Point-of-care and traditional erythrocyte sedimentation rate, point-of-care rheometry, and cell-free DNA concentration in dogs with or without systemic inflammation

Abstract

RBC aggregation and deformability characteristics are altered by inflammatory, microcirculatory, and hemorheologic disease. These changes can be indirectly evaluated using the erythrocyte sedimentation rate (ESR). Newer point-of-care devices employ syllectometry to evaluate RBC rheology, which can give information beyond the ESR. We evaluated 2 point-of-care rheometers (iSED and MIZAR; Alcor Scientific) in 52 dogs presented to a university teaching hospital. Whole blood samples were analyzed for correlation between the ESR using the Westergren (ESRw) method (measured at 1 h and 24 h) and the predicted ESR using iSED. Plasma fibrinogen and cell-free DNA concentrations were also measured as probable markers of inflammation. The iSED-predicted ESR was positively correlated to the ESRw method at 1 h (r = 0.74; p < 0.001) and 24 h (r = 0.62; p < 0.001). Comparing dogs with or without inflammation (defined as plasma fibrinogen concentration >3.5 g/L [350 mg/dL]), significant differences were seen in the MIZAR parameters of base point, amplitude, integral, and half-time. Median cell-free DNA concentrations were higher in the group of dogs with inflammation (117 [range: 51–266] ng/mL vs. 82.7 [range: 19–206] ng/mL; p = 0.024). The iSED-predicted ESR is a good predictor of the ESRw and was obtained more rapidly. Rheometric parameters measured by MIZAR may be useful in detecting inflammation and monitoring secondary morphologic and functional changes in canine RBCs.

RBC aggregation, defined as the reversible clumping of RBCs, usually through weak interactions with positively charged plasma proteins, can occur at low flow rates, effectively increasing blood viscosity (this is unrelated to RBC aggregation mediated by strong antigen–antibody interactions, typically happening in immune-mediated hemolytic anemia, which will not be discussed here). RBCs are highly deformable in health; they can reversibly change shape in response to external forces, allowing for unobstructed transit through the microcirculation. ¹ Blood rheology is the study of the biophysical properties that permit RBC flow, and RBC deformability is a key property that allows laminar blood flow and decreased blood viscosity at high flow rates, which is especially important in the microcirculation. ⁷ Pathologic conditions, such as inflammation, microcirculatory failure, and hemorheologic disease, are associated with increased RBC aggregation and decreased RBC deformability, resulting in decreased microcirculatory flow and tissue oxygen delivery.^8,10 In humans with sepsis, RBCs are less deformable and more likely to aggregate compared to RBCs in healthy individuals.^10,11,35 The measurement of RBC rheology may be valuable in the diagnosis, monitoring, or prognostication of critical illness in veterinary species.^10,3,6

The RBC membrane contains proteins and glycoproteins embedded in a lipid bilayer. In health, sialylated glycoproteins on the RBC surface result in a net negative charge, causing RBCs to repel one another. ¹⁴ In inflammatory states, increased concentrations of plasma proteins, such as IgM, alpha-2 macroglobulin, and fibrinogen, coat the RBC membrane, neutralizing its negative surface charge. ³⁶ When the aggregation force is greater than the force of repulsion between cells, RBCs aggregate, and subsequently, sediment. ¹⁴ The tendency for RBC aggregation in the presence of inflammation is the basis for the erythrocyte sedimentation rate (ESR) test. ¹ The ESR test measures the degree to which anticoagulated RBCs sediment in a narrow vertical tube over the course of an hour, reported in mm/h. An accelerated ESR reflects increased RBC aggregation caused by increased concentrations of acute-phase proteins. ³⁶ Although ESR is nonspecific in the identification of any particular disease, it is a monitoring tool in inflammatory, neoplastic, or immunomodulatory disorders and remains one of the most widely used blood tests in human medicine. ¹

The most common ESR methodology is the Westergren (ESRw) method, which places a citrated and EDTA-anticoagulated whole blood sample in a tube with a 2.5-mm internal diameter graduated from 0 to 200 mm. ²⁰ The tube is placed upright and left undisturbed for an hour; the RBCs in the sample sediment in response to gravity and interactions with plasma proteins. The ESRw method was adopted as a gold standard by the International Council for Standardization in Haemotology (ICSH) and the Clinical Laboratory Standards Institute (CLSI) given its reliable, reproducible, and sensitive results. ²⁰

The ESR is not measured frequently in canine patients, although sporadic reports exist.^16,18,30,33 Lack of canine data is likely partially because of the lack of an established gold standard test in veterinary medicine, as well as the ready availability of other blood tests to indicate inflammation (e.g., plasma fibrinogen and C-reactive protein [CRP] concentrations).^4,31 Elevated levels of cell-free DNA (cfDNA) have been correlated with inflammation, given that cfDNA is associated with DNA release from increased neutrophil extracellular trapping. ¹⁵ Serial analysis of cfDNA may have prognostic value in critically ill animals. ²⁶ Although cfDNA may be a possible biomarker of inflammation, at least 2 methodologies for sample preparation have been described in the veterinary literature,^10,17,24 one being more straightforward to perform than the other (single- vs. double-centrifugation methods). It is unclear if both methods provide similar assessment of cfDNA or if one is superior.

The measurement of ESR in dogs may not be performed frequently because of some practical considerations. Canine RBCs have a smaller mean corpuscular volume (MCV) than human RBCs, conferring an increased RBC surface area:volume ratio, and, subsequently, an altered ESR.^22,28 This difference raises concerns about the assumption that the ESRw will be an appropriate clinical test in dogs. Interpretation is further challenged by the various methodologies available for ESR measurement. ²⁰ The only available report of a RI using the ESRw method in canine blood gave a range of 0–5 mm/h ³⁰ ; 3 other veterinary studies evaluated the ESR using different methods.^12,18,33

In addition to the ESR test, syllectometry is a method available to assess RBC aggregation and additional rheologic parameters. In syllectometry, light is incident on a layer of whole blood before and after the application of shearing stress, and the resulting change in light intensity is recorded (Figs. 1, 2). Light transmittance data over time is represented graphically as a syllectogram (Fig. 3), which is characterized by different chronologic stages: application of shear stress, RBC shape recovery and redistribution, and aggregation. ⁹

Figures 1, 2.

Light transmission optical flow cell technology in iSED and MIZAR instruments. Figure 1. Basic schematic of assembly. Figure 2. Schematic of the reading cell subassembly. Figures courtesy of Francesco Frappa, Alcor Scientific.

Figure 3.

Syllectogram (light transmittance vs. time) generated by a MIZAR instrument. AI = aggregation index; Amp/2 = half of amplitude; Bp = base point; ED = end transmission; Amp = amplitude; INT = integral; OT = optical transmission; OTF = optical transmission flow; T₀ = time of zero amplitude; T_1/2 = time to reach ½ the amplitude. Figure courtesy of Francesco Frappa, Alcor Scientific.

The iSED (Alcor Scientific) is a point-of-care (POC) instrument that evaluates RBC aggregation using syllectometry to assess rheometric properties of RBCs. iSED uses quantimetric photometry to identify rouleaux formation, which is an early phase of RBC sedimentation. ⁴ This permits the iSED to generate a prediction of the ESR in whole blood samples. In a prospective study evaluating human blood at 2 test sites, ESR predicted by the iSED exhibited a strong correlation with the ESRw method (0.862 and 0.916 for sites 1 and 2, respectively). ²³

The MIZAR analyzer (Alcor Scientific), a related POC instrument, gives a more in-depth analysis of the effects of shear stress on RBC aggregation and deformability in anticoagulated whole blood samples using syllectometry. Rather than a single predictive output, the MIZAR provides objective data for the entire syllectogram. As shear stress is applied, optical transmission during flow (OTF) is recorded, denoting sample transport into the reading cell. Minimum optical transmission (OT) is recorded after shear stress cessation. Delta OTF-OT (∆OTF-OT) defines the optical variation from shear stress application and stasis. Base point (BP), a calculated time based on the first and last minimum OT measured, is the time necessary to complete RBC redistribution and shape recovery; BP low (BPL) and BP high (BPH) represent the time at which the first OT and last OT values are detected. End transmission (ED) is the final value of OT recorded. Information collected from the integration of the syllectogram during the aggregation stage (integral [I], amplitude [AMP]) allows for the calculation of the aggregation index (AI), which reflects the speed of sample aggregation. The time to reach one-half of AMP is noted is the half time (T_1/2). The parameters measured by the MIZAR analyzer are AI, AMP, BP, BPH, BPL, ED, I, OT, OTF, ∆OTF-OT, and T_1/2.

Both instruments, iSED and MIZAR, generate results within 3 min, but neither instrument has been evaluated or validated for use in dogs. We hypothesized that 1) the iSED-predicted ESR in canine whole blood would be positively correlated to the ESR measured by the ESRw method; 2) the ESRs measured by the iSED and ESRw method would be significantly different between dogs with and without inflammation; 3) MIZAR parameters for dogs without inflammation would be significantly different among dogs with inflammation, dogs with inflammation caused by sepsis, and dogs with postoperative inflammation; and 4) a cfDNA preparation utilizing 2 sequential centrifugation steps would result in a lower measured cfDNA concentration, but would not impact the clinical value of the data.

Materials and methods

Study design

We conducted a prospective observational study of dogs hospitalized in the University of Georgia Veterinary Teaching Hospital (Athens, GA, USA) from December 2020 to May 2021. Our study was approved by the University of Georgia Clinical Research Committee (CRC-664). Informed owner consent was obtained before sample collection.

Venous whole blood samples were obtained from subjects and aliquoted into EDTA and sodium citrate collection tubes. An aliquot from the EDTA sample tube was used to perform the ESRw, as described below. The remaining EDTA sample was evaluated using the automated iSED and MIZAR analyzers. The citrated sample was processed into plasma for cfDNA and fibrinogen analysis, as described below. Blood samples were kept on a tube rocker pending analysis with the POC machines. All assays requiring fresh samples were initiated within 90 min of collection.

Venous whole blood samples were collected via direct venipuncture or from a central venous catheter following removal of a pre-sample of at least 5 times the prime volume of the catheter lumen. At least 4.7 mL of blood were collected from each subject and divided into aliquots that were anticoagulated with EDTA for ESRw and POC analyzer testing (2 mL), or 3.2% sodium citrate (2.7 mL).

Experimental animals

Recruited dogs were hospital patients and apparently healthy animals (based on thorough history and physical examination) owned by veterinary hospital staff. Dogs were divided into 2 main experimental groups: an inflammatory (INF) group and a non-inflammatory (NI) group. Groups were stratified post hoc, with the NI group having a plasma fibrinogen <3.5 g/L (350 mg/dL). Nine of the dogs were presented for elective orthopedic procedures and were sampled both preoperatively and 1 d postoperatively, accounting for 18 blood samples. No attempt was made to exclude dogs based on age, sex, or breed. Dogs were excluded if they weighed <10 kg, if owner consent was not obtained, or if the dog’s condition or temperament precluded sampling. Primary disease process was recorded. Dogs were classified as septic if they had a confirmed septic focus and evidence of end-organ dysfunction. ¹³ When available, WBC count, PCV, mean corpuscular volume (MCV), and mean corpuscular hemoglobin concentration (MCHC) were recorded. PCV was elected over hematocrit because not all dogs had a full CBC available. Samples were excluded if a fibrinogen concentration was not obtained.

Analytical methods

The ESRw protocol followed established guidelines of the ICSH and the CLSI. ²⁰ EDTA-anticoagulated whole blood (800 µL) was added to a 1-mL polypropylene tube containing 200 μL of 3.2% sodium citrate solution (Polymedco). Immediately before testing, the tube was inverted 8 times. A plastic pipette with a 2.5-mm internal diameter graduated from 0 to 200 mm (Polymedco) was inserted into the top of the polypropylene tube and placed upright in a designated balanced pipette rack, and allowed to sit at room temperature (24–25°C). Sample readings were initially taken only at the 60-min interval (results labeled ESRw1h), but an additional reading was taken after 24-h (labeled ESRw24h) after several samples did not result in measurable ESR within an hour.

The ESR was predicted by the iSED analyzer, and rheometric parameters were measured using the MIZAR analyzer. Both instruments were used following the manufacturer’s instructions. Bi-level quality control kits (Alcor Scientific) were evaluated daily on both analyzers and verified to meet the manufacturer’s standards before the analysis of any experimental samples. EDTA-anticoagulated samples were analyzed using the iSED and MIZAR without specific order for the initial machine, and then analyzed immediately using the other machine.

The protocol for handling and processing of samples intended for cfDNA extraction and measurement followed established reports, which varied in the specification of centrifugation rates.^12,19,26 Consequently, both a single and a double spin were evaluated on the samples to compare the most widely published techniques. Sodium citrate tubes were centrifuged at 2,500g for 20 min without braking (Sorvall Legend X1R; Thermo Scientific). Without disturbing the buffy coat, plasma was removed and pipetted into polypropylene tubes, from which a 20-µL sample was removed for cfDNA analysis. This “single spin” sample was designated as cfDNA_a. The remaining plasma was then centrifuged again at 16,000g for 10 min (Eppendorf 5415C; Brinkmann) and a second 20-µL aliquot removed for cfDNA analysis. This “double spin” sample was designated as cfDNA_b. The remainder of the plasma was stored frozen at –80°C until batched measurement of clottable fibrinogen concentration at a commercial laboratory (Cornell Animal Health Diagnostic Center, Ithaca, NY, USA).

Cell-free DNA analysis was performed (Qubit 3.0 fluorometer; Thermo Fisher). Control and calibration were performed per the manufacturer’s recommendations. The centrifuged plasma sample was mixed with 180 µL of working solution (Qubit dsDNA [high sensitivity] assay kit; Thermo Fisher) and then analyzed. Three readings were taken for each sample. Each reading was verified to be within 10% of the other two; if outliers were identified, they were removed, and another reading was obtained. For analysis, an average of the 3 values was utilized.

Statistical analysis was performed using SigmaStat v.14.0 (Systat), and data were graphed using Prism v.8.0 (GraphPad). Normality of the data was assessed using the Shapiro–Wilk test. Variance was assessed with the Brown–Forsythe method. Descriptive statistics were presented as median and range, given that not all datasets were normally distributed. Pearson correlations were performed to compare ESRw and iSED-predicted ESR measurements, as well as to compare both of those values to plasma fibrinogen and cfDNA concentrations. The results of analysis between NI and INF groups were compared using unpaired t-tests (AI, AMP, BP, BPH, cfDNA_b, MCV, MCHC, T_1/2,) or the Mann–Whitney rank sum test (BPL, cfDNA_a, ED, ESRw1h, ESRw24h, fibrinogen, I, iSED, OT, OTF, ∆OTF-OT, PCV, WBC count) depending on the normality and variance of the data. In addition, receiver operator characteristic (ROC) curves were prepared using ESR and iSED measurements, WBC count, and MIZAR analysis results (AI, AMP, BP, BPH, BPL, ED, I, OT, OTF, ∆OTF-OT, T_1/2) to identify any parameter that would distinguish between NI and INF groups with high sensitivity and specificity. Parameters for septic dogs (AI, AMP, BP, BPH, BPL, cfDNA_a, cfDNA_b, ED, ESRw1h, ESRw24h, fibrinogen, I, iSED, OT, OTF, ∆OTF-OT, T_1/2) were compared to other dogs with inflammation using 2-tailed independent t-tests. Separately, pre- and postoperative results (AI, AMP, BP, BPH, BPL, cfDNA_a, cfDNA_b, ED, ESRw1h, ESRw24h, fibrinogen, I, iSED, OT, OTF, ∆OTF-OT, T_1/2) from the same dog were compared using paired t-tests. These samples were also included as individual data points in the unpaired analysis. For all statistical analyses, significance was set at p ≤ 0.05.

Results

We collected 61 samples from 52 different dogs. Seven other dogs were excluded because of missing fibrinogen measurements. Breeds included mixed breed (13), Labrador Retriever (13), Pitbull (6), German Shepherd (4), Boxer (4), Golden Retriever (2), Siberian Husky (2), and 1 each of the following: German Shorthair Pointer, Great Dane, Doberman Pinscher, Bassett Hound, Gordon Setter, Airedale Terrier, Australian Shepherd, and Border Collie. The median age was 5.5 y (range: 6 mo to 12 y). Thirty-one dogs were male (26 castrated, 5 intact) and 21 were female (20 spayed, 1 intact).

All 61 samples from the 52 included dogs had measurements recorded for plasma fibrinogen concentration, ESRw1h, and MIZAR rheometric parameters. Thirty-one dogs had a plasma fibrinogen concentration <3.50 g/L (350 mg/dL), comprising the NI group. Twenty-one dogs were categorized as INF (plasma fibrinogen > 3.50 g/L [350 mg/dL]). Three dogs (2 NI, 1 INF) did not have an ESRw24h value recorded. One dog did not have an iSED value recorded because of inadequate sample volume, and 1 dog did not have cfDNA values recorded as a result of user error. Thirteen NI dogs and 6 INF dogs did not have WBC counts recorded. Seven dogs had a septic focus confirmed by bacterial culture, by fungal culture, or by the presence of intracellular bacteria on abdominal effusion cytology (Table 1). Nine dogs had pre- and postoperative samples collected before and after an elective orthopedic procedure.

Table 1.

Primary disease processes affecting NI (non-inflammatory) and INF (inflammatory) dogs determined by serum fibrinogen concentration.

Non-inflammatory	Inflammatory
Apparently healthy animal (18)	EHBO and cholecystectomy (1)
Heartworm positive (1)	Disseminated aspergillosis (1)
Orthopedic disease	Splenic HSA and pulmonary carcinoma (1)
• Tarsal bone luxation (1)	Multicentric LSA (1)
• Pre-op TPLO (7)	Septic peritonitis (1)
• Post-op TPLO (2)	Pyothorax (1)
• Cranial cruciate rupture (2)	Severe pancreatitis and abscessation (1)
• Osteoarthritis (1)	Orthopedic disease
Cancer	• Post-op TPLO (6)
• Hepatosplenic LSA (1)	• Pre-op TPLO (1)
• Multicentric LSA (3)	• Post-op tarsal arthrodesis (1)
• Osteosarcoma (1)	Severe ITP (1)
• Chemodectoma (1)	Mediastinal mass and aspiration pneumonia (1)
• Malignant melanoma (1)	Mediastinal mass and severe pyrexia (1)
	Fracture and incisional dehiscence (1)
	Open femoral fracture (1)
	Retrobulbar abscess (1)

EHBO = extra-hepatic biliary obstruction; HSA = hemangiosarcoma; ITP = immune-mediated thrombocytopenia; LSA = lymphoma; TPLO = tibial plateau leveling osteotomy.

Pearson product moment correlations between the ESRw and the iSED-predicted ESR generated correlation coefficients were 0.740 for ESRw1h (p < 0.0001) and 0.621 for ESRw24h (p < 0001). The correlation coefficient between the measured fibrinogen concentration and the ESRw1h was 0.534 (p < 0.0001) and to the ESRw24h was 0.602 (p < 0.001); the correlation coefficient for the fibrinogen concentration to the iSED ESR was 0.710 (p < 0.0001).

Inflamed and non-inflamed dogs

The ESRw measured at 1 and 24 h was greater in INF dogs than in NI dogs (ESRw1h, p < 0.001; ESRw24h, p < 0.001). The iSED-predicted ESR was greater in INF dogs than in NI dogs (p < 0.001). INF dogs had a significantly higher fibrinogen than NI dogs (p < 0.001; Table 2).

Table 2.

Comparison of blood samples from dogs classified as non-inflammatory (NI, n = 39) or inflammatory (INF, n = 22) based on serum fibrinogen concentration. Median (range) for ESR testing, serum fibrinogen, PCV, CBC values, MIZAR values, and cfDNA were compared using 2-tailed independent t-tests or Mann–Whitney rank sum tests.

	NI dogs	INF dogs	p
AI	0.539 (0.433–0.670)	0.604 (0.530–0.684)	<0.0001*
AMP	38.0 (15.0–68.0)	72.8 (33.0–136)	<0.0001*
BP, ms	313 (0–678)	191 (0–310)	<0.001*
BPH, ms	381 (1.00–1,010)	191 (1.00–392)	<0.0001*
BPL, ms	142 (0–678)	152 (0–279)	0.364
cfDNAa, ng/mL	82.7 (19.0–206)	117 (50.7–266)	0.021*
cfDNAb, ng/mL	70.6 (12.7–182)	104 (5.98–206)	0.024*
ED	1,580 (1540–1,790)	1,660 (1,590–1,760)	<0.0001*
ESRw1hr, mm/h	0 (0–7.00)	45.0 (0–58.0)	<0.0001*
ESRw24hr, mm/h	10.0 (0–43.0)	54.0 (5.00–130)	<0.0001*
Fibrinogen, g/L	1.70 (0.30–3.25)	5.15 (3.46–8.50)	<0.001*
I	6,700 (12,800)	1,530 (6,310–29,800)	<0.0001*
iSED ESR, mm/h	3.0 (0–15.0)	26.0 (1.00–118)	<0.0001*
MCHC, g/L	345 (320–365)	343 (318–361)	0.408
MCV, fL	72 (63–78)	68 (60–73)	0.006*
OT	1,550 (1,510–1,750)	1,580 (1,550–1,680)	<0.0001*
OTF	1,560 (1,520–1,770)	1,600 (1,560–1,680)	<0.0001*
∆OTF-OT	12.0 (–6.00 to 27.0)	16.6 (2.00–52.0)	0.0474*
PCV, L/L	0.46 (0.23–0.59)	0.44 (0.24–0.54)	0.147
T1/2, s	2.34 (0.95–3.02)	2.14 (1.32–2.76)	0.002*
WBC, ×109/L	8.6 (4.8–22.1)	12.9 (3.7–29.2)	0.013*

AI = aggregation index; AMP = amplitude; BP = base point; BPH = time from pumping media stop on which the last OT value is detected; BPL = time from pumping media stop on which the first OT value is detected; CfDNA_a = cell-free DNA obtained with a single round of centrifugation; CfDNA_b = cell-free DNA obtained with 2 centrifugation steps; ED = end transmission; ESR = erythrocyte sedimentation rate; ESRw = Westergren ESR; ESRw1h = ESRw at 1 h; ESRw24h = ESRw at 24 h; I = integral; iSED ESR = point-of-care erythrocyte sedimentation rate; MCHC = mean corpuscular hemoglobin concentration; MCV = mean corpuscular volume; MIZAR = point-of-care rheometer; OT = optical transmission; OTF = optical transmission flow; ∆OTF-OT = difference between OTF and OT; PCV = packed red blood cell volume; T_1/2 = time to reach ½ the amplitude; WBC = white blood cell count.

^* p ≤ 0.05.

PCV was not significantly different between INF dogs and NI dogs (p = 0.147; Table 2). However, the MCV was lower in INF dogs than NI dogs (p = 0.006). No difference was seen in the MCHC between groups (p = 0.408). Where measured (NI, n = 24; INF, n = 15), INF dogs had a higher median WBC concentration than NI dogs (Table 2).

INF dogs had higher median values than NI dogs for the following MIZAR rheometric parameters: AI, AMP, ED, I, OT, OTF, and ∆OTF-OT, and lower values compared to NI dogs for BP, I, and T_1/2 (Table 2). The mean cfDNA concentrations in the INF group were higher than the NI group regardless of centrifugation methodology (Table 2).

ROC analysis

ROC curve analysis compared ESRw, iSED-predicted ESR, MIZAR rheometric parameters, fibrinogen, and cfDNA regarding their ability to distinguish between dogs in NI and INF groups (Table 3). An ESRw1h >1.5 mm had a sensitivity of 73% and specificity of 79.5%; an ESRw24h value >13 mm had a sensitivity of 90.5% and specificity of 70.3%. An iSED value >6.5 mm/h had a 95.2% sensitivity and an 87.2% specificity for distinguishing between INF and NI animals. A MIZAR ED value >1,600 or an I value > 9,720 both corresponded to a sensitivity of 95.5% and specificity of 79.5% for distinguishing between groups.

Table 3.

ROC curve analysis comparing ESR, fibrinogen, MIZAR values, and cfDNA in 39 non-inflammatory and 22 inflammatory canine blood samples based on serum fibrinogen concentration.

	Area	SE	95% CI	p
AI	0.833	0.052	0.732, 0.934	<0.0001*
AMP	0.914	0.037	0.841, 0.987	<0.0001*
BP, ms	0.749	0.061	0.630, 0.869	0.0013*
BPH, ms	0.804	0.055	0.696, 0.912	<0.0001*
BPL, ms	0.571	0.073	0.428, 0.714	0.3635
ED	0.934	0.003	0.869, 0.999	<0.0001*
ESRw1h, mm/h	0.822	0.061	0.703, 0.941	<0.0001*
ESRw24h, mm/h	0.877	0.051	0.776, 0.978	<0.0001*
I	0.921	0.034	0.853, 0.988	<0.0001*
iSED ESR, mm/h	0.934	0.037	0.862, 1.00	<0.0001*
OT	0.878	0.044	0.792, 0.965	<0.0001*
OTF	0.905	0.040	0.826, 0.984	<0.0001*
∆OTF-OT	0.654	0.076	0.506, 0.802	0.0474*
T1/2, s	0.753	0.076	0.623, 0.882	0.0011*
WBC, ×109/L	0.740	0.089	0.567, 0.914	0.013*

See Table 2 notes.

^* p ≤ 0.05.

Dogs with sepsis

All 7 dogs with a confirmed septic focus were also in the INF group. Compared to the other 15 INF dogs, there was no difference in plasma fibrinogen concentration (p = 0.321). There was no difference in the ESRw1h, ESRw24h, or iSED values (p = 0.915, 0.921, 0.709, respectively). Similarly, there were no significant differences in any rheometric parameters measured by the MIZAR. Concentrations of cfDNA were not different between the septic and non-septic dogs in the INF group when the single centrifugation methodology was used, but a difference was revealed when cfDNA underwent 2 centrifugation steps (p < 0.012; Table 4).

Table 4.

Comparison of ESR, MIZAR, and cfDNA in inflamed dogs based on serum fibrinogen concentration with a confirmed septic focus (n = 7) and without (n = 15). All parameters are listed as median (range) and are compared using 2-tailed independent t-tests.

	Septic dogs	Inflamed dogs	p
AI	0.602 (0.530–0.671)	0.621 (0.537–0.684)	0.341
AMP	80.0 (33.0–135)	79.0 (42.0–136)	0.935
BP, ms	195 (91.0–262)	166 (0–310)	0.374
BPH, ms	255 (119–340)	177 (1.00–392)	0.072
BPL, ms	136 (63.0–209)	155 (0–279)	0.557
cfDNAa, ng/mL	121 (56.5–265)	125 (50.7–266)	0.508
cfDNAb, ng/mL	146.7 (54.1–206)	87.6 (5.98–149)	0.012*
ED	1,670 (1,610–1,730)	1,670 (1,590–1,760)	0.945
ESRw1h, mm/h	15.0 (0–58.0)	14.0 (0–50.0)	0.915
ESRw24h, mm/h	60.0 (14.0–130)	62.0 (5.00–127)	0.921
Fibrinogen, g/dL	5.84 (4.14–7.96)	5.16 (3.46–8.50)	0.321
I	17,000 (6,310–29,800)	16,800 (9,770–29,700)	0.954
iSED ESR, mm/h	38.4 (1.00–86.0))	33.3 (7.00–118.0)	0.709
OT	1,590 (1,560–1,640)	1,590 (1,550–1,680)	0.698
OTF	1,600 (1,580–1,660)	1,610 (1,560–1,680)	0.806
ΔOTF-OT	17.0 (10.0–34.0)	18.0 (2.00–52.0)	0.832
T1/2, s	2.10 (1.54–2.76)	2.00 (1.32–2.74)	0.372

See Table 2 notes.

^* p ≤ 0.05.

Comparison of dogs before and after orthopedic surgery

Eighteen of 61 samples were paired and represented 9 dogs with 1 sample taken before an orthopedic procedure and 1 taken 24 h following the procedure (Table 4). There were no differences in the ESR for either group. Postoperative samples had a higher plasma fibrinogen concentration (p < 0.001) and a higher iSED result (p = 0.017). Postoperatively, dogs had significant increases in MIZAR rheometric parameters AI (p = 0.001), I (p = 0.006), and AMP (p = 0.016) compared to preoperative values; MIZAR rheometric parameters BP (p < 0.0001) and T_1/2 (p < 0.001) were lower after surgery. Cell-free DNA concentrations were not different between pre- and postoperative measurements. Dogs undergoing orthopedic procedures had only one recorded CBC from the preoperative period, precluding the comparison of pre- and post-samples (Table 5).

Table 5.

Comparison of healthy dogs before and after elective surgical orthopedic procedures. Median (range) for MIZAR, ESR, fibrinogen, and cfDNA are compared using paired t-tests.

	Pre-op dogs	Post-op dogs	p
AI	0.537 (0.433–0.615)	0.621 (0.588–0.684)	<0.001*
AMP	44.0 (23.0–56.0)	59.0 (40.0–102)	0.016*
BP, ms	310 (145–489)	143 (106–258)	0.001*
BPH, ms	392 (201–636)	167 (106–274)	0.001*
BPL, ms	142 (90.0–403)	125 (45.0–243)	0.203
cfDNAa, ng/mL	153 (53.1–206)	117 (83.8–173)	0.366
cfDNAb, ng/mL	97.3 (53.8–182)	85.5 (5.98–160)	0.434
ED	1,600 (1,590–1,650)	1,628 (1,580–1,690)	0.071
ESRw1h, mm/h	1.0 (0–5.0)	1.0 (0–15.0)	0.313
ESRw24h, mm/h	16.0 (5.0–25.0)	15.0 (5.0–100)	0.479
Fibrinogen, g/L	1.91 (0.30–4.00)	5.05 (2.60–8.50)	<0.001*
I	8,090 (3,400–11,800)	12,800 (8,040–22,600)	0.006*
iSED ESR, mm/h	4.0 (1.0–9.0)	13.0 (6.0–56.0)	0.017
OT	1,560 (1,550–1,600)	1,569 (1,530–1,590)	0.610
OTF	1,580 (1,560–1,660)	1,590 (1,540–1,600)	0.479
ΔOTF-OT	12.0 (8.0–52.0)	15.0 (5.0–21.0)	0.734
T1/2, s	2.50 (1.94–2.98)	1.94 (1.32–2.14)	<0.001*

See Table 2 notes.

^* p < 0.05.

Discussion

In dogs with and without inflammation, we found that the iSED-predicted ESR was more indicative of inflammation than either the 1 h or 24 h ESR measured by the ESRw method. Our results mirror a report that evaluated 109 canine blood samples and compared a different POC ESR analyzer to the ESRw ³⁰ ; the POC ESR measurement and the ESRw method were correlated, but the inflammatory state of the dogs was not characterized. The dogs without inflammation in our study had ESRw1h of 0–7 mm/h, with a median of 0 mm/h. This is consistent with a report that noted a RI of 0–5 mm/h ³⁰ ; however, a 24-h sample was not evaluated in that investigation. Another report evaluated the ESR in healthy animals using a slightly different methodology and noted that ESRs in dogs were lower than the other species investigated (horse, sheep). ³³

At our institution, the MCV RI is 42–56 fL for horses, 28–30 fL for sheep, and 64–75 fL for dogs. The MCV RI in humans is 80–100 fL. Differences in MCV may contribute to the variable performance of the ESRw among species. RBCs with a higher MCV may have a lower surface-to-volume ratio than those with a low MCV, leading to reduced negative surface charge and increased RBC sedimentation.^3,39 Interestingly, MCV was significantly lower in dogs with inflammation in our study, and this group had a higher ESRw than dogs without inflammation, indicating that RBC size is only one factor contributing to the ESR value.

Decreased erythrocyte deformability in critical illness can contribute to impaired oxygen delivery as a result of congestion of RBCs in the microvasculature.^6,8,29,34 This phenomenon has been identified in dogs with septic shock. ³⁴ Syllectometry evaluates RBC deformability. Changes in light intensity after shear stress reflect RBC shape recovery. Recovery time, a parameter similar to BP in our study, describes the time necessary for RBCs to return to a normal shape. ²¹ Shorter recovery times are associated with less-deformable human RBCs, ²¹ which is consistent with the shortened BP seen in INF compared to NI dogs in our study. These findings are paired with a significantly shorter BP in the INF group. Rheometric parameters measured by the MIZAR may give information about the relative deformability of canine RBCs in various disease states that has not been available previously.

Increased RBC aggregation can also contribute to impaired microcirculatory flow. ⁸ The MIZAR assesses RBC aggregation in the latter part of the syllectogram. The I value reflects the surface area of the kinetic graphic, indicating the magnitude of the aggregation phenomena. In our study, I value was significantly elevated in dogs with inflammation, as were the AMP and ED parameters that describe the kinetic graphic. Together, these results suggest increased RBC aggregation in the INF group. The AI reflects the speed at which RBC aggregation occurs, and T_1/2 is the time necessary to reach half of the amplitude of the aggregate. Quantitative correlation analysis has demonstrated a strong linear association between AI and T_1/2, suggesting that larger aggregates of RBCs form more quickly. ²⁵ In our study, INF dogs had a significantly lower T_1/2 and a significantly higher AI compared to NI dogs, which further supports increased RBC aggregation in inflammation in dogs.

Plasma concentrations of inflammatory markers (e.g., positive acute-phase proteins, such as CRP, haptoglobin, ceruloplasmin, serum amyloid A, and fibrinogen^5,31,40,41) are more commonly used to characterize inflammation in small animal medicine than ESRw, for reasons highlighted above. Fibrinogen can promote RBC aggregation through nonspecific protein binding to RBC membranes and through fibrinogen-specific binding sites on the RBC surface. ²⁷ We used elevated fibrinogen concentration as a criterion for systemic inflammation, but we may have also influenced our ESR results by grouping dogs that were more likely to show RBC aggregation. Stratification using another validated criterion or biomarker (e.g., CRP or cfDNA) could be used in future studies to assign study groups.

We found increased cfDNA concentration and elevated WBC counts in dogs with inflammation. Cell-free DNA is elevated in dogs with systemic inflammatory response syndrome (SIRS) and sepsis, and markedly elevated cfDNA in animals with critical illness has been correlated with non-survival. ²⁶ A 2020 report described increased cfDNA concentrations in septic dogs relative to those with SIRS, and the cfDNA:neutrophil ratio was increased in non-survivors. ²⁶ We also evaluated the effect of centrifugation techniques for preparation of cfDNA for analysis. The second, higher-speed centrifugation significantly decreased the measured cfDNA, but the median values were relatively close to each other, and it is unclear if higher-speed centrifugation would result in clinically significant differences, especially if additional DNA extraction or purification steps were performed. Additionally, it is unknown if the higher concentration read in the single-spin samples reflects actual cfDNA or a contaminant. The second-spin technique is recommended for cfDNA measurement in humans and may offer an advantage in sample purity. ³⁸ Comparison of cfDNA concentrations through an alternate method would be necessary to answer this question.

When we compared INF dogs with and without a confirmed septic focus, RBC aggregation (ESRw and iSED methodologies) and RBC deformability (MIZAR analysis), in addition to cfDNA and fibrinogen concentration, were not different between the 2 populations. This result differs from that of a prospective study in which cfDNA values were significantly different between septic dogs and healthy dogs. ²⁶ The small number of septic dogs in our study increases the likelihood of type II error and may explain the discrepancies with other published reports. Because testing to characterize patient populations as “septic” was left to the discretion of the primary clinician, it is also possible that there were additional undiagnosed septic patients classified as inflamed.

Tissue injury induced during surgery induces release of inflammatory cytokines and systemic inflammation, and we found that postoperative fibrinogen concentrations increased significantly compared to preoperative values. Similarly, rheometric parameters measured by MIZAR reflected increased aggregation and decreased deformability in postoperative samples. In addition, the iSED-predicted ESR was significantly elevated in postoperative patients, although no significant differences were seen in the ESRw tests. Elevated ESR has been demonstrated in humans in the immediate postoperative period (<5 d) following total knee or total hip arthroplasty.^2,24,32 Intriguingly, one study of humans undergoing revision of knee surgery also found that elevated ESR and CRP were sensitive, specific, and accurate in predicting the presence of infection. ¹⁷ Although the diagnosis of infection per se was beyond the scope of our study (and was likely underpowered to identify differences in the population with sepsis), the relevance of ESR and CRP in detecting infection in human medicine may support more routine measurement of these inflammatory markers or POC assessment in veterinary patients.

Significant increases in fibrinogen occur in people following total hip replacement surgery and other orthopedic surgeries, ³⁷ mirroring our results. Interestingly, we found no significant differences between pre- and postoperative cfDNA concentrations, which may be the result of the relatively short half-life of cfDNA (~6 h). ⁴³ It is possible that the locoregional inflammation induced by an elective orthopedic procedure may not be sufficient to justify a persistently elevated cfDNA concentration. Blood sampling performed closer to the end of surgery may have demonstrated changes in cfDNA concentration associated with surgical trauma. ¹⁴

Our study has several limitations. In instruments such as the MIZAR analyzer, parameters connected to speed and magnitude of aggregation may change based on the amount of saturated hemoglobin present in the sample, given that hemoglobin has a characteristic absorbance spectrum influenced by oxygen saturation. ⁴² One study demonstrated that data were more likely to be influenced with wavelengths <700 nm. ⁴² We did not evaluate the degree of oxygenation in blood samples, so it is unknown if this variable bears relevance to our investigation.

Our study was further limited by small sample size, and by inconsistent availability of baseline bloodwork (CBC, serum biochemistry) to further support the presence or absence of ongoing systemic disease. Further, repeated analysis for the same patient was not attempted, but may have provided more information regarding normal variation between samples and may have been useful in monitoring patient progression. Additionally, because our study subjects were not anemic, it is difficult to say whether ESR results would have been affected by anemia, as demonstrated by previous work. ¹⁸ This is a factor that should be assessed when using the ESRw clinically (PCV is not expected to influence ESR predictions from the iSED analyzer).

POC rheology analyzers such as iSED (predicted ESR) and MIZAR (rheometric parameters) represent a practical and rapid way to evaluate rheologic and inflammatory changes in hospitalized dogs. We found that both analyzers produced results rapidly (especially compared to the ESRw methodology) and were intuitive to use. Future prospective studies evaluating MIZAR and iSED analysis in individual patients throughout a hospitalization period may provide perspective into disease progression and may alert clinicians to subtle changes in patient status.