James L. S., Adamsons K. Jr.: Lac-
tate and pyruvate gradient between red blood cells and plasma during acute asphyxia. JAppiPhysiol 19: 1100—1104, 1963. Yeh M. P., Gardner R. M., Adams T. D., Yanowitz F. G., Crapo R.
0.: "Anaerobic threshold": problems of determination and validation.JApplPhysiol5S (4): 1178—1186,1983. 6 Yellow Springs Instruments: YSI model 23L lactate analyzers instruction manual. Scientific Division, Yellow Springs, OH, 1985.
Phillip Bishop, Ed. D.
P0 Box 870312 Tuscaloosa, AL 35487-03 12 USA
Downloaded by: University of Pennsylvania Libraries. Copyrighted material.
study. We observed very good agreement between the lactate concentration measured in blood lysed with TX in buffer and between lactate concentrations measured with an enzymatic method using samples treated with perchloric acid (1).
I
Abstract
P. A. Bishop, M May, I F. Smith, I Kime, J. Mayo and M. Murphy, Influence of Blood Handling Techniques on Lactic Acid Concentrations. mt j Sports Med, Vol 13,Nol,pp56—59, 1992. Accepted after revision: April 25, 1991
lytic agents, treatment with stabilizing elements for long-term storage, dilution with lysing agents, and separation of plasma by centrifugation. Although these procedures are frequently employed in different combinations, there is limited knowledge of the influences of blood handling techniques on the re-
sulting measured lactic acid concentrations. Yeh et a!. (5) point out in the introduction of their paper that the blood sampling site might influence the point of "anaerobic threshold". Likewise, the variations in blood handling techniques could
influence the lactate concentrations and subsequent judge-
Despite the popularity of measuring blood lactic acid concentrations, many of the common variations in technique have not been evaluated. The purposes of this study were to: 1) establish the relationship between plasma and blood lactate concentrations, 2) determine the interanalyzer reliability, and 3) assess the stability of lactate concentration in blood stored for up to one week. Blood was sampled from 26 volunteers before exercise, at 80 Yo of esti-
mated maximum heart rate, and 5 minutes after a treadmill run to exhaustion. Inter-machine reliability was tested between two Yellow Springs Instruments analyzers with buf-
fer treated with a lysing agent and between two without. Blood lactate levels at all three levels could be predicted from plasma with R2> .95. Correlations between duplicates on the same machine were greater than .96 for blood and .97 for plasma. In the worst cases, between duplicate differences and between machine differences were 2%. Lactate in stored blood was in some cases significantly different after 24 hours of storage. Moderate and high lactate concentrations in plasma were not significantly altered after 2 days of storage.
ments made.
The purposes of this study were: 1) To compare blood lactate levels of whole blood, blood treated with Triton X-l00 as a lysing agent, and plasma, at low, moderate, and high physiological concentrations, 2) to determine the effects of storage (with a preservative agent) on lactate concentration for whole blood and plasma at low, moderate and high physiological concentrations, 3) to determine differences between duplicate measures and between automated analyzers for blood and plasma at low, moderate, and high physiological concentrations.
Methods
Subjects for this study were male and female volunteers (18 M, 8 F) who provided written informed consent
in accordance with the standards of our institution. Subjects had a mean (± SD) age of 24 (5) years, were 177 (9) cm, weighed 70.1 (14.1) kg, with a V02 of 3.48 (0.99) 1/mm (49.5 ml/kg), and a maximal heart rate of 196 (10) bpm. Subjects re-
ported to the laboratory and remained standing for 20 Key words
Measurement of lactic acid, reliability, stability of lactic acid, blood handling, plasma lactate concentration
minutes. An 18 gauge catheter was inserted into an antecubital vein and approximately 14 ml of blood were withdrawn and transferred into two open 7 ml vacutainers each containing 14 mg potassium oxalate and 17.5 rug sodium floride. The blood was immediately placed on ice and subdivided into 0.5 ml and 1.5 ml aliquots for immediate and delayed analyses with Yellow Springs Instruments Inc Model 23L automated analyzers (YSI). The balance of the blood was spun at 3200 rpm for 5 mm
whereupon the serum was removed and respun for an addiIntroduction
Measuring blood lactic acid concentration is
tional five minutes to ensure serum was free of red blood cells. The serum was subsequently sub-divided in the same volumes as the blood.
common in physiological research and training athletes. Blood for lactate analysis can be handled a number of ways, in-
Subjects performed a continuous graded
cluding treatment with various anti-clotting and anti-glyco-
treadmill run at either 9.7 or 11.2 km/hr (6 or 7mph) with a 2% elevation increase every two minutes. When subjects achieved
mt. J. SportsMed. 13(1992)56—59 GeorgThieme Verlag Stuttgart New York
approximately 80% of estimated maximum heart rate (HR), they were stopped and blood immediately withdrawn. HR's
Downloaded by: University of Pennsylvania Libraries. Copyrighted material.
P. A. Bishop1, M. May2, IF. Smith1, J. Kime1 , Mayo1, M Murphy' Human Performance Laboratory, The University of Alabama Department of Chemistry, Ohio State University
ml. J. Sports Med. 13(1992) 57
Influence of Blood Handling Techniques on Lactic Acid Concentrations
were recorded on a Quinton 630A ECG. Blood was immediately removed and treated as previously described. Subjects were then connected to gas collection equipment (Sensormedics MMC-l). Subjects continued the treadmill run to maximal volitional fatigue. Four to six minutes following the termination of the treadmill run, blood was again removed and han-
Table 1 Mean (SD) lactate concentrations (mmol/l) by type (NON, TX), and by level (LOW, MODERATE, HIGH) for whole blood (B) and
dled as previously described. After the first two blood removals, the catheter was kept patent with heparinized saline which was discarded prior to blood collection.
B P
1.30 (0.64) 1.76 (0.89)
B P
1.19 (0.67) 1.73 (0.89)
B P
1.25 (0.56) 1.63 (0.711*
B P
1.40 (0.63)*
B P
1.44 (0.71)* 1.77 (0.87)
Machine
LOW Initial
.05. That is, p was adjusted by dividing by the number of
1.51 (0.90) 1.75 (0.93) MODERATE Initial
2.66 (1.22) 3.08 (1.17)
2.18 (0.98) 3.04 (1.29)
P
24 Hour P
1.92 (1.12)* 3.06 (1.82)
B P
2.14 (1.29) 3.04 (1.76)
B P
2.35 (1.16)* 3.32 (1.56)
2.59 (1.43) 3.18 (1.77) 2 Day
2.66 (1.42)* 3.16 (1.82) 4 Day 2.74 (1.34)
3.44 (1.61) 7 Day 2.57 (1.19) 3.20 (1.48)
2.42 (1.09)* 3.12 (1.451*
B P
HIGH Initial P
7.31 (2.20) 10.86 (3.31)
B P
7.26 (2.72) 10.62 (3.80)
B
8.09 (2.93)* 10.62 (3.80)
B
9.42 (2.99) 11.33 (3.39)
24 Hour
9.87 (345)* 11.29 (3.76) 2 Day
P
9.77 (3.39) * 11.39 (3.95)
4 day P
8.71 (2.62) * 10.99 (3.16)
B P
8.60 (2.81)* 10.36 (3.12)*
B
were grouped to maximize statistical power and multiple MANOVA's were adjusted to yield family-wise alpha levels of
1.47 (0.57)* 1.67 (0.72)
1.71 (0.72) *
7 Day
B
MANOVA analysis is particularly sensitive to missing values. Occasional samples were missed due to inability to obtain a full blood sample or loss of whole or partial samples in handling. Lost data represented less than one percent of all data collected. Consequently, to ensure maximal efficacy of data analysis, omnibus MANOVA's could not be run. Instead variables
1.51 (0.56) 1.71 (0.72) 4 Day
through the normal range of exercise lactate values.
Data were analyzed with MANOVA and mul-
1.62 (0.73) * 1.82 (1.04) 2 Day
B
tiple regression using the SAS GLM procedure. The
1.48 (0.73) 1.80 (0.89)
24 Hour
throughout the test day a 15 mmol/l lactate standard (YSI lithium lactate solution) was analyzed to verify linearity Sixteen blood and plasma samples were stored at room temperature (19 °C) and read with the automated analyzers approximately 24 hours after initial collection. Samples were stored at room temperature to emulate field collections in sports and other situations wherein immediate refrigeration is often not possible. Samples were then refrigerated at 5 °C and read the next day (24 hour), two days later (two day), and four days later (four day) and one week (seven day) after initial collection. Stored blood was preserved with 0.015 micrograms of citric acid and 0.015 micrograms of potassium citrate per 1 .5 ml of blood or plasma.
TX
NON
9.92 (2.70) 11.56 (3.14) 7 Day
9.42 (3.10) 10.89 (3.36)* _______
* Indicates a significant (p < .05) difference from initial value of lac-
MANOVA's.
tate concentration.
Results The 20 separate {5 points in time x 2 machine
comparisons x 2 treatments (blood and plasma)] repeated measures MANOVA's revealed that with only two exceptions,
neither duplicate measures nor between machine measures were significantly different (p > .05/20) for whole blood, lysed blood, or plasma. Only the analyses of initial whole blood with no Tritzon X-lOO in the buffer (NON) showed that
there was significant interaction between lactate concentration levels (moderate, high) and duplicate samples for blood analyzed on TX machines. Additionally, there was a sig-
nificant interaction between blood concentration levels and
also between NON machines for plasma at all lactate concentration levels. Follow-up tests indicated that differences existed in both instances at medium and high lactate levels; but, inspection of the mcans suggests that in the worst case (high
concentrations), initial lactate concentration differences amounted to only 0.17 mmol/1 (2%) for duplicates and only 0.21 mmol/l (2%) between machines. Correlations between duplicates on the same machines 3 & 4 (TX) were all greater than .96 for all lactate concentrations for whole blood at Day 0 and Day 7. For plasma, the lowest correlation between these
Downloaded by: University of Pennsylvania Libraries. Copyrighted material.
Shortly after the blood was drawn, it was analyzed with four YSI automated analyzers. Two of these analyzers were supplied with a buffer solution (YSI #2357) (NON) and two were supplied with buffer to which 0.25% Triton X100 (TX) was added as a lysing agent. Analyzers were read at approximately the same time, in alternating order. Analyzers were calibrated with a 5.0 mmol/liter lactate standard (YSI, lithium lactate solution) between each reading. All readings were made in duplicate yielding 8 analyses per blood sample. Plasma was read in the same way. All analyzers were zeroed before each reading. At the start of each day and periodically
plasma (P), at initial, 24 hour, 2, 4, and 7 day
mt. J. Sports Med. 13(1992)
P. A. Bishop, M May, I F Smith, J. Kime, J. Mayo, M Murphy
____
Table 2 Regression equations for estimating whole blood (B) lactate concentrations from plasma (P) lactate and initial hematocrit (Hct) B Status
P Status
P Coeff
Hct Coeff
NON
NON TX
0.678 0.787
0.0080
TX
NON TX
0.667 0.768
NON TX
NON TX
0.643 0.860
Intercept
R2
Sxy
—0.24 0.29
.91
.95
.20 .16
0.04 —0.18
.98 .97
.14 .15
0.63 —2.60
.96 .98
.43 .42
LOW Lactates TX
0.0004
MODERATE Lactates NON
0.0004
0.0070
0.0085 0.0574
measures was .98. Correlations were greater than .99 for many
of the plasma-plasma comparisons. Because of the lack of practical differences, for subsequent analyses, duplicate and between machine values were averaged (1. e. TX'd machines were combined, and NON machines were combined). TX'd blood mean lactate concentrations were significantly higher (p < .05) than NON blood mean lactate concentrations at all lactate levels. TX plasma lactate concentrations were also sig-
or TX) for YSI analyses. The larger R squares for the higher lactate concentrations is at least partially attributable to the larger range of lactate concentrations. The inclusion of initial hematocrit levels contributed little to the equation indicated by the relatively low coefficients and inspection of the partial regression plots. This is not surprising in that the range of hematocrit was limited. Cross-validation of these equations is needed.
nificantly higher than NON plasma concentrations at all levels. Actual mean lactates by sample and machine for all collection times are shown in Table I.
The changes in lactate concentration relative to the initial value are also shown in Table 1. At least one lactate
concentration for blood was significantly different by 24 hours. Plasma values were unchanged after 24 hours. Plasma values were unchanged after 2 days for moderate and high lac-
tate concentrations. The occurrence and direction of lactate concentration shift followed no clear pattern.
Whole blood and lysed blood lactate levels were predicted from plasma lactates and hematocrits. Linear regression equations, R squares, and standard errors of the estimate are displayed in Table 2.
Discussion
In the present study we compared lactate concentrations in whole blood, lysed blood, and plasma at three levels and also examined within machine and between machine reliability as well as the effects of storage with a preservative.
Previous studies have reported differences between lactate levels inside and outside the erythrocytes (3, 4). When red blood cells are lysed, the intercellular lactate raises the measured concentration per unit of sample (3, 4). When Triton X-1 00 was added to plasma, the apparent lactate level was 0.04 mmol/l higher for low and moderate lactate concentrations. Such a difference would generally not be detectable since the analyzer reads only to one decimal. At the high lactate levels, plasma with Triton X-l00 showed mean concentrations 0.47 mmol/l higher (p < .05) than untreated plasma. This 4% increase is not attributable to the analyzer, because the zero, 5 and 15 mmol/l calibrations were made with buffer containing Triton X- 100. Instead, the Triton X- 100 must react with some plasma constituent to produce a lactate or peroxide (6) analog. The use of citric acid and potassium citrate as a preservative of blood lactate concentration appears only partially effective. The stability of the lactate concentration appeared greater for plasma than for nonlysed or lysed blood. The mean lactate concentrations appeared to vary randomly, which was disconcerting. With this preservative it appears that only plasma is reliably stable but only for one day. Blood stability follows no clear pattern. Since TX blood is frequently
stored for varying time periods without any preservative, Both within machine and between machine re-
liability was high. In two instances, significant interactions were identified, but inspection of the actual means suggests that, in most applications, these differences would not be important. Significant differences were not observed in the other 18 comparisons.
further research is needed to examine this practice. In the present study, blood was stored at room temperature for 24 hours following the initial measurement to simulate field conditions. Better results might have been obtained if the blood had been collected on ice and then refrigerated or frozen. Buono (2) has reported that deproteinized blood can be stored up to 90 days
by freezing. We have not yet been successful in analyzing
Differences in lactate concentration between whole blood, lysed (TX'd) blood and plasma were significant
blood treated with perchloric acid with the YSI analyzers.
but predictable. The derived regression equations yielded consistently high R squares with reasonable low standard errors for the derivation sample. It appears that whole blood can be predicted from plasma lactates of the same type (either NON
when blood cannot be immediately analyzed. Incomplete hemolysis could be problematic. In another study we com-
TX or other lysing agents are often employed pared blood treated with TX in the same way as in the present
Downloaded by: University of Pennsylvania Libraries. Copyrighted material.
HIGH Lactates
mt. J. Sports Med. 13(1992) 59
Influence of Blood Handling Techniques on Lactic Acid Concentrations
It appears that if storage of blood for later lac-
tate measurement is required, plasma would be the best storage form, and storage should not exceed two days with our preservation technique. Assuming adequate cross-validation, the equations derived should allow interconversion of plasma, whole and lysed blood lactate values, suggesting that the technique for handling blood samples can be selected depending upon collection conditions.
In summary, both duplicate and inter-machine reliability is good at low, moderate, and high lactate levels. The
use of TX does influence lactate at all lactate concentrations, even in plasma. YSI whole blood lactates are predictable from YSI plasma values. And, moderate and high concentrations of lactate in plasma appear stable in acidified samples for up to two days, but this was not true for whole blood. Acknowledgement This study was supported in part by a grant from Yellow Springs Instruments INC.
References I
BishopP.A.,SmithJ.F.,KimeJ.C.,MayoJ.,TinY.H.:Comparison of a manual and an automated enzymatic technique for determining blood lactic acid concentrations. mt j Sports Med 1991 (In
2 Press). Buono M. J.: Freezing provides a viable method for the prolonged,
stable storage of blood lactate. Can J App! Spt Sd 11(2): 80—81, 1986.
Buono M. J., Yeager J. E.: Intraerythrocyte and plasma lactate concentrations during exercise in humans. Eur JAppi Physiol 55: 326—329,1986.
Daniel S. S., Morishima H. 0., James L. S., Adamsons K. Jr.: Lac-
tate and pyruvate gradient between red blood cells and plasma during acute asphyxia. JAppiPhysiol 19: 1100—1104, 1963. Yeh M. P., Gardner R. M., Adams T. D., Yanowitz F. G., Crapo R.
0.: "Anaerobic threshold": problems of determination and validation.JApplPhysiol5S (4): 1178—1186,1983. 6 Yellow Springs Instruments: YSI model 23L lactate analyzers instruction manual. Scientific Division, Yellow Springs, OH, 1985.
Phillip Bishop, Ed. D.
P0 Box 870312 Tuscaloosa, AL 35487-03 12 USA
Downloaded by: University of Pennsylvania Libraries. Copyrighted material.
study. We observed very good agreement between the lactate concentration measured in blood lysed with TX in buffer and between lactate concentrations measured with an enzymatic method using samples treated with perchloric acid (1).
