• Open Access

Cyclooxygenase Expression and Platelet Function in Healthy Dogs Receiving Low-Dose Aspirin


  • Abstract presented in part at the 2011 ACVIM Forum, Denver, CO

Corresponding author: J. Thomason, Department of Clinical Sciences, College of Veterinary Medicine, Mississippi State University, PO Box 6100, Mississippi State, MS 39762-6100; e-mail: thomason@cvm.msstate.edu



Low-dose aspirin is used to prevent thromboembolic complications in dogs, but some animals are nonresponsive to the antiplatelet effects of aspirin (“aspirin resistance”).


That low-dose aspirin would inhibit platelet function, decrease thromboxane synthesis, and alter platelet cyclooxygenase (COX) expression.


Twenty-four healthy dogs.


A repeated measures study. Platelet function (PFA-100 closure time, collagen/epinephrine), platelet COX-1 and COX-2 expression, and urine 11-dehydro-thromboxane B2 (11-dTXB2) were evaluated before and during aspirin administration (1 mg/kg Q24 hours PO, 10 days). Based on prolongation of closure times after aspirin administration, dogs were divided into categories according to aspirin responsiveness: responders, nonresponders, and inconsistent responders.


Low-dose aspirin increased closure times significantly (62% by Day 10, P < .001), with an equal distribution among aspirin responsiveness categories, 8 dogs per group. Platelet COX-1 mean fluorescent intensity (MFI) increased significantly during treatment, 13% on Day 3 (range, −29.7–136.1%) (P = .047) and 72% on Day 10 (range, −0.37–210%) (P < .001). Platelet COX-2 MFI increased significantly by 34% (range, −29.2–270%) on Day 3 (P = .003) and 74% (range, −19.7–226%) on Day 10 (P < .001). Urinary 11-dTXB2 concentrations significantly (P = .005, P < .001) decreased at both time points. There was no difference between aspirin responsiveness and either platelet COX expression or thromboxane production.

Conclusions and Clinical Importance

Low-dose aspirin consistently inhibits platelet function in approximately one-third of healthy dogs, despite decreased thromboxane synthesis and increased platelet COX expression in most dogs. COX isoform expression before treatment did not predict aspirin resistance.


11-dehydro-thromboxane B2


2,3-dinor-thromboxane B2


adenosine diphosphate


analysis of variance




fluorescence-activated cell sorting – phosphate buffered saline


fluorescein isothiocyanate


immunoglobulin G


immune mediated hemolytic anemia


mean fluorescence intensity


Platelet Function Analyzer-100




thromboxane A2

Thromboembolic complications in dogs are associated with multiple diseases, including immune-mediated hemolytic anemia (IMHA), protein-losing nephropathies, and hyperadrenocorticism.[1, 2] Antiplatelet treatment plays an important role in the prevention of thromboembolic complications. One of the most commonly used antiplatelet medications, aspirin, is highly effective at inhibiting platelet function in dogs and humans.[3-7] Anti-inflammatory doses of aspirin inhibit platelet function, decrease inflammation, and provide analgesia, but can cause important adverse effects.[8] Lower doses of aspirin, in contrast, maintain the antiplatelet properties of the medication without adverse effects, making low-dose aspirin an appealing treatment for the prevention of thromboembolic complications. The administration of low doses of aspirin has been associated with an improved survival rate in dogs with IMHA.[9]

Unfortunately, despite preventative treatment, some dogs treated with aspirin develop thromboembolic complications. A similar phenomenon occurs in humans receiving low doses of aspirin, indicating that these patients are poorly responsive to the antiplatelet effects of aspirin (“aspirin resistant”). The true incidence of aspirin resistance in humans is unknown, but estimates range from 8 to 45% of human patients.[3, 10-12] Because dogs appear to inconsistently respond to low-dose aspirin, it is possible that, similar to humans, some dogs are aspirin resistant. The incidence of aspirin resistance in dogs is, however, unknown. The exact mechanism for aspirin resistance in both humans and dogs is also unknown, but possible explanations include genetic and acquired differences in cyclooxygenase (COX) enzyme activity or thromboxane receptors, increased platelet reactivity to other platelet agonists, and alternative pathways of thromboxane generation.[10]

Once synthesized and released by activated platelets, thromboxane A2 (TXA2) triggers vasoconstriction, enhances platelet aggregation, and induces further platelet activation.[5, 13] The COX enzyme, which exists in 2 major isoforms, COX-1 and COX-2, is primarily responsible for platelet generation of thromboxane A2.[14, 15] Previously, circulating platelets were thought to contain only the COX-1 isoform,[16-18] but recent studies have identified the COX-2 isoform in both human and canine platelets.[15, 17, 19] The COX-1 isoform was originally believed to be the sole source of platelet TXA2 production,[16] but platelet COX-2 could provide an alternate source of TXA2 production, offering a possible mechanism for aspirin resistance.

Aspirin functions as a permanent inhibitor of the COX enzyme via acetylation of a serine residue.[10, 20, 21] Permanent inhibition of the COX enzyme prevents conversion of arachidonic acid to prostaglandin H2 and many subsequent prostaglandins, including TXA2 and prostacyclin. While platelet-derived TXA2 plays a critical role in vasoconstriction and promoting platelet aggregation, prostacyclin originating from the vascular endothelium has the opposite effect by causing vasodilation and inhibiting platelet aggregation. Low-dose aspirin treatment, unlike higher anti-inflammatory doses, will alter platelet function without permanently inhibiting COX function in the vascular endothelium, allowing prostacyclin production to continue and contribute to the prevention of thrombus formation.[5]

High aspirin doses consistently inhibit platelet function[6, 7, 19] and thromboxane production[19, 22] while concurrently altering platelet COX-1 and COX-2 expression in dogs.[19] Platelet COX enzyme expression in dogs receiving low-dose aspirin has not been previously evaluated, but variations in COX isoform expression could provide an explanation for aspirin resistance. A better understanding of the relationship between platelet COX isoform expression and response to low doses of aspirin could potentially allow clinicians to predict which individuals will demonstrate aspirin resistance.

The objective of this study was to evaluate the effects of administration of a low dose of aspirin (1 mg/kg/day)1,[8] on platelet function and COX expression in healthy dogs. The goals of our study were to determine whether low-dose aspirin treatment would consistently affect platelet function and thromboxane production, and if not, whether pre-existing COX isoform expression or subsequent changes in isoform expression were associated with aspirin resistance. Our hypotheses were that low-dose aspirin would inhibit platelet function, decrease platelet thrombox-ane synthesis, and alter platelet COX isoform expression.

Material and Methods

Study Population

Healthy client-owned adult dogs were recruited from students, staff, and faculty associated with the Mississippi State University College of Veterinary Medicine. Owner consent was obtained before initiation of the study. The dogs were not exposed to any medications or vaccines for at least 2 weeks before initiation of the study, and no nonaspirin medications were administered during the testing period. All drug dosing was based on the body weight obtained at the beginning of the study. Normal health status was established via physical examination, complete blood count, serum chemistry, urinalysis, heartworm testing, and platelet function analysis with a point-of-care platelet function analyzer, the PFA-1002 (collagen/ADP cartridge3). Dogs were excluded from the study if they were found to have abnormal physical examination findings, abnormal complete blood count or serum chemistry results, heartworm or tick-borne disease, or platelet abnormalities based on either an abnormal platelet count or prolonged PFA-100 (collagen/ADP) closure time. Animal use was approved by the Mississippi State University Institutional Animal Care and Use Committee.

Study Design

Aspirin4 was administered orally to each dog at a dose of 1 mg/kg once daily for 10 days. Based on individual body weight, the aspirin dose was compounded by the Mississippi State University College of Veterinary Medicine Pharmacy. Blood and urine samples were collected from all study participants for platelet function evaluation (PFA-100 collagen/epinephrine cartridge,5) measurement of platelet cyclooxygenase expression, and urinary thromboxane analysis on Day 0 (before initiating aspirin), and again on Days 3 and 10 during aspirin administration.

Platelet Function Analysis

The PFA-100 has previously been evaluated for use in dogs[6, 7, 23] as a commercial point-of-care platelet function analyzer. The PFA-100 is an in vitro platelet function analyzer that records the closure time, in seconds, needed to form a platelet plug after activation by platelet agonists. Blood was collected via jugular venipuncture with a 20-gauge needle directly into a 5-mL vacutainer tube containing 3.8% sodium citrate.6 Samples were well mixed, and 800 μL of whole blood was transferred into either a collagen/ADP cartridge (establishment of normal platelet function before the study) or collagen/epinephrine cartridge (main study) for analysis. All samples were analyzed within 4 hours of collection and were kept at room temperature until analysis. The cut-off time for the instrument is greater than 300 seconds, and results were considered to be prolonged if the closure time was greater than 300 seconds. All cartridges were stored at 4°C and warmed to room temperature before analysis.

Based on PFA-100 collagen/epinephrine closure times after aspirin administration, all dogs were divided into one of 3 groups: aspirin responders (prolonged closure times at Day 3 and 10), aspirin nonresponders (no prolongation of closure times at either time point), and inconsistent aspirin responders (prolonged closure times on either Day 3 or 10, but not both). The initial Day 0 blood samples for platelet function analysis (collagen/epinephrine cartridge) were discarded because of an unanticipated instrument malfunction. Blood samples were obtained 4 weeks after completion of administration of aspirin to represent the baseline analysis for the PFA-100 collagen/epinephrine cartridge (all dogs included in the study had a baseline closure time of less than 300 seconds).

Flow Cytometry

Blood was collected via jugular venipuncture with a 20-gauge needle directly into a glass vacutainer tube containing 3.8% sodium citrate. Sample preparation was initiated within 1 hour of collection and completed within 4 hours.

COX-1 labeling: Platelet COX-1 expression was quantitated by utilizing a previously described protocol.[19] Briefly, 5 μL of whole blood was combined with 45 μL of FACS-PBS and incubated with a FITC-conjugated mouse antiovine-COX-1 monoclonal antibody.7 For platelet identification, samples were incubated with a mouse antihuman CD9:RPE monoclonal antibody.8 Samples were fixed for 10 minutes at 4°C in the dark with 1% paraformaldehyde.9

COX-2 labeling: A previously reported protocol[19] was used to quantitate platelet COX-2 expression. Briefly, 5 μL of whole blood was added to 45 uL of FACS-PBS and fixed with 1% paraformaldehyde at 4°C in the dark. Samples were washed and then incubated in 0.3% Triton X-100.10 After another wash, samples were incubated with a FITC-conjugated mouse antihuman-COX-2 monoclonal antibody.11 Platelets were identified with a mouse anti-pig CD61-purified monoclonal antibody12 and goat antimouse IgG:RPE antibody.13

Samples were stored in the dark at 4°C before analysis. Isotype-matched monoclonal antibodies were used for both COX antibodies. Analysis was performed with a flow cytometer14 by CellQuest Pro software.15 Platelet populations were displayed on log forward-scatter versus log side-angle light scatter plots. Gates were adjusted to baseline platelet populations, and a total of 5,000 gated events were recorded for each labeling. Expression was quantitated by the intensity of antibody fluorescence and expressed as mean fluorescence intensity (MFI). A histogram was created with MFI on the x-axis and events on the y-axis.

Urinary 11-dehydro-thromboxane B2 Analysis

Urinary 11-dehydro-thomboxane B2 (11-dTXB2) was measured with a multiplex analyzer16 and a commercial competitive enzyme immunoassay kit17 that has been previously validated for use with dog urine.[24] Urine was collected either by free catch or cystocentesis and stored at −80°C until analysis. Before analysis, urine samples were thawed and the specific gravity was measured. The assay buffer was used to standardize each sample to a urine specific gravity (range, 1.002–1.014), which was the optimal working range for the analyzer. Each urine sample was analyzed in duplicate and reported in picograms per milliliter of urine. A correction factor was applied to account for the sample dilutions. A 96-well plate was prepared by adding 100 μL of the urine sample and 50 μL of both the 11-dTXB2 Phycoerythrin Tracer and the 11-dTXB2 XMAPR beads to the appropriate well. Each plate was incubated at room temperature in the dark and on an orbital shaker before analysis. The urine creatinine concentration of each sample was determined with a biochemistry analyzer,18 and the urine 11-dTXB2 concentration was normalized to create a urinary 11-dTXB2 to creatinine ratio.[24, 25]


A power analysis was performed before study initiation, based on results from a previous study evaluating the effects of nonsteroidal anti-inflammatory drugs on platelet function.[26] The results of this analysis indicated that a sample size of 27 dogs would be needed to achieve a power value of 0.80 for detection of aspirin-induced changes in platelet function. A single population, repeated measures design was used in this study. After visual assessment of Q-Q plots and histograms, COX-2 MFI and percent of platelet expression appeared to be normally distributed; however, COX-1 MFI and percent of platelet expression, 11-dTXB2, and PFA-100 (collagen/epinephrine) closure times were not considered to be normally distributed. Consequently, nonparametric methods for analysis of repeated measures were utilized[27, 28] for all results. For each outcome, the data were ranked and then analysis of variance type statistics were obtained through PROC MIXED (SAS)19 by the ANOVAF option and the MIVQUE0 estimation method for the covariance parameters and a REPEATED statement specifying an unstructured covariance structure. With the cut-off time for the PFA-100 being greater than 300 seconds, all closure times that were considered to be prolonged (greater than 300 seconds) were analyzed with a closure time value of 300 seconds. Sample time was included in the model as a fixed effect. Differences in least square means were used for multiple comparisons of the 3 time points for the analyses; the simulation option within SAS was used for adjustment of P-values for the pair-wise comparisons. After classification into response groups based on PFA-100 collagen/epinephrine closure times, the effect of responder classification was assessed by the addition of response group and the time by response group interaction into the COX-1, COX-2, and 11-dTXB2 models. A P-value of less than .05 was considered to be significant for all analyses.


Study Group Characteristics

A total of 25 dogs satisfied the inclusion criteria and participated in the study. There were 14 different breeds with Labrador Retrievers being the most commonly represented breed (n = 6). Other breeds included English Springer Spaniels (n = 2), Australian Shepherds (n = 2), Shar-Pei (n = 1), Pug (n = 1), Yorkshire Terrier (n = 1), Chesapeake Bay Retriever (n = 1), Chihuahua (n = 1), Boxer (n = 1), Border Collie (n = 1), German Shepherd (n = 1), and mixed breed dogs (n = 7). The study population consisted of 8 neutered males, 12 spayed females, and 5 intact females. The mean age was 3.7 years (range, 1–9 years old) and the mean body weight was 23.2 kg (range, 5.4–48.9 kg). One dog, a female intact Labrador Retriever, was excluded from the study after detection of a prolonged baseline PFA-100 collagen/epinephrine closure time. The remaining 24 dogs received aspirin for 10 days with no reported adverse effects.

Platelet Function Analysis

The median PFA-100 collagen/epinephrine closure time of all 24 dogs on Day 0 was 135 seconds (range, 72–248 seconds), whereas the median closure times of all dogs on Days 3 and 10 were 268 seconds (range, 91–300 seconds) and 260 seconds (range, 64–300 seconds), respectively, an increase of 70% from baseline values. When compared with Day 0, there was a significant increase in closure times at both Day 3 (P < .001) and Day 10 (P < .001). There was no difference in closure times between Days 3 and 10 (P = .991). There was an equal distribution among the 3 groups of aspirin responsiveness, with a total 8 dogs per group.

In the aspirin nonresponder group, median closure time on Day 0 was 113 seconds (range, 91–203), whereas the median closure times on Days 3 and 10 were 159 seconds (range, 91–235) and 174 seconds (range, 92–219), respectively. In the incomplete aspirin responder group, median closure time on Day 0 was 135 seconds (range, 72–155). Four dogs had closure times of greater than 300 seconds on Day 3, and the other 4 dogs had closure times of greater than 300 seconds on Day 10. In the aspirin responder group, the median closure time on Day 0 was 167 seconds (range, 90–248). Median closure times during aspirin administration could not be calculated in the incomplete aspirin responder and the aspirin responder groups because, at each time point, a number of dogs exceeded the machine cut-off closure time.

Platelet COX-1 and COX-2 Expression

Platelet expression of COX-1 and COX-2 was measured before and at all tested time points during aspirin administration. There was a significant increase in platelet COX-1 expression from Day 0 to Day 3 with a median increase of 13% (range, −29.7–136.1%) (P = .047) and from Day 0 to Day 10 with a median increase of 72% (range, −0.37–210.36%) (P < .001) (Fig 1). There was also a significant (P < .001) increase in COX-1 expression by 52% (range, −24.7–241.4%) from Day 3 to Day 10. There were no significant (P = .958) differences in COX-1 expression among the three groups of aspirin responsiveness. Before drug administration, the percentage of platelets that expressed COX-1 was 55% (range, 50.2−63%). During aspirin administration, there was no significant change in the percentage of platelets that expressed COX-1 on Day 3 compared with baseline; however, this percentage increased significantly (P < .001) to 65% (range, 57.2–77.8%) on Day 10. In addition, there was a significant (P < .001) difference in the percentage of platelets that expressed COX-1 from Day 3 to Day 10 (Fig 2). There were no significant (P = .600) differences in the percentage of platelets that expressed COX-1 among the three groups of aspirin responsiveness.

Figure 1.

Platelet COX-1 expression during treatment of 24 healthy dogs with low-dose aspirin (1 mg/kg PO every 24 hours for 10 days). There was a significant increase in COX-1 expression from Day 0 to Day 3 (P = .047) (*) and from Day 0 to Day 10 (P < .001) (**). There was also a significant (P < .001) increase between Day 3 and Day 10. The box and whiskers plot demonstrates the median (line), interquartile range (box), and total range (whiskers).

Figure 2.

Representative flow cytometry histograms demonstrating platelet COX-1 expression on Day 0 (black line), Day 3 (dark gray line), and Day 10 (light gray line).

Platelet COX-2 expression was significantly (P = .003, P < .001, respectively) increased from Day 0 to Day 3 by 34% (range, −29.2–270.4%) and from Day 0 to Day 10 by 74% (range, −19.7–226.2%) (Fig 3). COX-2 expression was also significantly (P = .014) increased between Days 3 and 10 by 21% (range, −27.3–117.2%). Similar to the COX-1 expression, there were no significant (P = .479) differences in COX-2 expression among the three groups of aspirin responsiveness. Before drug administration, the percentage of platelets that expressed COX-2 was 78% (range, 66.6–82.6%), whereas during aspirin administration, there was a significant (P < .002) increase in the percentage of platelets that expressed COX-2 on Day 3 (to 85%; range, 74.3–94%), but no significant (P = .896) difference from baseline on Day 10. There was a significant (P < .001) difference in the percentage of platelets that expressed COX-2 when Day 3 was compared with Day 10 (Fig 4). There were no significant (P = .213) differences in the percentage of platelets that expressed COX-2 among the 3 groups of aspirin responsiveness.

Figure 3.

Platelet COX-2 expression during treatment of 24 healthy dogs with low-dose aspirin. There was a significant increase in COX-2 expression from Day 0 to Day 3 (P = .003) (*) and from Day 0 to Day 10 (P < .001) (**). There was also a significant (P = .014) increase between Day 3 and Day 10. The box and whiskers plot demonstrates the median (line), interquartile range (box), and total range (whiskers).

Figure 4.

Representative flow cytometry histograms demonstrating platelet COX-2 expression on Day 0 (black line), Day 3 (dark gray line), and Day 10 (light gray line).

Urinary Thromboxane Concentration

The median urine 11-dTXB2 to creatinine ratio on Day 0 was 8.8 (range, 4.0–38.5), whereas the median ratios for Day 3 and Day 10 were 7.5 (range, 3.2–18.2) and 5.2 (range, 1.2–11.8), respectively. There was a 29% decrease in the urine 11-dTXB2 to creatinine ratio between Day 0 and Day 3, whereas there was a 56% decrease between Day 0 and Day 10 and 35% decrease between Day 3 and Day 10 (Fig 5). Significant decreases in the urinary 11-dTXB2 to creatinine ratio occurred at all time points, from Day 0 to Day 3 (P = .005), from Day 0 to Day 10 (P < .001), and from Day 3 to Day 10 (P = .014). There were no significant (P = .885) differences in the urine 11-dTXB2 to creatinine ratio among the 3 groups of aspirin responsiveness.

Figure 5.

Median urinary 11-dTXB2 to creatinine ratios during treatment of 24 healthy dogs with low-dose aspirin. There was a significant decrease in the urinary 11-dTXB2 to creatinine ratio from Day 0 to Day 3 (P = .005) (*) and from Day 0 to Day 10 (P < .001) (**). There was also a significant (P = .014) decrease between Day 3 and Day 10. The box and whiskers plot demonstrates the median (line), interquartile range (box), and total range (whiskers).


Our study revealed that low-dose aspirin, in contrast to administration of anti-inflammatory doses, does not inhibit platelet function in all dogs as measured by the PFA-100 collagen/epinephrine cartridge, indicating that aspirin resistance occurs in dogs as well as in humans. Chronic low-dose aspirin administration will consistently inhibit thromboxane synthesis in dogs. The inhibition of thromboxane by low-dose aspirin, however, does not appear to correlate with aspirin responsiveness as measured by the PFA-100. With no differences in urinary thromboxane concentrations among aspirin responders, nonresponders, and inconsistent responders, our study suggests that thromboxane synthesis that is independent of COX-1 activity, either through the COX-2 enzyme pathway or alternative pathways, is unlikely to be the mechanism for aspirin resistance in dogs.

One possible explanation for aspirin resistance in dogs could be variable responsiveness of platelets to thromboxane. Previous studies have suggested that about 70% of canine platelets are insensitive to thromboxane stimulation.[29-31] The thromboxane insensitivity observed in some canine platelets could be caused by impaired platelet thromboxane A2 receptor-linked G proteins.[31] If only 30% of dogs have platelets that are sensitive to the effects of thromboxane, this could explain why only one-third of the animals in our study responded to low-dose aspirin with a consistent increase in PFA-100 closure time. In those dogs that have platelets that are insensitive to thromboxane-induced activation, other agonists such as ADP or serotonin might play a more important role in platelet activation, and such dogs could therefore be relatively resistant to the inhibitory effects of low-dose aspirin. A recent study of humans suggests that resistance to aspirin could be caused by a COX independent pathway and that, in affected people, inhibition of platelet function is more readily achieved by an ADP antagonist.[12] If platelet activation pathways are comparable in dogs and humans, ADP receptor antagonists could provide more reliable inhibition of platelet function than low-dose aspirin, although additional studies are required to fully evaluate the effectiveness of this drug class in dogs.

We measured urine levels of the thromboxane metabolite 11-dehydro-thromboxane B2 to evaluate the effects of aspirin on thromboxane synthesis. Measurement of this metabolite has been used in previous studies as an indicator of aspirin-induced COX inhibition.[32-34] After synthesis by platelets, thromboxane A2 has an immediate physiologic effect, and is then metabolized to thromboxane B2 and further metabolized into 11-dehydro-thromboxane B2 and 2,3-dinor-thromboxane B2 (2,3-dinorTXB2) before being eliminated in the urine.[8, 34, 35] Concentrations of thromboxane metabolites in plasma or serum can be artifactually increased because of platelet activation during collection of blood and sample processing. The use of urine to measure thromboxane metabolites avoids the problem of in vitro platelet activation associated with blood collection, and is therefore considered to be a more reliable indicator of platelet thromboxane synthesis.[34, 35]

Although most dogs in this study had a persistent decrease in urine 11-dTXB2 to creatinine ratios after administration of low-dose aspirin, and every dog had a decreased ratio for at least 1 time point, a few dogs demonstrated an increased ratio at a single time point. Similar inconsistent inhibition of urine 11-dTXB2 levels has previously been reported in dogs in a study using the same low dose of aspirin.[8] This previous study also evaluated urinary 2,3-dinor-thromboxane B2 and demonstrated that, compared with 11-dTXB2, urine 2,3-dinorTXB2 levels were a more sensitive indicator of aspirin-induced thromboxane inhibition in dogs.[8] However, although the urine 11-dTXB2 assay used in our study is less sensitive for detecting aspirin-induced thromboxane inhibition compared with the urine 2,3-dinorTXB2 assay, our assay was still sufficiently sensitive to detect a significant decrease in urine thromboxane metabolites at both measured time points after aspirin administration.

While cyclooxygenase function was inhibited by low-dose aspirin as indicated by prolongation of PFA-100 collagen/epinephrine closure times and decreased urine 11-dTXB2 concentrations, platelet expression of both COX-1 and COX-2 progressively increased throughout the course of aspirin administration. The COX-1 enzyme isoform was originally thought to be strictly constitutively expressed, but recent studies have shown that COX-1 expression can be induced in certain tissues, including human megakaryoblasts during thrombocytopoiesis.[15, 36, 37] Similar to the results of this study, anti-inflammatory doses of aspirin also cause increased platelet COX-1 expression, decreased thromboxane synthesis, and inhibition in platelet function.[19] One possible explanation for increased platelet COX-1 expression could be a compensatory upregulation of production of the COX-1 isoform by megakaryocytes in response to a reduction in platelet thromboxane A2 production. Platelet COX-2 expression also increased during administration of low-dose aspirin, a finding that contrasts with a previous study in dogs treated with an anti-inflammatory dose of aspirin, in which platelet COX-2 expression decreased in most treated dogs.[19] Because COX-2 expression is typically induced by inflammatory mediators,[15, 38, 39] it is feasible that the anti-inflammatory effects of high doses of aspirin are sufficient to decrease COX-2 expression, whereas the low dose of aspirin used in our current study was insufficient to exert the same effect. In addition, platelet COX-2 expression could have been upregulated as a compensatory response to the inhibitory effects of aspirin on COX-1 function and subsequent thromboxane synthesis. Finally, the increase in platelet COX expression could have originated within the circulating platelets themselves, and could therefore be independent of effects on megakaryocytes. Platelets contain rough endoplasmic reticulum and polyribosomes, both of which allow platelets to biosynthesize protein from cytoplasmic mRNA. These intraplatelet organelles could provide a source of COX-1 and COX-2 mRNA that would enable platelets to increase cyclooxygenase expression.

Pretreatment patient-to-patient variations in platelet COX-1 and COX-2 expression have been proposed as 1 potential mechanism for variable aspirin responsiveness. Our current study, however, did not demonstrate a difference in pretreatment platelet COX-1 or COX-2 expression among the 3 categories of aspirin responsiveness. Our results suggest that differences in pretreatment platelet COX expression do not play a major role in aspirin resistance in dogs, and indicate that pre-existing COX expression cannot be used to predict aspirin responsiveness.

Platelet aggregometry has long been considered the gold standard methodology for the evaluation of platelet responsiveness to aspirin.[7, 40] Platelet aggregometry, however, requires specialized equipment and training, thus making it a less commonly utilized technology in a clinical setting. The PFA-100, in contrast, is an instrument that is practical for routine clinical use, and prolongation of the PFA-100 closure time as measured with the collagen/epinephrine cartridge has become well established as a sensitive indicator of aspirin-induced platelet dysfunction in both humans and dogs.[7] The PFA-100 closure times in this study revealed that only one-third of healthy dogs responded adequately and consistently to low-dose aspirin treatment. If our study is an accurate representation of the canine population, the administration of a 1 mg/kg per day dose of aspirin may not be adequate to create a consistent and reliable preventive treatment for thromboembolic complications in a majority of dogs. Anti-inflammatory doses of aspirin (10 mg/kg bid), in contrast, have been shown to consistently inhibit platelet function in dogs.[19] Commencement with low-dose aspirin at 1 mg/kg per day and subsequent utilization of periodic measurement of platelet function via the PFA-100 to individually dose escalate “to effect” could increase the population of aspirin responders.[41] The aspirin responders in our study had inhibition of platelet function by Day 3 of treatment, suggesting that platelet function analysis after 3 days of aspirin should be adequate to detect whether a patient will respond to the current dose of aspirin, or require a higher dose. However, as the aspirin dose increases, so does the chance of developing adverse effects. Furthermore, as aspirin dosage approaches anti-inflammatory doses, the likelihood of concurrent inhibition of prostacyclin production by the vascular endothelium increases, potentially negating the antiplatelet effects of aspirin.[42, 43]

There were several limitations to our study. Although our study population consisted of a variety of breeds and ages, we evaluated aspirin effects only in healthy dogs. Responsiveness to aspirin could vary in the presence of endogenous platelet activation, and a similar study in dogs with proinflammatory or prothrombotic diseases would be necessary to determine the incidence of aspirin resistance in sick patients. In addition, our study used no methods to ensure that all dogs received the appropriate daily dose. Although it is unlikely, a lack of owner compliance might explain why some dogs either did not respond or inconsistently responded to the aspirin dose. Because of unexpected dropout of several owners at the start of the study, and the exclusion of 1 dog because of a prolonged baseline PFA-100 collagen/epinephrine closure time, the study was completed with 24 dogs rather than the minimum of 27 dogs suggested by preceding power analysis. However, despite this slight reduction in the number of enrolled dogs, animal numbers were still sufficient to detect significant aspirin-induced changes in platelet function, platelet COX expression, and thromboxane production.

Another limitation of this study is the absence of Day 0 closure times for the PFA-100 collagen/epinephrine cartridge before aspirin administration. On Day 0, blood was collected from all enrolled animals for analysis, and then owners were sent home with the prescribed course of aspirin. Unfortunately, we then experienced an instrument malfunction, and, because most dogs had already received their 1st dose of aspirin, we could not repeat the testing. Instead, we elected to repeat “baseline” samples 4 weeks after completion of aspirin treatment, which is supported by a 14-day washout period being adequate for closure times to return to normal in dogs after administration of various nonsteroidal anti-inflammatory drugs.[26] Similarly, after administration of an anti-inflammatory dose of aspirin (10 mg/kg BID for 10 days) to normal dogs, collagen/epinephrine cartridge closure times returned to normal in all dogs within 14 days of discontinuing the medication (unpublished data).[19] Four weeks therefore should have been more than sufficient time after the completion of aspirin treatment to determine each dog's normal platelet function.

In humans, the PFA-100 collagen/epinephrine cartridge is typically utilized to detect aspirin-induced platelet dysfunction, and the collagen/ADP cartridge is recommended to evaluate for congenital and acquired platelet function defects.[44] In healthy dogs, collagen/epinephrine cartridges have been shown to generate a wider range of closure times, and in some studies,[45] a higher prevalence of results that exceed machine cut-off values when compared with collagen/ADP cartridges. This has prompted some authors to suggest that the collagen/epinephrine cartridge might not be useful in dogs.[45] However, in dogs, as in humans, administration of anti-inflammatory doses of aspirin will consistently inhibit platelet function based on prolongation of closure times when using collagen/epinephrine cartridges, but not when collagen/ADP cartridges are utilized.[7, 19] Platelet aggregometry studies have similarly confirmed that anti-inflammatory doses of aspirin also do not inhibit ADP-induced platelet aggregation in dogs.[46] Given that aspirin doses that are considerably greater than the dose of aspirin used in our study do not appear to inhibit ADP-induced platelet aggregation, we believe that it is highly unlikely that we would have been able to detect any inhibition of platelet function using collagen/ADP cartridges. Therefore, although collagen/epinephrine cartridge might provide greater variability in closure times in dogs compared with the collagen/ADP cartridge, we believe that in dogs, as in people, the collagen/epinephrine cartridge is a more sensitive indicator of aspirin-induced platelet dysfunction.

Determination of aspirin responsiveness in our study hinged on the results of a single test methodology, the PFA-100 using the collagen/epinephrine cartridge. Because platelet aggregometry using agonists such as epinephrine also appears to be a sensitive indicator of aspirin-induced platelet dysfunction, future studies would be strengthened by adding aggregometry to augment the results of PFA-100 analysis. Further studies would also be strengthened by analyzing baseline and posttreatment PFA-100 closure times in duplicate or triplicate to better account for inherent test variability. In our study, aspirin responsiveness was defined as a PFA-100 collagen/epinephrine closure time that exceeded the cut-off time for the instrument (greater than 300 seconds). This definition has previously been established in studies of aspirin resistance in human patients.20,[47] An alternative approach to defining aspirin responsiveness would have been to calculate the ratio between baseline and postaspirin closure times for each individual dog, and use an arbitrary change in ratio (for example, doubling) as evidence of treatment response. However, because the response after aspirin administration exceeded the instrument cut-off time in half of the samples analyzed, and we had no means of determining a quantitative closure time above this value, in our opinion using by default a posttreatment value of “300 seconds” for all results that exceeded instrument cut-off times would have led to inaccurate ratio calculations in a large number of the samples analyzed.

In our original study group of 25 dogs tested with the collagen/epinephrine cartridge, only 1 dog had a closure time of greater than 300 seconds in the absence of aspirin. Based on previous studies conducted by the authors,[19, 26] this finding is not unexpected, because in our experience, it is relatively uncommon for normal dogs to have collagen/epinephrine cartridge closure times that exceed the machine cut-off time of 300 seconds. Similar findings have been previously reported in other comparable studies in dogs.[48] Other researchers, however, have reported a higher rate of prolonged collagen/epinephrine cartridge closure times in normal dogs: in 1 study, for example, 5 of 29 dogs tested had closure times of over 300 seconds.[45] Variability in the proportion of normal dogs that exceed machine cut-off times may be breed-related. In 1 study of collagen/EPI cartridge closure times in 4 different dog breeds, closure times remained below cut-off times in 3 breeds, but exceeded 300 seconds in a few individual dogs of the 4th breed (Labrador Retrievers).[7] Interestingly, the 1 normal dog in our study that had a baseline closure time of over 300 seconds was also a Labrador Retriever, although another 5 dogs of the same breed had normal closure times. Within the population of normal dogs investigated in our study, because baseline prolongations of collagen/EPI cartridge closure times are uncommon, the dramatic increase in the rate of closure times above the machine cut-off time of 300 seconds after commencement of aspirin (24 of 48 tests performed exceeded 300 seconds, compared with only 1 in 25 in the absence of the drug) is undoubtedly attributable to the effects of the drug rather than normal individual-to-individual variations in test results.

Our study demonstrates that low-dose aspirin inhibits platelet thromboxane synthesis and increases platelet expression of both COX-1 and COX-2, whereas inhibition of platelet function is less consistently observed in individual dogs. Platelet COX expression before and during treatment was not related to aspirin responsiveness. Additional research is needed to better define the mechanism of aspirin resistance in dogs.


This work was funded by the Mississippi State University College of Veterinary Medicine Salsbury Research Grant, the WH Lindley Fund for Excellence, and the Dr Hugh G. Ward Endowment.

The authors thank Jodi Griswold and Angela Briggs for their assistance.

Conflict of Interest: Authors disclose no conflict of interest.


  1. 1

    Shearer L, Kruth SA, Wood D. Effects of aspirin and clopidogrel on platelet function in healthy dogs. J Vet Intern Med 2009; 23: 745 (abstract)

  2. 2

    PFA-100, Siemens Healthcare Diagnostics, Deerfield, IL

  3. 3

    PFA Collagen/ADP Test Cartridge, Siemens Healthcare Diagnostics, Duluth, GA

  4. 4

    Aspirin, Major Pharmaceuticals, Livonia, MI

  5. 5

    PFA Collagen/Epinephrine Test Cartridge, Siemens Healthcare Diagnostics

  6. 6

    3.8% sodium citrate, Vacutainer tube, Becton Dickinson, Franklin Lakes, NJ

  7. 7

    FITC-conjugated monoclonal COX-1, Clone CX111, Cayman Chemical Co, Ann Arbor, MI

  8. 8

    Monoclonal anti-human CD9:RPE, Clone MM2/57, AbD Serotec, Raleigh, NC

  9. 9

    Paraformaldehyde, Biolegend Inc., San Diego, CA

  10. 10

    Triton X-100, Sigma-Aldrich, St. Louis, MO

  11. 11

    FITC-conjugated monoclonal COX-2, Clone CX299, Cayman Chemical Co

  12. 12

    Monoclonal CD61-purified, Clone JM2E5, Accurate Chemical, Westbury, NY

  13. 13

    Goat anti-mouse IgG:RPE, AbD Serotec, Raleigh, NC

  14. 14

    FACSCalibur, Becton Dickinson, San Jose, CA

  15. 15

    CellQuest software, Becton Dickinson

  16. 16

    Luminex 200 System xMAP Technology, Luminex Corporation, Austin, TX

  17. 17

    Luminex 11-dehydro Thromboxane B2 Kit, Cayman Chemical Co

  18. 18

    ACE Alera Clinical Chemistry System, Alfa Wasserman, Inc, West Caldwell, NJ

  19. 19

    SAS for Windows version 9.2, SAS Institute, Inc, Cary, NC

  20. 20

    Leone M, Mezzasoma, A, Davi G, Gresele P. Detection of a critical degree of platelet TxA2-production inhibition by the PFA100 in aspirin-treated subjects: Possible relevance for the monitoring of aspirin resistance. J Thromb Haemost 2005;3:P0964 (abstract)