Relationship between assays of inflammation and coagulation: A novel interpretation of the canine activated clotting time

Abstract

The processes of inflammation and coagulation are known to be interconnected through several mechanisms; however, the influence of inflammation on the interpretation of coagulation assays remains unknown. Blood was collected from 87 dogs admitted to a tertiary referral intensive care unit (ICU) and 15 control dogs. The association between 2 markers of inflammation [mature neutrophil count and C-reactive protein (CRP)] and 5 coagulation parameters [activated clotting time (ACT), prothrombin time (PT), activated partial thromboplastin time (aPTT), antithrombin (AT), and platelet count (plt)] were evaluated through correlation analysis. The study population was then divided into 4 groups based on severity of ACT prolongation with comparisons to all other variables assessed through an analysis of variance (ANOVA) test. A strong correlation for a biological system was demonstrated between ACT and CRP (r = 0.66; P < 0.0001). Statistically significant results were also found between aPTT and AT with the markers of inflammation, but the correlations were weaker. Within ACT groups of increasing severity, higher CRP concentrations (P < 0.0001) and lower AT activities (P < 0.0001) were identified. This study provides evidence for an association between assays of inflammation and coagulation and suggests that modification of our traditional interpretations of coagulation assays may be required. As a point-of-care test, ACT is a simple and inexpensive tool that can be used to assess an underlying inflammatory or hemostatic process.

Introduction

Veterinarians are frequently presented with patients along a wide spectrum of illness severity and duration. Some of these patients may manifest a systemic response to their underlying illness and fulfill the criteria defining the systemic inflammatory response syndrome (SIRS) (1,2). The close link between inflammation and coagulation is gaining increasing attention and the 2 systems are thought to intersect through 3 main processes: 1) coagulation activation, 2) down-regulation of natural anti-coagulants, and 3) inhibition of fibrinolysis (3–6). The outcome of the inflammatory response on the coagulation system is the formation of a pro-coagulant environment with potential clinical importance. Overly exuberant inflammation with ongoing activation of coagulation can result in widespread microvascular thrombosis, which may lead to microvascular failure if unimpeded. Regional microvascular thrombosis is thought to be a contributing factor to organ dysfunction (7).

Laboratory tests serve to provide key information on specific pathologic processes occurring in patients. However, the interpretation of diagnostic tests influences decision-making and subsequently, patient outcome. The initial database may identify a specific problem; however, the degree of an associated inflammatory response is not always clear in a clinical setting. Knowing the status of the inflammatory response may be valuable in formulating a treatment plan and monitoring the patient’s response to therapy, as unregulated inflammation is known to be detrimental. Given that the inflammatory and coagulation systems are closely integrated, it is possible that traditional tests of coagulation are influenced by inflammation and that the clinical interpretation of these assays should be modified.

The objective of this study was to determine if a relationship exists between coagulation and inflammatory parameters in dogs affected with an illness requiring admission to an intensive care unit (ICU). The activated clotting time (ACT) has traditionally been used as a simple, quick, and inexpensive test of coagulopathy; however, based on anecdotal observations of patients admitted to the Ontario Veterinary College (OVC) ICU, it may also potentially indicate an inflammatory process. It is these observations that form the basis for the current study. To the authors’ knowledge, there are no previous veterinary studies examining the influence of inflammation on routine laboratory and point-of-care coagulation assays.

Materials and methods

Selection criteria

This study was prospective and observational in design. All procedures were approved by the Animal Care Committee of the University of Guelph. The study population consisted of canine patients admitted to the OVC ICU. Inclusion criteria required that dogs weighed more than 10 kg and that the primary clinicians had requested tests that were used in the present study (see Diagnostic assays). Any dog with a known primary coagulopathy based on history and physical examination (such as vitamin K antagonist poisoning) was excluded. The control group consisted of staff- and clinician-owned pet animals, and purpose-bred dogs enrolled in an unrelated research trial prior to any intervention. Control dogs were deemed healthy based on normal history, physical examination, complete blood (cell) count (CBC), serum biochemistry, and coagulation profile.

Blood collection procedure and tests

Tests of inflammation included C-reactive protein (CRP) and the mature neutrophil count obtained from the CBC. Coagulation assays included prothrombin time (PT), activated partial thromboplastin time (aPTT), antithrombin activity (AT), platelet count (plt) obtained from the CBC, and ACT. Collected blood was processed as follows: 2 mL was placed immediately into a warmed ACT tube containing siliceous earth (BD Vacutainer Evacuated Glass Tube REF 366522; Becton, Dickinson and Company, Franklin Lakes, New Jersey, USA), 1.8 mL was placed into a citrated tube containing 3.2% sodium citrate, 1.8 mL was placed into an ethylenediamine tetra-acetic acid (EDTA) tube containing 3.6 mg of potassium EDTA, and 2 mL, or the remaining amount of blood, was placed into a plain tube for serum collection. Activated clotting time was completed by a trained ICU technician or either of 2 investigators (TC or KM), with all operators using the same protocol. The citrated tube was centrifuged and plasma aliquoted for determination of PT, aPTT [performed by the Animal Health Laboratory (AHL), University of Guelph, Ontario] and AT (performed by the Hemostasis Reference Laboratory, Hamilton, Ontario). Plasma for AT assays were batched and stored at −70°C until analysis. Blood in the plain tube was allowed to stand at room temperature for 10 min before centrifugation and aliquoting for serum biochemical profile and CRP determination. Five hundred mL of serum for CRP was stored at −70°C and batched for later analysis. The remaining serum and EDTA tube were submitted to the AHL for a biochemistry profile and CBC, respectively.

Diagnostic assays

ACT — The methodology for ACT determination is as previously published using the human axilla as a heat source (8). Whole blood (2 mL) is injected, through vacuum flow, into an ACT tube that has been pre-warmed in the samplers axilla. The timing begins immediately upon injection of blood into the tube. The tube is gently inverted to ensure mixing of blood with the contact activant (siliceous earth) and immediately following mixing, the tube is replaced within the axilla for 60 s. The tube is then inspected for visual clot formation at 60 s and every 10 s thereafter until clot formation is visualized with the tube being immediately replaced in the axilla between inspection times. The canine range of reference values for this protocol used at the OVC is 80–100 s.

PT — The patient plasma was incubated before addition of thromboplastin (Dade Thromboplastin C Plus; Dade Behring, Marburg, Germany) in buffer and time in seconds to clot formation was determined by using a Sigma Amelung KC4 coagulation instrument. The canine reference interval is 9–15 s.

aPTT — The patient plasma was incubated before addition of Alexin (AMAX Alexin; Trinity BioTech, St. Louis, Missouri, USA) in buffer, followed by the addition of calcium chloride. Time in seconds to clot formation was then determined by using a Sigma Amelung KC4 coagulation instrument. The canine reference interval is 15–23.5 s.

AT — A factor Xa-based functional assay (Electrachrome Antithrombin; Instrumentation Laboratory Company, Lexington, Massachusetts, USA) was performed. The diluted patient plasma was incubated with Factor Xa in a 96-well microplate. Chromogenic substrate (para-Nitroaniline) was added and the mixture incubated until stopping the reaction with 20% acetic acid. The plate was then read at 405 nm using an AMAX 190 Plus coagulation analyzer (Trinity BioTech). The residual quantity of Factor Xa is inversely proportional to the AT level of the test sample and is determined from a standard curve. The canine reference interval is 85–135%.

CRP — Serum samples for CRP were assayed using a commercially available canine specific solid-phase sandwich immunoassay according to manufacturer’s instructions (PHASE RANGE Canine C-reactive protein assay; Tridelta Development, Kildare, United Kingdom). The reported canine reference value is < 10 μg/mL.

Data analysis

Coagulation test results (ACT, PT, aPTT, AT, and plt) were correlated to markers of inflammation (mature neutrophil count and CRP concentrations). The patient population was then subdivided into groups based on severity of ACT prolongation. Four groups were generated: group 1 representing healthy control subjects (ACT 80–100 s); group 2, clinical patients with low to normal ACT (70–100 s); group 3, clinical patients with mild to moderately elevated ACT (105–140 s); and group 4, clinical patients with marked elevation in ACT (> 140 s). The coagulation parameters and markers of inflammation were then compared for significance between each ACT group.

Statistical analysis

Data were entered into a commercially available statistical package (SAS OnlineDoc 9.1.3; SAS Institute, Cary, North Carolina, USA). Distribution was improved for all parameters following logarithmic transformation as confirmed with a Shapiro-Wilk test on the residuals. A Pearson’s correlation analysis was used to determine significant association between ACT and CRP. Spearman’s correlation was used to determine significant association between the remaining nonparametric coagulation parameters (PT, aPTT, AT, plt) and markers of inflammation (neutrophil count and CRP). Correlations are reported as the correlation coefficient (r) with associated P-value. A Kruskall-Wallis rank sum test with Tukey adjustment for multiple comparisons was applied to identify significant differences among the ACT groups and all other measured coagulation and inflammatory variables. Normally distributed results are reported as mean ± standard deviation (s). Results that are skewed are reported as median (range). The level of significance was set at P < 0.05.

Results

Study population

A total of 102 dogs were included in this study; 87 clinical patients and 15 control subjects. The clinical patients represented 31 breeds and had a mean age of 7.0 ± 3.1 y and a mean weight of 41.6 ± 22.3 kg, with 41% (n = 36) being male castrated, 8% (n = 7) male intact, 42% (n = 37) female spayed, and 8% (n = 7) female intact dogs. The control subjects represented 6 breeds and had a mean age of 5.0 ± 5.5 years and a mean weight of 32.9 ± 27.5 kg, with 13% (n = 2) being male castrated, 47% (n = 7) male intact, 20% (n = 3) female spayed, and 20% (n = 3) female intact dogs. There was no difference between the ages (P = 0.22) and weights (P = 0.20) of the control and clinical dogs.

Correlation data

A summary of the correlation results is presented in Table I. Activated clotting time and aPTT were found to have significant positive correlation to both CRP and neutrophil count. Antithrombin demonstrated a significant negative correlation with both inflammatory markers. Although these correlations were all found to be significant, there is only a low to moderate degree of correlation among them. The exception is the strong correlation found between ACT and CRP (r = 0.66, P < 0.0001) (Figure 1) (9). Prothrombin time and plt count were not significantly correlated with either CRP or neutrophil count.

Table I.

Correlation coefficients (r) between markers of inflammation and coagulation variables of our study population

	CRP	Neutrophil count
ACT	0.66	0.39
P-value	< 0.0001	< 0.0001
n	90	92
PT	−0.12	0.12
P-value	0.2869	0.2506
n	85	87
aPTT	0.54	0.43
P-value	< 0.0001	< 0.0001
n	87	89
AT	−0.54	−0.43
P-value	< 0.0001	< 0.0001
n	85	88
plt	−0.19	0.09
P-value	0.0818	0.3836
n	81	91

ACT — activated clotting time; PT — prothrombin time; aPTT — activated partial thromboplastin time; AT — antithrombin; plt — platelet count; CRP — C-reactive protein.

Figure 1.

Scatter plot of the correlation between the natural log (ln) of the activated clotting time (ACT) and C-reactive protein (CRP) in the study population (n = 90).

ACT grouped data

Overall, ANOVA was significant between ACT groups and CRP, neutrophil count, AT, and aPTT (all P ≤ 0.0001) (Table II). Following Tukey adjustment for multiple comparisons, 3 comparisons had 1 ACT group that was not different from another ACT group (Figures 2 and 3). The first ACT grouping unable to demonstrate a difference was between the CRP concentration and ACT groups 2 and 3 (P = 0.11). Due to increased clinical interest between these groups, a direct comparison was completed using a Wilcoxon-Mann-Whitney test, which demonstrated a statistical difference in CRP concentration between ACT groups 2 and 3 (P = 0.04). The second ACT grouping that did not demonstrate a difference was found between groups 3 and 4 when AT activity (P = 0.35) was compared to ACT severity. No difference was demonstrated between aPTT and ACT groups 2 and 3 (P = 0.07). The neutrophil counts in clinically ill patients were significantly different compared with the control dogs; however, there were no differences among the groups based on ACT severity.

Table II.

Comparison of medians (range) between clinical ACT groups and statistically significant inflammatory and coagulation variables

	Group 1 (normal)	Group 2 (ACT 70–100 s)	Group 3 (ACT 105–140 s)	Group 4 (ACT > 140 s)	Reference values
CRP (μg/mL)	1.82 (0.93–4.41)	24.6a (0.85–151)	57.3a (3.2–395)	178a,b,c (71.53–389)	< 10
Neutrophil (× 109/L)	4.97 (3.2–9.8)	13.5a (3.1–26.8)	16.3a (0.02–90)	17.9a (0.06–47.9)	2.9–10.6
AT (%)	120 (100–156)	101a (60–138)	79.5a,b (31–135)	72a,b (42–106)	85–135
aPTT (s)	19 (13.9–22.8)	20.5a (16.6–30)	23.4a (16.4–49.5)	28.6a,b,c (21.1–34.3)	15–23.5

^aIndicates a significant difference from group 1.

^bIndicates a significant difference from group 2.

^cIndicates a significant difference from group 3.

Figure 2.

Bar graph representing the ACT group comparisons with CRP concentrations. The overall ANOVA is significant (P < 0.0001). The top number in each bar represents the median CRP concentration in each group. * denotes significance from group 1; † denotes significance from group 2; § denotes significance from group 3.

Figure 3.

Bar graph representing the ACT group comparisons with AT activity. The overall ANOVA is significant (P < 0.0001). The top number in each bar represents the median AT value in each group. * denotes significance from group 1; † denotes significance from group 2.

Discussion

The results from this study demonstrate an association between several coagulation parameters and inflammatory mediators. Dogs with naturally occurring disease requiring admission to a tertiary referral ICU were selected so as to exclude the limitations of an artificially created study model intended to mimic the in vivo intricacy of the interaction between the inflammatory and coagulation systems. Two markers of inflammation were evaluated; one that is routinely measured (neutrophil count) and another that is known to be a sensitive indicator of inflammation (CRP). Interpretation of a single CBC, including the neutrophil count, can be influenced by the simultaneous effects of glucocorticoids, catecholamines, neoplasia, or antibiotic therapy. Therefore, the neutrophil count as a marker of inflammation lacks specificity. On the other hand, CRP is one of several positive acute phase proteins (APP) that can increase dramatically in dogs with inflammation (10). C-reactive protein is elevated in dogs with infections (11–17), and with noninfectious inflammation (11,18–22). Currently, CRP determination is not widely available in veterinary medicine and primarily limited to the research setting.

In the present study, the strongest correlation was found between ACT and CRP (r = 0.66). Although a positive correlation between aPTT and inflammation (assessed by CRP) was also recognized, its association was weaker than that of ACT. The enhanced association noted between the coagulation parameters and CRP is supported by reports of the superiority of APP’s over the leukogram in distinguishing inflammation (23). Although from a coagulation perspective, ACT and aPTT both reflect dysfunction in the intrinsic or common coagulation pathways, the results of this study indicate that ACT is more closely correlated with inflammation in our particular patient population. Consequently, ACT may reflect and be influenced by inflammation. This observation and alteration of the interpretation of ACT has been previously reported in human patients during cardiopulmonary bypass (24). The stronger correlation found between ACT and CRP, as opposed to aPTT and CRP, may be due to the ability of the ACT to indicate a patient’s global state of coagulation (25). It is suspected that global tests of coagulation, like ACT, which rely not only on the importance of factor activity but also considers the necessity of platelets and cells bearing TF may be influenced by inflammation to a greater degree.

The negative correlation found between AT and inflammation in this study is in agreement with previous reports investigating effects of sepsis on AT (26). Antithrombin has also been shown to be a negative acute phase protein in both animal (27) and human studies (28). While AT may be decreased in sepsis, DIC, protein-losing renal disease, or gastrointestinal diseases (29); a potential decrease in AT should also be considered with any cause of inflammation without overt protein loss.

Our results demonstrate that with higher concentrations of CRP, which reflect an increasing severity of inflammation (11,19,30), the associated ACT values were prolonged, while functional AT concentrations were lower. These results are not initially intuitive as both inflammation and a low AT concentration predispose to a hypercoagulable condition. In humans, ACT has been reported to be sensitive enough to reflect hypercoagulation in patients with trauma (25) and patients undergoing cardiopulmonary bypass (31). Therefore, a potential explanation as to why ACT is prolonged with inflammation is to consider the nature of this assay. Like PT and aPTT, the ACT is a clot-based assays and the clot-based assays do not adequately assess the absolute capability of a patient to generate thrombin, as only < 5% of the total thrombin potential is required to convert fibrinogen to fibrin with subsequent formation of a clot (32). Clot-based assays are known to be prolonged in severe inflammatory conditions represented by sepsis (26,33). Therefore, the strong positive correlation between ACT and CRP supports that inflammation delays thrombin generation as reflected by the prolonged ACT and aPTT. Inflammation and coagulation both occur over a continual spectrum of illness severity and the precise moment at which inflammation-induced activation of coagulation advances to systemic hypocoagulation is unknown. Being a tertiary referral hospital, it is possible that our patients had progressed from an initial hypercoagulable state to hypocoagulation prior to their presentation at the OVC ICU.

Several important results were demonstrated when ACT was grouped according to severity. Median CRP concentrations within ACT groups were all significantly greater than control subjects. The clinical significance of this result can be demonstrated between groups 1 and 2. Both groups have patients with ACT values between 80–100 s, except for 1 patient in group 2 with a slightly shortened ACT of 70 s. This emphasizes the importance of continued monitoring of ACT in the global assessment of patients with an identified or vague illness. Although a significant difference was not demonstrated between the median values for CRP in patients of ACT groups 2 and 3, broad ACT groups may reflect the severity of CRP concentration. The results of this study are consistent with our anecdotal observations, and previous reports that ACT prolongation coincides with severity of illness (34).

Median AT activity was significantly different between ACT groups 1, 2, and 3. When assessed strictly from a coagulation perspective, the lower AT activity within higher ACT groups is counterintuitive. However, as previously discussed, AT is likely acting as a negative acute phase protein while ACT prolongs with inflammation.

A limitation of this study may be the small number of patients, making it possible that our dogs do not reflect the total canine population presenting to a veterinary ICU. However, when reviewing the clinical diagnoses of our study population, the authors are in agreement that the diversity of cases included in this study is likely representative of patients requiring ICU care. A second limitation to consider is the generation of an institution-dependent range of reference values for ACT. As ACT results are known to vary when different contact activants are used (35), each institution must establish reference values based on individual hospital ACT protocols as different heating methods may be employed. In the authors’ opinion, multi-reagent point-of-care ACT assays that require the addition of exogenous phospholipids, stabilizers, and buffers may not reflect the influence of inflammation as the results of this study were drawn from the use of the conventional tube ACT which singly contains siliceous earth as the contact activant to the patients’ blood.

In conclusion, this study demonstrates a direct correlation between coagulation parameters and markers of inflammation. The clinical consequence of these results is the alteration of our traditional views of ACT and AT. In canine patients, a persistently elevated ACT, without pharmacologic cause or overt coagulopathy, should raise suspicion of an underlying inflammatory process warranting further investigation. Antithrombin can be decreased with inflammation and should be considered a negative acute phase protein. Ultimately, novel interpretation of coagulation parameters may lead to earlier detection of diseases causing systemic dysfunction.