Primary hemostatic function in dogs with acute kidney injury

Abstract Background Bleeding tendencies can occur with uremia. Objectives To characterize primary hemostatic function in dogs with acute kidney injury (AKI). Animals Ten dogs with International Renal Interest Society AKI grade III or above and 10 healthy controls. Methods Prospective study comparing PCV, platelet count, platelet aggregometry (Multiplate), and von Willebrand factor antigen to collagen binding activity ratio (vWF:Ag:vWF:CBA) in 2 groups of dogs (AKI group versus controls). Buccal mucosal bleeding time was measured in the AKI group only. Data are presented as median [25th, 75th percentile] unless otherwise stated. Significance was set at P < .05. Results Mean PCV was significantly lower in the AKI (34.7%; ±SD, 8.8) than in the control (46.1%; ±SD, 3.6; P < .001) group. Platelet count was significantly higher in the AKI (350.5 × 103/μL [301, 516]) than in the control (241 × 103/μL [227, 251]; P = .01) group. Collagen‐activated platelet aggregometry measured as area under the curve was significantly lower in the AKI (36.9 ± 17.7) than in the control (54.9 ± 11.2; P = .05) group. vWF:Ag:vWF:CBA was significantly higher in the AKI (2.2 [1.9, 2.6]) than in the control (1.1 [1.1, 1.2]; P = .01) group. There was a strong correlation between vWF:Ag:vWF:CBA and creatinine (r = 0.859; P < .001), but no other variables. Conclusions and Clinical Importance Dogs with AKI had decreased collagen‐activated platelet aggregation and appear to have a type II von Willebrand disease‐like phenotype as indicated by the high vWF:Ag:vWF:CBA.

treatment. 1 Other invasive procedures such as placement of central venous catheters, dialysis catheters for renal replacement treatment, peritoneal catheters for peritoneal dialysis, 2,3 or feeding tube placement for longer term management might also be required. 4 Uremic bleeding occurs in people with chronic kidney disease (CKD) due to abnormalities in primary hemostasis. [5][6][7] Intrinsic factors including platelet secretion defects, abnormal platelet-vessel wall interactions including von Willebrand Factor (vWF) abnormalities, increased platelet inhibitors, uremic toxins, as well as factors such as anemia, if present, contribute to bleeding. [5][6][7] Extrinsic factors might also be involved, including medications, comorbidities, and iatrogenic factors such as extracorporeal circulation. [5][6][7] Although these mechanisms are well described in people with CKD, the understanding of uremic bleeding in dogs with AKI is limited. 8,9 It is known that platelet function is altered in uremic dogs; however, the mechanisms are not clearly understood. In experimental models of kidney injury, prolongation in buccal mucosal bleeding time (BMBT) and decreased platelet glass bead retention percentage occurs without significant changes in arachidonic acid (AA)-, adenosine diphosphate (ADP)-, collagen (COL)-, and epinephrine-induced light transmittance aggregometry (LTA). 10 In addition, vWF antigen concentrations (vWF:Ag) increases, 11 suggesting 1 of the mechanisms of bleeding to be due to abnormalities in platelet adhesion independent of vWF or platelet agonists. In dogs with naturally occurring CKD, significant changes in AA-, ADP-, and COL-induced multiple electrode impedance platelet aggregometry (MEPA) does not occur despite clinical bleeding. 12 However, platelet dysfunction as measured by the platelet function analyzer-100 does occur. 13 Despite wide acceptance that platelet dysfunction occurs in people with CKD, [5][6][7] there is not a comprehensive understanding of hemostatic defects in dogs with naturally occurring AKI. Therefore, the primary objective of this study was to characterize primary hemostatic function measured by AA-, ADP-, and COL-induced MEPA, and vWF assays in dogs with naturally occurring AKI. The secondary objective was to determine if BMBT can be used to predict primary hemostatic dysfunction in dogs with AKI when compared to MEPA and vWF assays. We hypothesized that (1) there will be no difference in AA-, ADP-, and COL-induced MEPA between dogs with AKI and healthy dogs; (2) there will be no difference in vWF:Ag, vWF collagen binding activity (vWF: CBA), and vWF:Ag to vWF:CBA ratio (vWF:Ag:vWF:CBA) between dogs with AKI and healthy dogs; and (3) there will be strong correlation between BMBT and these primary hemostatic variables measured.  Once a diagnosis of AKI was made, client consent for study enrolment was obtained, and diagnostic tests and treatment were performed at the discretion of the attending clinician. Age, breed, sex, and minimum database as described above were collected before or during the study, and additional information including evidence of clinical bleeding and final diagnosis was obtained from medical records after the study was concluded. When routine blood sampling was required at the discretion of the attending clinician, a 21-gauge blood collection system (BD Safety-Lok, Becton Dickinson, UK) was inserted in the jugular or saphenous vein, and 1.8 mL of blood was collected in a 3.2% less than 0.98 or greater than 20%, respectively. Reference intervals for MEPA were generated from sampling of 34 blood donors that were considered healthy based on history, clinical examination, hematology, and biochemistry. As the sample size was small, Horn's method was used for outlier detection and robust method was used for the calculation of reference intervals. 15 This was implemented using refLimit function from <referenceIntervals> package in the R software (R Studio, version 1.1.456, Boston, Massachusetts).

| MATERIALS AND METHODS
Sodium citrate anticoagulated blood was centrifuged at 2500g for 20 minutes at 4 C within 30 minutes of sample collection. After centrifugation, plasma was collected and stored at −80 C for no longer than 6 months. Von Willebrand Factor antigen concentrations, vWF:CBA, and vWF:Ag:vWF:CBA were measured from the citrated plasma by an external laboratory as previously described. 16

| RESULTS
During the recruitment period, 58 dogs presented with AKI, and from these, 10 dogs met the inclusion criteria for the AKI group and 10 control dogs were enrolled to match this number. Within the AKI group, 2 were male entire, 4 were male neutered, and 4 were female neutered. Within the control group, 4 were male entire, 4 were male neutered, 1 was female entire, and 1 was female neutered. Breeds in the AKI group included a German Shepherd Dog cross, Jack Russell Terrier, Cairn Terrier, Labrador Retriever, Golden Retriever, Border Collie, Boston Terrier, English Springer Spaniel, and 2 crossbreed dogs.
Breeds included in the control group included 4 Labrador Retrievers, 2 German Shepherd Dogs, 2 German Shepherd Dog crosses, 1 Boxer, and 1 Pyrenean Mountain Dog. There were significant differences in age and body weight between the 2 groups ( Table 1).
The final diagnosis of AKI included 5 dogs with AoCKD for which a reason for the acute deterioration was not identified, and 1 dog with AKI secondary to each of the following diagnoses: hypertensive nephropathy secondary to pheochromocytoma, leptospirosis, cutaneous and renal glomerular vasculopathy, pancreatitis, and gentamicin administration with associated pancreatitis.
Invasive procedures were performed in 6 dogs in the AKI group, including nasoesophageal tube placement, esophagostomy tube placement, renal biopsy, bone marrow biopsy, and prostatic wash with no evidence of clinical bleeding. Three dogs had gastrointestinal hemorrhage; however, the underlying cause was not determined. One dog was anemic at the time of admission with a PCV and total protein (TP) of 28% and 4.3 g/dL, respectively, with no clinical bleeding despite nasoesophageal tube placement. One dog was anemic with a PCV and TP of 14% and 4.8 g/dL, respectively, at the time of hospital admission with no clinical sign of bleeding despite esophagostomy tube placement. This dog was included in the study after receiving a blood transfusion. Two dogs had decreasing PCV with 1 requiring 2 blood transfusions after the study period, despite no evidence of clinical bleeding.
Packed cell volume (P = .001), COL AUC (P = .05), and vWF:CBA (P = .04) were significantly lower in the AKI group compared to the control group, although COL AUC remained within reference interval.
Platelet count (P = .01), creatinine (P < .001), urea (P < .001), and vWF:Ag:vWF:CBA (P = .001) were significantly higher in the AKI group compared to the control group. There was no significant difference between the 2 groups in other measured variables (Table 1).
There was no significant difference in any variable between the dogs with AoCKD and dogs with no underlying CKD (data not presented). There was no significant correlation between other variables (Table 2).

| DISCUSSION
This study assessing primary hemostatic function in dogs with AKI found that COL-, but not AA-or ADP-, induced platelet aggregation as measured by MEPA was significantly decreased in dogs with AKI compared to a control population of healthy dogs; however, aggregation remained within reference intervals. Although there was no difference in vWF:Ag concentration, vWF:Ag:vWF:CBA was significantly higher in dogs with AKI, indicating that less vWF was bound to COL possibly due to a reduction in higher molecular weight (MW) vWF multimers in dogs with AKI. This was supported by the correlation between vWF:Ag:vWF:CBA and both creatinine and urea. Although BMBT was marginally prolonged in dogs with AKI, there was no correlation between BMBT and any other variables. However, these aforementioned results must be interpreted in light of the difference in the lower PCV and higher platelet count in the dogs with AKI.
Our study suggests that COL-inducted platelet aggregation is impaired in dogs with AKI. Collagen found in vessel walls and basement membranes are important in normal platelet function as it directly binds to glycoprotein (GP) VI receptors and integrin α 2 β 1 , and indirectly via vWF binding to GPIb-IX-V complex which are integral processes for normal platelet function. 20 Dogs with CKD also have impairment in COL-induced platelet function as measured by platelet function analyzer's COL-and ADP-induced platelet closure time. 13 However AA-, ADP-, or COL-induced platelet aggregation as measured by LTA are normal in dogs with CKD 12 and experimentally induced kidney injury. 10 One case report describes a dog with CKD with decreased AA-and ADP-induced LTA; however, COL-induced aggregometry was not performed. 21 The difference in results might be explained by the different nature of kidney injury and the difference in methodology of testing platelet function.
There are differences between MEPA and LTA. Light transmittance aggregometry has been traditionally considered the gold standard of platelet function tests, 22 and like MEPA, platelet aggregation is measured after activating platelets with agonists including AA, ADP, and COL. The advantages of MEPA, however, are that it is a point-ofcare test which utilizes whole blood, as opposed to platelet-rich T A B L E 1 Measured variables in the acute kidney injury (AKI) and control groups vWF Ag, von Willebrand factor antigen; vWF:Ag:vWF:CBA, von Willebrand factor antigen to collagen binding activity ratio; vWF CBA, von Willebrand factor collagen binding activity. *r > ±0.5 and P < .05.; **r > ±0.7 and P < .05. plasma (PRP) which is required for LTA, minimizing preanalytical and analytical variables. The use of whole blood reduces the time to analysis, obliterates the need to produce PRP in which platelets could be activated in the process, and the sample used is more physiological having erythrocytes present which augments platelet function by increasing AA release and expression of α IIb β 3 receptors. 20 Multiple electrode impedance platelet aggregometry has also been validated 23,24 and used in dogs to assess platelet function in the setting of sepsis 14 and hemorrhagic shock. 25 When assessing platelet function, erythrocyte concentration must be taken into consideration, as it affects platelet function by 1 or more mechanisms. Erythrocytes release ADP and thromboxane which augment platelet activation and aggregation and scavenge nitric oxide which is a platelet antagonist. 26 Anemia results in decrease in AA-and ADP-induced, but not COL-induced, platelet aggregation in dogs 25 and in people. 27,28 Therefore, it is unlikely that the lower PCV in the AKI group would have influenced the decrease in COL-induced platelet aggregation in this study. There is positive linear correlation between platelet count and platelet aggregation measured by MEPA. [29][30][31][32] Therefore, it is possible that a type-II error might have occurred with AA-and ADP-induced platelet aggregation due to the significantly higher platelet count in the AKI group in the current study. The proposed mechanism of AVWS in people is due to increased proteolytic degradation of vWF by specific proteases. 33,34 This proposed mechanism has been supported by findings of low vWF ristocetin (vWF:RCo) to vWF:Ag ratio, which is another method for detecting decreased large MW vWF in people, as well as a lack of large MW vWF on multimeric analysis in CKD. 35 However, factors other than vWF:Ag concentration or vWF multimeric pattern might play a role in bleeding tendencies in uremic people with CKD, as normal, increased, or decreased vWF:Ag, vWF:RCo, and multimeric pattern occurs with and without prolonged bleeding times. [35][36][37] 10,11 There are several limitations to our study. The AKI group had lower AA-and ADP-induced platelet aggregation compared to the control group which was not statistically significant. As this is a pilot study, the small sample size might have contributed to a type II error.
A sample size of 26 dogs in each group would be required to detect a difference in platelet aggregation between the 2 groups, should such a difference exist. The small sample size might have also contributed to a type II error in vWF:Ag concentrations, in which 115 dogs would be required to detect a difference between the 2 groups, should such a difference exist. There were many variables with strong correlation which did not reach significance, in which 24 dogs would be required in each group to reach significant difference if present. In addition, as already discussed, the difference in PCV and platelet count might have also influenced our results. If PCV and platelet count were standardized between the 2 groups, there might be decreased or increased difference in platelet aggregation respectively.
We aimed to characterize platelet function in dogs with AKI; however, we also included dogs with AoCKD. This inclusion of dogs with CKD might limit our ability to purely state that the findings are specific to AKI, as primary hemostatic function might have been contributed from mechanisms associated with CKD also. Although our study design excluded known causes of AKI, other unknown comorbidities might have influenced the results.
In conclusion, we found that dogs with AKI have abnormal primary hemostatic function, with decreased COL-induced MEPA and abnor-