Correspondence P. Boutet, GIGA-Research, Cellular and Molecular Physiology Unit, B34 (2nd floor), Avenue de l'Hôpital, 1, B-4000 Sart-Tilman, Belgium E-mail: firstname.lastname@example.org
Background: D-dimer measurement in dogs is considered the most reliable test for detecting disseminated intravascular coagulation or thromboembolism.
Objectives: The purposes of this study were to compare 2 D-dimer assays, a quantitative immunoturbidimetric and a semiquantitative latex agglutination assay, and to assess the effect of hemolysis and storage conditions on D-dimer concentration using the quantitative assay.
Methods: The immunoturbidimetric assay was validated using canine citrated plasma samples containing different concentrations of D-dimer. The effect of storage at various temperatures and times was assessed. Hemolysis was produced by adding lysed RBCs to the samples for a final hemoglobin concentration of 0.35 g/dL.
Results: For clinically relevant values (>250 μg/L), intra-assay and interassay coefficients of variation were 6.8% and 7.2%. The assay was linear (r2=1.00), and the tests had good agreement (κ=0.685, P<.001). Storage at 4 °C and −20 °C and hemolysis had no significant effect on D-dimer concentrations. In hemolyzed samples stored at room temperature for ≥48 hours, fine clots were noted and often resulted in falsely increased D-dimer concentrations.
Conclusions: Our findings suggest that the immunoturbidimetric assay validated in this study is reliable and accurate for the measurement of D-dimer in canine plasma. Samples can be stored for up to 1 month at −20 °C and moderate hemolysis does not significantly affect the D-dimer concentration in frozen or refrigerated samples.
D-dimer measurement in dogs is considered the most reliable and sensitive diagnostic test for detecting and monitoring disseminated intravascular coagulation (DIC) or thromboembolism (TE).1–5 Assays for D-dimer antigen are based on monoclonal antibodies binding to the fibrin fragment, D-dimer. They are therefore specific for the degradation of cross-linked fibrin, and thus for active coagulation and fibrinolysis.6,7
Various D-dimer assays have been validated in dogs, including immunoturbidimetric and latex agglutination tests. The D-dimer concentration for healthy dogs is <250 μg/L or <300 μg/L.2 The sensitivity of D-dimer concentration for predicting TE varies according to the cut-off value used. Nelson and Andreasen2 found 100% sensitivity and 70% specificity at a 500 μg/L cut-off and 36% sensitivity and 98.5% specificity at a 2000 μg/L cut-off. They suggested that higher concentrations of D-dimer were more likely to be associated with TE compared with lower concentrations of D-dimer. Identification of an appropriate threshold value is a key point in early identification of DIC or TE. Several diseases other than DIC or TE may cause D-dimer concentration to increase. Hence, a quantitative test would help determine an accurate clinical decision level, allow more effective therapy, and potentially reduce the high morbidity and mortality that results from this severe complication of many diseases.
DIC is precipitated by an underlying clinical disorder, such as bacterial or parasitic infections, severe inflammatory diseases such as pancreatitis or hepatitis, neoplasia, severe trauma, or burns, venoms, toxins, and intravascular hemolysis, including immune-mediated hemolytic anemia (IMHA).8 Although many laboratory tests are affected by the presence of free hemoglobin in the sample, to the best of the authors' knowledge, the effect of hemolysis on D-dimer concentration according to the storage condition has not been addressed previously.
The objectives of this study were to (1) compare 2 D-dimer assays, a quantitative immunoturbidimetric and a semiquantitative latex agglutination method, both of which use the same monoclonal antibody; and (2) assess the effects of hemolysis and storage conditions on the D-dimer concentration measured by the immunoturbidimetric test in citrated plasma.
Materials and Methods
D-dimer was measured by using a semiquantitative latex agglutination assay (Accuclot, D-dimer, Trinity Biotech, Bray, UK) and a quantitative immunoturbidimetric assay (Miniquant D-dimer, Biopool, Trinity Biotech). The latter assay has a working range of 75–3200 μg/L and samples with higher concentrations were diluted 1:1 in saline. Results were reported as the mean of duplicate assays. For the latex agglutination assay, D-dimer concentration was categorized as follows: 1, <250 μg/L; 2, 250–500 μg/L; 3, 500–1000 μg/L; 4, 1000–2000 μg/L, and 5, >2000 μg/L.
Validation of the immunoturbidimetric assay
The immunoturbidimetric assay was compared with the latex agglutination assay currently used at the Central Diagnostic Services of the Queen's Veterinary School Hospital (QVSH) of the University of Cambridge. This semiquantitative test has been validated previously in dogs.2,8 Both tests use the same mouse monoclonal antibody (clone MA-8D3), which is specific for D-dimer.9 The manufacturers indicate that they selected a hybridoma that secretes antibodies with a 1000-fold greater affinity for purified D-dimer and fragment D of fibrin, than for native fibrinogen.
To validate the new immunoturbidimetric assay, within-assay (intraassay) variability was calculated by determining D-dimer concentrations in 2 samples of low and high levels, repeated 10 times. The coefficient of variation (CV) for each sample was determined. The assay-to-assay (interassay) CV was determined on D-dimer concentrations in samples from 8 dogs (D-dimer concentration 24.6–3928.5 μg/L) that were split into 6 equal aliquots and stored frozen at −20 °C. D-dimer concentrations were measured in an aliquot on each of 6 consecutive days.
Linearity was measured over 5 dilutions of each of 2 samples with D-dimer concentrations >3000 μg/L. The samples were used undiluted and diluted with saline to obtain samples containing 75%, 50%, 25%, and 12.5% of the analyte. The measured concentrations were plotted against the dilutions in a linear regression analysis.
The relationship between the 2 assays was assessed by parallel analysis of 42 samples of different concentrations. The samples were collected from 25 healthy dogs and 17 dogs presented to the QVSH with various diseases, including hepatitis, pancreatitis, neoplasia, trauma, DIC (5 dogs), and TE (1 dog diagnosed at necropsy). The diagnosis of DIC or TE was based on supportive laboratory findings (prolonged activated partial thromboplastin time and/or prothrombin time, and thrombocytopenia), clinical signs, thoracic radiographs, ultrasonography, histopathology, and postmortem examination. For statistical analysis, the quantitative values were converted to the equivalent semiquantitative category, and agreement between the assays was assessed by the κ statistic.
Effect of sample storage and hemolysis using the immunoturbidimetric assay
Blood samples were collected from the jugular vein into sodium citrate-containing tubes (3.8% wt/vol). Some samples were collected from hospital patients referred to the QVSH, while other samples were mailed to the diagnostic laboratory from external practices. Samples were no older than 24 hours when received. Samples were centrifuged at 2000 g for 10 min and the plasma harvested.
To assess the effect of storage, 5 samples with D-dimer concentrations >500 μg/L were divided into 11 aliquots of 100 μL each. The determination of D-dimer was carried out within 2 hours, or after storage at room temperature (∼20 °C) for 24 and 48 hours, after refrigeration (4 °C) for 24 and 48 hours, or after freezing (−20 °C) for 2, 5, 8, 11, 14, and 30 days. The recommendations provided by the manufacturer were carefully followed for the stability and storage of reagents.
To assess the effect of hemolysis, approximately 10 μL of a solution of lysed RBCs were added to 5 citrated plasma samples with D-dimer concentrations >500 μg/L to obtain a final hemoglobin concentration of 0.35 g/dL. This concentration was considered as equivalent to moderate to marked hemolysis. The final hemoglobin concentration of the plasma was determined with a CELL-DYN 3500 analyzer (Abbott, Maidenhead, UK) using a modified cyanide method. The hemoglobin solution was prepared from red cells washed twice in saline and lysed by several cycles of freezing and thawing. Ten samples were divided into 100-μL aliquots, and the D-dimer concentration was measured after storage for 24, 48, and 72 hours at room temperature; after 24 and 48 hours of refrigeration; and after 2, 3, and 30 days of freezing.
Stability of stored samples and the effect of hemolysis were assessed by the mean percentage deviation from the initial D-dimer concentration in the sample. The 5 mean deviations from each storage and hemolysis condition were compared with zero by a 1-sample Student's t-test. The criteria for significance was P<.05.
The mean intra-assay CV for both samples in the immunoturbidimetric assay was 10.5% (Table 1). The interassay CV was 11.1% when calculated only with values within the declared working range of the assay (75–3200 μg/L) and 7.2% for clinically relevant values (>250 μg/L) (Table 2). Using linear regression analysis to determine the linearity of the assay, an r2 value of 1.00 was obtained for a regression line passing through the origin, with slopes of 0.99 and 1.00, respectively, for the 2 samples.
Table 1. Intraassay variability for the immunoturbidimetric assay using citrated plasma samples with low and high D-dimer concentrations.
Table 2. Interassay variability of the immunoturbidimetric assay.*
Interassay Sample No.
D-dimer was measured on 6 separate days in 8 citrated plasma samples of different D-dimer concentrations.
The medians of the mean percentage change after storage and with hemolysis are shown in Table 3. Storage conditions of nonhemolyzed samples did not significantly affect D-dimer concentrations. The mean percentage change for hemolyzed samples was not significant, although fine clots often formed in samples stored at room temperature for at least 48 hours. These fine aggregates were associated with falsely increased D-dimer concentrations (reported as >5000 μg/L). Storage of hemolyzed samples at room temperature for 48 and 72 hours led to the largest percentage changes, ranging from −24% to +69% and −32% to +194%, respectively (Table 3).
Table 3. Effect of hemolysis and storage for up to 30 days on samples with D-dimer concentrations >500 μg/L.*
Nonhemolyzed Samples (n=5)
Hemolyzed Samples (n=5)
Data are the median (minimum–maximum) of the mean percentage change for each storage condition.
−11 (−5 to −24%)
−13 (−21 to 2%)
−8 (−13 to 10%)
−3 (−11 to 7%)
−7 (−28 to 5%)
−4 (−22 to 5%)
−4 (−9 to 3%)
−5 (−24 to 69%)
−2 (−11 to 4%)
−3 (−5 to 3%)
0 (−32 to 194%)
−6 (−15 to 9%)
−1 (−10 to 5%)
−4 (−13 to 9%)
−3 (−14 to 5%)
−9 (−12 to 18%)
−3 (−12 to 11%)
−3 (−28 to 2%)
−1 (−3 to 8%)
The 2 assays were in good agreement (κ=0.685, P<.001) (Figure 1). All healthy dogs had D-dimer concentrations of <250 μg/L by the latex agglutination assay (mean concentration by the immunoturbidimetric assay, 39 μg/L, range 0–163 μg/L). All 5 dogs diagnosed with DIC had D-dimer concentrations >2000 μg/L for both tests (mean immunoturbidimetric concentration, 3191 μg/L; range, 2567–3765 μg/L). One dog diagnosed with TE (thrombus in iliac artery) had a D-dimer concentration >1000 μg/L (immunoturbidimetric concentration, 868 μg/L).
The immunoturbidimetric assay used in this study seems to accurately measure D-dimer in canine citrated plasma. We compared this assay with the latex-agglutination test currently used in our laboratory, which has been demonstrated to be sensitive and specific for the diagnosis of TE in dogs.2 Both tests are based on the same monoclonal antibody, but the immunoturbidimetric assay offers the advantage of speed and eventual automation. For D-dimer concentrations ranging from 250 to 3200 μg/L, within-assay and assay-to-assay CVs were <10%, as reported by the manufacturer. These results are in agreement with CVs reported for another immunoturbidimetric assay.3 In the present study, the immunoturbidimetric assay had good concordance with the latex-agglutination test.
DIC is a subacute, acute, or chronic thrombohemorrhagic disorder that occurs as a secondary complication of a variety of diseases.10 In dogs, DIC has been reported with protein-losing disease, sepsis, pancreatitis, cancer, IMHA, increased endogenous or exogenous corticosteroids exposure, and vascular disease.2,4,8 Glomerulonephropathy, neoplasia, and IMHA were the most frequent diagnoses in dogs with TE.2 Assessment of interference in D-dimer measurement by hemolysis is important since TE is a frequent cause of death in dogs with primary IMHA.11,12 No significant effect of hemolysis (hemoglobin=0.35 g/dL) was observed when samples were stored refrigerated or frozen. However, when left at room temperature for at least 48 hours, we noted small clots in some samples. These clots were often associated with falsely high concentrations (>5000 μg/L) of D-dimer. The clots likely increased turbidity by interfering with scattered light. The fine clots might be due to fibrin-hemoglobin aggregates that increase with time at room temperature, as seen in hemolyzed samples in this study. Centrifugation at 2000 g may be insufficient to clear the plasma of fine particles. However, hemolysis did not interfere with D-dimer concentration if room temperature samples were measured within 24 hours or if they were appropriately refrigerated or frozen. In human plasma, the manufacturer did not find any interference by hemoglobin at concentrations <6.7 g/L. However, the samples were not assessed repeatedly over time.
Storage of nonhemolyzed samples did not affect D-dimer concentration in any of the conditions reported in this study. D-dimer was stable in samples stored at −20 °C for up to 1 month after collection. These findings are in agreement with those of Stokol et al.3
In conclusion, the immunoturbidimetric assay validated for the measurement of D-dimer concentration in canine citrated plasma seems to be as reliable and accurate as the latex-agglutination test. The effect of storage of nonhemolyzed plasma samples is not significant and hemolysis (at hemoglobin concentrations of 0.35 g/dL) does not affect D-dimer concentration when samples are refrigerated or frozen. Further work with a larger number of dogs is needed to reliably assess the clinical sensitivity and specificity of these tests, and to determine an accurate threshold for clinical decisions regarding the presence or absence of significant fibrinolysis.
The authors are grateful to G. E. McGowen and the technicians of the Central Diagnostic Services for their excellent technical assistance, constant helpfulness, and overall kindness. We warmly thank Dr. K. Freeman for critical reading of the manuscript. This study was partly supported by the University of Liège (Belgium), its Faculty of Veterinary Medicine, and the “Prix Edmond Huynen de Médecine vétérinaire.”