Aberrant heartworm migration to the abdominal aorta and systemic arteriolitis in a dog presenting with vomiting and hemorrhagic diarrhea

Case description

A 5.0 kg, 2-year-old, castrated male Dachshund was referred to the Texas A&M University Veterinary Medical Teaching Hospital for a 6-day history of vomiting and a 3-day history of hematochezia. The dog had been treated by the referring veterinarian with famotidine, maropitant citrate, penicillin, and metronidazole without clinical improvement. The patient had been previously diagnosed as heartworm positive when it was adopted 2 mo prior to presentation, and had received 2 once-monthly doses of ivermectin heartworm preventative. On physical examination, the patient was depressed and icteric. A grade II/VI, left apical systolic heart murmur was ausculted. The remainder of the physical examination was within normal limits.

A complete blood (cell) count (CBC) revealed a normocytic, normochromic, non-regenerative anemia with a red blood cell count of 3.48 × 10^{/L [reference interval (RI); 5.5 to 8.5 × 10/L], leukocytosis (23.6 × 10/L; RI: 6.0 to 17.0 × 10/L), neutrophilia (16.76 × 10/L; RI: 3.0 to 11.5 × 10/L) with a left shift (1.65 × 10/L bands; RI: 0 to 0.3 × 10/L), monocytosis (1.65 × 10/L; RI: 0.15 to 1.25 × 10/L), and severe thrombocytopenia (7 × 10/L; RI: 200 to 500 × 10/L). A chemistry profile revealed total hypocalcemia (1.93 mmol/L; RI: 2.33 to 2.95 mmol/L), hypoproteinemia (44 g/L; RI: 57 to 78 g/L), hypoalbuminemia (17 g/L; RI: 24 to 36 g/L), hyperbilirubinemia (150.5 μmol/L; RI: 0 to 13.7 μmol/L), hyponatremia (136 mmol/L; RI: 139 to 147 mmol/L), hypokalemia (2.9 mmol/L; RI: 3.3 to 4.6 mmol/L), and hypochloremia (105 mmol/L; RI: 107 to 116 mmol/L). A coagulation profile revealed an increased partial thromboplastin time (16.4 s; RI: 7.1 to 10.0 s), and low antithrombin (91.7 μg/L; RI: > 114% NHP). D-dimers could not be measured due to icterus.}

Thoracic radiographs revealed dilation and blunting of the caudal and right middle pulmonary arteries. The pulmonary parenchyma appeared normal. Mild enlargement of the right heart and main pulmonary artery was noted. Radiographic findings were consistent with pulmonary hypertension and right-sided cardiomegaly caused by heartworm disease. An echocardiogram was performed to rule out caval syndrome. No heartworms were observed.

Abdominal ultrasound revealed heartworms within the abdominal aorta from the level of the diaphragm to both femoral arteries (Figure 1). A segment of small intestine was thickened, hypoechoic, and displayed loss of layering (Figure 2). No blood flow could be demonstrated in this segment with Doppler ultrasound, and it was considered to be avascular. This segment of small intestine, along with the stomach, was fluid-filled and hypomotile, consistent with functional ileus. Infarcts were visualized in the spleen and left kidney.

Figure 1

Ultrasound image showing D. immitis worms (arrows) within the aorta.

Figure 2

Avascular, thickened segment of small intestine with loss of wall layering highlighted by the black arrows. The white arrow shows an unaffected intestinal loop with visible wall layering.

The dog was euthanized and submitted for necropsy. At necropsy, four 8- to 16-cm long heartworms were present in the abdominal aorta from just cranial to the right renal artery to the iliac bifurcation (Figure 3). These were unassociated with thrombi or intimal fibroelastosis. Three similar filariae were present in the right ventricle and pulmonary artery, also unassociated with thrombi and with minimal intimal fibroelastosis. Ischemic necrosis was noted in the myocardium of the left ventricle. Multiple splenic and left renal infarcts were noted, and the pancreas was hemorrhagic. The mucosa of the urinary bladder had multifocal hemorrhages, and the urine was red. The lungs were rubbery and clear fluid oozed from cut sections, indicating pulmonary edema. Small thrombi were macroscopically visualized in several branches of the pulmonary arteries and this was confirmed histologically. The gastric and intestinal mucosa and serosa were petechiated. Necrohemorrhagic colitis was noted in 2 discrete sections of colon, with a necrohemorrhagic typhlitis. Histologically, mural necrosis of small arterioles was seen with hemorrhagic infarcts in the liver, lung, kidney, pancreas, lung, and heart. This arteriolitis was especially severe in the submucosa and muscularis of the entire gastrointestinal (GI) tract. The arterioles affected were 1 to 3 leiomyocytes thick and the outside diameter of the arterioles never exceeded 120 μm in early stages. Arteriolitis began with fibrinoid necrosis of the media that was later infiltrated by leukocytes (Figures 4 and ​and5),5), but minimal inflammation, beyond edema, was seen. Most unusual, was that the lumina of necrotic arterioles did not contain thrombi and remained lined by swollen and hypertrophied endothelium. No macroscopic obstruction was noted, except by the small thrombi in the lungs. Few microfilariae were seen histologically. The renal infarcts were acute on histologic evaluation. Glomeruli and tubules were unremarkable with no protein or cellular casts. The brain had small thrombi in capillaries, but no arteriolitis was seen.

Figure 3

Necropsy image showing D. immitis worms within the abdominal aorta (arrow). The urinary bladder (UB) and left kidney (LK) can also be seen.

Figure 4

Photomicrograph of a small intestinal arteriole with early arteriolitis and fibrinoid degeneration with some karyorrhectic nuclear debris of segments of the medial wall. Swollen endothelial cells (arrows) line the lumen (asterisk) and mild edema and swollen adventitial cells surround the arteriole. [Hematoxylin & eosin (H&E), Bar = 50 μm]

Figure 5

Photomicrograph of a small intestinal arteriole in late stage of the arteriolitis with the media replaced by fibrinoid degeneration and karyorrhectic nuclear debris. The arteriole maintains a lining of swollen endothelial cells (arrows), and has surrounding edema, some neutrophils and swollen macrophages, and adventitial cells. The lumen (asterisk) is still patent with some erythrocytes and a few leucocytes (H&E, Bar = 50 μm).

Discussion

Aberrant migration of D. immitis has previously been reported to involve the eye, brain, and systemic circulation. With typical heartworm infection in the dog, a mosquito bite allows entry of L3 larvae which travel within the subcutaneous tissues and molt into the L4, and then L5 stages. After approximately 100 d, the L5 larvae enter the vascular system and migrate to the pulmonary arteries, where they cause inflammation that leads to endothelial damage and changes within the vascular walls. These changes are characterized by villous myointimal proliferation which leads to narrowing of the vessel. Endothelial damage leads to thrombosis (5). The proposed theory behind the aberrant ocular migration is via direct penetration into the globe by 4th stage larvae (6). Aberrant migration to the central nervous system can present with a multitude of neurologic signs depending on the vessel obstructed by the filaria. One dog with a 3-month history of neurologic signs (seizures, blindness, ataxia, and circling) following administration of arsenamide sodium and diathiazine iodide for heartworm infection was found to have a D. immitis worm in the right posterior communicating artery on necropsy; however, no theories were given for the aberrant migration in this case (8). There have been few reports of aberrant migration to the left side of the circulatory system (7,9–11). Most reports on D. immitis in the femoral arteries of dogs describe lameness of 1 or both hind limbs (9–11). The largest case series evaluated 5 dogs infected with D. immitis in the abdominal aorta and/or femoral arteries (9). All dogs in the study presented with lameness, paresthesia, and/or self-mutilation of the pelvic limbs. A right to left shunt was diagnosed in 2 of these dogs with aberrant migration, and the shunt was thought to be the cause of the systemic involvement. The shunt was diagnosed more specifically as a patent foramen ovale in 1 dog and an aorticopulmonary window in the other (9). Our case is unique in that D. immitis worms were in the abdominal aorta and femoral arteries without clinical pelvic limb signs, and the primary clinical signs were GI in origin and associated with widespread arteriolitis.

Necrotizing vasculitis has been reported in a case of aberrant migration of D. immitis to the left side of the circulatory system, but was limited to areas where the worms were located (7). In our case, vessel inflammation was limited to arterioles and was widespread. The connection between arteriolitis and the heartworm infection is unclear. In human medicine, a link has been found between various infectious agents and the presence of vasculitis (12–14). Mechanisms postulated for the development of vasculitis from an infectious agent include direct invasion of endothelial cells, immune complex deposition leading to vessel wall damage, and stimulation of autoreactive lymphocytes (12–14). The presence of D. immitis in human lungs has been shown to cause pulmonary vasculitis, but this is usually localized (15). In dogs, antigen-antibody complexes formed during heartworm infection can cause glomerulonephritis, leading to proteinuria (16). Presumably, in our case, arteriolitis caused the severe thrombocytopenia, with vessel leakage causing hypoalbuminemia and loss of clotting factors, and may have resulted in the development of the widespread thrombosis and hemorrhagic infarcts in the GI tract that ultimately led to the dog’s clinical signs. Also, it was suspected that the heartworms in the aorta of the dog herein also contributed to thrombosis or obstruction of vasculature to portions of the GI tract and viscera, as typically occurs in the pulmonary arteries with heartworm infection (5).

Directly linking the unique vessel lesion to D. immitis is difficult because most cases of vessel inflammation in human and veterinary literature are referred to as vasculitis, and the anatomic site and nature of the vasculitis in reported cases is usually not described adequately. The term vasculitis may refer to large or small venous or arterial vessels or both in many published reports. Because arteriolitis is uncommonly described, it is difficult to link it to pathogenesis or agents. Many would assume that any D. immitis-associated vessel lesions would be linked to pulmonary hypertension arteriopathy (PHA) described in pulmonary arteries of humans and dogs, and caused by a variety of processes leading to pulmonary hypertension (17). Also, PHA also can cause some arteritis (18). However, the lesion of PHA is different from that of our case, and our subject had multi-organ arteriolitis. The possibility of a second, circulating, infectious agent causing the acute-onset lesion of our dog was considered. Leishmaniasis is associated with widespread “systemic vasculitis” that is seen as necrosis of arterioles in dogs; however, no Leishmania organisms were seen, and the lesions in leishmaniasis are associated with pleocellular cuffing not seen in our dog (19). Novel agents such as canine circovirus and canine bocavirus are described to cause vasculitis of small vessels and hemorrhagic enteritis, but again, their lesion is limited to the bowel (20,21). Interestingly, the vessel lesion of canine circovirus is similar to that of our case. Classic, type III immune reactions have been described in dogs exposed to repeated doses of human serum albumin (22,23). This results in lesions similar to the arteriolitis we saw. The lesions seen in our case could have resulted from episodic release of filarial antigens. It is possible that release of filariae on the systemic side permitted delivery in a more noxious manner. Experiments with soluble dirofilarial antigens may show this to occur, but have not been performed to our knowledge. Finally, filarial antigen release may have led to an autoimmune condition like that of Henoch-Schonlein purpura (HSP), a transient, though severe, and treatable disease of children that is characterized by widespread arteriolitis similar to our case (24). Again, being aware that a treatable HSP-like disease syndrome could result from D. immitis may provide a more favorable outcome for prepared practitioners. These theories linking D. immitis with arteriolitis are as yet unproven and would require further research to substantiate.

How the heartworms crossed to the left side of the circulatory system in this case is unclear. Previous reports have documented probable migration through a cardiovascular shunt (9,10,25) although other reports of aberrant heartworms have not demonstrated evidence of a cardiovascular shunt (8–10,25–27). In the dog in this report, heartworms were in the heart and lungs at necropsy, demonstrating the dog had a right-sided heartworm infection, but no cardiovascular shunts were identified. Therefore, presumably aberrant worm migration resulted in the presence of heartworms in the aorta. Arteriolitis may have played a role in aberrant migration, or may have been the result of aberrant migration.

To our knowledge, this is the first case of aberrant migration of D. immitis with systemic arteriolitis and presenting primarily with GI signs. The primarily GI presentation of the dog in this report was atypical for the clinical signs commonly seen with heartworm infection. Aberrant worm migration may have caused thromboembolism that resulted in ischemia and hemorrhage in the GI tract. Thus, it should be known that migration of heartworms in the aorta may cause thrombus formation and may present with primary clinical signs of hemorrhagic diarrhea, as in this case. This dog also had systemic arteriolitis diagnosed on necropsy that further contributed to the coagulation changes and thrombosis. Therefore, heartworm disease with aberrant migration should be considered in dogs with possible heartworm disease that demonstrate signs of hemorrhagic gastroenteritis. CVJ