Fatal Intraoperative Pulmonary Fat Embolism During Cemented Total Hip Arthroplasty in a Dog

Case Report

A 46-kg, 3-year-old, castrated male German shepherd dog was admitted to the University of Florida Veterinary Medical Teaching Hospital (UFVMTH) for unilateral cemented total hip arthroplasty. The dog had a history of lameness and a radiographic diagnosis of severe coxofemoral degenerative joint disease secondary to hip dysplasia. The results of a presurgical complete blood count, serum biochemical analyses, and urinalysis were within reference ranges. The left hind limb of the dog was prepared for surgery.¹²

The acetabular component was implanted. The femoral canal was reamed, lavaged, and suctioned in preparation for implantation of the femoral prosthesis. Polymethylmethacrylate (PMMA)^a was mixed under suction and allowed to cure to a doughy consistency. Multiple, small cylinders of PMMA were packed into the femoral medullary cavity with a stainless steel rod. The PMMA began to harden during implantation of the femoral prosthesis, and adequate seating of the femoral component was difficult. As the femoral prosthesis was being impacted with the stem impactor and mallet, the expired carbon dioxide pressure (pECO₂) of the dog decreased precipitously from 40 mm Hg to approximately 20 mm Hg. The marked decrease in pECO₂ was followed by sinus bradycardia. The arterial blood oxygen saturation (SpO₂) as measured by a pulse oximeter decreased to <90%, and cardiac arrest ensued. Anesthesia was stopped, and cardiopulmonary resuscitation was initiated. Ventricular asystole was diagnosed on an electrocardiograph monitor. The cardiac rhythm converted to a sinus rhythm following three doses of intravenous epinephrine^b (0.1 mg/kg). Atropine^c (0.05 mg/kg) was administered intravenously to increase the heart rate. After approximately 1 minute, the heart stopped again, and cardiopulmonary resuscitation was reinstituted. Epinephrine was administered intravenously (0.1 to 0.2 mg/kg) q 3 minutes. Attempts at cardiopulmonary resuscitation were stopped after 25 minutes.

A complete necropsy was performed the next day. At necropsy, significant findings were limited to the lungs, kidneys, liver, spleen, and coxofemoral joints. All lung lobes were diffusely glistening, dark red, and heavy. On cut surfaces, there was a homogeneous red discoloration, and airways exuded abundant, pink, frothy fluid. Pulmonary arterial or venous thrombi were not evident grossly. Detailed examination of the heart revealed no cardiac malformations or anomalies. Both kidneys were severely congested, enlarged, and the cut surfaces were homogeneously dark red with poor corticomedullary demarcation. Both the liver and spleen were severely congested, mildly enlarged, and exuded abundant blood from the cut surfaces. The right acetabulum was shallow, and the acetabular articular surface was roughened. The articular cartilage of the right femoral head was extensively eroded with eburnation of the underlying bone. The acetabular prosthetic component was present on the left side. The femoral component was removed after the dog expired. Longitudinal sectioning of the left femur revealed a firm, brown material (i.e., PMMA) occupying the distal femoral cortex. The gross lesions observed at necropsy were suggestive of acute pulmonary insufficiency as the cause of death.

At necropsy, tissue samples were collected in 10% neutral-buffered formalin. In addition, samples of lung, liver, spleen, and kidney were frozen (without a fixative). Formalin-fixed tissue samples were embedded in paraffin, cut into 5 μm-thick sections, and stained with hematoxylin and eosin (H&E). Frozen tissues were sectioned and stained with an oil red O stain for demonstration of lipid.

Histopathological examination of the H&E-stained sections of lung revealed severe congestion of the alveolar capillaries, larger pulmonary veins, small arterioles, and arteries. Alveoli were extensively flooded with edema fluid accompanied by hemorrhage. The most remarkable histopathological change in the lung was the presence of intravascular lipid droplets within many veins and, to a lesser extent, within the capillaries and small arterioles. The intraluminal lipid appeared as either single or multiple, well-defined, clear, circular spaces that displaced the erythrocytes and precursor cells to the periphery of the vessel lumen [Figure 1]. Clusters of hematopoietic precursor cells (erythroid and myeloid) were also present in many pulmonary vessels. In addition to the lipid deposits and hematopoietic cells, a few larger pulmonary veins contained small fragments of an amorphous, pale, eosinophilic material. The pale, eosinophilic material was presumed to be PMMA, but this could not be confirmed. The presence of intravascular fat was confirmed by histopathological examination of the frozen sections stained with oil red O. The frozen sections revealed variably sized, bright red globules within pulmonary vessels [Figure 2]. In some smaller vessels, the lipid globules occupied most of the vascular lumen.

The liver was severely congested, and clear spaces consistent with lipid were observed within central veins and occasionally within hepatic sinusoids. Similarly, the kidneys were moderately congested with intravascular fat within veins. Frozen sections stained with oil red O revealed intravascular lipid in the liver and kidneys. In addition, fragments of an amorphous, pale, eosinophilic material (similar to that seen in the lung) were observed within veins of the liver and kidneys. Congestion was present in the brain and heart, but intravascular lipid was not found in these tissues. The death of the dog was attributed to pulmonary insufficiency associated with fat embolization.

Discussion

Numerous intraoperative complications are associated with cemented total hip arthroplasty in humans, including systemic hypotension, reduced cardiac output, increased pulmonary arterial pressure, and hypoxemia.^3–9 Similar complications have been observed in dogs used as animal models.^6131516 Cardiac arrest and death following total hip arthroplasty in humans have a reported incidence of 0.6% to 10% and 0.02% to 0.5%, respectively. Reports of cardiac arrest or mortality in dogs undergoing experimental or clinical total hip arthroplasty are exceedingly rare.^10–12 Researchers at the University of Munich reported four deaths associated with total hip arthroplasties performed in 375 dogs.¹³¹⁴ Details concerning the clinical course or gross postmortem findings in those dogs were not given; however, fat and hematopoietic precursor cells were observed to be occluding pulmonary vessels on histopathology in three of the dogs.¹³¹⁴

The pathogenesis of complications associated with total hip arthroplasty is often unclear and may be multifactorial. Possible causes include the pharmacological and vasoactive effects of PMMA, pulmonary embolization of fat or bone marrow particles, and activation of the clotting cascade with resultant disseminated intravascular coagulation. Polymethylmethacrylate has been shown to decrease peripheral vascular resistance and arterial blood pressure when injected intravenously in dogs.¹⁵¹⁶ Despite proven vasoactive effects of PMMA, several experimental studies have disputed PMMA’s role as a causative factor of the intraoperative complications associated with total hip arthroplasty.⁶⁸¹⁷¹⁸ The amount of PMMA necessary to cause decreased pulmonary function is 35 times the amount released into the blood in humans during total hip replacement.¹⁶¹⁸ In addition, serum levels of the methylmethacrylate monomer were undetectable in dogs that experienced intraoperative complications during experimental total hip arthroplasty.⁶ Experiments using bone wax and saline as a substitute for PMMA have proven that the presence of PMMA is not necessary for the development of systemic and vasoactive abnormalities during total hip surgery.⁶¹⁷¹⁹

Embolization of fat and bone marrow is believed to be the primary cause of intraoperative complications associated with total hip arthroplasty in humans.^5920–22 The emboli cause obstruction of pulmonary and cerebral blood vessels. The source of the embolized fat is the bone marrow. The embolization of fat and bone marrow particles during arthroplasty occurs from an increase in pressure in the femoral medullary canal during insertion of PMMA and the femoral prosthesis.^20–22 A study in rabbits determined that a pressure of 5 to 10 cm water (3.75 to 7.5 mm Hg) within the medullary cavity was sufficient to produce pulmonary embolization.²² Although direct comparisons cannot be made between rabbits and dogs, medullary canal pressures of 250 to 900 mm Hg have been measured during total hip surgery in dogs.²¹ Reindl and Matis observed emboli of unknown origin within the right ventricle via transesophageal echocardiography in 62 of 62 dogs undergoing total hip arthroplasty.¹⁴

Mechanical obstruction of blood vessels by fat, bone marrow particles, or both, is not the only cause of physiological effects associated with total hip replacement. Damage to endothelium and activation of intravascular coagulation may also play a role. Free fatty acids released from fat globules are toxic to endothelium.²³ Damaged endothelium results in platelet aggregation, release of inflammatory cytokines, and leukocyte recruitment. Thromboplastin is released from the damaged endothelium, resulting in activation of the extrinsic coagulation cascade.¹⁴²³ The combination of these factors increases the prothrombotic potential within blood vessels, resulting in occlusion of blood vessels by thrombi in addition to those already occluded by fat or bone marrow.

The reason for the development of severe embolism and cardiac arrest in this dog was unclear. The surgical protocol used in this dog has been used at the UFVMTH for >15 years, with only one other intraoperative death. The death of that dog was also suspected to be a result of a fat embolism, but the cause was not confirmed at necropsy. In the dog reported here, the PMMA had begun to cure as the femoral prosthesis was being implanted, making it necessary to forcibly seat the femoral component with a stem impactor and mallet. This may have produced high intramedullary pressures, thereby potentiating the embolization of fat, marrow particles, and the foreign material presumed to be PMMA. Orsini et al. found that the intramedullary pressure must exceed a critical threshold in order to generate enough pulmonary emboli to become physiologically important.⁶ In that study, 10 times the number of pulmonary emboli were produced with intramedullary pressures >150 mm Hg, in contrast to pressures <150 mm Hg.⁶

Although mortality from fat embolization is rare in dogs undergoing total hip arthroplasty, measures can be taken to reduce fat embolism. High-pressure, high-volume pulsatile lavage reduced the number of fat emboli and the resulting cardiorespiratory changes produced during experimental total hip arthroplasty in dogs.⁸ Venting the distal medullary canal may reduce intramedullary pressure and the subsequent embolization of fat.¹⁷ A drawback to the venting procedure in dogs is that the femur may fracture at the vent hole. Experimentally, the use of noncemented femoral prostheses significantly reduced intramedullary pressures compared to cemented prostheses.⁶ The use of noncemented prostheses generated lower mean pressures, resulted in fewer pulmonary emboli, and was not associated with significant cardiorespiratory changes.⁶

Conclusion

Deaths from fat embolization in dogs are exceedingly rare. Pulmonary fat embolism should be suspected when dogs undergoing cemented total hip arthroplasty develop acute respiratory distress or die suddenly during the surgical procedure. Gross necropsy findings may be suggestive of pulmonary embolism, and the diagnosis may be confirmed by microscopic examination of formalin-fixed tissues or by oil red O staining of frozen tissues. Hematopoietic elements may be intermixed with the fat emboli in blood vessels. Increased intramedullary pressures raise the likelihood of fat embolization, and measures can be taken during surgery to reduce the severity of fat embolism and the subsequent systemic response.