Examination of four different instruments for measuring blood lactate
Contents
1. For the YSI 23L the relationship for this reduced data set is largely the same as for the full set Fig 2A Thus also within this range of values the YSI 23L showed values that were too low The error of regression was roughly half that found for the full data set Hemolyzed blood samples measured on the YSI 1500 showed average values 13 too high p lt 0 001 while for YSI 15002 there was no sign of a systematic deviation Fig 2B C The values from LP8 were on average 25 too high p lt 0 001 Fig 2D The error of regression found for this instrument was about twice that found for the other instruments p lt 0 001 The Lactate Pro showed no sys tematic deviations from the reference method for samples with a blood lactate concentration less than 6mmol L Fig 2E For the instruments mentioned so far the systematic errors were less in both absolute and relative terms for samples with less than 6 mmol lactate L blood than for samples covering a larger span The values from measurements by the Accusport were systematically higher than those from the reference method p lt 0 001 Fig 2F The LP8 showed a larger variability than the other instruments see Figs 1 and 2 We therefore took three parallel measurements on a number of blood samples with this instrument The median of each measurement was related to measurements with the reference method Fig 3A This approach did not give considerable better values than sin2. The temperature in the bag was measured with a digital thermometer It appeared that repeated opening and closing of the bag lowered the temperature and if it dropped to 5 C or lower the experiments were stopped Statistics The values are given as means SD unless otherwise stated explicitly Different instru ments or methods were compared by linear regression using the approach geometric mean that gives equal weight to errors in both series Blood lactate measured by different instruments 371 of measurements 8 9 standard linear regres sion assumes that all error or imprecision is caused by measurements of the ordinate Y For the regression analyses carried out here the approach geometric mean gives the same results as a more demanding robust and distribution free approach 10 the method of Bland and Altman 11 also gives corresponding results for our data If two instruments or methods differ systematically the slope will differ from 1 00 and the data points will not be randomly distributed around the line of identity Thus possible differences were tested by examining the residual and by standard t tests on the slope As stated above we also kept in mind the fact that our reference method may have had a systematic error of 1 SEM The error of regression scatter around the regression line is used as a measure of the random errors Different errors of regression were tested by Fisher tests 12 Errors can
3. 5 mmol L Even for blood sam ples taken at rest and with a known concentration less than 1 mmol L the Accu sport reported values around 2 mmol L When it reported a value of 3 0 mmol L the true value was 2 5 mmol L in our experiments and that error may be important when testing athletes Our data suggest that there may be an inbuilt error in this instru ment On examining the Accusport versus a 378 J I Medbo et al reference method RoBkopf et al 17 did not find the bias that we registered but their random errors were considerably larger than ours Two different Accusport instruments gave similar results when the same blood samples were measured The instrument did not work adequately at simulated altitudes above 6 km but it did work well in our cold experiments The Accusport was more difficult to use than the Lactate Pro and we got a number of incorrect measurements with it Some errors could be excluded on grounds of uneven coloring seen through the window on the underside of the strip or too little blood applied but we also experienced errors that could not be detected without comparisons with other ana lyses Several measurements were lost because we accidentally hit the on off button when the analysis was started by closing the cover We had technical problems with both instruments One reported low battery and stopped work ing even when new batteries with the proper voltage were used The
4. ak ba es cy eee 5 0 79 S 0 020 n 78 Y 0 294 1127 1 0 992 0 814 x r 0 991 o o S 0 56 S 0 013 n 115 Sy 0 41 S 0 017 n 45 0 1 a 0 I dee 1 i 1 0 7 a i 1 0 5 10 15 0 5 10 15 0 5 10 15 La enz photofluorometry mmol L La enz photofluorometry mmo Ly La enz photofluorometry mmol Ly JL e blood lactate concentration measure enzymatic photofluorometr an six differen Fic 1 The blood lactat trat d by enzymatic photofi try d by six different instruments Y The samples measured by the YSI 23L panel A were on non hemolyzed blood and these measurements do not record lactate inside the red blood cells YSI 1500 and YSI 1500 panels B and C refer to measurements on two different YSI 1500 instruments LP 8 panel D is a lactate analyzer from Dr Lange Lactate Pro panel is from Arkray in the KDK Corporation while Accusport panel F is from Boehringer Mannheim Data on the regressions are given for each set of data and the regression lines are shown as solid lines Sy is the error of regression and is expressed in mmol LT Sy is the error of the slope The thin dashed lines show the lines of identity for each set of data e refers to accepted measure ments while O refers to outliers that were formally rejected because of an error of measurement with the instrument examined these values were therefore not used to calculate the regression lines ing data from Fig 1 were reanalyzed Fig 2
5. deviations given later and is thus regarded in this study as without significance YSI instruments The lactate concentration in samples of 25 pl was measured on an YSI 23L Yellow Springs Instruments Yellow Springs OH USA on non hemolyzed blood and on plasma samples Further blood analyses were done on five YSI 1500 instruments on either hemolyzed or unhemolyzed 25 ul blood samples All YSI instruments were calibrated by 5 mmol lactate L solutions YSI 2327 As an additional control a standard solution of 15 mmol lactate L was used YSI 2328 on some occasions we also used a 30 mmol L solution YSI 1530 In accordance with the instruments manuals we required that readings with these solutions gave 5 0 0 1 mmol L 15 0 0 7 mmol L and 30 0 1 5 mmol L respectively before any analyses could be carried out YSI instruments assume that a known and constant amount of blood is used for each measurement and the instruments are equipped with two types of pipettes one with a needle YSI 1501 syringepet and one using capillary tubes YSI 1502 pipette In agreement with the manufacturer s instructions we used the same pipette for both the calibrations and the subsequent measurements When analyses were done on more than one YSI instrument per day the same syringes and calibration solutions were used for all instruments and all instruments were in addition handled by the same operator In blood the lactate is
6. measurements of umbilical artery lactate levels in the prediction of perinatal outcome Am J Obstet Gynecol 1995 173 1416 22 Passoneau JV Lowry OH Enzymatic analysis Humana Press Totowa New Jersey 1993 p 188 93 Siggaard Andersen O The acid base status of the blood Munksgaard Copenhagen 4th ed 1976 p 79 Gambke B Berg A Fabian K Francaux M Haber P Hartman U Kamber M Lormes W Rosskopf P Schwarz L Multicenter evaluation of a new portable system for determining blood lactate In Ramstetter E Zieres Nauth C Mack M editors Workshop Report Accusport Z rich 1994 Boehringer Mannheim 1995 p 29 32 Emmert J Analytical requirements for the mea surement of lactate In Ramstetter E Zieres Nauth C Mack M editors Workshop Report Accusport Z rich 1994 Boehringer Mannheim 1995 p 26 8 Riggs DS Guarnieri JA Addleman S Fitting straight lines when both variables are subject to error Life Sci 1976 22 1305 60 Brace RA Fitting straight lines to experimental data Am J Physiol Integrative Comp Physiol 1977 233 R94 R99 Passing H Bablok W A new biometrical proce dure for testing the equality of measurements from two different analytical methods J Clin Chem Clin Biochem 1983 21 709 20 Bland JM Altman DG Statistical methods for assessing agreement between two methods of clinical measurements Lancet 1986 i 8476 307 10 Owen DB Handbook of statistical tables Read ing MA Addi
7. out as bicycling on ergometers Blood samples were taken during treadmill running rollerskiing bicycling outdoors and as canoeing or rowing on ergometers In some experiments described in more detail elsewhere 2 blood was drawn from catheters in the femoral artery and vein Otherwise capillary blood was taken from a finger after the hand had been warmed for at least 30 s in tempered water This procedure increases the perfusion of the hand and makes it easier to get enough blood for several samples in sequence Sweat contains lactate and wash ing in water will also dissolve and thus remove lactate on the skin that otherwise could contaminate the blood The skin was punctured using a lancet and the first drop of blood was wiped off Thereafter blood was taken as described in detail below In some experiments blood samples were taken in conjunction with intensive exercise to exhaustion in others as part of testing subjects and finding their blood lactate threshold and maximal O uptake Finally blood samples were taken in some experiments with strenuous exercise with a significant anaerobic energy release but where the exercise was stopped before exhaustion We took around 800 blood or plasma samples and measured them using at least two different methods or instruments Altogether nearly 3000 single measurements of the lactate concentration in blood or plasma samples were carried out METHODS Enzymatic photofluorometr
8. the underside of each strip after each analysis and examined the coloring if uneven the analysis was rejected We also checked that enough blood was added to each strip When blood is added some of it penetrates the surface and thus reaches chemi cals that start the reactions processing the lactate We required that there should be liquid blood on the top of the strip s pad after each analysis Otherwise the result was rejected We thus used 25 50ul of blood for each analysis Blood was added to the strip by letting it drip from a finger and in accordance with the instrument s instructions we never let the finger touch the strip s pad This instrument only measures the lactate in plasma according to the manufacturer built in equations calculate the concentration in whole blood from the measured value in plasma 7 For our measurements we used strips from Boehringer Mannheim with the code number 512 In concurrence with the instructions for each instrument all measurements for all instru ments were carried out by experienced test leaders who had each done at least a thousand analyses of the blood lactate concentration Measurements at simulated altitude Control solutions from Boehringer Mann heim article no 1447335 with a reported lactate concentration of 3 6 mmol L7 low BM control Lactate 1 and 9 mmol L high BM control Lactate 2 were measured in a chamber where the O pressure was less than 10 kPa co
9. value even on samples taken at rest This instrument should therefore perhaps not be used for that kind of testing It has also been proposed that an intensity giving a lactate concentration of rest value plus 1 5 mmol L should be sought 13 That definition could be less sensitive to variations between instruments Effect of simulated altitude Control solutions with given lactate concen trations were measured on the Lactate Pro and the Accusport when the O pressure was less than 10 kPa corresponding to that found at altitudes above 6km Readings with the Lactate Pro were not affected by the reduced O2 pressure suggesting that this instrument may be well suited for testing at altitude The Accusport reported too high values in our measurements Our measurements do not allow us to conclude how this instrument will behave at the O pressures of altitudes of 1500 3000 m that are more typical of athletes in training Measurements in the cold According to the manufacturers the Lactate Pro and the Accusport may both be unreliable for testing at temperatures below 10 C Our data suggest that if the instruments and the strips are shielded from the cold except for the 20 30 s when the strips are inserted in the instrument and blood or test solutions are applied to the strips the reported values are not affected by surrounding temperatures down to 20 C Thus both of these pocket sized instruments may be just as well suited t
10. Scand J Clin Lab Invest 2000 60 367 380 Examination of four different instruments for measuring blood lactate concentration J I MEDB A MAMEN O HOLT OLSEN amp F EVERTSEN National Institute of Occupational Health Oslo Norway tSogn og Fjordane College Teacher Training Faculty Sogndal Norway Norwegian Olympic Training Centre Oslo Norway Norwegian University for Sports and Physical Education Oslo Norway Medb JI Mamen A Holt Olsen O Evertsen F Examination of four different instruments for measuring blood lactate concentration Scand J Clin Lab Invest 2000 60 367 380 Information on the performance of different instruments used to measure blood lactate concentration is incomplete We therefore examined instruments from Yellow Springs Instruments YSI 23L and YSI 1500 and three cheaper and simpler instruments Dr Lange s LP8 Lactate Pro from Arkray in the KDK corporation and Accusport from Boehringer Mannheim First a number of blood samples were analysed by standard enzymatic photofluorometry our reference method and in addition by one or more of the instruments mentioned above Second measurements using two or more identical instruments were compared Third since Lactate Pro and Accusport are small 100 g pocket size battery driven instruments that could be used for outdoor testing the performance of these instruments was examined at simulated altitudes O pressure of lt 10 kPa and at tem
11. The values from the YSI 1500 were on average 20 higher than those of the reference method p lt 0 001 For the YSI 15002 the values were on average 5 higher than those of the reference method these latter values did not deviate significantly from the line of identity Fig 1B C Blood samples were also measured on three different simpler instruments The values from the LP8 were on average 26 higher than those of the reference method p lt 0 001 Fig 1D These measurements also showed a larger error of regression than the others p lt 0 05 The relationship to the reference method was convex and a second order curve fit was better than a linear one p lt 0 001 Values from measurements on the Lactate Pro were on average 12 higher than those from the reference method p lt 0 001 Fig 1E For all of the instruments mentioned so far all the regression lines found lie close to the origo This means that possible systematic errors in the measurements were small for low blood lactate concentrations and rose roughly proportional to the lactate concentration The Accusport showed two kinds of deviations First the Y intercept was close to 1 0 mmol L p lt 0 001 versus an intercept of 0 0 In line with this the instrument reported blood lactate concentration around 2 mmol L even for blood samples taken at rest that is in samples with a true concentration less than 1 mmol L The slope found for this instrument was 0 81 wh
12. YSI 1500 2 nonhemolyzed blood mmol L D La YSI 1500s nonhemolyzed blood mmo L p gt La Lactate Pro mmol L La enzymatic photofluorometry mmol i La Lactate Pro mmol L Fic 3 A The blood lactate concentration measured on LP8 versus corresponding measurements using enzymatic photofluorometry The value given for the LP8 is the median of three parallel measurements B The lactate concentration measured on non hemolyzed blood samples on the YSI 1500 apparatus versus measurements with the Lactate Pro e shows accepted measurements O shows measurements that were rejected because of an incorrect measurement with the instrument examined and therefore not used in calcu lating the regression parameters C the lactate concentration measured on non hemolyzed blood samples on the YSI 15003 e YSI 15004 and YSI 1500 instruments versus measurements using the Lactate Pro Regression data are given for each set of data and the regression lines are shown as solid lines Syx is the error of regression and is expressed in mmol lactate L blood Sp is the error of the slope The thin dashed lines are the lines of identity The thin long dashed line in C is a copy of the regression line in B corresponding measurements at normal O pressure and temperature Fig 4 The Lactate Pro showed no effect of reduced O pressure The values found with the Accusport at redu ced O pressure were 1 85 0 08 mmol ie mean SD higher
13. ample may not be processed 4 This latter effect although quantitatively less important will add further curvature to the standard curve There is no reason to assume that our use of second order curve fits has caused the possible non linear trends in Fig 1 since the second order component of our standard curves was small and in addition tended to reduce rather than increase non linearities We used a commercial 1 00 mol lactate L stock solution to establish the standard curve and separate laboratory controls showed that the concentration of this stock solution was accurate within 2 95 confidence interval of the control analyses A possible bias of less than 2 or random variation less than 0 3 mmol L of our reference method has no influence on the conclusions drawn in this study Do differences between instruments matter When the same sample was measured on two different instruments or by two different methods we found differences of between 10 and 25 To examine how this may have affected the so called blood lactate threshold that entity here taken as the exercise intensity giving a blood lactate concentration of 3 00 mmol L was established for one subject during bicycling at 1 5 Hz and found to be 3 42 W kg Fig 5 To each measured lactate concentration we added and subtracted 10 and 25 which caused the 3 00 mmol L threshold to appear at 2 and 5 lower higher powers respectively Systemat
14. asurements do not record lactate inside the red blood cells YSI 1500 and YSI 15002 panels B and C refer to measurements on two different YSI 1500 instruments LP 8 panel D is a lactate analyzer from Dr Lange Lactate Pro panel is from Arkray in the KDK Corporation while Accusport panel F is from Boehringer Mannheim Data on the regressions are given for each set of data and the regression lines are shown as solid lines Sy is the error of regression and is expressed in mmol L S is the error of the slope The thin dashed lines show the lines of identity for each set of data e refers to accepted measurements while O refers to outliers that were formally rejected because of an error of measurement with the instrument examined these values were therefore not used to calculate the regression lines We did further comparisons of measurements on unhemolyzed blood samples measured using the YSI 1500 and the Lactate Pro The values found with the YSI 1500 were on average 67 of those found with the Lactate Pro Fig 3B Correcting for a 12 bias in the values from the Lactate Pro see Fig 1E suggests that on non hemolyzed blood the YSI 1500 gives values 25 less than the true value of whole blood For a further evaluation of the YSI 1500 blood samples were measured using three other instruments here called YSI 15003 YSI 15004 and YSI 1500s and by the Lactate Pro Fig 3C The values given with these three YSI 1500s did n
15. d to measure the lactate concentration in a small blood sample within a few minutes or less We have examined the properties and qualities of different instruments All measurements are subject to some random error even if minimized as much as possible Systematic errors can also occur One aim of our study has been to compare the imprecision of different instruments and to look for possible systematic errors In addition we have looked for possible differences between instruments of the same type by measuring the same blood sample on two or more similar instruments Two of the instruments examined the Lactate Pro from Arkray in the KDK Corporation and the Accusport from Boeh ringer Mannheim are battery driven pocket sized instruments that might be suitable for outdoor testing too Thus the performance of these instruments was examined at simulated altitude and in the cold We also examined instruments from Yellow Springs Instruments YSI 23L and YSI 1500 and the LP8 from Dr Lange As our reference we used a standard enzymatic photofluorimetric method SUBJECTS EXPERIMENTS AND METHODS Subjects Healthy young to middle aged trained men and women served as subjects in these experi ments All were told that they were serving as volunteers in our experiments They were also told that they were free to leave at any stage and that they could do so without giving any reason Experiments Most of the exercises were carried
16. d lactate concentration above 10 mmol L We saw no systematic differ ence when the same blood sample was ana lyzed on different instruments of this type The random variations in measurements with the Lactate Pro were similar to those of the YSI instruments and the Accusport Lactate Pro gave reliable results at simulated altitude and in the cold Shimojo et al 16 have also examined this instrument and although they found no bias at high lactate concentration their random variation was larger than ours There is no information on how this instru ment accounts for lactate in the red blood cells If it measures lactate in plasma only and adds an assumed value for the lactate in the red blood cells one would expect too high a reading when fluids without red blood cells are mea sured Measurements of the lactate concentra tion in plasma samples and in control solutions not shown with Lactate Pro gave reliable results for solutions with a lactate concentration less than 10 mmol L Thus this instrument must lyze the red blood cells and thus measures the cell lactate For non blood fluids with a lactate concentration above 10 mmol L the Lactate Pro gave too low values in our measurements This instrument may therefore not be suitable for use when fluids other than blood are analyzed and the lactate concentra tion is high Accusport The Accusport gave correct values only for blood samples with a lactate concen tration of
17. e bias versus the reference method and was therefore regarded suitable as a control when comparing different YSI 1500s These three YSI 1500s showed a larger lactate concentration in unhemolyzed blood than the YSI 15005 judging from the comparisons with the measurements by the Lactate Pro We have no explanation for why different YSI instruments gave different results in our experiments As stated in the Methods section the instruments were operated by experienced test leaders and in accordance with the instruction manual Dr Lange s LP amp This instrument showed values 25 30 too high and the error of regression was nearly twice that of the others Using the median of three parallel measure ments did not reduce the random error much This instrument costs about the same as the Lactate Pro and Accusport and only 10 15 of that of an YSI 1500 Apart from the lower price than for an YSI 1500 we see no advan tage with this instrument compared to the others This instrument performed less well than the others It was more difficult to use in terms of handgrips caps repeatedly off and on to obtain precise measurements fill the capillaries completely without spilling blood on the outside and in terms of not mixing samples in larger series Lactate Pro This instrument showed little bias when blood samples with low to moder ate lactate concentrations were measured while it recorded values 12 too high on sam ples with a bloo
18. found in both plasma and the red blood cells and in most cases 75 of the lactate in blood is found in the plasma compartment 1 2 According to the manuals of the YSI instruments only lactate in plasma is measured unless the blood samples are hemolyzed Dr Lange miniphotometer LP8 Blood was taken using 10 ul end to end capillaries and placed in a reagent solution hemolyzing the blood lactate was processed in a reaction producing quinonimin in proportion to the amount of lactate in the sample and the concentration of quinonimin was read off in an LP8 apparatus Dr Bruno Lange GmbH Berlin Germany at 540 nm 576 THz after a 3 min reaction time 370 J I Medbo et al Lactate Pro The lactate concentration has been measured in blood and plasma samples on altogether seven different LT 1710 Lactate Pro analyzers Arkray Factory Inc KDK Corporation Shiga Japan This instrument measures lactate on a tiny drop of Sul of fluid The lactate concentration is read off after 60 s For our measurements we used strips with production numbers L8D04A F 4 and L8L07B F 7 Accusport The blood lactate concentration was mea sured on two different Accusport portable lactate analyzers type 1488767 Boehringer Mannheim A drop of no less than 15 ul of blood is applied to the strip 6 and 60 s later the lactate concentration is read off In accordance with the manufacturer s instructions we checked the window on
19. gle measurements as judged from the errors of regression in Figs 1D and 3A Blood lactate measured by different instruments 373 YSI 23L YSI 1500 1 YSI 1500 2 6r 6 Z 3 Z EON amp T E 4 F 3 F F 5 EVER E E E Z 4 a RT 3 a z 2 25 2p a aL g 2h F Ci a Y 0 11 40 776 x r 0 987 5 Y O0 214 1 13 x r 0 974 S Y 0 06 1 00 x r 0 989 Sy 0 20 S Da n 91 Sy 7 0 33 S 0 03 n 98 S 0 23 S 0 03 n 32 0 ks j 4 0 us ofp ae tg 0 ye CA aa o 2 4 6 o 2 4 6 0 2 4 6 La enz photofluorometry mmol t La enz photofluorometry mmol 1 La enz photofluorometry mmol Ly Accusport LP 8 Lactate Pro P o E 6b Z rE oo a E f E E E E a Z4 E 7 A 5 E 2 eo 2b a2 5 J H rie 7 7 Y 02 1 25x r 0 91 5 0 13 1 05 x r 0 975 7 Y 0 81 t O88 x r 0 974 Sya 0 51 S 0 08 n 57 O Sy 0 31 S 0 03 n 88 Z Syp 0 23 S 0 04 n 33 0 0 0 1 1 P 1 J 0 2 4 6 0 2 4 6 0 2 4 6 Lal enz photofluorometry mmol L 5 La enz photofluorometry mmol L h La enz photofluorometry mmol L 5 Fic 2 The blood lactate concentration measured by enzymatic photofluorometry X and by six different instruments Y The data are the same as in Fig 1 but only samples with a blood lactate concentration less than 6 mmol L as measured by enzymatic photofluorometry are included The samples measured by the YSI 23L panel A were on non hemolyzed blood and these me
20. ic differ ences of that magnitude may be important when judging possible changes for an athlete Thus for precise testing of possible changes by time instruments known to have similar properties should be used There are other definitions of the lactate threshold For blood lactate concentrations higher than 3 mmol L7 the curve is steeper and if for example the 4 mmol L threshold is used systematic differences between instru ments have less effect For lower lactate concentration the curve is less steep and small variations in the measured value will have a larger effect on the reported threshold For long distance running bicycling and skiing the intensity corresponding to a blood lactate concentration of 2 mmol L may be of inter 376 J L Medbo et al Blood lactate concentration mmol L an T EA 00 25 3 0 3 5 4 0 4 5 Power W kg Fic 5 Measurements of the blood lactate concen tration by stepwise increments of the power during bicycling e and thick solid line Each step of con stant power lasted 5 min and a 1 min rest separated each step to allow blood sampling and setting a higher power To each measured value first 10 thin dashed dotted lines and 25 was added and subtracted thin dashed lines The dotted line shows the level of 3 00 mmol L and the power corresponding to that value is here taken as the lactate threshold est We found that the Accusport showed this
21. ich is considerably less than 1 00 p lt 0 001 Thus this instrument showed correct values for blood samples with a lactate concentration of 5 mmol L For samples with less lactate the reported values were systematically too high while for samples with higher concentra tion the reported values were too low When the so called blood lactate threshold is sought in sports testing particularly interest ing is the performance of the instruments at low to moderate blood lactate concentrations The same is true for most clinical testing Therefore blood samples with a lactate concentration above 6 mmol L as measured by enzymatic photofluorometry were left out and the remain 372 J I Medbo et al YSI 23L YSI 1500 1 YSI 1500 2 15 20 r o 15 Z e fee alt ca a n Zast z a 3 E 3 10 E E 10 7 S a F S 5 a 5 x z T Sor a Y 0 07 0 786 x r 0 995 3 s Y 0 10 1 204 x r 0 992 Ss Y 0 17 1 056 x r 0 993 Sy 0 40 S 0 006 n 160 Sy 0 66 S 0 013 n 15 Sy 0 52 S 0 020 n 44 0 Paes career 1 1 0 1 POER 1 1 0 A is ae ae as 0 5 10 15 0 5 10 15 0 5 10 La enz photofluorometry mmol L La enz photofluorometry mmol L La enz photofluorometry mmol L LP 8 Lactate Pro Accusport aii D e E is F H A 15 ae a f SP 3 3 5 3 z 10 E T E E E z 10 7 lh a 2 z z 3 a f 3 25 we sL A a5 a Y O 17 1 264x r 0 991 F E
22. ightly larger errors found for the YSI 1500s the Lactate Pro and the Accusport are a consequence of small variations between blood samples rather than less precise measurements by these instruments Statistical considerations suggest that the random error of measurements with each of these three instru ments is roughly half of the reported errors of regression in Figs 1 and 2 Blood lactate measured by different instruments 377 YSI instruments We used one YSI 23L and five YSI 1500s On non hemolyzed blood sam ples the YSIs showed values 20 25 less than the reference method This finding agrees with data of Foxdal et al 14 and Lormes et al 15 It is clearly stated in the YSI instru ment user manual that if unhemolyzed blood is measured lactate in the red blood cells is not recorded It is well known that after a few minutes of equilibration around 25 of the lactate in blood is found in the red blood cells e g 1 2 Thus one must expect too low apparent blood lactate concentrations when unhemolyzed blood samples are mea sured We did parallel measurements on two differ ent YSI 1500 instruments Both had fresh membranes were calibrated using the same solutions and by the same pipette and handled by the same test leader Nevertheless these instruments showed a 12 systematic difference on the same samples Three other YSI 1500s were examined and tested versus the Lactate Pro which showed a systematic but reproduci bl
23. ndard curve was 0 1 0 2 mmol lactate L blood or plasma A standard solution of 1 00 mol lactate L7 L lactat e standard 125 440 batch 66762401 from Boehringer Mann heim was used for making the samples for the standard curve The total imprecision of the method is 0 1 0 3 mmol L depending on the lactate concentration in the measured sample The volume fraction of water in whole blood was taken as 84 and that of plasma as 94 5 Since we found systematic differences between the results given by the tested methods and our reference method the standard solution from Boehringer Mannheim was calibrated independently by an enzymatic reaction similar to that described above 4 The increase in the NADH concentration was read off in a Shimadzu MPS 2000 spectrophotometer at 340 nm 882 THz in quartz cuvettes type 18 B Starna Essex UK where the light traversed 10 00 mm The lactate concentration was taken from the measured increase in the NADH concentration using a coefficient of extinction of NADH of 6270 m7 mol These measure ments showed that the lactate concentration of the standard solution did not differ from the reported value and the imprecision SE in this calibration was 1 That imprecision was included when we examined whether there was a systematic difference between a tested instrument and the reference method Otherwise the possible error in the reference method is considerably less than the systematic
24. o outdoor testing as for use in the laboratory Admittedly the variations in the measurements being larger in the cold was probably caused by technical problems since it was more difficult to add a drop of the test solutions from the nipple of the flasks than to add a drop of blood from a fingertip to the strips particularly for the Lactate Pro in the cold Evaluation of the different instruments Common comments The error of regression is a simple and usually the best measure of how well the model fits the data In Fig 1 the random variations rise by the lactate concen tration and the reported error of regression in that figure is a mean of the error at low and high concentrations In Fig 2 where only samples with a concentration less than 6 mmol L are included the error of regres sion is less and within the range of data in that figure the random variations seem inde pendent of the concentration With the exception of the data on the YSI 23L Figs 1A and 2A parallel measurements on the same blood sample were done by taking two or more samples from the same finger in sequence that is 10 30 s apart and in random order The reported error of regression depends on the imprecision in the measurements by enzymatic photofluorometry X in the instrument examined Y and in possible variations between blood samples taken some seconds apart The error of regression was least for the YSI 23L It is likely that the sl
25. occur in all measurements When measurements with two methods or instruments turn out as expected the values will fall close to a line If there is something wrong with one of the measurements the corresponding point will deviate from the linear relationship A simple scatterplot alone cannot be used to decide whether the deviation seen is due to an error in the abscissa X or in the ordinate Y For a number of the samples analyzed here each sample was measured using at least three different instruments or methods here called X Y and Z If a point appears as an outlier in the XY plot and the YZ plot but not in the XZ plot it is likely that the deviation is caused by an error in the Y measurement We used this principle to unravel possible errors of measure ments and outliers RESULTS Blood samples were taken during different experiments and the blood lactate concentration was measured using at least one of the instruments to be tested Y and in addition by enzymatic photofluorometry X serving as our reference method Fig 1 Non hemolyzed blood samples measured on the YSI 23L showed on average a 22 lower value than the reference method p lt 0 001 Fig 1A Since the YSI 23L does not hemolyze the blood it does not record lactate in the red blood cells and this may have caused the systematic deviation Hemolyzed blood samples were measured on two different YSI 1500 here referred to as the YSI 1500 and the YSI 15005
26. ot differ systematically but for a given value as measured by the Lactate Pro they gave a 13 higher value than the instrument we called YSI 15005 The instruction manual gives no information on how the Lactate Pro reflects lactate in the red blood cells and how it responds to plasma Therefore plasma from blood samples taken after intense bicycling was measured by enzy matic photofluorometry and by the Lactate Pro For samples with a lactate concentration less than 10 mmol L the Lactate Pro responded as for blood samples that is the values found by the Lactate Pro were up to 10 higher than those of the reference method For samples with a higher plasma lactate concentration the curve leveled off which means that the Lactate Pro showed too low values not shown Measurements at simulated altitude The lactate concentration in two control solutions was measured in a chamber with an O2 pressure less than 10 kPa corresponding to that found at an altitude of more than 6 km above sea level The values were compared with 374 J I Medbo et al LP 8 vs enzymatic photofluorometry YSI 1500 2 vs Lactate Pro YSI 1500s vs Lactate Pro A Y 0 18 1 28 x Sy 0 86 o S 0 04 n 59 r 0 98 S w 20 o Y 0 14 0 668 x n 142 Sy 0 19 S 0 010 r 0 993 Y 0 25 0 758 x n 70 Sy 0 23 S 0 012 r 0 983 La Dr Langes LP8 mmol L 0 5 10 15 w La
27. other one could not read the code on the strips used even when the instrument s window was properly cleaned Switching batteries and strips between the two instruments showed that the problems lay in the instruments and not in the strips or batteries SUMMARY AND CONCLUSIONS Different instruments gave different values when the lactate concentration of a blood sample was measured The differences were usually in the range 10 25 and a bias of this amount may have some effect when athletes are tested Of the four instruments examined the Lactate Pro was best The Lactate Pro cost only 15 of an YSI 1500 and is much easier to use Different YSI instruments gave different values The LP8 was inferior to the others Both Lactate Pro and Accusport can be used for outdoor testing in the cold Lactate Pro can also be used for testing at altitude ACKNOWLEDGMENT We are grateful to Jorid Thrane Stuenzs for her skilled technical assistance REFERENCES i w nn co 10 1 er 12 13 14 15 Smith EW Skelton MS Kremer DE Pascoe DD Gladden LB Lactate distribution in the blood during progressive exercise Med Sci Sports Exerc 1997 29 654 60 Medb JI Hanem S Noddeland H Jebens E Arterio venous differences of blood acid base status and plasma sodium caused by intense bicycling Acta Physiol Scand 2000 168 311 26 Westgren M Divon M Horal M Ingemarsson I Kublickas M Shimojo N Nordstr m L Routine
28. peratures below 20 C while screening the instruments as much as possible from the cold Most of the different instruments showed systematically too high or too low values 10 25 deviation The observed differences between instruments may affect the blood lactate threshold by 2 5 We found different readings between equal YSI 1500 instruments while we could see no difference when comparing the other instruments of the same type Lactate Pro gave reliable results at both 21 1 C and at simulated altitude Accusport gave reliable results in the cold but 1 85 0 08 mmol L7 mean SD too high readings at the simulated altitude Of the three simpler instruments examined the Lactate Pro was at least as good as the YSI instruments and superior to the other two Key words Bicycling blood exercise lactate lactate threshold plasma testing training Jon Ingulf Medbo National Institute of Occupational Health PO Box 8149 dep N 0033 8149 Oslo Norway Tel 47 23 19 51 00 fax 47 23 19 52 01 e mail Jon Medbo stami no 367 368 J I Medbo et al INTRODUCTION Blood lactate concentration is often measured in connection with the training and testing of athletes 1 2 and also of patients 3 Tradi tional methods in the laboratory are time consuming but now faster and simpler methods have been developed as a consequence of technological development Today there are several different instruments that can be use
29. rresponding to an altitude of more than 6 km above sea level In practice No gas was led into a bag serving as the chamber The O2 pressure in the bag was measured using a Metamax 1 O analyzer Cortex Biophysic GmbH Leipzig Germany and no analysis was done unless the O pressure was less than 10 kPa Each standard solution was measured at least 10 times with each of the Lactate Pro and Accusport instruments As a control the same control solutions were measured at least 10 times in normal laboratory conditions that is at a temperature of 22 1 C and an air pressure of 100 kPa corresponding to an O3 pressure of 21 kPa Measurements in the cold Each of the control solutions from Boehrin ger Mannheim 3 6 and 9 mmol L7 nominal lactate concentration was measured at least 10 times on the Lactate Pro and Accusport instruments in a freezing storage room that is at a temperature of 21 1 C The test leader wore a thick sweater and coat To shield the instruments as much as possible from the cold the test leader wore a bag under the sweater and thus close to the body While the strip was inserted into the instrument and the control solution was applied to the strip actions that took 20 30 s the instrument and the strip were exposed to the cold As soon as the analysis started the instrument was placed in the bag and was thus no longer exposed to the cold The instrument and the strips lay in the bag between each analysis too
30. sample with a known lactate concentra tion varies from day to day due to interassay variations and consequently a new standard curve is made for each day of analysis While standard curves are usually fitted by linear regression we used second order regres sion First a slight curved relationship was readily visible samples with a high lactate concentration showed a smaller fluorescence than a linear extrapolation dictates Second NADH is produced in proportion to the amount of lactate in the sample There may Blood lactate measured by different instruments 375 1 BM 3 6 mmol L 5 Control mmm Lypoxia Po lt 10kPa 4 zzz Cold 21 C Lactate concentration mmo L Lactate Pro Accusport BM 9 mmol L Lactate concentration mmol L Lactate Pro Accusport Fic 4 The lactate concentration of two control solutions from Boehringer Mannheim with a nominal lac tate concentration of 3 6 and 9 mmol L7 as measured by the Lactate Pro and the Accusport in control experiments at room temperature and normal air and O pressures at room temperature but with an O pres sure below 10 kPa and at normal air and O pressures but at 21 1 C be a 5 quenching of the signal in a cuvette with a NADH concentration of 10 umol L 4 and since the effect is proportional to the concentration a parabolic relationship is expected Finally in samples with a high lactate concentration 1 2 of the lactate in the s
31. son Wesley 1962 p 63 87 Helgerud J Ingjer F Stromme S Sex differences in performance matched marathon runners Eur J Appl Physiol 1990 61 433 9 Foxdal P Bergqvist Y Eckerblom S Sandhagen B Improving lactate analysis with the YSI 2300 GL hemolyzing blood samples makes results comparable with those of deproteinizing whole blood Clin Chem 1992 38 2110 4 Lormes W Steinacker JM Stauch M Lactate determination with the Accusport system and a fully enzymatic photometric method in an incre mental stage test and in prolonged exercise In Ramstetter E Zieres Nauth C Mack M editors Workshop Report Accusport Z rich 1994 Boehringer Mannheim 1995 p 37 41 Blood lactate measured by different instruments 379 16 Shimojo N Naka K Uenoyama H Hamamoto K Yoshioka K Okuda K Electrohemical assay system with single use electrode strip for measur ing lactate in whole blood Clin Chem 1993 39 2312 4 17 RoBkopf P Lamprecht W Liesen H The Accu sport analyzer and its operation In Ramstetter E Zieres Nauth C Mack M editors Workshop Report Accusport Z rich 1994 Boehringer Mannheim 1995 p 33 6 Received 29 February 2000 Accepted 10 May 2000
32. than those found in the control experiments p lt 0 001 Measurements in the cold The lactate concentration of the control solutions was measured at 21 1 C Fig 4 The instruments were exposed to the cold while a strip was inserted and the control solution was applied to the strips Otherwise the instruments were well shielded from the cold The mean values did not differ from those of the control experiments p gt 0 5 for each instrument but the random variations were larger p lt 0 01 DISCUSSION The main results in this study are first that for all the instruments examined the reported blood lactate concentration rose roughly linearly by the value given by the reference method Most of the instruments showed either systematically too high or too low values and the deviations were mainly in the range 10 25 Apparently equal YSI 1500 instruments showed different values on the same blood samples Lactate Pro gave reliable results at simulated altitude and in the cold experiments Accusport gave too high readings at simulated altitude while the values in the cold were reliable The LP8 was the most biased instrument and in addition showed the largest random variations Evaluation of the enzymatic photofluorometry method used here We measured the lactate concentration in blood and plasma by a method that has been in use for more than 30 years in numerous studies It is common knowledge that the fluorescence from a
33. y A volume of 25 or 50 ul of whole blood or plasma was taken by Acupette capillary tubes P4518 50 Dade Diagnostics Inc Puerto Rico USA according to the manufacturer the accuracy of the tubes volume is better than 0 5 and the coefficient of variation better than 1 The fluid was transferred to a tube containing 500 ul of 0 4 mol L7 perchloric acid PCA that lyzes the red blood cells and thus frees lactate inside the cells These tubes were stored frozen at 20 C until later analyses of the lactate concentration by a method according to Passoneau and Lowry 4 which uses the increase of NADH concentration in the LDH reaction lactate dehydrogenase EC 1 1 1 27 from beef hearts L2625 type III Blood lactate measured by different instruments 369 from Sigma St Louis MO USA Pyru vate produced was further processed by the glutamate pyruvate transaminase reaction l alanine 2 oxoglutarate aminotransferase EC 2 6 1 2 from swine heart Boehringer Mann heim GmbH Mannheim Germany thus allowing almost complete processing of all lactate in the sample The fluorescence was read off in an RF 5000 spectrofluorophoto meter Shimadzu Kyoto Japan at 460 nm 652 THz after 45 min incubation at 23 5 C For each analysis a new second order standard curve covering the whole range of fluorescence values was made The error of regression scatter around the regression line the statistical error in reading off from the sta
Download Pdf Manuals
Related Search