Introduction

Protein-losing enteropathy (PLE) is defined as non-selective protein intestinal loss due to intestinal disease, resulting in hypoalbuminemia.^1,2 In dogs, PLE is frequently observed as a consequence of chronic small intestinal disease, such as severe parasitism and fungal enteritis, food-responsive enteropathy, idiopathic inflammatory bowel disease (IBD), intestinal neoplasia (particularly intestinal lymphoma), and lymphangiectasia.^1–3

In human medicine, thromboembolism (TE) and other vascular events are well-recognized complications of PLE, especially when secondary to IBD, such as Crohn's disease (CD) and ulcerative colitis, and the prevalence of TE in humans with ulcerative colitis and CD has been increasing.^4–6 Humans with IBD have a threefold increase in the risk of developing deep venous thrombosis (DVT) or pulmonary thromboembolism (PTE) compared with other hospitalized patients, with TE being identified postmortem in 7 to 41% of such cases.^6,7

It has also been proposed that PLE is associated with hypercoagulability in dogs, and that such dogs are at increased risk of thromboembolic events.^3,8–14 However, with the exception of familial PLE and nephropathy of soft-coated wheaten terriers, and PLE-associated intestinal crypt lesions, TE in dogs with PLE, particularly in idiopathic IBD, has been reported only infrequently.^{8,10–13,15,16}

The pathogenesis of TE in human IBD is incompletely understood, and is likely multifactorial.¹⁷ Antithrombin (AT) loss, thrombocytosis, platelet hyper-aggregation, hyperfibrinogenemia, endothelial lesions due to inflammation and immune-complex deposition, increased concentrations of plasminogen activator inhibitor, fibrin degradation products, factor XIII, and tissue plasminogen activator, as well as hyperhomocysteinemia, have all been demonstrated in humans with IBD.^4,7,17,18 The risk factors of developing TE in human IBD include the inherent inflammatory process per se, corticosteroid therapy, central IV catheter use, prolonged immobilization, and the presence of hyperhomocysteinemia or vitamin deficiency.^5,19

This retrospective observational study describes the clinical presentation in eight dogs with thromboembolic disease associated with PLE, secondary to IBD or lymphangiectasia, and discusses the potential risk factors for thrombosis, raising awareness of this potentially fatal complication.

Materials and Methods

This study includes dogs definitely diagnosed with small intestinal enteropathy, PLE, and TE in three academic veterinary hospitals (four dogs in Hebrew University Veterinary Teaching Hospital, Hebrew University of Jerusalem; three in the Hospital for Small Animals, Royal (Dick) School of Veterinary Studies, University of Edinburgh; and one in the Queen's Veterinary School Hospital, University of Cambridge). Thrombosis was demonstrated using imaging methods or at necropsy. The diagnosis of small intestinal enteropathy and PLE was based on presence of compatible history and clinical signs (i.e., acute or chronic small intestinal diarrhea, weight loss, intermittent vomiting, anorexia, and abdominal distension), as well as hypoalbuminemia, deemed secondary to enteropathy, after exclusion of albuminuria (by urinalysis and urine protein to creatinine ratio) and hepatic failure (by serum bile acids measurement). The electronic database of Hebrew University Veterinary Teaching Hospital (which saw four dogs) and the Internal Medicine and Pathology Services records of the other two institutions involved in the study (four dogs seen between the Hospital for Small Animals, Royal (Dick) School of Veterinary Studies Hospital, and the Queen's Veterinary School Hospital), were searched for dogs diagnosed with TE (and in these, a search for concurrent PLE/IBD was done) that presented during the following time periods: between 2007 and 2010 (Hebrew University Veterinary Teaching Hospital), between 2007 and 2009 (Hospital for Small Animals, Royal (Dick) School of Veterinary Studies), and between 2001 and 2008 (Queen's Veterinary School Hospital). Data were then retrieved from the medical records, including the signalment, history, physical examination, laboratory, imaging and histopathological findings, therapy, and outcome. Risk factors for thrombosis, as previously described in the veterinary and human medicine literature, were identified and reviewed, including glucocorticoid therapy, parenteral nutrition, and previous surgery.

Results

Signalment, History, Clinical Signs, Laboratory Findings, and Diagnosis

As the medical records were only partially computerized, it was impossible to determine the incidence of PLE secondary to IBD or lymphagiectasia in two of the hospitals involved in this study. In the third, during the study period (2007–2010), 56 cases of PLE in dogs were diagnosed, of which 4 (7.1%) developed TE. The study included eight dogs (Table 1) with median age of 7 yr (range 3 to 10 yr) and median body weight of 20.9 kg (range 5.3 to 45 kg) of the following breeds: two cross-breed dogs, one Pomeranian, one golden retriever, one Dalmatian, one English springer spaniel, one Labrador retriever, and one Cane Corso. There were five males (two neutered) and three spayed females. The history and clinical signs included chronic small intestinal diarrhea (seven dogs), intermittent vomiting (four dogs), weight loss (four dogs), abdominal distension (three dogs), anorexia (two dogs) and acute small intestinal diarrhea (one dog).

TABLE 1 Signalment and Clinical Data of Eight Dogs with Protein-Losing Enteropathy and Thromboembolism at the Time of the Thromboembolic Event

View Fullscreen

When PLE was diagnosed, all eight dogs had hypoalbuminemia (median serum albumin 16.45 g/L, range 13–22.5, reference interval [RI] 25–40) and five had concurrent hypoglobulinemia (median serum globulin 17.3 g/L, range 11.1–23, RI 18–37). Hypocobalaminemia was present in four of the seven dogs in which serum cobalamin concentration was measured (median 178 ng/L, range <150 to 774, RI >200), and of these seven dogs, folate concentration was above RI in two and below RI in one (above 24 ng/L, and 4.1 ng/L respectively, RI 7.1–14).

Endoscopic and full-thickness surgical gastrointestinal biopsies were obtained and evaluated microscopically. Histopathology revealed lymphoplasmacytic enteritis (LPE, seven dogs) with lymphoplasmacytic gastritis (two of seven dogs), and primary lymphangiectasia (one dog).

Findings at the Time of Thrombotic Events and Outcome

When TE occurred, seven of eight dogs had persistent gastrointestinal clinical signs, despite therapy for chronic small intestinal disease, including vomiting (in five dogs), diarrhea (in four dogs), recurrent pleural and/or abdominal effusion (in two dogs), exercise intolerance (in two dogs), anorexia (in five dogs), weight loss (in two dogs), and depression (in one dog). Poor body condition was recorded in five dogs, and in four of these, their body condition score markedly decreased compared to their first presentation.

Thrombotic events were noted between 6 and 85 wk (median 38 wk) from the onset of gastrointestinal signs. These included PTE (six dogs), splenic vein thrombosis (two dogs), and arterial thromboembolism (ATE) (one dog). TE was diagnosed ante mortem in two dogs (two had splenic vein thrombosis, of which one had concurrent ATE) based on abdominal ultrasound findings (confirmed on exploratory laparotomy in one dog), while the remaining six dogs had TE diagnosed at necropsy. Regarding the dogs diagnosed with PTE, several additional tests were performed. Thoracic radiographs were performed in two dogs. In one dog, radiographs showed collapse of the right lung lobes, and were unremarkable in the other. Pulse oxygen saturation (pulse oxymetry) was 84% in the first dog. In another dog that became acutely dyspneic after removal of the jugular catheter, pulse oxymetry oxygen saturation was 65%. In three dogs, venous blood gas analysis was performed. Data, even though incomplete, were available for two of them. In both these cases, metabolic acidosis was noted.

Two of the six dogs with PTE died suddenly (one after a walk and one in the hospital), while four presented with dyspnea of acute onset, one of which died shortly after presentation. The remaining three dogs developed acute respiratory distress in the hospital, and died despite intensive care. The latter included positive pressure ventilation (one dog), oxygen (two dogs), unfractionated heparin (one dog, 1000 international unit [IU] IV once) and dobutamine (one dog, dose unavailable). Two dogs had splenic thrombosis, one of which had ascites and abdominal pain, while the other had concurrent ATE and was presented due to bilateral hind limb weakness, which worsened with exercise. One dog with splenic thrombosis underwent splenectomy, recovered uneventfully, and later received preventive low-dose aspirin (0.5 mg/kg per os [PO] q 24 hr) in addition to the medical management of LPE, and was discharged three days post-op. The dog with splenic thrombosis and ATE was treated with low-dose aspirin (0.5 mg/kg PO q 24 hr), discharged, and was lost for follow-up.

Laboratory findings after the occurrence of TE were available in six dogs (Table 2). Hematological abnormalities included anemia (four dogs, median haematocrit 0.30 L/L, range 0.21–0,38 RI 0.37–0.55), neutrophilia (four dogs, median neutrophil count 23.93x10⁹/L, range 14.48–27.73, RI 3–11.5x10⁹/L), and thrombocytopenia (two dogs, platelets 155x10⁹/L and 83x10⁹/L, RI 175–500 x10⁹/L). Serum biochemistry abnormalities included hypoalbuminemia (five dogs, median albumin 18.75 g/L, range 12.2–27.7, RI 25.0–40.0), hypocalcemia (four of four dogs, median total calcium 1.85 mmol/L, range 0.99–1.99, RI 2.3–2.8), hypomagnesemia (two of two dogs, total magnesium 0.47 and 0.35 mmol/L, RI 0.69–1.18), and decreased ionized calcium concentration (one of three dogs, 1.12 mmol/L, RI 1.18–1.4). Serum cobalamin and folate concentrations were measured in three dogs. Serum cobalamin was subnormal in all three (median serum cobalamin 164 ng/L, range 150–178, RI >200 ng/L). Serum folate was subnormal in one dog (3.8 μg/L, RI 8.2-13.5 μg/L). Prothrombin time and activated partial thromboplastin were measured in three dogs and were within RI in two, while in one activated partial thromboplastin was mildly prolonged (21.4 s, RI 10.0–20.0 s). Fibrinogen, AT, and D-dimer concentration were measured in two dogs, and abnormalities included hyperfibrinogenemia (one dog, 4.15 g/L, RI 2.0–4.0), hypoantithrombinemia (one dog, 45.6%, RI 87.0–140.0%), and positive D-dimer concentration (one dog, >250 ng/mL). The latter occurred concurrently with hypoantithrombinemia.

TABLE 2 Clinicopathological Data of Eight Dogs with Protein-Losing Enteropathy and Thromboembolism at the Time of the Thromboembolic Event

View Fullscreen

At the time of TE (Table 1), seven dogs were being treated. Therapy included oral prednisolone or prednisone (six dogs, doses varying between 1 mg/kg PO q 12 hr and 0.5 mg/kg q 48 hr), budesonide (one dog, 0.3 mg/kg PO q 12 hr), azathioprine (one dog, 1.8 mg/kg PO q 48 hr), metoclopramide (one dog, 0.4 mg/kg subcutaneous q 12 hr), ranitidine (two dogs, 2 mg/kg IV q 12 hr), famotidine (one dog, 0.7 mg/kg PO q 24 hr), tylosin (three dogs, 10–20 mg/kg PO q 12 hr), ampicillin (one dog, 20 mg/kg IV q 8 hr), amoxicillin-clavulanate (one dog, 15 mg/kg PO q 12 hr), doxycycline (one dog, 10 mg/kg PO q 24 hr), metronidazole (one dog, 14mg/kg PO q 12h), cobalamin (two dogs, 1000 μg subcutaneous weekly), folate (one dog, 0.01 mg/kg PO q 24 hr), probiotics¹ (one dog, 3 mL PO q 12 hr), low-dose aspirin (two dogs, 0.5 mg/kg PO q 24 hr), phenobarbital (one dog, 3 mg/kg PO q 12 hr), IV magnesium-sulfate (one dog, dose unavailable), calcitriol (one dog, 5 ng/kg PO q 24 hr), and calcium-phosphate (one dog, dose unavailable). Management with commercial low-fat and hydrolyzed protein (one dog), hydrolyzed protein (two dogs), and novel protein (one dog) diets was also included. Repeated thoraco- and abdomino-centesis were performed in one dog, who also received IV colloids. Parenteral nutrition² (two dogs) was administered via a jugular catheter, up to one week prior to the occurrence of TE in one dog, while the other developed PTE immediately post-removal of the jugular catheter.

One dog underwent surgery 1 day prior to TE for obtaining duodenal biopsies. Another dog underwent an exploratory laparotomy at the day the TE had occurred due to a suspected gastro-intestinal foreign body, which was ruled out. Therapy was instituted after occurrence of TE in six dogs. This included unfractionated heparin (one dog, 188 IU/kg IV once), oxygen therapy (two dogs), ventilation (one dog), oral low-dose aspirin (two dogs, 0.5 mg/kg PO q 24 hr), human albumin (two dogs, 1.25 g/kg IV at constant rate infusion over 48 hr administered to one dog and 0.7g/kg IV administered to the other), hetastarch (one dog, 1 mL/kg/h IV for 24 hr), blood component transfusion (one dog, whole blood; one dog, packed red blood cells and plasma), IV adrenaline (one dog, dose unavailable), IV furosemide (one dog, dose unavailable), IV dobutamine (one dog, dose unavailable), and various opioids (five dogs).

Necropsy Findings

Necropsy was performed in the six non-survivors (Table 3), revealing pulmonary artery thrombi in all six (Table 3, Figure 1). Microscopic examination of pulmonary tissue sections showed additional smaller thrombi diffused throughout the lung parenchyma in two dogs. Additional thrombi were noted in the jugular vein and right atrium in the dog that died following removal of a jugular catheter. Histopathology of the small intestines was performed in five of the six dogs. Compared to previous microscopic findings (examined at the time when intestinal disease was diagnosed), abnormalities at the time of TE were considered more severe in two dogs, of similar magnitude in two dogs, and milder in one dog.

TABLE 3 Gross and Microscopic Postmortem Findings in Six Non-Survivor Dogs With Inflammatory Bowel Disease or Lymphangiectasia and Thromboembolism

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

Discussion

This study describes eight dogs diagnosed with TE and PLE, secondary to LPE or lymphangiectasia. Pulmonary TE was the most common TE, in agreement with findings in humans with IBD, although DVT is also common in such patients.²⁰ Arterial thrombosis, observed in one of our dogs, was reported less frequently than DVT in humans with IBD.²¹ It is mostly associated with previous surgery and premature arteriosclerosis in young CD patients.²⁰ Systemic arterial thrombosis occurs uncommonly in dogs, but it has been reported in association with chronic gastrointestinal disease.^10–22

Although frequently cited in the literature as a cause of TE in dogs, reports of non-neoplastic PLE with concurrent TE are limited mostly to single case reports.^8,9,10–13 The paucity of such reports in the literature may reflect the fact that TE is an infrequent complication of canine IBD. Alternatively, this may reflect the low index of suspicion for TE in the disease; its variable clinical presentation, even in severe cases; and its challenging antemortem diagnosis, especially of PTE.^23–25 The latter commonly does not manifest by classic signs of tachycardia or dyspnea, and often leads to sudden death, which may occur after discharge from the hospital, thereby remaining undiagnosed.²³ Occurrence of comorbidities and complications (e.g., iatrogenic hyperadrenocorticism, dyspnea secondary to pleural effusion, or anemia) might make interpretation of subtle clinical signs difficult. Interestingly, in human IBD patients, chronic, subtle, and, often, silent TE is now frequently recognized, possibly reflecting increased awareness of such clinical presentations of TE as well as improved screening for TE in these patients.²⁶

In dogs, even when the clinical suspicion of PTE is high, antemortem diagnosis can be challenging, particularly in the absence of advanced diagnostic imaging modalities, such as computed tomography (CT), and due to poor sensitivity and specificity of other imaging techniques.^25–27 In the largest retrospective study of canine PTE, in 6 of 47 cases, there was no clinical suspicion of PTE.²⁵ In human medicine, D-dimer concentration measurement has emerged as the most useful laboratory marker of TE, especially as a negative predictor. Studies in dogs have suggested similar findings.^28–30 Interestingly, D-dimer concentration was increased in only one of two dogs in this study. Failure to detect increased D-dimer concentration in the other dog might have resulted from a late sampling in relation to the thromboembolic event since D-dimer concentration increases transiently following acute TE.³¹

Due to the retrospective nature of this study and the differences between institutional protocols and availability of certain laboratory tests, the clinicopathological evaluation in this study varied, and was sometimes incomplete. This fact, along with the limited number of cases included, makes drawing definite significant conclusions difficult. Nonetheless, several clinicopathological observations are worthy of comment. Hypoantithrombinemia was recorded in only one of two cases, supporting the notion that decreased AT activity is unlikely to be the sole causal factor for TE in this cohort of patients.¹³ Thrombocytosis was not recorded in any of the cases, and was therefore unlikely to be involved in the pathogenesis of TE. Hyperfibrinogenemia, known to contribute to hypercoagulable states in dogs, was observed in one of two dogs in which fibrinogen was measured.³²

It is noteworthy that cobalamin concentration at the time of TE was subnormal in all dogs in which it was measured. Cobalamin and folate play a role in homocysteine metabolism, and their deficiency leads to hyperhomocysteinemia, a known risk factor for thrombosis in humans with IBD.^17,19,33 This is likely to happen in dogs as well, being supported by a recent study that demonstrated a negative relationship between serum cobalamin and homocysteine concentrations in dogs diagnosed with gastrointestinal disorders.³⁴ We can, therefore, cautiously speculate that serum homocysteine concentration was increased in our dogs, predisposing them to thrombosis. Also of interest is that hypomagnesemia, observed in two of the dogs in which serum magnesium was measured, has also been associated with thrombotic predisposition.³⁵

In human IBD patients, thromboembolic risk appears to be directly correlated with disease activity, partly due to the activation of coagulation factors by inflammatory mediators and leucocyte-derived tissue factor-expressing microparticles.^5,17,36,37 The correlation between thromboembolic risk and disease severity possibly also reflects the impact of intensive treatment protocols and recumbency.³⁸ It is of note that most of our dogs had poorly controlled IBD and PLE at the time of the TE. Several of the established risk factors for TE in humans with IBD were identified in our dogs, including corticosteroid therapy, central venous catheterization, and surgical intervention.¹⁸ These were also recognized as risk factors for TE in other conditions in dogs, highlighting the likely multifactorial pathogenesis of TE in dogs with chronic intestinal disease.^12,27 The majority of the dogs in the present study were chronically treated with glucocorticosteroids at the time of the occurrence of TE. Corticosteroid administration is known to induce a state of hypercoagulability, as manifested by abnormalities in thromboelastography (TEG) parameters.^40–42 It is also noteworthy that one of the dogs developed TE-related clinical signs immediately after removal of a central venous catheter. The use of such catheters is a well-established risk for development of TE in humans affected with IBD, and has been associated with PTE in dogs as well.^18,27 In this particular case, it is therefore difficult to define the relative weight of each of these different risk factors in the pathogenesis of TE, and the presence of the central venous catheter may have played a role in the development of PTE as well, in addition to the PLE.

The overall outcome of the dogs in this study was poor, especially of those with PTE. Given this poor outcome, once TE occurs, early identification of dogs at risk for TE and initiation of thromboprophylaxis are important, although standardized treatment protocols for dogs are lacking. TEG, which gives an integrated assessment of coagulation, shows great promise in identifying patients in a hypercoagulable state, and has been used to demonstrate hypercoagulability in dogs with PLE.^9,13,39 However, it is unclear from previous veterinary studies which of the measured thromboelastographic variables relates to the actual thromboembolic risk.²⁷ Unfortunately, TEG was unavailable in our institutions. In light of the observed hemostatic abnormalities in some of our dogs, comprehensive hemostatic tests seem to constitute useful information for assessment of the risk for TE and appropriate therapy institution.

This study has several limitations. It included a rather small number of dogs and it was retrospective. Therefore, some data (e.g., certain laboratory test results) were missing, thereby limiting interpretation of the results. In addition, assessment of the severity of disease using uniform pre-set criteria was impossible. Lastly, this study included dogs treated in three different institutions, using several different treatment protocols, thereby introducing variance. Due to these limitations, our conclusions must be interpreted and applied to other cases with caution. Nevertheless, with the paucity of reports of TE in dogs with IBD and PLE, this case series provides useful information.

Conclusion

This case series highlights the association between PLE due to IBD or lymphangiectasia, and the development of TE, in dogs. Clinicians should be aware of this potentially serious life-threatening complication in these conditions. Clinical signs depend on location of the thrombus and may be subtle. Early diagnosis might facilitate early therapeutic intervention, potentially improving the outcome. Further studies are required to establish the true prevalence of TE, assess the usefulness of screening tests, and perform a comprehensive risk factor analysis in dogs with IBD and PLE.