Sodium citrate is the anticoagulant recommended for collection of blood samples from humans and other animals for coagulation tests such as PT, aPTT, and AT and for measurement of concentrations of fibrinogen, D-dimer, and specific coagulation factors. In those situations, a 1:9 ratio of sodium citrate (3.2 g/dL) solution to blood is typically used because the chelating effect of sodium citrate is easily reversed by the addition of ionized calcium.1 This means that an additional and specific tube containing sodium citrate must be used for sample collection when assays of secondary hemostasis are required. These tubes are not the same as the EDTA tubes used for hematologic analysis or the plain or heparinized tubes used for biochemical analyses, thereby increasing the costs and difficulties associated with adequate sample collection.
In humans, blood samples collected in EDTA tubes are reportedly adequate for determination of PT.2,3 Benefits of using an EDTA tube for both coagulation and hematologic assays in routine veterinary practice include low volumes of blood required for testing, saving of time and resources, and reduction of sources of error associated with sample collection. However, to the authors' knowledge, there are no published reports in the veterinary literature regarding the use of EDTA, which is also a chelating agent, for coagulation assays. The purpose of the study reported here was to evaluate the use of potassium-EDTA tubes for collection of blood samples for assays of secondary hemostasis. We also sought to determine intra-assay and interassay precision and correlation coefficients for all hemostatic tests on plasma from potassium-EDTA tubes and sodium citrate tubes and to measure the effect of storage conditions on assay results for the 2 types of plasma samples.
Materials and Methods
Animals—One hundred eight client-owned dogs were enrolled in the study, including 59 males and 49 females. Ages ranged from 4 months to 16 years (mean ± SD, 7 years and 10 months ± 4 years and 3 months). Various breeds such as German Shepherd Dog, Boxer, Poodle, and Golden Retriever as well as mixed-breed dogs were represented. Nineteen of the 108 dogs were healthy dogs from which blood had been collected during a routine evaluation and for which results of hemostatic testing were within the reference ranges established at the San Marco laboratory. The remainder (n = 89) were dogs with at least 1 result of hemostatic testing that was outside the reference range; some had clinical signs related to secondary hemostatic abnormalities, such as hemorrhages or hematomas. The study protocol complied with guidelines regarding research on animals in Italy (Legislative Decree 116/92). Consent was obtained from all dog owners.
Sample collection and processing—Whole blood was collected from all dogs via cephalic venipuncture by use of an evacuated tube collection system.a Two milliliters of blood was subsequently added to potassium-EDTA tubesb containing 0.04 mL of 7.5% EDTA solution; then 3.5 mL of blood was added to citrate tubesc containing 0.35 mL of buffered 0.109M (3.2%) sodium citrate solution. The EDTA and citrate tubes were immediately centrifuged at 3,000 × g for 10 minutes, and plasma was immediately separated and stored in plastic tubesd at room temperature (20° to 24°C). Hemostatic tests of plasma harvested from the EDTA tubes (EDTA-treated plasma) and sodium citrate tubes (citrated plasma) were performed within 30 minutes after blood sample collection. Hemolyzed or lipemic blood samples were discarded. When multiple blood samples collected at various time points were available for some dogs with hemostatic disorders, these samples were included in the study, yielding a total of 130 blood samples (111 samples from 89 dogs with hemostatic diseases; 19 samples from 19 healthy dogs).
Prothrombin time, aPTT, AT, and fibrinogen concentration were measured by use of an automated coagulation analyzere with commercially available reagents.f Concentration of D-dimer was measured by use of an immunoturbidimetric method.4 Calibration was performed with aliquots of pooled citrated plasma samples that had been obtained from 3 healthy dogs and stored at −80°C.
Precision assays—Within the same run of 5 replicates, intra-assay precision was determined for all hemostatic tests via analysis of an EDTA-treated plasma sample and a citrated plasma sample obtained from 1 healthy dog and 1 dog with DIC. To evaluate interassay precision, aliquots of plasma from these 2 dogs were stored at −80°C, thawed at 37°C, and tested on 5 different days.
Storage assays—Aliquots of EDTA-treated and citrated plasma obtained from 5 healthy dogs were analyzed immediately after collection, at 12 and 24 hours of storage between 20° and 24°C (room temperature) and at 4°C (refrigerator), and at 24 hours of storage at −80°C. To evaluate the effects of short-term storage on EDTA-treated plasma samples, plasma samples from 3 healthy dogs were stored at 4°C and between 20° and 24°C, and hemostatic measurements were made immediately after collection and at 1, 2, and 3 hours of storage. In addition, EDTA–whole blood aliquots from 5 healthy dogs were stored between 20° and 24°C and centrifuged 1 hour after collection.
Statistical analysis—Statistical analyses were performed by use of standard computer software.g Intra- and interassay precision was determined via calculation of CV values.5 Ordinary regression analysis was used to obtain linear regression equations and correlation coefficients for relationships between results of each hemostatic test for citrated versus EDTA-treated plasma samples. In addition, Bland-Altman plots were used to evaluate the clinical agreement between results for the 2 types of plasma samples and reveal any bias.6 Bias was computed as the mean difference between scores for the 2 methods. Values for the 95% LOA were computed as mean difference ± 1.96 SD of the difference. Additionally, correlation between the mean of the methods and their differences was evaluated. Results are expressed as mean ± SD and CV.
A paired Student t test was used to compare results of hemostatic tests for EDTA-treated versus citrated plasma samples. For each anticoagulant and storage condition, results obtained at various times during storage of plasma samples were compared with those of plasma samples analyzed immediately after collection. Values of P < 0.05 were considered significant.
Results
Mean values of hemostatic tests of 130 plasma samples that were obtained from whole blood of 108 dogs (19 healthy and 89 with secondary hemostatic disorders) and transferred to potassium-EDTA and sodium citrate tubes were compared (Table 1). Mean values of PT, AT, and D-dimer were significantly greater for all EDTA-treated plasma samples combined, compared with values for all citrated plasma samples combined (increase of 21.7%, 14.3%, and 9.75%, respectively). On the other hand, the mean concentration of fibrinogen was significantly lower for all EDTA-treated plasma samples combined, compared with the mean concentration for all citrated plasma samples combined (decrease of 9%). Mean values of aPTT were not significantly different between the 2 types of plasma samples. Similar results were obtained when data for EDTA-treated and citrated plasma samples from healthy dogs and dogs with hemostatic disorders were analyzed separately, although concentrations of D-dimer and fibrinogen were not significantly different between EDTA-treated and citrated plasma samples from healthy dogs.
Mean ± SD results of hemostatic tests of plasma harvested from canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA and sodium citrate tubes (n = 130 samples [19 from 19 healthy dogs and 111 from 89 dogs with hemostatic abnormalities]).
| Hemostatic test | All samples | Samples from healthy dogs | Samples from dogs with hemostatic abnormalities | |||
|---|---|---|---|---|---|---|
| Citrated plasma | EDTA-treated plasma | Citrated plasma | EDTA-treated plasma | Citrated plasma | EDTA-treated plasma | |
| aPTT(s) | 14.77 ± 16.55 | 14.71 ± 16.59 | 11.51 ± 0.50 | 11.45 ± 1.24 | 15.32 ± 17.86 | 15.09 ± 17.51 |
| PT(s) | 9.58 ± 14.16 | 11.66 ± 13.89‡ | 7.37 ± 0.32 | 9.50 ± 0.68‡ | 9.96 ± 15.30 | 12.03 ± 15.01‡ |
| AT(%) | 108.9 ± 28.9 | 124.5 ± 33.4‡ | 122.7 ± 9.62 | 139.7 ± 11.4‡ | 106.5 ± 30.4 | 121.8 ± 35.2‡ |
| D-dimer (μg/ml_) | 0.41 ± 1.2 | 0.45 ± 1.19† | 0.06 ± 0.07 | 0.08 ± 0.11 | 0.47 ± 1.29 | 0.52 ± 1.27* |
| Fibrinogen (mg/dL) | 377.18 ± 210.8 | 342.2 ± 169.7‡ | 196.1 ± 29.15 | 184.84 ± 37.6 | 408.1 ± 213.1 | 369.1 ± 168.9‡ |
Value for EDTA-treated plasma samples differs significantly (P < 0.05
Value for EDTA-treated plasma samples differs significantly (P < 0.01
Value for EDTA-treated plasma samples differs significantly (P < 0.001
Mean differences between the 2 types of plasma samples with respect to the various hemostatic tests were as follows: aPTT, −0.05 seconds (95% LOA, −1.9 to 1.79 seconds); PT, 2.07 seconds (95% LOA, 0.23 to 3.9 seconds); D-dimer concentration, 0.03 μg/mL (95% LOA, −0.33 to 0.41 μg/mL); AT, 15% (95% LOA, 1.68% to 29.4%); and fibrinogen concentration, −34 mg/dL (95% LOA, −163.7 to 93.8 mg/dL). Bland-Altman plots revealed that bias increased with higher values of D-dimer, AT, and fibrinogen (Figures 1–5). Evaluation of correlations between means of analytic values for the 2 types of plasma samples and the magnitude of bias for the respective tests revealed a significant positive correlation for D-dimer (r = 0.52; P < 0.01), AT (r = 0.70; P < 0.001), and fibrinogen (r = 0.77; P < 0.001). Linear regression analyses revealed that results of the various hemostatic tests for EDTA-treated plasma samples (y) were associated with results of respective tests for citrated plasma samples (x) as follows: aPTT, y = (1.001 × x) − 0.065 (r = 0.99); PT, y = (0.979 × x) + 2.27 (r = 0.99); D-dimer, y = (0.978 × x) + 0.049 (r = 0.98); fibrinogen, y = (0.775 × x) + 49.83 (r = 0.96); and AT, y = (1.137 × x) + 0.589 (r = 0.98).
Bland-Altman plot illustrating the 95% LOA for aPTT values determined by use of an automated coagulation analyzer for plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. Of the 130 blood samples, 111 were obtained from 89 dogs with hemostatic diseases and 19 were obtained from 19 healthy dogs. Top and bottom horizontal dotted lines represent 95% LOA values (calculated as mean difference ± 1.96 × SD of the difference); middle horizontal dotted line represents mean difference (bias).
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for aPTT values determined by use of an automated coagulation analyzer for plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. Of the 130 blood samples, 111 were obtained from 89 dogs with hemostatic diseases and 19 were obtained from 19 healthy dogs. Top and bottom horizontal dotted lines represent 95% LOA values (calculated as mean difference ± 1.96 × SD of the difference); middle horizontal dotted line represents mean difference (bias).
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for aPTT values determined by use of an automated coagulation analyzer for plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. Of the 130 blood samples, 111 were obtained from 89 dogs with hemostatic diseases and 19 were obtained from 19 healthy dogs. Top and bottom horizontal dotted lines represent 95% LOA values (calculated as mean difference ± 1.96 × SD of the difference); middle horizontal dotted line represents mean difference (bias).
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for PT values determined by use of an automated coagulation analyzer for plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for PT values determined by use of an automated coagulation analyzer for plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for PT values determined by use of an automated coagulation analyzer for plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for AT values determined by use of an automated coagulation analyzer for plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for AT values determined by use of an automated coagulation analyzer for plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for AT values determined by use of an automated coagulation analyzer for plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for D-dimer concentrations determined by use of an immunoturbidimetric method in plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for D-dimer concentrations determined by use of an immunoturbidimetric method in plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for D-dimer concentrations determined by use of an immunoturbidimetric method in plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for fibrinogen concentrations determined by use of an automated coagulation analyzer for plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for fibrinogen concentrations determined by use of an automated coagulation analyzer for plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Bland-Altman plot illustrating the 95% LOA for fibrinogen concentrations determined by use of an automated coagulation analyzer for plasma harvested from 130 canine blood samples that were collected via cephalic venipuncture and transferred to potassium-EDTA tubes and sodium citrate tubes. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 69, 9; 10.2460/ajvr.69.9.1141
Intra-assay CV values for plasma samples obtained from whole blood in potassium-EDTA and sodium citrate tubes were < 10% in all cases, with the exception of the concentration of D-dimer in EDTA-treated and citrated plasma samples of the healthy dog and in the EDTA-treated plasma sample of the dog with DIC (Table 2). With 2 exceptions (AT in the healthy dog and concentration of D-dimer in the dog with DIC), interassay imprecision for all hemostatic tests was higher for EDTA-treated plasma samples, compared with corresponding values for citrated plasma samples, with imprecision values > 10% for aPTT, PT, D-dimer, and fibrinogen for EDTA-treated plasma samples (Table 3).
Values of intra-assay precision* for hemostatic tests of plasma harvested from blood samples that were obtained from a healthy dog and a dog with DIC and transferred to sodium citrate and potassium-EDTA tubes.
| Hemostatic test | Healthy dog | Dog with DIC | ||||||
|---|---|---|---|---|---|---|---|---|
| Citrated plasma | EDTA-treated plasma | Citrated plasma | EDTA-treated plasma | |||||
| Mean ± SD | CV(%) | Mean ± SD | CV(%) | Mean ± SD | CV(%) | Mean ± SD | CV(%) | |
| aPTT(s) | 11.9 ± 0.12 | 1.02 | 12.5 ± 0.40 | 3.24 | 19.9 ± 0.34 | 1.74 | 18.7 ± 0.43 | 2.31 |
| PT(s) | 7.8 ± 0.08 | 1.06 | 9.8 ± 0.65 | 6.68 | 13.1 ± 0.07 | 0.53 | 17.5 ± 0.39 | 2.22 |
| AT(%) | 108.8 ± 3.40 | 3.14 | 127.0 ± 2.12 | 1.67 | 67.2 ± 2.38 | 3.55 | 77.0 ± 2.64 | 3.43 |
| D-dimer (μg/ml_) | 0.07 ± 0.05 | 82.40 | 0.04 ± 0.03 | 84.70 | 0.34 ± 0.02 | 6.86 | 0.36 ± 0.04 | 11.09 |
| Fibrinogen (mg/dL) | 162.6 ± 5.1 | 3.15 | 143.0 ± 2.8 | 1.97 | 169.2 ± 4.9 | 2.90 | 160.4 ± 8.1 | 5.03 |
Five replicates were performed within the same run for each type of assay.
Values for interassay precision* for hemostatic tests of plasma harvested from blood samples that were obtained from a healthy dog and a dog with DIC and transferred to sodium citrate and potassium-EDTA tubes.
| Hemostatic test | Healthy dog | Dog with DIC | ||||||
|---|---|---|---|---|---|---|---|---|
| Citrated plasma | EDTA-treated plasma | Citrated plasma | EDTA-treated plasma | |||||
| Mean ± SD | CV(%) | Mean ± SD | CV(%) | Mean ± SD | CV(%) | Mean ± SD | CV(%) | |
| aPTT(s) | 12.0 ± 0.38 | 3.18 | 16.3 ± 2.22 | 13.61 | 21.0 ± 1.52 | 7.24 | 23.8 ± 3.27 | 13.76 |
| PT(s) | 7.8 ± 0.08 | 1.05 | 11.9 ± 1.21 | 10.14 | 13.1 ± 0.33 | 2.53 | 20.8 ± 3.38 | 16.18 |
| AT(%) | 110.7 ± 4.60 | 4.22 | 129.4 ± 5.22 | 4.03 | 61.6 ± 3.85 | 6.25 | 70.4 ± 6.14 | 8.73 |
| D-dimer (μg/ml_) | 0.06 ± 0.01 | 23.89 | 0.07 ± 0.05 | 64.90 | 0.32 ± 0.04 | 14.64 | 0.34 ± 0.04 | 11.75 |
| Fibrinogen (mg/dL) | 158.9 ± 13.7 | 8.62 | 148.6 ± 28.1 | 18.90 | 159.8 ± 9.7 | 6.06 | 136.2 ± 20.0 | 14.70 |
Aliquots of plasma stored at −80°C were thawed at37°C and assayed on 5 different days.
Storage at various conditions did not significantly affect results of hemostatic tests for citrated plasma samples, with the exception of a mild increase in AT detected after storage for 24 hours between 20° and 24°C or at 4°C (Table 4). On the other hand, results of hemostatic tests for EDTA-treated samples were significantly greater for aPTT and PT after storage for 12 hours between 20° and 24°C or at 4°C or after storage for 24 hours at −80°C. An increase in AT was also detected after storage of EDTA-treated plasma samples for 24 hours between 20° and 24°C. A significant decrease in fibrinogen concentration was detected after storage of EDTA-treated plasma samples for 24 hours between 20° and 24°C. In short-term storage conditions, aPTT and concentration of D-dimer were significantly greater for EDTA-treated plasma stored between 20° and 24°C for 3 and 2 hours, respectively, compared with values for EDTA-treated plasma samples that were assayed immediately (Table 5). Values for AT in EDTA-treated plasma samples stored for 3 hours between 20° and 24°C were significantly less than those for samples tested immediately. Storage of EDTA-treated plasma or EDTA-treated whole blood between 20° and 24°C for 1 hour had no significant effect on results of hemostatic tests (Table 6).
Mean ± SD effects of various temperatures and durations of storage on the results of hemostatic tests performed on aliquots of EDTA-treated and citrated plasma harvested from blood samples of 5 healthy dogs.
| Hemostatic test | Plasma treatment | 0 hours | 20 to 24°C | 4°C | − 80°C | ||
|---|---|---|---|---|---|---|---|
| 12 hours | 24 hours | 12 hours | 24 hours | 24 hours | |||
| aPTT(s) | Citrate | 11.24 ± 0.70 | 11.54 ± 0.80 | 11.42 ± 0.50 | 12.20 ± 13.00 | 11.40 ± 0.39 | 11.42 ± 0.60 |
| EDTA | 12.04 ± 1.23 | 25.40 ± 2.68‡ | 30.86 ± 1.60‡ | 23.70 ± 7.38* | 20.40 ± 4.17† | 15.26 ± 1.85† | |
| PT(s) | Citrate | 7.68 ± 0.52 | 7.62 ± 0.41 | 7.60 ± 0.52 | 7.56 ± 0.41 | 7.80 ± 0.38 | 7.66 ± 0.56 |
| EDTA | 9.80 ± 0.47 | 12.50 ± 0.74‡ | 14.66 ± 1.10† | 10.80 ± 0.63‡ | 11.20 ± 0.57† | 11.0 ± 0.94† | |
| AT(%) | Citrate | 126.2 ± 8.6 | 125.6 ± 8.8 | 132.2 ± 10.4* | 125.2 ± 8.7 | 131.8 ± 8.7† | 131.2 ± 10.6 |
| EDTA | 145.2 ± 8.2 | 146.2 ± 9.1 | 152.6 ± 9.8† | 146.6 ± 10.6 | 152 ± 10.2 | 153.2 ± 11.8 | |
| D-dimer (μg/mL) | Citrate | 0.11 ± 0.08 | 0.15 ± 0.30 | 0.12 ± 0.20 | 0.16 ± 0.25 | 0.11 ± 0.21 | 0.14 ± 0.20 |
| EDTA | 0.19 ± 0.39 | 0.22 ± 0.30 | 0.19 ± 0.30 | 0.22 ± 0.34 | 0.15 ± 0.28 | 0.16 ± 0.27 | |
| Fibrinogen (mg/dL) | Citrate | 190.8 ± 60.0 | 199.8 ± 68.3 | 189.6 ± 67.0 | 199.2 ± 65.1 | 196.8 ± 71.4 | 193.8 ± 70.3 |
| EDTA | 182.8 ± 71.2 | 171.4 ± 62.4 | 146.8 ± 52.8* | 186.0 ± 67.8 | 185.0 ± 67.4 | 177.6 ± 68.4 | |
See Table 1 for key.
Mean ± SD effects of storage at 4°C or between 20° and 24°C and assaying immediately (0 hours) or after storage for 1, 2, or 3 hours on hemostatic test results for EDTA-treated plasma samples obtained from 3 healthy dogs.*
| Hemostatic test | Temperature (°C) | Time after sample collection† | |||
|---|---|---|---|---|---|
| 0 hours | 1 hour | 2 hours | 3 hours | ||
| aPTT(s) | 20–24 | 12.06 ± 0.66 | 12.73 ± 0.20 | 14.86 ± 0.09 | 16.70 ± 0.65† |
| 4 | 12.80 ± 0.26 | 13.80 ± 0.78 | 15.60 ± 1.91 | ||
| PT(s) | 20–24 | 9.90 ± 0.62 | 9.93 ± 0.64 | 10.30 ± 0.43 | 10.60 ± 0.90 |
| 4 | 9.90 ± 0.50 | 9.93 ± 0.60 | 10.40 ± 1.57 | ||
| D-dimer (μg/mL) | 20–24 | 0.28 ± 0.18 | 0.24 ± 0.23 | 0.30 ± 0.18† | 0.26 ± 0.20 |
| 4 | 0.21 ± 0.05 | 0.28 ± 0.16 | 0.24 ± 0.20 | ||
| AT (%) | 20–24 | 154.6 ± 6.6 | 156.6 ± 8.1 | 152.6 ± 6.3 | 149.3 ± 9.7‡ |
| 4 | 156.6 ± 8.6 | 151.0 ± 5.5 | 150.3 ± 9.2 | ||
| Fibrinogen (mg/dL) | 20–24 | 153.0 ± 21.1 | 145.0 ± 7.2 | 154.0 ± 12.1 | 155.6 ± 12.6 |
| 4 | 142.6 ± 13.6 | 156.0 ± 5.2 | 162.6 ± 17.6 | ||
Aliquots of plasma harvested from blood samples in potassium-EDTA tubes were stored at 4°C or between 20° and 24°C, and hemostatic testing was performed immediately after collection (0 hours) and at 1, 2, and 3 hours after collection.
Significantly (P < 0.05) different from value obtained at 0 hours.
Significantly (P < 0.001) different from value obtained at 0 hours.
Mean ± SD effects of immediate assaying (0 hours) or storage between 20° and 24°C for 1 hour on hemostatic test results for plasma that was harvested from blood samples of 5 healthy dogs and transferred to EDTA tubes.*
| Hemostatic test | Time of assay after blood collection | |
|---|---|---|
| 0 hours | 1 hour | |
| aPTT(s) | 12.04 ± 0.49 | 11.88 ± 0.14 |
| PT(s) | 9.78 ± 0.54 | 9.88 ± 0.81 |
| D-dimer (μg/mL) | 0.26 ± 0.16 | 0.27 ± 0.18 |
| AT(%) | 149.4 ± 8.61 | 150.4 ± 6.84 |
| Fibrinogen (mg/dL) | 159.0 ± 21.2 | 153.4 ± 22.6 |
For the immediate assay, plasma was harvested immediately after blood sample collection; for stored samples, plasma was harvested after 1 hour of storage.
Discussion
Although the use of EDTA-treated plasma samples for measurement of PT to diagnose hemostatic disorders or monitor the effects of anticoagulant therapy in humans has been reported,2 to our knowledge, the use of EDTA as anticoagulant for hemostatic tests as an alternative to sodium citrate in veterinary medicine has not been explored. The anticoagulant EDTA acts as a chelating agent by forming complexes with calcium ions (essential for coagulation), thereby decreasing the amount of ionized calcium available for coagulation. In canine blood, the decrease in ionized calcium induced by EDTA is comparable to that induced by sodium citrate.7 One may postulate that the calcium ions contained in the reagents added during the performance of various hemostatic assays should compensate for the low concentration of ionized calcium in the EDTA-treated plasma sample and allow the coagulation reaction to occur, but whether this happens is influenced by the testing method used. For example, in the PT assay used in the study reported here, only approximately 5% of the reaction mixture consisted of plasma sample, whereas another method may use up to 33% plasma.8 With that other method, it may be more difficult to achieve an adequate concentration of ionized calcium to allow development of the appropriate reactions.
The results of all hemostatic tests of citrated and EDTA-treated plasma samples obtained from dogs were highly correlated (r > 0.96 in all situations; r > 0.99 for PT and aPTT). A good correlation between results for citrated plasma and EDTA-treated plasma with citrate plasma calibrators is also reported for humans.8 Despite this high correlation, the results of the hemostatic tests performed in our study differed with respect to direction of differences between EDTA-treated and citrated plasma. For PT, D-dimer concentration, and AT, values were significantly higher for EDTA-treated plasma samples, compared with values for citrated plasma samples. On the other hand, fibrinogen concentrations were significantly lower in EDTA-treated versus citrated plasma samples. One possible explanation for these discrepancies is that the dilutions of collected blood samples differed depending on whether EDTA or sodium citrate was used as an anticoagulant. Theoretically, the low dilution (2%) of EDTA-treated plasma samples may cause higher values for results of immunologic and chromogenic assays (ie, D-dimer concentration and AT, respectively), compared with results for citrated plasma samples (10% dilution). In our study, values of D-dimer and AT, which were measured by immunologic or chromogenic assays, were approximately 10% higher in EDTA-treated versus citrated plasma samples. For tests based on the activity of coagulation factors (ie, PT, aPTT, and fibrinogen), a less dilute sample will contain a higher concentration of coagulation factors, which should result in faster coagulation times, a subsequent increase in fibrinogen concentration, and a decrease in values for PT and aPTT. However, in our study, fibrinogen concentrations were actually lower, and values of PT and aPTT were not lower, in EDTA-treated versus citrated plasma samples. In fact, PT values were significantly higher for EDTA-treated plasma samples, compared with PT values for citrated plasma samples. Therefore, EDTA may interfere with some coagulation proteins. It is recognized that molecular configuration of coagulation proteins such as Factor VIII is maintained by ionized calcium and that configuration may be altered by a strong chelator such as EDTA, which would cause a loss in procoagulant activity.9
Another source of variation between results for the EDTA-treated and citrated plasma samples could have been the substance used to calibrate the automated coagulation analyzer (citrated plasma). Use of the same anti-coagulant at the same concentration as that of the analyzed plasma samples for calibration of the machine may have reduced the differences between EDTA-treated and citrated plasma samples detected in our study, as has been reported for analysis of human plasma samples.2
The CV values for citrated plasma samples were < 10% for most hemostatic tests with the exception of D-dimer concentration, the CV for which was > 10%, likely because of the low concentrations of the plasma samples, which reportedly can affect test results for dogs.10 However, higher interassay CV values were obtained for values of PT and aPTT of EDTA-treated plasma samples, which was likely attributable to significant changes in the results of these tests when samples were frozen. Regardless of temperature of storage, there was a significant increase in results after 12 hours for the PT and aPTT tests of EDTA-treated plasma samples. To the authors' knowledge, this effect has not been reported and may indicate that storage induced instability in some coagulation proteins. Type of blood collection tube used can also influence results of hemostatic tests and may have affected the results of our stability tests.3,11 The instability was less evident in our shortterm storage experiment, for which results correspond with those of a human study3 in which changes in PT were not significant after 4 or 6 hours of storage between 20° and 24°C. Changes in our study were almost negligible when plasma or whole blood samples were stored for 1 hour on the counter or in a refrigerator (4°C). On the basis of the aforementioned findings, we recommend performance of hemostatic tests within 1 hour after blood sample collection when EDTA tubes are used.
Results of our study suggested that a common EDTA tube may be used for hematologic and hemostatic analyses. Because the size of the blood sample required for a CBC and blood smear is usually not > 250 μL, the sample can be centrifuged to obtain plasma for coagulation tests after hematologic analyses are performed, provided that the coagulation tests are performed within 1 hour after sample collection. This order of events may have many advantages over a 2-tube collection system, including speedier and easier blood collection, smaller volume of blood required (important in very young or small dogs), lower cost of materials, lower consumption of natural material, and lower production of waste.8 Blood samples collected into EDTA tubes could also be used for other types of analyses, such as glycosylated hemoglobin measurements or PCR assays in which the use of EDTA as anticoagulant is preferable.12 The small amount of anticoagulant in an EDTA tube dilutes the blood minimally, which maintains the blood closer to its physiologic state and decreases the risk of analytic errors attributable to dilution, compared with larger amounts of anticoagulant in a sodium citrate tube, in which variability in dilutions of blood (eg, an under-filled tube) are more likely to occur.3,13
Results of the study reported here should be interpreted with caution because they were obtained via specific equipment and commercial kits, and important variations in results may be obtained with other collection tubes, reagents, techniques, or analyzers.2,14,15 For example, a study9 of human plasma samples revealed that readdition of calcium to plasma samples before analysis results in better analytic performance and increases the stability of coagulation factors. Therefore, complete validation of the analytic protocol, including establishment of reference ranges, should be performed for each laboratory before EDTA tubes are used for coagulation tests. Moreover, additional studies should be performed to compare clinical sensitivity and specificity of EDTA-treated and citrated plasma samples for different hemostatic disorders in dogs such as DIC, rodenticide ingestion, thrombosis, or hemophilia.
ABBREVIATIONS
| aPTT | Activated partial thromboplastin time |
| AT | Antithrombin activity |
| CV | Coefficient of variation |
| DIC | Disseminated intravascular coagulation |
| LOA | Limits of agreement |
| PT | Prothrombin time |
Vacuette system, Greiner Bio-One, Frickenhausen, Germany.
Becton-Dickinson, Plymouth, UK.
Vacuette coagulation tubes, Greiner Bio-One, Frickenhausen, Germany.
Kaltek, Padova, Italy.
STA Compact, Diagnostic Stago, Roche, Basilea, Switzerland.
STA reagents, Diagnostic Stago, Roche, Basilea, Switzerland.
SPSS, version 15.0, SPSS Inc, Chicago, Ill.
References
- 1.↑
Young DS, Bermes EW. Specimen collection and other preanalytical variables. In: Burtis CA, Ashwood AR, eds. Tiezt fundamentals of clinical chemistry. 5th ed. Philadelphia: WB Saunders Co, 2001;30–54.
- 2.↑
Horsti J. Use of EDTA samples for prothrombin time measurement in patients receiving oral anticoagulants. Haematologica 2001;86:851–855.
- 3.↑
Horsti J. EDTA samples are stable for prothrombin time measurement by combined thromboplastin reagent. Clin Chem 2001;47:1731–1733.
- 4.↑
Caldin M, Furlanello T, Lubas G. Validation of an immunoturbidimetric D-dimer assay in canine citrated plasma. Vet Clin Pathol 2000;29:51–54.
- 5.↑
Büttner J, Borth R, Boutwell JH, et al. International Federation of Clinical Chemistry. Committee on Standards. Expert Panel on Nomenclature and Principles of Quality Control in Clinical Chemistry. Approved recommendation (1978) on quality control in clinical chemistry. Part 2. Assesment of analytical methods for routine use. Clin Chim Acta 1979;98:145F–162F.
- 6.↑
Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–310.
- 7.↑
Ceron JJ, Martinez-Subiela S, Henneman C, et al. The effects of different anticoagulants on routine canine plasma biochemistry. Vet J 2004;167:294–301.
- 8.↑
Horsti J. Measurement of prothrombin time in EDTA plasma with combined thromboplastin reagent. Clin Chem 2000;46:1844–1846.
- 9.↑
Wang W, Wang J, Kelner D. Coagulation factor VIII: structure and stability. Int J Pharm 2003;259:1–15.
- 10.↑
Furlanello T, Caldin M, Stoco A, et al. Stability of stored canine plasma for hemostasis testing. Vet Clin Pathol 2006;35:204–207.
- 11.
Toulon P, Ajzenberg N, Smahi M, et al. A new plastic collection tube made of polyethylene terephtalate is suitable for monitoring traditional anticoagulant therapy. Thromb Res 2007;119:135–143.
- 13.
International Council for Standardization in Hematology. Expert Panel on Cytometry. Recommendations of the International Council for Standardization in Hematology for ethylenediaminetetraacetic acid anticoagulation of blood for blood cell counting and sizing. Am J Clin Pathol 1993;100:371–372.
- 14.
Stokol T, Brooks MB, Erb HN. Effect of citrate concentration on coagulation test results in dogs. J Am Vet Med Assoc 2000;217:1672–1677.
- 15.
Tseng LW, Hughes D, Giger U. Evaluation of a point-of-care coagulation analyzer for measurement of prothrombin time, activated partial thromboplastine time, and activated clotting time in dogs. Am J Vet Res 2001;62:1455–1460.
