Arterial thromboembolism (ATE) is a devastating condition in cats, mostly associated with underlying cardiac diseases, less frequently with hyperthyroidism and neoplasia, and in a minority of cases, the etiology is undetermined. In cats, ATE is clinically manifested mostly by an acute onset of unilateral or bilateral paraparesis or paraplegia, weak or absent arterial pulses, pain, pale or cyanotic footpads and nail-beds, and cold extremities of the affected limbs.1, It most commonly affects both hind limbs, although a single hind limb, or the front limbs, might be affected.1, ATE occurs less commonly in dogs than in cats and is associated with a wider range of predisposing conditions.[2-6]
The diagnosis of ATE is based primarily on the clinical presentation, which might be insufficient to allow confirmation. A history of predisposing conditions might support the diagnosis, but a definitive diagnosis requires advanced imaging tools. Selective iodine contrast medium angiography is the gold standard for diagnosing aortic thrombosis. Conventional angiography has excellent spatial resolution, but is invasive and technically challenging.[8, 9] Radionuclide angiography allows assessment of the location and severity of aortic thrombosis, but requires special facilities, whereas anatomic detail is poor. Abdominal ultrasonography might be useful for identifying aortic thrombi, but might require technical expertise, and can be challenging, because the ultrasonographic characteristics of thrombi vary depending on their age.[8, 9, 11] In acute ATE, the size of the thrombus might be insufficient for a diagnosis.[7, 12] Doppler-flow evaluation is usually necessary for the assessment of thrombi.[8, 11] Contrast enhanced magnetic resonance angiography (MRA) was anatomically accurate for detection the full extent of aortic thrombi in 4 dogs with ATE; however, MRA is often unavailable, is cost-prohibitive, and requires general anesthesia.
In ATE, the blood supply to the area distal to the thrombus is diminished, with decreased nutrient and oxygen delivery.2, Our hypothesis was that ATE results in diminished local venous glucose concentration in the affected, paralyzed limbs, whereas systemic glucose concentration remains unaffected by the ATE, and will therefore remain within reference interval (RI), or might be increased because of stress, pain, and excitement. The difference between the systemic and the affected limb, peripheral blood glucose concentration in cases of ATE is therefore expected to be higher than that between the systemic glucose concentration and its peripheral concentration in an unaffected limb, and higher than this difference in ambulatory animals and those with paralysis of other causes. This difference may therefore serve as a useful, readily available indicator of acute ATE in cats and dogs.
- Top of page
- Materials and Methods
The study groups included 22 cats and 9 dogs diagnosed with ATE, whereas the corresponding nonambulatory control groups included 10 cats and 11 dogs, and the ambulatory control groups included 10 cats and 9 dogs. There were significant age differences among the 3 study groups, of both cats and dogs (P = .03 and P = .01, respectively). The median age of the ATE cat group was 6 years (range 1.5–15.0) compared to 1 year (range 0.3–11.0) in the nonambulatory control group, and 1.7 years (range 0.3–15.0) in the ambulatory control group. The median age of the ATE canine group was 10 years (range 7–14), compared to 4.5 years (range 0.5–8.0) in the nonambulatory control group, and 7 years (range 1–14) in the ambulatory control group. There were significant age differences among groups, in both cats and dogs, and post hoc analysis showed these resulted because of differences between the ATE groups and their respective nonambulatory control groups.
In the ATE groups, 15/22 cats (68%) and 4/9 dogs (44%) did not survive to discharge, of which 12/15 cats (80%) and 4/4 dogs (100%) were euthanized.
In the ATE feline group, 19 cats (86%) presented with bilateral hind limb paralysis and 3 with a single limb paralysis (2 with a right hind leg paralysis and 1 with right front limb paralysis). Doppler ultrasonography, confirming the ATE, was performed in 3 cats, whereas in 2 additional cats, ATE was confirmed post mortem. Therefore, in 17/22 cats ATE was diagnosed based only on their clinical presentation. Underlying cardiac diseases were diagnosed in 9/22 cats (41%), based on physical examination findings, thoracic radiography and echocardiography, performed by a board-certified cardiologist, including hypertrophic cardiomyopathy (7 cats), left ventricular enlargement (1), and restrictive cardiomyopathy (1). In 9 additional cats (41%) in which the owners declined a board-certified cardiologist echocardiogram because of financial constraints, underlying cardiac diseases were suspected based on physical examination findings (eg, presence of cardiac murmur or gallop rhythm), thoracic radiography and echocardiography, performed by an emergency clinician, revealing left atrial enlargement and a left atrium/aorta diameter ratio >1.5. An underlying cardiac disease was therefore diagnosed in 18/22 of the cats (82%). In the remaining 4 cats (18%), no underlying disease was diagnosed, despite an investigation that included CBC, serum chemistry, urinalysis, and abdominal ultrasonography and radiography. Moreover, in 1 of these 4 cats necropsy did not reveal the underlying cause of ATE.
The final diagnoses in the nonambulatory control cats group included pelvic fracture (3 cats), vertebral fracture and brachial plexus avulsion (2 each), spinal neoplasia, infectious meningomyelitis, and intervertebral disc herniation (1 each). The ambulatory control cats group included healthy, staff-owned cats (5), a cat presented postovariohysterectomy because of dehiscence of the surgical incision, as well as cats with feline panleukopenia virus infection, acute kidney injury, chronic gastritis, and an infected skin laceration (1 each).
In the ATE dog group, 6/9 dogs presented with bilateral hind limb paralysis, of which 1 had left front limb paralysis as well, whereas 3 dogs had single hind limb paralysis. ATE was confirmed using Doppler ultrasonography in all dogs. In 3 dogs, compatible post mortem findings of ATE were recorded as well. The primary concurrent diagnoses included adrenal mass with mitral and tricuspid regurgitation, severely hypertensive acute kidney injury with pancreatitis, normal pregnancy, ruptured splenic hemangiosarcoma with disseminated intravascular coagulation, and granulomatous meningoencephalitis (1 each). The latter was definitely diagnosed with ATE at necropsy. In 4 dogs, no underlying disease was diagnosed despite an extensive workup, similar as described in cats. In 1 of these dogs, a necropsy was performed and revealed GME, as well as the saddle ATE; however, the cause of the ATE was undetermined.
The nonambulatory control dogs group included traumatic vertebral fracture (6 dogs), intervertebral disc herniation (4) and traumatic pelvic fractures (1). The ambulatory control dog group included healthy blood donors (2 dogs), and dogs with hemoperitoneum, idiopathic epilepsy, esophageal spirocercosis, cervical vertebral disc herniation, granulomatous meningencephalitis, post-operative gastric dilatation, and volvulus and traumatic mandibular fracture (1 each).
In the animals with ATE, peripheral glucose concentration, obtained from the affected limb, was significantly lower compared to systemic glucose concentration in both cats (median 45 mg/dL; range 17–182, versus 181 mg/dL; range 98–394, respectively; P < .0001) and dogs (median 46 mg/dL; range 10–161, versus median 94 mg/dL; range 58–298 mg/dL, respectively; P = .008). There were no significant differences between systemic and peripheral blood glucose concentrations in all animals in both control groups (Table 1).
Table 1. Systemic and peripheral, limb blood glucose concentrations in dogs and cats with acute arterial thromboembolism and their respective negative control groups.
|Group||Source of Blood Sample||N||Blood Glucose Concentration (mg/dL)||P Value|
There were significant ΔGlu and %ΔGlu differences among the cat groups and among the dog groups (P < .0001 and P < .001, respectively, for both variables) (Table 2; Fig 1). Post hoc analysis showed that both ΔGlu and %ΔGlu, in both cats and dogs with ATE, were significantly higher compared to their respective 2 negative control groups (P < .0001 and P < .001, respectively, for both variables), whereas there were no differences in these parameters among the 2 respective negative control groups in each animal species (P > .05 for both variables, in both species) (Table 2).
Table 2. Absolute and relative differences between systemic and peripheral, affected limb blood glucose concentration in dogs and cats with acute arterial thromboembolism and in their respective negative control groups.
|Group||N||ΔGlu (mg/dL), Median (Range)||%ΔGlu (%), Median (Range)||P Valuea|
|ATE||22||107a (51–354)||74a (35–95)||<.0001|
|Nonambulatory controls||10||8b (−10 to 101)||5b (−13 to 35)|
|Ambulatory controls||10||0b (−13 to 7)||0b (−7 to 6)|
|ATE||9||46a (18–187)||46a (31–95)||<.001|
|Nonambulatory controls||11||−2b (−23 to 13)||−3b (−32 to 9)|
|Ambulatory controls||9||1b (−10 to 8)||1b (−13 to 6)|
Figure 1. Box and whiskers plots of the absolute difference (calculated as systemic blood glucose – peripheral blood glucose) between systemic and peripheral glucose concentrations, and the relative difference (calculated as the absolute difference/systemic blood glucose concentration) in the study groups (A, C: cats, B, D: dogs). The line within the boxes represents the median, whereas the boxes represent the interquartile range, the whiskers represent the range, extending to a maximum of 1.5 times the interquartile range, and outliers are marked as dots.
Download figure to PowerPoint
The ATE group of cats was divided further into 2 subgroups: confirmed (5 cats) and clinically diagnosed (17 cats) subgroups. These subgroups and the 2 control groups were then compared. There was no significant difference in ΔGlu between the 5 definitely diagnosed cats with ATE and the 17 diagnosed clinically (median 72 mg/dL [range 53–147] and median 110 mg/dL [range 51–354], respectively; P = .22). There were significant ΔGlu and %ΔGlu differences when these 2 subgroups and the 2 control groups were compared (P < .001, for both). Post hoc analysis, with Bonferroni's correction for α, showed no significant ΔGlu and %ΔGlu differences between the 2 ATE subgroups (P > .0125 for both). There were significant ΔGlu and %ΔGlu differences between both ATE subgroups and each of the control groups (P < .007 for all comparisons). The median ΔGlu for the definitely diagnosed ATE feline subgroup and the clinically diagnosed ATE feline subgroup were 72 mg/dL (range 53–147) and 110 mg/dL (range 51–354), respectively, compared to 8 mg/dL (range 10–101) in the nonambulatory control group and median 0 mg/dL (range −13 to 7) in the ambulatory control group.
Peripheral glucose concentration was significantly lower (P = .006) in the affected limb in the ATE cat group (22 cats combined; median 45 mg/dL, range 17–182) compared to both nonambulatory (median 176 mg/dL, range 87–218) and ambulatory control (median 127 mg/dL, range 23–196) groups. In dogs, peripheral glucose concentration was significantly lower (P = 0.006) in the affected limb in the ATE group (median 46 mg/dL, range 10–161) compared to its concentration in both the nonambulatory (median 90 mg/dL, range 46–136) and ambulatory control (median 90 mg/dL, range 73–134) groups. There were no statistical systemic glucose concentration differences among the cat or dog groups.
Receiver operator characteristics analyses for ΔGlu and %ΔGlu as predictors of ATE in cats showed AUCs of 0.98 (95% confidence interval [CI95%] 0.90–1.00), and 1.00 (CI95% 0.94–1.00), respectively. In dogs, ROC analyses for the ΔGlu and the %ΔGlu as predictors of ATE showed AUCs of 1.00 (CI95% 0.89–1.00), and 1.00 (CI95% 0.89–1.00), respectively. A ΔGlu cutoff of 30 mg/dL corresponded to sensitivity and specificity of 100% and 90%, respectively, for predicting ATE in cats. A %ΔGlu cutoff of 23% corresponded to a sensitivity and specificity of 100% and 90%, respectively in cats. In dogs, corresponding cutoffs of 16 mg/dL for ΔGlu, and 20% for %ΔGlu, corresponded to sensitivity and specificity of 100%, for both.
- Top of page
- Materials and Methods
The results of this study show that in both cats and dogs with ATE, peripheral (affected limb) blood glucose concentration is significantly lower compared to systemic glucose concentration. These results are in agreement with previous findings, showing that decreased local venous glucose in samples obtained from affected limbs of 11 cats and 2 dogs with acute ATE is a sensitive marker of ATE.2
The peripheral blood glucose concentration in samples obtained from a paralyzed limb in cases of ATE is expected to be low in face of systemic normoglycemia or hyperglycemia. The latter has been attributed to stress, anxiety, pain, and hyperlactatemia. Anxiety, pain, and stress lead to increased concentrations of diabetogenic hormones, including catecholamines, glucagon, and cortisol, which inhibit insulin secretion and increase insulin resistance, gluconeogenesis, and glycogenolysis.[1, 3] Myocyte anaerobic glycolysis in the affected limbs results in lactate production, a major gluconeogenic precursor, which is delivered through the unaffected venous blood supply in the affected limb to the systemic circulation, and to the liver, thereby contributing to increased systemic glucose concentration. Since the underlying conditions may affect blood glucose concentration, analyzing single samples from affected limbs may not be sufficient for a diagnosis of ATE. However, when systemic and peripheral blood glucose samples are analyzed concurrently, and compared, glucose concentration difference becomes apparent, and is therefore diagnostic of ATE.
This study showed that ΔGlu and %ΔGlu are significantly higher in both cats and dogs of the ATE groups compared to their corresponding ambulatory as well as nonambulatory control groups, in which ATE was absent. Cutoffs of ΔGlu ≥30 mg/dL and %ΔGlu ≥23%, respectively, in cats, and of ΔGlu ≥16 mg/dL and %ΔGlu ≥20%, respectively, in dogs, showed very high sensitivity and specificity for prediction of ATE. Therefore, this diagnostic procedure seems to be extremely accurate. It has additional advantages, including high applicability and availability, minimal blood volume requirement, and high cost-efficiency, especially when compared to other more technically challenging, often cost-prohibiting (ie, conventional and Doppler ultrasonography[7, 10]) or invasive diagnostic procedures (ie, nuclear scintigraphy, contrast angiography,[7, 8] and MRA). Increased plasma D-dimer concentration was positively associated with ATE in dogs and cats,[16-18] however, although the test is simple, it is often unavailable, it requires a higher blood volume, and is performed in citrated plasma, thereby consuming more time. In addition, it most often cannot be performed bedside, since whole blood centrifugation is required. Finally, the test is much more expensive compared to blood glucose measurement.
Routine serum biochemistry testing was also investigated in dogs and cats with ATE, and the common abnormalities include hyperglycemia, azotemia, hyperkalemia, and increased serum muscle enzyme activity.[1, 4, 15] All these abnormalities are nonspecific of ATE, and cannot be solely used to discriminate between damage to a limb because of ATE and other, nonvascular lesions (eg, trauma, hematoma, rhabdomyolysis, infection, or neoplasia). In addition, some of such abnormalities may result from the primary underlying condition, or from secondary complications (eg, azotemia and hyperkalemia in acute kidney injury,[19, 21] or increased creatine kinase activity in sepsis and bacterial infections).
Consistent with the pathogenesis of ATE, and similar to the behavior of systemic and peripheral glucose concentrations herein, differences in systemic and peripheral concentrations of other laboratory analytes may occur as well, including lactate, potassium, pH, and blood gases, which may also be considered as diagnostic indicators of ATE and were not investigated herein. This is because the arterial blood supply to the affected limb is diminished, with resultant increased local anaerobic metabolism. Lactate concentration in blood samples obtained from affected limbs in 11 cats and 2 dogs with acute ATE was shown to be a sensitive indicator of ATE.2 Future studies of ATE may include these analytes, as well as physical parameters (eg, toe temperature) as predictors of acute ATE. Herein, investigation of glucose concentration differences was selected because of the very high availability, low cost, and the simplicity of blood glucose testing. Results are quickly obtained, and the use of a glucometer requires as little as 0.6 μL of fresh whole blood.
Obtaining blood samples from affected limbs, in which arterial blood supply was diminished or absent, raised concerns regarding a risk of infection, and infliction of pain. We therefore elected to sample only a single paralyzed limb in each animal. However, infections were not noted in the sampled affected limb in any animal. As for infliction of pain, in all of the animals with ATE in this study, deep pain sensation was either absent, or markedly diminished, and no indication of pain was observed during blood sampling from the affected limbs in any of these animals. Since glucose concentration measurement using the glucometer requires only a single blood drop, we strongly believe that the procedure was safe, causing no discomfort to the animals with ATE and very little discomfort to the negative control animals.
This study has several limitations. First, the size of study groups was rather limited, although it met the numbers determined by the a priori power analysis. Larger groups might have allowed further substantiation of these results. Second, this study was not designed to investigate the extent of the damage to the affected limbs, the size of the thrombus, or the prognosis of acute ATE. It would be extremely interesting to perform such future studies, with further laboratory testing (eg, muscle enzyme activity and D-dimer concentration), and to investigate the correlation of such tests with systemic and peripheral blood glucose concentration differences, to further characterize the severity of the lesions and the prognosis of acute ATE. Third, in most of the cats with ATE herein, the diagnosis of ATE was based only on clinical signs, and unconfirmed by advanced imaging or necropsy. This subset of cats is not an ideal population to be used for confirmation of a diagnostic test. Nevertheless, in our clinical setting, cat owners most often decline further testing when given a diagnosis of ATE and its grave prognosis. Nonetheless, the results show that ΔGlu and %ΔGlu in this subset of cats did not differ statistically from those of the definitely confirmed ATE cases in cats. Fourth, this study has investigated only cases of acute ATE with paralysis. Dogs with ATE of chronic onset, and cats with long-standing (days) ATE were excluded. Therefore, these results should not be applied to this set of animals, or to paretic animals with ATE, and future investigation of blood glucose in such cases is warranted. In such cases, collateral arterial circulation does occur, which will likely positively affect glucose concentration in peripheral, affected limb blood samples. Fifth, there were significant age differences between the study groups, in both dogs and cats. In both animal species, both the ATE and the nonambulatory control groups animals were older compared to the ambulatory negative controls. Although age may potentially have affected blood glucose concentration, no blood glucose concentration differences were observed between the control groups in both dogs and cats. We therefore believe that these results were not affected by the age differences among groups. Sixth, the study utilized particular glucometer models, and therefore the ΔGlu and %ΔGlu cutoff values recommended herein may be somewhat different when other hand-held glucometers or autoanalyzers are utilized. Our results should be therefore confirmed using other analytical methods. In addition, specific RIs should preferably be set. Seventh, during the study period, 2 different glucometers were utilized. Although theoretically differences in their reading might exist, both are based on the same ISO standard and measurement method, and are produced by the same manufacturer. We therefore believe that such potential differences did not affect the validity of our results. Lastly, this study included glucose samples obtained from a single affected limb. In animals in which more than a single limb was affected, it would have been interesting to test samples from all of the affected limbs, and possibly, compare blood glucose concentrations among samples from all 4 limbs, and compare these to a jugular vein blood sample. This was not done in this study to minimize the potential risks (ie, infection) and animal discomfort; however, it should be considered in future studies, in light of the lack of complications and pain inflicted by the procedure in this study.
In conclusion, there is a significant difference between systemic and peripheral, affected-limb glucose concentrations in cats and dogs presented with paralysis because of acute ATE, providing an accurate diagnostic tool for ATE in such cats and dogs. With clear ΔGlu and the %ΔGlu cutoff points, the study demonstrates this test to be highly sensitive (100%) and specific (90% in cats and 100% in dogs). The test it is cost-efficient, minimally invasive, highly available in every clinical setting, and issues instant, clear-cut results.