THROMBOSIS RESEARCH 60; 477-487,199O 0049-3848190 $3.00 + .OOPrinted in the USA. Copyright (c) 1990 Pergamon Press pk. All rights reserved.
THE INFLUENCE OF SAMPLE AGE ON COLLAGEN-INDUCED PLATELET AGGREGATION IN WHOLE BLOOD
MiillerMR, Schreiner W, Wohlfahrt A, Salat A, Wolner E Department of Surgery 2, University of Vienna Spitalgasse 23, A-1090 Vienna, Austria (Received 28.9.1990; accepted
ABSTRACT
inrevised form4.10.1990 by Editor H.Vinazzer)
Whole blood electrical aggregometry (WBEA) has become an accepted method to gain quick information on platelet disorders. Compared to the optical method WBEA is closer to physiology and less complicated, but on the other hand more difficult to standardize. Different approaches have been attempted in the past to improve the reliability and practicability of this technique. The influence of sample age has not been defined so far. In a first step a mathematical modelling programm was established, which is able to characterize the aggregation curves obtained after collagen stimulation. In a mathematical analysis various characteristics of the curve function were calculated and their sensitivity for aging investigated. Regression was performed for each characteristic, and correction factors defined.
Our results indicate, that whole blood specimen for collagen induced aggregation can be used without correction factor up to 30 minutes. Data obtained with an age exeeding half an hour have to be corrected following a quadratic regression.
Key words: platelet aggregation, impedance method, curve-analysis, computer 477
sample-age,
478
THE INFLUENCE OF SAMPLE AGE...
Vol. 60, No. 6
INTRODUCTION Since the introduction of whole blood electrical aggregometry (WBEA) in the year 1980 by David C. Cardinal and Roderick J. Flower (1) this method has developed to a valuable clinical tool for studying platelet behaviour. The technique bases on the change of impedance between a pair of electrodes, caused by an accumulation of aggregating platelets on them. In the years before this innovation the optical method using platelet rich plasma (PRP) has been extensively used to study platelet aggregation. This technique was described by Born in 1962 (2) and bases on the increase of light transmittance of platelet rich plasma caused by the formation of aggregates and a decrease in the volume occupied by the platelets in a photoelectric cell (3). Although of proved diagnostic value, this technique has certain limitations. Since only PRP can be used, centrifugation is required for sample preparation, which is time consuming on one hand and probably influences platelet function. Red cells (RBC) and white cells (WBC) together with abnormal heavy platelets are removed from the final platelet suspension to be tested. These cells probably contribute as aggregation modulators through adenine nucleotide uptake, release of prostanoides and binding of prostacyclin (PGI2) (4,5). In the beginning, the new technique was mainly used as a quick screening method for detection of platelet function disorders (6). For more detailed studies, still the Born aggregometer was prefered. Although the results obtained by the two methods were similar, the parameters measured with the novel device were not clearly understood at that time. (7) The main disadvantage of the whole blood method is the problem of standardizing the condition of the specimen. Various studies have described the influence of the platelet count and hematocrit on rate and extent of in vitro aggregation in whole blood. (8,9) Platelet 8 ount has been reported to be without influence beyond 50x10 /l (8) and the aggregation of human platelets (WBEA) was enhanced at lower values of hematocrit. (9) Optimal responses occured after dilution of the samples to hematocrit of 30%. Physiologic saline solution (PSS) was found to be the diluent of choice in this connection. (8) The technique has been further investigated and improved during the first years following its invention. (10,11,12) Nevertheless unphysiological conditions of the specimen must be taken into account, which additionally might influence the obtained data. We wanted to investigate the influence of sample age on various characteristics of the aggregation curve after collagen induced platelet aggregation using the impedance method with whole blood. This aspect might be of interest, when numerous samples have to be investigated within limited time. The study was planned in two steps. Within a first step we represented the aggregometry plots mathematically and performed curve analysis. In the second step, the parameters obtained from curve analysis were investigated for their sensitivity for the aging of the samples.
Vol. 60, No. 6
THE INFLUENCE OF SAMPLE AGE...
479
MATERIAL AND METHODS Whole blood electrical aggregometry (WBEA): In this study an impedance aggregometer from CHRONO-LOG CORPORATION, 2 West Park Road, Havertown, PA, was used. Collagen (COLL) was used as aggregation trigger in titrated whole blood (CWB). Na-citrate was choosen, because heparin is reported to lead to a slight platelet activation and release of platelet factor 4 (PF4). On the other hand citrate anticoagulated blood contains less free calcium ions, although the remaining calcium should be sufficient to allow a normal platelet response. Blood was taken by puncture of the cubital vein from 10 healthy volunteers (5 men, 5 women) at the age of 23-33 years (mean=29.4), all non-smokers and without any medical treatment for at least 14 days. Blood collection was performed with the same type of cannula and by the same person directly into a vacutainer tube containing 10% of Na-citrate. After incubation of the specimen in the aggregometer, platelet aggregation was triggered by adding 1.5 ul of 0.1% type-l-collagen, suspended in isotonic glucose solution of pH 2.7 (= 1.5 ug of collagen), to 1000 ul of whole blood. Collagen reagents were from CHRONO-LOG CORPORATION (Chrono-Par No. 3801, 2 West Park Road, Havertown, PA, 190834691. The change in impedance over time was registered in a strip chart recorder (Chrono-Log) and the plotted aggregometer curves were then digitized and analysed on an IBM XT PC. Data Acquisition: Aggregation curves were digitized on a graphics tablet (Summagraphics) connected to an IBM-XT computer. Together with the digitized values for impedance, Y, the following supplementary parameters were stored for each curve: a 5 Ohm calibration signal / patient-ID / age of sample Calculation of Characteristic Quantities for Impedance Curves: By use of the 5 Ohm calibration signal, the original data were scaled and then analyzed along the following lines. (figure 1) The maximum (Y-MAX) and minimum (Y-MINI impedance as well as the time at which impedance was minimum (T-YMIN) were detected. The amplitude (AMPL) was obtained as AMPL=Ymax-Ymin. Next, three integrals were obtained: (1) The impedance was integrated over time between initiation of aggregation by adding the agonist (t=O) and the end of each measurement (t=tmax) using the trapezoidal rule. Note that this integral (ITOT) corresponds to the area between the impedance curve and the time-axis. (2) The value of ITOT was subtracted from the rectangular area defined by the duration of measurement and the maximum impedance found: IPART = Tmax * Ymax - ITOT. Note that the integral IPART represented the area above the impedance curve, with Y = Ymax as its upper border, and was thus in some respect "complementary" to ITOT. (3) Impedance was integrated from the time where the minimum impedance occured (Tymin) to the end of each measurement (tmax) to yield the quantity "IMINI".
480
THE INFLUENCE OF SAMPLE AGE...
Vol. 60, No. 6
.j. -i 1
0
1
2
3
4
6 5 TIME (minutes)
7
8
9
Tmax
10
11
FIGURE 1 Scheme of characteristic quantities of collagen-triggered WBEA-curves. Calibration signal indicated as vertical bar. In a final step the impedance curve was analyzed by an adaptive fitting procedure in order to locate the time (TINFL) and the slope (SINFL) of the inflection point. Since each impedance curve is characterized by a minimum followed by an asymptotic increase, polynomials could not be used for the fit. Instead, data were transformed according to mathematical functions, which intrinsically exhibit the correct shape: yb (t) = 1 / (1 + b * (t - Tymin) ** 2) For the parameter b we choose five different values (0, 0.01, 0.02, 0.1 and 0.5) in order to generate five independent transformations for the original impedance values. Transformed data were considered as five new variables and were subjected to a multiple least squares fit, linear in the fitting parameters. For each impedance curve, the best subset of transformed variables was determined in a stepwise procedure, using the goodness-of-fit as a criterion to include or exclude one of the transformed variables (F-test). The performance of the criterion for the selection of the best fit was found capable of adequately found various shapes curves of representing impedance experimentally. From the fitted curve, timewise location and impedance of the inflection point were calculated analytically. The calculations were coded in FORTRAN, and routines of the CERN-ScientificSubroutine-Package were used for the matrix in-version necessary to obtain the parameters of the curvilinear fits.
Vol. 60, No. 6
THE INFLUENCE OF SAMPLE AGE...
481
Statistical Evaluation: Each of the calculated parameters described above was obtained from each aggregation curve. Means and standard deviations were calculated for each sample age examined. Results were then plotted over sample age, see figures 2 through 7. Additionally, the dependence of each parameter on the age of the blood sample was analyzed statistically. For each calculated parameter a polynomial regression was performed, involving an intercept, a linear and a quadratic term. (Tab. 1+2) RESULTS Parameters calculated from aggreuation curves and their dependence on the age of the blood sample: Figures 2 trough 7 give calculated means and standard deviations for each age-ofsample, plotted over sample-age. The results (see also table 1) revealed that, for most of the parameters, a straight line was insufficient to represent the dependence on sample-age, cf. the significance of the p-values for the quadratic terms. Inclusion of the quadratic terms provides adequate representation of the original data except for 30-minutes values. This is caused by two outliers, which influence the mean. The plots show, that up to 60 minutes only slight changes of the parameters occur, after another 60 minutes the difference against t=O is 25-30%, and the 150-minutes value is 40-50% of the initial. Parameters increasing with time changed to the same extent.
FIGURE 2 Dependence of parameter AMPL (Max-Ymin) on sample age
482
THE INFLUENCE OF SAMPLE AGE...
e
Vol. 60, No. 6
100
50
AGE OF SNIPLE
(mini
FIGURE 3: Dependence of parameter ITOT (integral between initiation of aggregation 't=O' and end of measurement 'Tmax') on sample age.
-----_______ “e+T 150 138 ‘____------f _’
118.. 9070
AGE OF SAMPLE
bin)
FIGURE 4: Dependence of parameter IMINI (integral between Tymin and Tmax) on sample age.
THE INFLUENCE OF SAMPLE AGE...
Vol. 60, No. 6
--_--------_______________
2I
AGE OF SfflPLE Cmin)
FIGURE 5 Dependence
of the parameter SINFL (slope at the inflection point) on sample age
0
58
158
100 AGE
OF WIPLE
200
bin)
FIGURE 6 Dependence
of parameter TINFL (time of the inflection point) on sample age
484
THE INFLUENCE
OF SAMPLE
AGE OF SAMPLE
AGE...
Vol. 60, No. 6
hi n)
FIGURE 7: Dependence of parameter WIN (time of minimum impedance) on sample age. Statistical analysis of the dependence of each parameter on the age of the blood sample: For each calculated parameter a polynomial regression was performed, involving an intercept, a linear and a quadratic term. (Table 2) The linear terms provided only insufficient representation of the original data. The R-squares for the fitting of the quadratic terms with the parameters are given in Table 1. TABLE 1: R-Squares
for
the Fitting of quadratic with Curve-Parameters
Terms
R-square Parameter ____________________~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 0.789 AMPL 0.874 ITOT 0.666 IPART 0.393 TINFL 0.649 SINFL 0.790 IMINI 0.602 TYMIN ____________________~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Vol. 60, No. 6
THE INFLUENCE OF SAMPLE AGE...
485
TABLE 2: Dependence of calculated Parameters of WHEA-Curves on Sample-Age For each parameter, the first line gives the fitted value, the second line gives the estimated standard deviation (sd) and the p-value, indicating if the the associated line third corresponding term (intercept, linear component, quadratic component) is significantly different from zero. intercept linear Parameter quadratic ____________________~--~~---------~~~~~~~~~--~~~~~~--~~~~~~~~~~~ -0.0005 0.014 21,844 AMPL 0.0001 0.025 sd 1.448 0.0008 0.585 P 0.0001 -0.0049 0.151 113.982 ITOT 0.0010 0.193 sd 10.855 0.0001 0.438 P 0.0001 -0.0021 0.177 63.676 IPART 0.0006 0.108 sd 6.086 0.0014 0.112 P 0.0001 -0.00009 -0.010 1.779 TINFL 0.00003 0.006 sd 0.380 0.012 0.117 P 0.0001 -0.0002 0.015 8.229 SINFL 0.00009 0.017 sd 0.987 0.0067 0.398 P 0.0001 -0.0044 0.129 159.514 IMINI 0.0011 0.209 sd 11.710 0.0006 0.540 P 0.0001 -0.00003 -0.0006 0.681 TYMIN 0.00001 0.0021 sd 0.122 0.028 0.759 P 0.0001 ___________-________~~~~~~~~~~~~~~~~~~~~~~~-~~~~~~~~~~~~________ DISCUSSION We have studied the influence of sample age on various parameters of the aggregometry curves obtained with the electrical impedance aggregometer. The electrical method offers results obtained under physiological conditions, time renders consuming sample preparation unnecessary, and thus can easily be used as part of a routine screening procedure. Correct interpretation of the obtained data requires standardized conditions. testing Influences of hematocrit and platelet count have been described by several authors (7,8,9), but the importance of the age of sample has not been mentioned so far. Although one blood sample can be analysed within 15 minutes using WHEA, delays in processing may occur with larger numbers of specimen. Knowledge of the time dependence of the results is necessary to standardize this method. The use of micro-computers for aggregation instruments has been described by other authors (13,141. However, the influence of varying sample conditions have not been clearly defined so far.
486
THE INFLUENCE OF SAMPLE AGE...
Vol. 60, No. 6
Our results show, that blood samples for whole blood aggregometry undergo an aging process, which influences the obtained data obviously after 30 minutes. Specimen exceeding an age of one hour should be investigated as usual, and the data corrected according the calculated quadratical terms. If longer delays in investigating the samples are to be expected, the specimen should be stored at 6OC and rewarmed in the heater block of the aggregometer for about 15 minutes before processing. The calculated parameter ITOT was found most sensitive to aging. Only 20% of the value at t=O was observed after 150 minutes. All other parameters with decreasing tendency were 45% to 50% at that time, whereas T-YMIN and T-INFL had increased by 100%. At the moment we are not able to explain the aging process physiologically. Several mechanisms are possible: Platelets may be altered by acidosis of the specimen.
unphysiological blood gases and
Platelets and other blood cells may undergo slight release reactions and leakage during storage under physiological temperature. Slightly elevated plasma ADP levels might desensitize ADP receptors, resulting in reduced response to different agents. Other cells may play a role in increasing the base excess through products of metabolism. Proteins may attach to the siliconized test vials resulting in a change of the percentual plasma distribution and influencing the activity of important modulators of platelet aggregation. (151 Acidosis can change surface ions of plasma proteins, which might influence normal aggregation. Substances which regulate platelet function such as prostacyclin and cyclic AMP may degrade to biologically inactive products. To evaluate the role of different blood cells, plasma proteins and blood gases on the described aging process, further investigations are necessary. Additionally, the effect of sample age on WBEA with triggers other than collagen will be of interest. Furthermore the influence of hematocrit, platelet count and other variables modulating platelet function have to be clearly defined, to be able to standardize WBEA by means of software rather than dilution of the samples. REFERENCES 1.
The electronic aggregometer: A Cardinal DC, Flower RJ. novel device for assessing platelet behavior in blood. J Pharm Meth;3:135-158,1980
2.
Born GVR. Quantitative investigation into the aggregation of blood platelets. J Physiol London;16:67-68,1962
Vol. 60, No. 6
THE INFLUENCE OF SAMPLE AGE...
467
3.
Effects of the numbers and sizes of Born GVR, Hume M. platelet aggregates on the optical density of plasma. Nature 215:1027-1029,1967
4.
Flower RJ, Russell-Smith N, Salmon JA, Blackwell GJ, Thorogood PB, Vane JR. Prostacyclin is produced in whole blood. Br J Pharmaco1;64:436,1978
5.
Willems C! Stel HV, van Aken W, van Mourik JA. Binding and inactivation of prostacyclin (PGI2) by human erythrocytes. Br J Haemato1;54:43-52,1983
6.
A quick method for Ingerman-Wojenski C, Silver MJ. screening platelet dysfunctions using the whole blood lumiaggregometer. Thromb. Haemostas.;51:154-156,1984
7.
Evaluation of Ingerman-Wojenski C, Smith B, Silver MJ. comparison with optical electrical aggregometry: aggregometry, secretion of ATP, and accumulation of radiolabeled platelets. J Lab Clin Med;101:44-52,1983
a.
Platelet impedance Mackie IJ, Jones R, Machin SJ. blood its inhibition by aggregation in whole and antiplatelet drugs. J Clin Patho1;37:874-878,1984
9.
Galvez A, Badimon L, Badimon JJ, Fuster V. Electrical aggregometry in whole blood from human, pig and rabbit. Thromb. Haemostas.;56:128-132,1986
10.
Difficulty in Ingerman-Wojenski C, Smith B, Silver MJ. detecting inhibition of platelet aggregation by the impedance method. Thromb. Res.;28:427-432,1982
11.
Russell-Smith N, Flower RJ, Cardinal DC. Measuring platelet and leucocyte aggregation/adhesion responses in very small volumes of whole blood. J Pharm Meth;6:315-333,1981
12.
Ingerman-Wojenski C. Simultaneous measurement of platelet aggregation and the release reaction in platelet-rich plasma and in whole blood. J Med Technol;l:697-701,1984
13.
Huzoor-Akbar, Romstedt K, Manhire B. Computerized aggregation instruments: A highly efficient and versatile system for aquisition, quantitation, presentation and management of platelet aggregation data. Thromb. Res.;32:335-341,1983
14.
Francis JL. Improved assessment of impedance aggregometry using a microcomputer. Thromb. Res.;45:851-855,1987
15.
Orchard MA, Robinson C. Stability of prostacyclin in human plasma and whole blood: studies on the protective effect of albumin. Thromb. Haemostas.;46:645-647,198l
THE INFLUENCE OF SAMPLE AGE ON COLLAGEN-INDUCED PLATELET AGGREGATION IN WHOLE BLOOD
MiillerMR, Schreiner W, Wohlfahrt A, Salat A, Wolner E Department of Surgery 2, University of Vienna Spitalgasse 23, A-1090 Vienna, Austria (Received 28.9.1990; accepted
ABSTRACT
inrevised form4.10.1990 by Editor H.Vinazzer)
Whole blood electrical aggregometry (WBEA) has become an accepted method to gain quick information on platelet disorders. Compared to the optical method WBEA is closer to physiology and less complicated, but on the other hand more difficult to standardize. Different approaches have been attempted in the past to improve the reliability and practicability of this technique. The influence of sample age has not been defined so far. In a first step a mathematical modelling programm was established, which is able to characterize the aggregation curves obtained after collagen stimulation. In a mathematical analysis various characteristics of the curve function were calculated and their sensitivity for aging investigated. Regression was performed for each characteristic, and correction factors defined.
Our results indicate, that whole blood specimen for collagen induced aggregation can be used without correction factor up to 30 minutes. Data obtained with an age exeeding half an hour have to be corrected following a quadratic regression.
Key words: platelet aggregation, impedance method, curve-analysis, computer 477
sample-age,
478
THE INFLUENCE OF SAMPLE AGE...
Vol. 60, No. 6
INTRODUCTION Since the introduction of whole blood electrical aggregometry (WBEA) in the year 1980 by David C. Cardinal and Roderick J. Flower (1) this method has developed to a valuable clinical tool for studying platelet behaviour. The technique bases on the change of impedance between a pair of electrodes, caused by an accumulation of aggregating platelets on them. In the years before this innovation the optical method using platelet rich plasma (PRP) has been extensively used to study platelet aggregation. This technique was described by Born in 1962 (2) and bases on the increase of light transmittance of platelet rich plasma caused by the formation of aggregates and a decrease in the volume occupied by the platelets in a photoelectric cell (3). Although of proved diagnostic value, this technique has certain limitations. Since only PRP can be used, centrifugation is required for sample preparation, which is time consuming on one hand and probably influences platelet function. Red cells (RBC) and white cells (WBC) together with abnormal heavy platelets are removed from the final platelet suspension to be tested. These cells probably contribute as aggregation modulators through adenine nucleotide uptake, release of prostanoides and binding of prostacyclin (PGI2) (4,5). In the beginning, the new technique was mainly used as a quick screening method for detection of platelet function disorders (6). For more detailed studies, still the Born aggregometer was prefered. Although the results obtained by the two methods were similar, the parameters measured with the novel device were not clearly understood at that time. (7) The main disadvantage of the whole blood method is the problem of standardizing the condition of the specimen. Various studies have described the influence of the platelet count and hematocrit on rate and extent of in vitro aggregation in whole blood. (8,9) Platelet 8 ount has been reported to be without influence beyond 50x10 /l (8) and the aggregation of human platelets (WBEA) was enhanced at lower values of hematocrit. (9) Optimal responses occured after dilution of the samples to hematocrit of 30%. Physiologic saline solution (PSS) was found to be the diluent of choice in this connection. (8) The technique has been further investigated and improved during the first years following its invention. (10,11,12) Nevertheless unphysiological conditions of the specimen must be taken into account, which additionally might influence the obtained data. We wanted to investigate the influence of sample age on various characteristics of the aggregation curve after collagen induced platelet aggregation using the impedance method with whole blood. This aspect might be of interest, when numerous samples have to be investigated within limited time. The study was planned in two steps. Within a first step we represented the aggregometry plots mathematically and performed curve analysis. In the second step, the parameters obtained from curve analysis were investigated for their sensitivity for the aging of the samples.
Vol. 60, No. 6
THE INFLUENCE OF SAMPLE AGE...
479
MATERIAL AND METHODS Whole blood electrical aggregometry (WBEA): In this study an impedance aggregometer from CHRONO-LOG CORPORATION, 2 West Park Road, Havertown, PA, was used. Collagen (COLL) was used as aggregation trigger in titrated whole blood (CWB). Na-citrate was choosen, because heparin is reported to lead to a slight platelet activation and release of platelet factor 4 (PF4). On the other hand citrate anticoagulated blood contains less free calcium ions, although the remaining calcium should be sufficient to allow a normal platelet response. Blood was taken by puncture of the cubital vein from 10 healthy volunteers (5 men, 5 women) at the age of 23-33 years (mean=29.4), all non-smokers and without any medical treatment for at least 14 days. Blood collection was performed with the same type of cannula and by the same person directly into a vacutainer tube containing 10% of Na-citrate. After incubation of the specimen in the aggregometer, platelet aggregation was triggered by adding 1.5 ul of 0.1% type-l-collagen, suspended in isotonic glucose solution of pH 2.7 (= 1.5 ug of collagen), to 1000 ul of whole blood. Collagen reagents were from CHRONO-LOG CORPORATION (Chrono-Par No. 3801, 2 West Park Road, Havertown, PA, 190834691. The change in impedance over time was registered in a strip chart recorder (Chrono-Log) and the plotted aggregometer curves were then digitized and analysed on an IBM XT PC. Data Acquisition: Aggregation curves were digitized on a graphics tablet (Summagraphics) connected to an IBM-XT computer. Together with the digitized values for impedance, Y, the following supplementary parameters were stored for each curve: a 5 Ohm calibration signal / patient-ID / age of sample Calculation of Characteristic Quantities for Impedance Curves: By use of the 5 Ohm calibration signal, the original data were scaled and then analyzed along the following lines. (figure 1) The maximum (Y-MAX) and minimum (Y-MINI impedance as well as the time at which impedance was minimum (T-YMIN) were detected. The amplitude (AMPL) was obtained as AMPL=Ymax-Ymin. Next, three integrals were obtained: (1) The impedance was integrated over time between initiation of aggregation by adding the agonist (t=O) and the end of each measurement (t=tmax) using the trapezoidal rule. Note that this integral (ITOT) corresponds to the area between the impedance curve and the time-axis. (2) The value of ITOT was subtracted from the rectangular area defined by the duration of measurement and the maximum impedance found: IPART = Tmax * Ymax - ITOT. Note that the integral IPART represented the area above the impedance curve, with Y = Ymax as its upper border, and was thus in some respect "complementary" to ITOT. (3) Impedance was integrated from the time where the minimum impedance occured (Tymin) to the end of each measurement (tmax) to yield the quantity "IMINI".
480
THE INFLUENCE OF SAMPLE AGE...
Vol. 60, No. 6
.j. -i 1
0
1
2
3
4
6 5 TIME (minutes)
7
8
9
Tmax
10
11
FIGURE 1 Scheme of characteristic quantities of collagen-triggered WBEA-curves. Calibration signal indicated as vertical bar. In a final step the impedance curve was analyzed by an adaptive fitting procedure in order to locate the time (TINFL) and the slope (SINFL) of the inflection point. Since each impedance curve is characterized by a minimum followed by an asymptotic increase, polynomials could not be used for the fit. Instead, data were transformed according to mathematical functions, which intrinsically exhibit the correct shape: yb (t) = 1 / (1 + b * (t - Tymin) ** 2) For the parameter b we choose five different values (0, 0.01, 0.02, 0.1 and 0.5) in order to generate five independent transformations for the original impedance values. Transformed data were considered as five new variables and were subjected to a multiple least squares fit, linear in the fitting parameters. For each impedance curve, the best subset of transformed variables was determined in a stepwise procedure, using the goodness-of-fit as a criterion to include or exclude one of the transformed variables (F-test). The performance of the criterion for the selection of the best fit was found capable of adequately found various shapes curves of representing impedance experimentally. From the fitted curve, timewise location and impedance of the inflection point were calculated analytically. The calculations were coded in FORTRAN, and routines of the CERN-ScientificSubroutine-Package were used for the matrix in-version necessary to obtain the parameters of the curvilinear fits.
Vol. 60, No. 6
THE INFLUENCE OF SAMPLE AGE...
481
Statistical Evaluation: Each of the calculated parameters described above was obtained from each aggregation curve. Means and standard deviations were calculated for each sample age examined. Results were then plotted over sample age, see figures 2 through 7. Additionally, the dependence of each parameter on the age of the blood sample was analyzed statistically. For each calculated parameter a polynomial regression was performed, involving an intercept, a linear and a quadratic term. (Tab. 1+2) RESULTS Parameters calculated from aggreuation curves and their dependence on the age of the blood sample: Figures 2 trough 7 give calculated means and standard deviations for each age-ofsample, plotted over sample-age. The results (see also table 1) revealed that, for most of the parameters, a straight line was insufficient to represent the dependence on sample-age, cf. the significance of the p-values for the quadratic terms. Inclusion of the quadratic terms provides adequate representation of the original data except for 30-minutes values. This is caused by two outliers, which influence the mean. The plots show, that up to 60 minutes only slight changes of the parameters occur, after another 60 minutes the difference against t=O is 25-30%, and the 150-minutes value is 40-50% of the initial. Parameters increasing with time changed to the same extent.
FIGURE 2 Dependence of parameter AMPL (Max-Ymin) on sample age
482
THE INFLUENCE OF SAMPLE AGE...
e
Vol. 60, No. 6
100
50
AGE OF SNIPLE
(mini
FIGURE 3: Dependence of parameter ITOT (integral between initiation of aggregation 't=O' and end of measurement 'Tmax') on sample age.
-----_______ “e+T 150 138 ‘____------f _’
118.. 9070
AGE OF SAMPLE
bin)
FIGURE 4: Dependence of parameter IMINI (integral between Tymin and Tmax) on sample age.
THE INFLUENCE OF SAMPLE AGE...
Vol. 60, No. 6
--_--------_______________
2I
AGE OF SfflPLE Cmin)
FIGURE 5 Dependence
of the parameter SINFL (slope at the inflection point) on sample age
0
58
158
100 AGE
OF WIPLE
200
bin)
FIGURE 6 Dependence
of parameter TINFL (time of the inflection point) on sample age
484
THE INFLUENCE
OF SAMPLE
AGE OF SAMPLE
AGE...
Vol. 60, No. 6
hi n)
FIGURE 7: Dependence of parameter WIN (time of minimum impedance) on sample age. Statistical analysis of the dependence of each parameter on the age of the blood sample: For each calculated parameter a polynomial regression was performed, involving an intercept, a linear and a quadratic term. (Table 2) The linear terms provided only insufficient representation of the original data. The R-squares for the fitting of the quadratic terms with the parameters are given in Table 1. TABLE 1: R-Squares
for
the Fitting of quadratic with Curve-Parameters
Terms
R-square Parameter ____________________~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 0.789 AMPL 0.874 ITOT 0.666 IPART 0.393 TINFL 0.649 SINFL 0.790 IMINI 0.602 TYMIN ____________________~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Vol. 60, No. 6
THE INFLUENCE OF SAMPLE AGE...
485
TABLE 2: Dependence of calculated Parameters of WHEA-Curves on Sample-Age For each parameter, the first line gives the fitted value, the second line gives the estimated standard deviation (sd) and the p-value, indicating if the the associated line third corresponding term (intercept, linear component, quadratic component) is significantly different from zero. intercept linear Parameter quadratic ____________________~--~~---------~~~~~~~~~--~~~~~~--~~~~~~~~~~~ -0.0005 0.014 21,844 AMPL 0.0001 0.025 sd 1.448 0.0008 0.585 P 0.0001 -0.0049 0.151 113.982 ITOT 0.0010 0.193 sd 10.855 0.0001 0.438 P 0.0001 -0.0021 0.177 63.676 IPART 0.0006 0.108 sd 6.086 0.0014 0.112 P 0.0001 -0.00009 -0.010 1.779 TINFL 0.00003 0.006 sd 0.380 0.012 0.117 P 0.0001 -0.0002 0.015 8.229 SINFL 0.00009 0.017 sd 0.987 0.0067 0.398 P 0.0001 -0.0044 0.129 159.514 IMINI 0.0011 0.209 sd 11.710 0.0006 0.540 P 0.0001 -0.00003 -0.0006 0.681 TYMIN 0.00001 0.0021 sd 0.122 0.028 0.759 P 0.0001 ___________-________~~~~~~~~~~~~~~~~~~~~~~~-~~~~~~~~~~~~________ DISCUSSION We have studied the influence of sample age on various parameters of the aggregometry curves obtained with the electrical impedance aggregometer. The electrical method offers results obtained under physiological conditions, time renders consuming sample preparation unnecessary, and thus can easily be used as part of a routine screening procedure. Correct interpretation of the obtained data requires standardized conditions. testing Influences of hematocrit and platelet count have been described by several authors (7,8,9), but the importance of the age of sample has not been mentioned so far. Although one blood sample can be analysed within 15 minutes using WHEA, delays in processing may occur with larger numbers of specimen. Knowledge of the time dependence of the results is necessary to standardize this method. The use of micro-computers for aggregation instruments has been described by other authors (13,141. However, the influence of varying sample conditions have not been clearly defined so far.
486
THE INFLUENCE OF SAMPLE AGE...
Vol. 60, No. 6
Our results show, that blood samples for whole blood aggregometry undergo an aging process, which influences the obtained data obviously after 30 minutes. Specimen exceeding an age of one hour should be investigated as usual, and the data corrected according the calculated quadratical terms. If longer delays in investigating the samples are to be expected, the specimen should be stored at 6OC and rewarmed in the heater block of the aggregometer for about 15 minutes before processing. The calculated parameter ITOT was found most sensitive to aging. Only 20% of the value at t=O was observed after 150 minutes. All other parameters with decreasing tendency were 45% to 50% at that time, whereas T-YMIN and T-INFL had increased by 100%. At the moment we are not able to explain the aging process physiologically. Several mechanisms are possible: Platelets may be altered by acidosis of the specimen.
unphysiological blood gases and
Platelets and other blood cells may undergo slight release reactions and leakage during storage under physiological temperature. Slightly elevated plasma ADP levels might desensitize ADP receptors, resulting in reduced response to different agents. Other cells may play a role in increasing the base excess through products of metabolism. Proteins may attach to the siliconized test vials resulting in a change of the percentual plasma distribution and influencing the activity of important modulators of platelet aggregation. (151 Acidosis can change surface ions of plasma proteins, which might influence normal aggregation. Substances which regulate platelet function such as prostacyclin and cyclic AMP may degrade to biologically inactive products. To evaluate the role of different blood cells, plasma proteins and blood gases on the described aging process, further investigations are necessary. Additionally, the effect of sample age on WBEA with triggers other than collagen will be of interest. Furthermore the influence of hematocrit, platelet count and other variables modulating platelet function have to be clearly defined, to be able to standardize WBEA by means of software rather than dilution of the samples. REFERENCES 1.
The electronic aggregometer: A Cardinal DC, Flower RJ. novel device for assessing platelet behavior in blood. J Pharm Meth;3:135-158,1980
2.
Born GVR. Quantitative investigation into the aggregation of blood platelets. J Physiol London;16:67-68,1962
Vol. 60, No. 6
THE INFLUENCE OF SAMPLE AGE...
467
3.
Effects of the numbers and sizes of Born GVR, Hume M. platelet aggregates on the optical density of plasma. Nature 215:1027-1029,1967
4.
Flower RJ, Russell-Smith N, Salmon JA, Blackwell GJ, Thorogood PB, Vane JR. Prostacyclin is produced in whole blood. Br J Pharmaco1;64:436,1978
5.
Willems C! Stel HV, van Aken W, van Mourik JA. Binding and inactivation of prostacyclin (PGI2) by human erythrocytes. Br J Haemato1;54:43-52,1983
6.
A quick method for Ingerman-Wojenski C, Silver MJ. screening platelet dysfunctions using the whole blood lumiaggregometer. Thromb. Haemostas.;51:154-156,1984
7.
Evaluation of Ingerman-Wojenski C, Smith B, Silver MJ. comparison with optical electrical aggregometry: aggregometry, secretion of ATP, and accumulation of radiolabeled platelets. J Lab Clin Med;101:44-52,1983
a.
Platelet impedance Mackie IJ, Jones R, Machin SJ. blood its inhibition by aggregation in whole and antiplatelet drugs. J Clin Patho1;37:874-878,1984
9.
Galvez A, Badimon L, Badimon JJ, Fuster V. Electrical aggregometry in whole blood from human, pig and rabbit. Thromb. Haemostas.;56:128-132,1986
10.
Difficulty in Ingerman-Wojenski C, Smith B, Silver MJ. detecting inhibition of platelet aggregation by the impedance method. Thromb. Res.;28:427-432,1982
11.
Russell-Smith N, Flower RJ, Cardinal DC. Measuring platelet and leucocyte aggregation/adhesion responses in very small volumes of whole blood. J Pharm Meth;6:315-333,1981
12.
Ingerman-Wojenski C. Simultaneous measurement of platelet aggregation and the release reaction in platelet-rich plasma and in whole blood. J Med Technol;l:697-701,1984
13.
Huzoor-Akbar, Romstedt K, Manhire B. Computerized aggregation instruments: A highly efficient and versatile system for aquisition, quantitation, presentation and management of platelet aggregation data. Thromb. Res.;32:335-341,1983
14.
Francis JL. Improved assessment of impedance aggregometry using a microcomputer. Thromb. Res.;45:851-855,1987
15.
Orchard MA, Robinson C. Stability of prostacyclin in human plasma and whole blood: studies on the protective effect of albumin. Thromb. Haemostas.;46:645-647,198l
![The study of platelet aggregation in whole blood [proceedings].](/assets/img/hold.png)
