Measurement of platelet aggregation in ovine blood using a new impedance aggregometer


Reinhard Mischke, Klinik für Kleintiere, Tierärztliche Hochschule Hannover, Bischofsholer Damm 15, D-30173 Hannover, Germany


Background: Whole blood platelet aggregometry (impedance) is an important method to investigate platelet function disorders. Examination of hemostatic function in sheep is important with respect to their role as an animal model of human disease.

Objective: The aim of this study was to evaluate and optimize selected methodological aspects (anticoagulant, agonist concentration) of impedance aggregometry in ovine blood using the new Multiplate 5.0 analyzer.

Methods: Blood samples were collected in hirudin anticoagulant from 40 clinically healthy sheep. Samples from selected sheep were collected in citrate, with or without the addition of calcium chloride. The agonists adenosine diphosphate (ADP), collagen, ristocetin, arachidonic acid, and thrombin receptor-activating peptide (TRAP) were added in several concentrations to induce aggregation.

Results: Based on maximum aggregation values and internal precision, no significant difference was found between ADP concentrations of 3–10 μmol/L and collagen concentrations of 3–5 μg/mL (P>.05). The lowest interindividual variation of approximately 3–4-fold was seen with 4 and 5 μmol/L ADP and 4 and 5 μg/mL collagen. Ristocetin, arachidonic acid, and TRAP did not induce significant aggregation at any concentration. Aggregation results were significantly lower when measured in citrate- vs hirudin-anticoagulated blood, regardless of the presence of calcium chloride.

Conclusions: Our results indicate that the multiplate impedance aggregometer is suitable for the measurement of platelet aggregation in sheep using optimal agonist concentrations of 4–5 μmol/L ADP and 4–5 μg/mL collagen. Hirudin-anticoagulated blood is the preferred sample material.


Platelet aggregometry is one of the main in vitro standard techniques for the evaluation of platelet function, despite the fact that important aspects of platelet response such as adhesion or quality of platelet aggregates are not detected. There are 2 main principles of platelet aggregation measurements. The turbidimetric method described by Born1 measures the increase in light transmission in platelet-rich plasma (PRP) during the aggregation process induced by response to an agonist.2 Impedance aggregometry is based on the increase in electrical impedance across 2 metal wires that results when platelets aggregate in response to an agonist.2,3

The turbidimetric method has the advantage that the platelet count in PRP can be adjusted to a standard value.3 In contrast to the electrical method, it allows detection of shape change and reversible responses. The major limitations of the light transmittance platelet aggregation assays result from the necessity of preparing PRP, which requires appropriate equipment, extra time, large sample volume, and technologists experienced in preparation and cell counting techniques.4 In contrast, electrical impedance aggregometry does not require platelet separation and requires only a small sample volume and minimal preparation time.4–8 Impedance aggregation can be completed within 20–30 minutes after a blood sample is obtained and is an approved clinical method for evaluating platelet function.2,4,8 Centrifugation may result in the loss of large, hyper- or hypoactive platelets.2 Impedance aggregometry also may better reflect the in vivo situation, because it is measured in whole blood, allowing platelet interactions with other blood cells. The wires of the test cells deliver an artificial surface for platelet aggregation in the impedance method, whereas platelets aggregate with each other in the liquid phase of the turbidimetric method.2,3

Examination of hemostatic function in sheep is primarily important with respect to their role as an animal model of various human diseases and procedures, including surgical methods,9 surgical trauma,10–12 cardiovascular surgery,13–15 sepsis,16 and shock.17 However, only a few studies have been published regarding the methods used to investigate primary hemostasis in ovine blood,18–23 and only one of these utilizes whole blood aggregometry.23

The aim of this study was to investigate the use of a new impedance aggregometer for examining platelet aggregation in sheep, including optimization of agonist concentrations, evaluation of sample type, and determination of reference values for ovine blood.

Materials and Methods

Experimental animals and sample collection

Blood samples were collected from a total of 40 clinically healthy adult sheep of different breeds and sexes. The sheep were kept for experimental and educational purposes in stables in small groups at the Hannover Medical School or the Hannover School of Veterinary Medicine. The animals had platelet counts of 221,000–932,000/μL (median 627,000/μL), RBC counts of 8.44–15.83 × 106/μL (10.73 × 106/μL), and HCTs of 23.7–38.6% (31.5%) (ADVIA 120, Siemens Healthcare Diagnostics, Tarrytown, NY, USA). For the collection of blood, animals were manually restrained by an assistant. Samples were taken using sterile disposable needles (1.3 × 45 mm/18 G × 1.3/4 in.) from jugular veins raised slightly and briefly. Plastic tubes containing hirudin (Thrombin Inhibitor blood collection tube, Dynabyte Medical GmbH, Multiplate, Munich, Germany) were filled completely to the 4.5 mL mark. Graduated plastic tubes prefilled with 1 mL of 0.11 mol/L sodium citrate solution (PPS natrium citritum 3.13%, MediPac, Rheinbreitbach, Germany) were filled completely to the 10 mL mark (9 parts blood to 1 part sodium citrate). Immediately afterwards, the plastic tubes were rocked gently to thoroughly mix the blood and anticoagulant. Human hirudin blood was collected as a control for selected experiments from one of the authors (K.K.), with consent. Experiments were performed in accordance with the German Animal Welfare law; experimental design was approved by the official animal health care officer of the university and by the ethics committee of the responsible national agency (Lower Saxony State Office for Consumer Protection and Food Safety).

Optimization of agonist concentration

Different concentrations of adenosine diphosphate (ADP) (1, 1.5, 2, 2.5, 3, 4, 5, 10, and 20 μmol/L), collagen (0.5, 1, 1.5, 2, 2.5, 3, 4, 5, 10, and 20 μg/mL), ristocetin (0.2 and 1 mg/mL), arachidonic acid (0.25, 0.5, 1, and 2 mmol/L), and thrombin receptor-activating peptide (TRAP; 32 and 160 μmol/L) were added to hirudin-anticoagulated blood samples from 6 (ristocetin, arachidonic acid, and TRAP) or 10 (ADP and collagen) healthy sheep and aggregometry was performed. To verify measurements with negligible signal, control measurements were performed with freshly collected human blood. Optimal agonists and concentrations were based on maximum measurement signal (the lowest agonist concentration leading to maximum aggregation values), minimal interindividual variation, and highest precision based on the internal measurements in duplicate.

Calculation of reference values

Based on the results of the agonist experiments, aggregometry was performed using hirudin-anticoagulated blood samples from an additional 30 sheep with ADP (3, 4, 5, and 10 μmol/L) and collagen (3, 4, and 5 μg/mL). Reference values were calculated using the results from all 40 sheep.

Evaluation of sample type

From 10 of the healthy sheep, additional blood samples were collected at the same time into citrate-anticoagulated tubes for comparison with hirudin-anticoagulated samples. Aggregometry measurements were performed using final concentrations of 5 and 10 μmol/L ADP and 5 and 10 μg/mL collagen. Aggregometry of the citrate-anticoagulated samples was performed with and without the addition of calcium chloride.

Impedance aggregometry


Whole blood aggregation was determined using a new generation impedance aggregometer (Multiplate 5.0 analyzer, Dynabyte Medical GmbH), developed for human samples. The device has 5 channels for parallel testing and a single-use test cell with duplicate impedance sensors, each consisting of 2 straight electrode wires (blood contact area, 3 mm length, 0.3 mm diameter for each sensor wire).3,5,6 The impedance change determined by each sensor is recorded independently. During analysis, the sample reagent mixture is stirred using a disposable polytetrafluoroethylene-coated magnetic stirrer at 800 rpm.

Test procedure

For the standard procedure (hirudin-anticoagulated whole blood), 300 μL of isotonic NaCl preheated to 37°C was pipetted into the test cells and 300 μL of sample was added. Citrate-anticoagulated blood was measured in the same manner (without recalcification) using 300 μL NaCl solution. For experiments on the addition of calcium, 300 μL of an isotonic NaCl solution containing 3 mmol/L CaCl2 (Dynabyte Medical GmbH) was added to the sample. After 3 minutes of incubation and stirring at 37°C, the measurement was started by adding 20 μL of the appropriate agonist solution.

Agonist preparation

Stock solutions of agonists (Dynabyte Medical GmbH) were prepared by dissolving the lyophilized reagents in double distilled water. ADPtest (0.2 μmoL ADP/vial) was dissolved in 322 μL to achieve a concentration of 620 μmol ADP/L (for a final assay concentration of 20 μmol/L); collagentest (100 μg collagen per vial) was dissolved in 161 μL (620 μg/mL; final concentration 20 μg/mL); ristocetintest (10 mg ristocetin per vial) was dissolved in 322 μL (31 mg/mL, final concentration 1 mg/mL); ASPItest (15 μM arachidonic acid per vial) was dissolved in 242 μL (62 mmol/L, final concentration 2 mmol/L); and TRAPtest (1 μmoL TRAP per vial) was dissolved in 202 μL (4960 μmol/L, final concentration 160 μmol/L). The TRAPtest reagent contains a shortened form of the human TRAP-peptide (TRAP-6: Ser–Phe–Leu–Leu–Arg–Asn). Agonist concentrations in the stock solution corresponded to the highest concentrations used in the study. Further dilutions were performed using double distilled water to achieve different agonist concentrations, as mentioned above. According to recommendations of the manufacturer, stock solutions of the agonists were stored in the refrigerator for a maximum of 7 days with the exception of arachidonic acid, which was stored frozen at −28°C.

Measurement parameters

The Multiplate analyzer continuously detects a change in impedance caused by the adhesion and aggregation of platelets on the electrode wires. Deviating from the standard test time for human blood of 6 minutes, a recording time of 12 minutes was used in the pilot study to allow the curves to nearly reach a plateau. The results registered by the 2 sensors provide 2 aggregation curves (Figure 1). If the 2 curves differ >20% or the correlation coefficient (r) is <0.98, the aggregometer signals a warning and indicates the need to repeat the measurement. Thus, the duplicate sensors serve as an internal control. The mean values of the 2 determinations are expressed in arbitrary “aggregation units” (AU) indicating the increase in impedance during the measurement. The results given automatically by the machine are the area under the aggregation curve (AUC) to express the aggregation response over the measured time (AU min), the maximum aggregation (expressed in AU), the velocity (AU/min), the percentage difference between the 2 curves, and the correlation coefficient of the 2 curves. When repeated measurements were made because of a warning signal, the percentage difference between the 2 curves was based on the manually calculated mean of all (usually 2) measurements both before and after the warning signal. The results of all other parameters refer exclusively to measurements obtained without warning signals, ie, only results accepted by the analyzer were used for calculation of AUC, maximum aggregation, and velocity.

Figure 1.

 Schematic diagram of the 2 aggregation curves from a single-use test cell of the Multiplate 5.0 impedance aggregometer.

Statistical analysis

Data were tested for normal distribution using the Kolmogorov–Smirnov test. Because data for all parameters had normal distribution, results from different experimental groups (agonist concentrations, anticoagulants) were compared using 1-way ANOVA with repeated measures. The percentage difference between curves was compared using 1-way ANOVA. Where indicated, paired t-tests were used to compare the results of individual agonist concentrations or anticoagulants. P-values <.05 were considered significant. Outliers were defined as values less than the lower quartile minus 1.5 × the interquartile range, or greater than the upper quartile plus 1.5 × the interquartile range; extreme (low) outliers were defined as values less than the lower quartile minus 3 × the interquartile range. Reference values were calculated as the 5% and 95% quantiles.


The range of concentrations tested for ristocetin (0.2 and 1 mg/mL), arachidonic acid (0.25, 0.5, and 1.0 μmol/L), and TRAP (32 and 160 μmol/L) did not induce measurable platelet aggregation in healthy sheep. Median AUCs were ≤56.5 AU min, whereas control measurements with human blood revealed adequate signals.

AUC values obtained with different concentrations of ADP (P=.004) and collagen (P<.001) in 10 healthy sheep were significantly different (repeated measures ANOVA) (Figures 2 and 3). AUC values obtained with ADP concentrations from 2 to 10 μmol/L did not differ significantly (P>.05, t-test), except that lower values were found with 2 than with 2.5 μmol/L (P=.028, t-test). Compared with platelet aggregation performed using 2–10 μmol/L ADP, AUC values were distinctly lower with ADP concentrations ≤1.5 μg/mL (P<.05, t-test), except between concentrations of 1.5 and 2 μmol/L ADP. Interestingly, 20 μmol/L ADP also induced lower AUC values than concentrations of 4, 5, and 10 μmol/L ADP (P<.05, t-test). Interindividual variation was higher with 2.0 and 2.5 μmol/L ADP compared with ≥3.0 μmol/L ADP. AUC values obtained with 3 to 20 μg/mL collagen were not significantly different (P>.05, t-test). In comparison, AUC values were significantly lower with collagen concentrations of ≤2.5 μg/mL (P<.05, t-test), with a few exceptions (2.0 and 2.5 μg/mL vs 3.0 μg/mL, 2.0 and 2.5 μg/mL vs 20 μg/mL collagen).

Figure 2.

 Areas under the curve (AUC) for impedance aggregometry measurements on blood samples from 10 healthy sheep using different concentrations of adenosine diphosphate (ADP). Boxes indicate 25% and 75% quantiles, horizontal lines indicate median values, vertical bars indicate minimum and maximum, and individual circles are outliers.

Figure 3.

 Areas under the curve (AUC) for impedance aggregometry measurements on blood samples from 10 healthy sheep using different concentrations of collagen. Boxes indicate 25% and 75% quantiles, horizontal lines indicate median values, vertical bars indicate minimum and maximum, and individual circles are outliers.

Reference values were calculated for selected concentrations of ADP and collagen in 40 sheep (Table 1). Significant differences were not observed between 3, 4, 5, and 10 μmol/L ADP or 3, 4, and 5 μg/mL collagen for the results of any of the parameters (P>.05, ANOVA). Marked interindividual variation was observed for all parameters. The lowest variation was seen at 4 and 5 μmol/L ADP and 4 and 5 μg/mL collagen, where, for example, AUC varied by 3- or 4-fold for ADP and collagen, respectively. In addition, samples containing 4 or 5 μg/mL collagen required fewer repeated measurements (0–1) than those with 3 μg/mL collagen (8). More repeated measurements were required with ADP (7–10 of 40) and differences were less marked between different ADP concentrations. In 4 samples (2 with 3 μg/mL collagen, 1 each with 4 and 5 μmol/L ADP), more than 1 repeated measurement was required to achieve a result accepted by the machine. In general, ADP produced higher AUC than collagen.

Table 1.   Reference values for parameters of impedance aggregometry in 40 healthy sheep using different concentrations of ADP and collagen.
 AUC (AU min)Maximum Aggregation (AU)Velocity (AU/min)Difference (%)
MedianIQR5–95% QuantileMedianIQR5–95% QuantileMedianIQR5–95% QuantileMedianMinimum–Maximum
  • *

    Significant differences were not observed between agonist concentrations (ANOVA).

  • †Percentage difference between AUCs of internal duplicate measurements.

  • AU, aggregation units; AUC, area under the curve; IQR, interquantile range (the difference between the 95% and 5% quantile); ADP, adenosine diphosphate.

ADP (μmol/L)*           
Collagen (μg/mL)*

Compared with hirudin-anticoagulated blood, samples anticoagulated with citrate had significantly lower AUCs when 5 and 10 μmol/L ADP and 5 and 10 μg/mL collagen were used, independent of the addition of calcium (P<.05, t-test) (Figure 4). The addition of calcium to citrate-anticoagulated blood samples led to higher AUC values when collagen was used as agonist (P<.05, t-test). There was no effect of added calcium on ADP-induced platelet aggregation in citrate-anticoagulated blood (P>.05, t-test).

Figure 4.

 Comparison of areas under the curve (AUC) for impedance aggregometry measurements on blood samples from 10 healthy sheep anticoagulated with hirudin or citrate (with or without the addition of calcium) using (A) 5 μmol/L adenosine diphosphate (ADP), (B) 10 μmol/L ADP, (C) 5 μg/mL collagen, and (D) 10 μg/mL collagen. Boxes indicate 25% and 75% quantiles, horizontal lines indicate median values, vertical bars indicate minimum and maximum, individual circles are outliers, and triangles are extreme outliers.


The results of this study indicate that ristocetin, arachidonic acid, and human TRAP-6, which induce platelet aggregation in human samples, are unsuitable agonists for induction of whole blood impedance aggregation in ovine platelets. Median measurement signals with the tested concentrations of ristocetin and arachidonic acid were <5% of the signal achieved with ADP or collagen. This was shown for agonist concentrations corresponding to those recommended for humans (ristocetin 0.2 mg/mL, arachidonic acid 0.5 mmol/L, TRAP 32 μmol/L) as well as for concentrations several times higher, up to 1 mg/mL ristocetin and 2 mmol/L arachidonic acid.

The lack of effect of TRAP, synthetic peptides that mimic many of the effects of thrombin, probably reflects species-specific differences in the thrombin receptor. Variability in cellular responsiveness to thrombin receptor-derived peptides has been reported in various species including rabbit, dog, pig, hamster, and rat.24 The unsuitability of ristocetin and arachidonic acid for the induction of platelet aggregation in ovine whole blood in the present study was partly in contrast to the results of previous studies on ovine blood using the turbidimetric (Born) method.20–22 In those studies, ristocetin was added in concentrations of 0.1, 0.2, 0.5, and 1 mg/mL, inducing aggregation of approximately <5%, 10%, 50%, and 93% of platelets.22 Arachidonic acid at concentrations of 0.5–7.5 mmol/L induced maximum aggregation of 25–100% of platelets.20–22 The agonist concentrations investigated in the present study were equal to or within the range of concentrations tested in those studies. However, the agonist concentrations in our study refer to the final concentration in whole blood; the plasma concentrations, which better reflect the situation in the cited studies, were approximately 2-fold higher. The reason for the discrepancy between our results and the literature remains unclear, but may be due to quality differences of the activating reagents provided by different manufacturers.

In the pilot experiment, the concentrations of ADP and collagen used were those recommended for human blood samples (ADP, 6.4 μmol/L; collagen, 3.2 μg/mL). Only those agonist concentrations that induced significant aggregation in all healthy animals were used for further experiments. It can be assumed that the sensitivity of aggregometry to detect platelet function disorders is near the threshold concentration, ie, the lowest agonist concentration leading to maximum aggregation values.25 Therefore, agonist concentrations (ADP, 20 μmol/L; collagen, 10 and 20 μg/mL) much higher than the suspected threshold level were also excluded from further measurements. Surprisingly, ADP added in the high concentration of 20 μmol/L induced lower AUC values than other ADP concentrations (4, 5, and 10 μmol/L). This may be the result of the formation of platelet aggregates already in the liquid phase, before reaching the electrode wires, and may explain why a similar effect was not observed with the light transmission technique.19,21

Based on the criteria of maximum measurement signal, minimum interindividual variation, highest precision, and reliability of measurement (ie, low number of required repetitions), 4–5 μmol/L ADP and 4–5 μg/mL collagen seem to be optimal for the investigation of platelet aggregation in ovine blood with the new device. Optimal ADP concentration for measurement of ovine platelets was slightly lower and the collagen concentration was slightly higher than the manufacturer's recommended agonist concentration for human platelets. Only 1 other study dealing with ovine whole blood platelet aggregation was found.23 In that study, which was performed with another instrument, ovine blood was compared with bovine, canine, and human blood at only 1 concentration of collagen (1 μg/mL) and ADP (10 μmol/L). These concentrations, which differed from the optimal agonist concentrations in the present study, resulted in electrical signals of nearly the same height (ADP, 23±10 Ω; collagen, 20±7.7 Ω). Aggregation studies on ovine blood with the light transmission (turbidimetric) technique used a wide range of ADP and collagen concentrations.20–22 The similar reactivity of ovine and human platelets to ADP and collagen may indicate that sheep models are suitable to study aspects of primary hemostasis in humans.

The significant 3–4-fold interindividual variation demonstrated in the present study was partly in accordance with a study on samples from healthy human blood donors in which measurements with ADP (5 μmol/L) varied by 4-fold and those with collagen (2.5 μg/mL) varied by <2-fold.3 This significant variation may indicate a low sensitivity for detecting reduced platelet function in individual patients, however, this conclusion requires further clinical verification. The present study included different breeds of sheep which may be an additional source of variation. Further studies with higher animal numbers are necessary to define the influence of breed on platelet aggregation results. However, it is debatable whether breed- and age-specific reference values are practicable in veterinary medicine, particularly as such ranges should preferably be laboratory specific.26

The results of this study demonstrate that aggregation is stronger in blood samples anticoagulated with hirudin compared with citrate. This finding is well in accordance with a human study, in which significantly lower AUC values were found in ADP- and collagen-induced whole blood aggregation of human platelets in citrated blood.3 The fact that the addition of calcium corrected only a small part of the discrepancy between anticoagulants, and only when collagen was used, may indicate that factors other than calcium binding to citrate may also play a role. However, it was not possible to restore calcium concentration to physiologic levels to prevent clotting during the test.

In conclusion, apart from the fact that approximately 20% of measurements with ADP have to be repeated, the method of whole blood impedance aggregometry investigated in this study seems appropriate for the examination of ovine platelets using the agonists ADP and collagen at optimal concentrations of 4–5 μmol/L and 4–5 μg/mL, respectively; however, sensitivity for detecting reduced platelet function must be verified in further clinical studies. Hirudin-anticoagulated blood is the preferred sample material.


This work was supported by a grant of the German Research Foundation/Deutsche Forschungsgemeinschaft (DFG).

Disclosure: The authors have indicated they have no affiliations or financial involvement with any organization or entity with a financial interest in, or in financial competition with, the subject matter or materials discussed in this paper.