Correspondence Harold Tvedten, Department of Clinical Sciences, Faculty of Veterinary Medicine and Animal Science, Swedish University of Agricultural Sciences, Box 7054, SE-750 07 Uppsala, Sweden E-mail: email@example.com
Background: Many Cavalier King Charles Spaniel (CKCS) dogs are affected by an autosomal recessive dysplasia of platelets resulting in fewer but larger platelets. The IDEXX Vet Autoread (QBC) hematology analyzer directly measures the relative volume of platelets in a blood sample (plateletcrit). We hypothesized that CKCS both with and without hereditary macrothrombocytosis would have a normal plateletcrit and that the QBC results would better identify the total circulating volume of platelets in CKSC than methods directly enumerating platelet numbers.
Objectives: The major purpose of this study was to compare the QBC platelet results with platelet counts from other automated and manual methods for evaluating platelet status in CKCS dogs.
Methods: Platelet counts were determined in fresh EDTA blood from 27 adult CKCS dogs using the QBC, Sysmex XT-2000iV (optical and impedance), CELL-DYN 3500, blood smear estimate, and manual methods. Sysmex optical platelet counts were reanalyzed following gating to determine the number and percentage of normal- and large-sized platelets in each blood sample.
Results: None of the 27 CKCS dogs had thrombocytopenia (defined as <164 × 109 platelets/L) based on the QBC platelet count. Fourteen (52%) to 18 (66%) of the dogs had thrombocytopenia with other methods. The percentage of large platelets, as determined by regating the Sysmex optical platelet counts, ranged from 1% to 75%, in a gradual continuum.
Conclusions: The QBC may be the best analyzer for assessing clinically relevant thrombocytopenia in CKCS dogs, because its platelet count is based on the plateletcrit, a measurement of platelet mass.
Many Cavalier King Charles Spaniel (CKCS) dogs have fewer and larger platelets than other breeds because of an inherited autosomal recessive dysplasia of platelets.1 Because the disorder is not associated with clinical abnormalities or bleeding tendency, recognition of hereditary macrothrombocytosis (macrothrombocytopenia) in CKCS is important to avoid unnecessary treatment of affected dogs for thrombocytopenia. A method is needed to identify when a CKCS dog has true thrombocytopenia (decreased platelet mass) that is clinically significant. Like other breeds of dogs, CKCS can develop Sertoli cell tumors, immune mediated thrombocytopenia, myelodysplasia, disseminated intravascular coagulation, and other diseases that can affect platelet production or removal and cause a bleeding tendency due to insufficient platelet numbers.
Total platelet mass is the most physiologically relevant parameter in hemostasis and in the regulation of thrombopoiesis.2 Plateletcrit in a blood sample is an indicator of platelet mass in the body, just as HCT is an indicator of total erythrocyte mass in the body. Platetcrit is determined by the QBC analyzer and subsequently converted to a calculated platelet count for ease of interpretation by veterinary practitioners. We hypothesized that the platelet results from the QBC analyzer would more accurately assess the platelet status of CKCS, compared with platelet counts from other analyzers or methods. The purpose of this study was to compare QBC platelet results with platelet counts obtained on the Sysmex and CELL-DYN automated analyzers and manual counts in CKCS with presumed normal circulating platelet mass. In addition, we regated the Sysmex optical platelet count to assess the percentage of large platelets, and evaluated blood smears for uneven distribution of large platelets.
Materials and Methods
Fresh blood was collected by jugular venipuncture using 0.8 × 19 mm butterfly needles (Neofly, Ohlmeda AB, Helsingborg, Sweden) and placed in 4.5 mL Vacutainer tubes containing EDTA (Becton Dickinson, Meylon, France). Samples were obtained from 27 adult CKCS dogs. Twenty-five of the dogs were being evaluated in a study of heart disease at the Swedish University of Agricultural Sciences and had blood samples collected for other purposes. The University and Swedish regulatory bodies for ethical animal usage approved this research. None of the 25 dogs received any medical therapy. None had evidence of organ-related or systemic disease other than mitral regurgitation. Two additional dogs were patients at the University hospital, which required hematologic evaluation of their clinical status. One was later euthanatized because of mitral valve regurgitation; it also had chronic nephrosis and lymphocytosis with atypical lymphocytes for 1½ years, which was diagnosed as lymphoid hyperplasia at necropsy. The other dog had coughing and sneezing with enlarged tonsils.
The samples were transported directly to the laboratory within 30 minutes of collection. All platelet counts were determined within 1–3 hours of collection. Platelet counts were determined using a VetAutoread (QBC) hematology analyzer (IDEXX Laboratories, Westbrook, ME, USA), Sysmex XT-2000iV (Sysmex Corporation, Kobe, Japan), and CELL-DYN 3500 (Abbott Laboratories, Abbott Park, IL, USA). The QBC platelet count is based on the instrument's measurement of the relative volume (percentage) of platelets in the blood sample (the plateletcrit). On the Sysmex, both optical (PLT-O) and impedance (PLT-I) platelet counts were determined. The CELL-DYN counts platelets using an impedance method. In addition, platelets were estimated in smears and counted manually. For manual counts, blood was diluted 1:20, stained with Leucoplate (Sobioda, Montbonnet, France), and counted in a glass counting chamber. Platelet estimates were determined by counting the number of platelets in 10 oil immersion × 100 objective fields in the monolayer area of a blood smear stained with modified Wrights stain (Ames Hematek, Miles Diagnostic Division, Elkhart, IN, USA). The average number of platelets per field was multiplied by 15 × 109/L to obtain the platelet estimate.3 The average number of platelets in 4 fields along the feathered edge of each blood smear was also determined in 22 of the 27 smears without aggregates to evaluate the effect of uneven distribution of platelets on platelet estimates.
Reference values for canine platelet counts for interpretation were defined as 164–510 × 109/L, based on values from Michigan State University.3 Reference values vary with different methods but a lower limit of 164 × 109/L is similar to that of many reference values. Linear regression analysis was performed with Microsoft Excel 2003 (Microsoft Corporation, Redmond, WA, USA). Precision of the Sysmex was determined by repeating the PLT-O and PLT-I analyses 10 times with 2 of the CKCS blood samples and then dividing the mean by the SD and multiplying by 100 to obtain the coefficient of variation (CV). The precision study was performed at the same time as other platelet analyses.
The Sysmex optical platelet count determines platelet size by forward laser light scatter and RNA content by lateral fluorescent light (SFL) (Figure 1). Platelets that are larger and have more RNA are located higher up and to the right in the PLT-O cytogram (Figure 2). The PLT-O cytogram was regated with a subjective algorithm based on the appearance of the PLT-O graphical display of 1 CKCS (Figure 1), which appeared to be unaffected with macrothrombocytosis and to have normal-sized platelets compared with other CKCS in the study, and based on the appearance of the cytogram in other normal dogs. The algorithm was then reapplied to all PLT-O results. The difference between the original PLT-O and the regated PLT-O, representing apparently normal-sized platelets, was interpreted to be the number of large platelets in that sample.
The dogs had QBC platelet counts ranging from 171–524 × 109/L, which were within the canine reference values (Figure 3).2 A prominent layer of platelets was visible in all QBC tubes (Figure 4). The width of this layer reflects the size of the plateletcrit (the percentage of blood volume comprised of platelets), and is converted by the QBC to a calculated platelet count. Because the QBC platelet counts should be proportional to the sample's plateletcrit, all plateletcrits were also considered normal. Of the 27 dogs, 14 (52%) to 18 (66%) had low platelet counts with other methods (Table 1).
Table 1. Range of platelet counts and frequency of macrothrombocytopenia in Cavalier King Charles Spaniels (n=27) using 5 methods.
Impedance, optical, and manual methods gave similar results when platelet counts were >164 × 109/L. These CKCS did not likely have hereditary macrothrombocytosis. More variation in results was seen with samples that had the lowest platelet counts. Results obtained by impedance counting (Sysmex PLT-I and CELL-DYN) were similar and distinctly lower than those obtained manually or with Sysmex PLT-O methods. Twelve of 27 (44%) dogs had counts of ≤10 × 109/L with the CELL-DYN and 4 had counts of ≤10 × 109/L with PLT-I. The lowest PLT-O count was 44 × 109/L and the lowest manual count was 30 × 109/L. The manual counts and Sysmex PLT-O counts were higher than impedance counts (PLT-I and CELL-DYN). The difference between impedance counts and optical or manual counts was more apparent at low platelet counts (Figure 3). When manual platelet counts were <164 × 109/L, the correlation (r) between PLT-O and manual platelet count was 0.89 and between PLT-I and manual platelet count was 0.76. The poorer correlation of manual counts with impedance counts in CKCS with low platelet counts supported the impression gained from visual inspection of the lower values in Figure 3.
Intra-assay precision in 2 CKCS dogs with low platelet counts was better with the Sysmex PLT-O (CV 3%, 6%) than with the Sysmex PLT-I (CV 19%, 33%) methods. These 2 dogs had PLT-O counts (mean±SD) of 69.4±3.9 and 125.7±4.3 and PLT-I counts of 14.0±4.69 and 46.8±9.0, respectively.
The number of normal-sized platelets in the blood samples was based on the subjective algorithm (regating) of the Sysmex PLT-O counts. Six of 27 dogs had 99% normal-sized platelets, and 5 additional dogs had >90% normal-sized platelets (Figure 5). The remaining dogs had 25–85% normal-sized platelets, comprising a gradual continuum without clear distinction of dogs with macrothrombocytosis. Expressing this in terms of the percentage of large platelets, the dogs had 1–75% large platelets in a gradual continuum.
Six CKCS dogs that had <50% normal-sized platelets, based on the PLT-O algorithm, had an average of 7.1 times more platelets near the feathered edge than in the monolayer area. Sixteen dogs with >50% normal-sized platelets had 2.4 times more platelets at the feathered edge than in the monolayer area.
Total platelet counts often cause misinterpretation of platelet status in CKCS patients with hereditary macrothrombocytosis. Platelet counts <10 × 109/L were found in 44% of our dogs based on CELL-DYN (impedance) results. Thus, even very severe decreases in platelet counts were seen in CKCS with hereditary macrothrombocytosis. None of the dogs had a reduced platelet count based on the QBC results.
The platelet counts reported by the QBC were actually incorrect in the CKCS with hereditary macrothrombocytosis. The QBC directly measures plateletcrit instead of platelet numbers. It measures the width of various layers of cells in the expanded buffy coat and because this is within a defined space in the tube, the volume of cells in the layers is measured. The relative volume of blood comprised of platelets is the plateletcrit (% or L/L). The widths of the layers are converted to a calculated number of cells. Veterinarians and others universally use platelet counts to interpret platelet status in a patient, so the plateletcrit measured by the QBC is converted into a platelet count. Even though the number of platelets reported by the QBC in CKCS with hereditary macrothrombocytosis was incorrect, it gave an estimate of platelet mass which is clinically more relevant. Conversion of plateletcrit by the QBC to a platelet count is only accurate when platelets in the sample are normal in size for dogs.
Total platelet mass is the most physiologically relevant platelet parameter.2 The QBC platelet count, which is really more an indicator of plateletcrit, should provide the best estimate of total platelet mass in CKCS dogs of the methods used. QBC platelet counts were usually normal in CKCS in another study.4 Only 1 of 17 CKCS dogs had a decreased platelet count by the QBC method while 14 of 41 had <100 × 109 platelets/L by an impedance cell counter and 11 of 41 had <100 × 109 platelets/L by a laser cell counter.
An estimate of plateletcrit was not available from the other methods. MPV is not reported by the Sysmex if the PLT-I histogram is abnormal. The MPV from the CELL-DYN was not considered validated in our laboratory. Therefore, plateletcrit could not be calculated from the MPV and platelet counts obtained from these instruments in our study. Bertazzolo et al4 found no significant difference in the MPV of CKCS with or without macrothrombocytes (11.6 and 12.1 fL, respectively) with an impedance instrument. The authors attributed that to the inability of an impedance instrument to detect large platelets. Thus, a plateletcrit calculated from an impedance hematology instrument would likely be erroneous in CKCS. The MPV (5.2 and 6.5 fL, respectively) from the laser hematology instrument used in that study (Bayer H1, Bayer Diagnostics, Tarrytown, NY, USA) was only half that obtained from the impedance instrument, suggesting the Bayer H1 also did not accurately estimate the plateletcrit. Only the QBC results indicated that the dogs in this study were free from a clinically relevant deficiency of platelets. We conclude that the QBC method is the best for detecting clinically relevant thrombocytopenia in this breed, of the instruments tested thus far. The ADVIA 2120 and 120 report plateletcrit in dogs, but this method has not yet been evaluated in CKCS.
The other 5 methods indicated a reduced number of platelets in more than half of the CKCS dogs. The impedance counts usually were lower than optical (laser) or manual counts, which included more of the larger platelets. The CELL-DYN had counts of ≤10 × 109/L in 12 of the dogs, sufficiently low as to raise concern even by veterinarians aware of the defect in CKCS. The Sysmex PLT-O and manual counts appeared more correct than impedance counts, but were also lower than reference values and could have lead to an erroneous diagnosis of thrombocytopenia. Thus, none of the other 5 methods could be expected to differentiate clinically relevant thrombocytopenia in CKCS dogs with macrothrombocytosis.
Earlier reports suggested low platelet counts in CKCS were due to an autoanalyzer error, and that blood smear estimates or manual counts showed a normal number of platelets.5 Our platelet estimates from blood smears were similar to platelet counts obtained with other counting methods and were in agreement regarding the presence of reduced platelet numbers. A common misconception in comparing methods of platelet counting is that a manual count is more accurate then an automated count. That was not true in this study, at least for the Sysmex PLT-O.
The frequency of macrothrombocytosis in CKCS varies among reports. Cowan et al6 reported that macrothrombocytes (>3 μm diameter) occurred in 33% of CKCS while 51% had a platelet count <100 × 109/L. Eksell et al7 reported a frequency of 31% in 102 CKCS dogs. Those dogs were considered affected if they had a platelet count <100 × 109/L. Prevalence of thrombocytopenia in our study was 52–69% depending on counting method and based on a cutoff of <164 × 109/L.
Because reference values for platelet counts depend on different instruments and methods, one cannot select a single platelet count as a lower reference limit between normal and affected CKCS dogs. A platelet count <100 × 109/L has been used in previous studies. It is too low to include all normal dogs, but may have been chosen as a “clinical” threshold for thrombocytopenia. Using this low threshold, fewer CKCS would have been considered as affected with macrothrombocytosis in this study. Because our data indicated a gradual continuum from few to many large platelets, no single platelet size cutoff would have clearly differentiated affected from unaffected CKCS dogs. Thus, the prevalence of macrothrombocytosis previously reported should be considered as approximate.
Brown et al5 reported that platelets in CKCS dogs had diameters of 2.5–3.75 μm while platelets of other breeds were 1.25–2.5 μm in diameter. The authors suggested a bimodal distribution of patients based on platelet diameter on blood smears. Our results indicated a gradual continuum in the percentage of large platelets without bimodal distribution and thus no clear division between normal and affected dogs. The PLT-O cytogram of affected CKCS also showed a gradual continuum of small to large sized platelets without any suggestion of a bimodal grouping. The flexible software of the Sysmex allowed easy estimation of the number of large platelets in each of the dogs.
The PLT-O system of the Sysmex shows affected CKCS platelets as larger and containing more RNA, detected as lateral fluorescent light (SFL). Our regating scheme seems similar to what the manufacturers of the Sysmex XE-21000 call the immature platelet fraction.8 In the Sysmex, platelets are stained with a proprietary nucleic acid dye containing polymethine.8 An increased number of these large RNA-rich platelets usually is interpreted to indicate increased thrombopoiesis.9 The assumption that they are immature platelets may not always be correct if hereditary macrothrombocytosis of CKCS does not affect platelet lifespan or rate of thrombopoiesis.
CKCS dogs with a high percentage of large platelets had a greater number of platelets along the feathered edge of the blood smear than dogs with a high percentage of normal-sized platelets. This supports the concept that larger cells are distributed unevenly to the distal edge of blood smears, which causes errors in counting.
In conclusion, platelet count, as determined via the plateletcrit on the QBC, was normal in CKCS dogs with and without hereditary macrothrombocytosis. Plateletcrit is physiologically the most relevant parameter of platelet status, and the QBC platelet count, although numerically incorrect in CKCS with hereditary macrothrombocytosis, should be useful for differentiating clinically relevant thrombocytopenia (decreased platelet mass) in this breed. Optical (PLT-O) and manual platelet counts likely included a larger percentage of large platelets in affected dogs and gave more accurate counts than impedance methods. However, dogs with hereditary macrothrombocytosis still had lower numbers of platelets even with manual and PLT-O counts. All methods except the QBC would indicate the affected CKCS had thrombocytopenia. The gradual continuum in the percentage of large platelets in CKCS dogs suggests the defect has variable expression.
The IDEXX VetAutoread hematology analyzer was on free loan from Kruuse, Sweden. The Sysmex XT-2000iV hematology system was on free loan from Sysmex, Kobe, Japan, Swedish Section. We acknowledge the technical assistance of Åsa Karlsson and Annika Thor-Asplund.