Marked Pleural Effusion Causing Right Atrial Collapse Simulating Cardiac Tamponade in a Dog

Introduction

Cardiac tamponade is the compression of one or more chambers of the heart from increased intrapericardial pressure caused by pericardial fluid accumulation.¹ Idiopathic, infective, and neoplastic etiologies are the most commonly diagnosed sources of pericardial fluid accumulation.¹ As fluid accumulates in the pericardial sac, intrapericardial pressure increases and is subsequently transmitted to the cardiac chambers, causing chamber collapse.¹ As intrapericardial pressure increases, it first equalizes with right ventricular filling pressure, resulting in right-sided collapse (cardiac tamponade).¹ With continued increases in intrapericardial pressure, left-sided chamber collapse occurs.¹ The hemodynamic effect of pericardial fluid accumulation depends on the rate and magnitude of change in the intrapericardial pressure, and it can be minimal to clinically very significant.¹

Significant hemodynamic effects are associated with right ventricular collapse, including signs of right-sided heart failure and potentially cardiogenic shock.¹ Right-sided cardiac chamber collapse compromises left ventricular output by decreasing left ventricular venous return, resulting in arterial congestion and venous hypertension.¹ Signs of right heart failure include decreased cardiac output, systemic venous congestion with caudal vena cava distension, hepatomegaly, ascites, and pleural effusion. ¹ The resulting clinical signs of cardiac tamponade are tachycardia, pulsus paradoxus (i.e., inspiratory reduction of arterial blood pressure of >10% compared to expiratory values), pleural effusion, ascites, and systemic hypotension.¹

Echocardiography is a sensitive and specific tool for the diagnosis of pericardial fluid accumulation.^2,3 Not only can the presence of pericardial fluid accumulation be identified, but echocardiography can also demonstrate the presence of chamber collapse. Both right atrial and right ventricular collapse have been sensitive and specific indicators of cardiac tamponade.^1,4–10

Rare case reports of pleural effusion causing cardiac chamber collapse have been reported in humans with hemodynamic consequences that have varied from no effect to severe derangement.^7,9,11–15 Experimental models of pleural effusion with minimal pericardial effusion have shown that pleural effusion can cause echocardiographic signs of cardiac tamponade, with both right atrial and right ventricle diastolic collapse.^16,17 This report describes a case of severe pleural effusion with subsequent right atrial collapse, an echocardiographic indicator of cardiac tamponade, in a dog. No pericardial fluid or gross pericardial disease was observed. Resolution of right atrial collapse occurred after the pleural effusion was resolved.

Case Report

A 16-month-old, 27.7-kg, intact female German shepherd dog was referred with a 14-day history of lethargy, anorexia, and weight loss, and a 2-day history of labored breathing, restlessness, abdominal distension, and sleeping upright. The referring veterinarian had diagnosed pleural effusion prior to referral.

On presentation to the emergency service, the dog was anxious and tachypneic, with marked abdominal efforts to breathe. The abdomen was pendulous and had a palpable fluid wave. Subcutaneous edema of the ventral thorax and emaciation (body condition score of 1/5) were also noted. Thoracic auscultation revealed muffled heart sounds; a left-sided, grade 3/6 systolic heart murmur over the mitral valve; diffusely harsh lung sounds; and decreased respiratory sounds ventrally. Femoral pulses were strong, mucous membranes were pink, and capillary refill time was <2 seconds. The dog was assessed to be 5% dehydrated. Rectal body temperature was normal.

A complete blood cell count (CBC) showed a leukocytosis (17,000 cells/μL, reference range 6200 to 14,400 cells/μL) consisting of a mature neutrophilia (13,600 cells/μL, reference range 3400 to 9700 cells/μL) and a monocytosis (1700 cells/μL, reference range 100 to 1000 cells/μL). Serum biochemical abnormalities included mildly low potassium (3.8 mEq/L, reference range 3.9 to 5.3 mEq/L) and chloride (100 mEq/L, reference range 107 to 117 mEq/L); hypoproteinemia (4.6 g/dL, reference range 5.6 to 7.1 g/dL), hypoalbuminemia (2.9 g/dL, reference range 3.1 to 4.1 g/dL), and hypoglobulinemia (1.7 g/dL, reference range 1.9 to 3.6 g/dL); slightly elevated bicarbonate (26 mEq/L, reference range 15 to 25 mEq/L) and blood urea nitrogen (BUN; 31 mg/dL, reference range 8 to 30 mg/dL); and elevations in phosphorus (6.4 mg/dL, reference range 2.8 to 5.3 mg/dL, likely caused by hemolysis of the sample), aspartate aminotransferase (71 U/L, reference range 16 to 50 U/L), and creatine kinase (391 U/L, reference range 58 to 241 U/L). Serum antithrombin III levels were evaluated because of the panhypoproteinemia and were normal (94%, reference range 75% to 120%).

Therapeutic thoracocentesis was performed, and 1.5 L of a tan-colored fluid was removed from the right side of the chest. The fluid was a pure transudate with a total protein of 1.4 g/dL, 0 to 1 segmented cells per high-power field (hpf), and 2 to 8 red blood cells per hpf. Abdominal fluid had similar characteristics. The dog’s anxiety and respiratory effort markedly improved after thoracocentesis. Subsequent thoracic and abdominal radiographs revealed severe pleural effusion, moderate abdominal effusion, and gas-filled intestinal loops within the thoracic cavity. An irregular diaphragmatic margin and cranial displacement of the right side of the diaphragm indicated the presence of a diaphragmatic hernia [Figures 1A, 1B]. The cardiac silhouette was within normal limits, and there was no continuity between the diaphragm and the cardiac silhouette. Thoracic ultrasonography demonstrated a large amount of pleural fluid and evidence of herniated hepatic tissue and gas-filled intestinal loops.

The dog was admitted to the intensive care unit with surgery planned for the following day. Blood pressure measurements ranged from systolic values of 117 to 140 mm Hg and diastolic values of 75 to 118 mm Hg. Evaluation of arterial blood gases^a demonstrated mild hypoxemia (oxygen pressure [PO₂] of 74 mm Hg, reference range 90 to 100 mm Hg) and a mild metabolic alkalosis (pH of 7.533, reference range 7.35 to 7.45). Carbon dioxide pressure (PCO₂) was 36.9 mm Hg (reference range 34 to 40 mm Hg); bicarbonate was 31 mmol/L (reference range 20 to 24 mmol/L); and base excess was 8 mmol/L (reference range 0 to 6 mmol/L). Until the time of surgery, the dog remained tachycardic (160 beats per minutes) and tachypneic (60 respirations per minute), with decreased lung sounds ventrally and increased respiratory effort.

Evaluation by the cardiology service confirmed a left-sided holosystolic murmur over the mitral valve and palpably weak pulses. Echocardiography revealed the left atrium and ventricle were normal in size limits and were functioning normally. Right atrial diastolic collapse was present, a usual indicator of cardiac tamponade [Figure 2; Video 1, http://www.jaaha.org/cgi/content/full/43/3/157/DC1]. No evidence of intrapericardial fluid accumulation was noted. Doppler examination showed mild mitral regurgitation consistent with the auscultated cardiac murmur. Pleural effusion was suspected to be secondary to the diaphragmatic hernia and was thought to be the cause of the right atrial collapse.

Thoracocentesis was repeated prior to anesthesia, and another 1.5 L of tan-colored fluid was removed from the left thorax. A plasma transfusion (210 mL) and hetastarch^b (5 mL/kg) were administered intravenously (IV) during surgery for oncotic support. Cefazolin^c (22 mg/kg IV) was administered preoperatively and repeated every 2 hours intraoperatively.

A ventral midline celiotomy was performed. Tan fluid (2 L) was removed upon entering the abdomen. Exploration of the abdomen revealed the presence of a large rent in the jejunal mesentery with organized edges, in addition to the presence of numerous other fibrous mesenteric adhesions. Initially, the intestines were dark pink in color, consistent with moderate circulatory compromise. A large (7 cm) diaphragmatic rent was present on the right dorsal aspect of the diaphragm, through the muscular portion and extending to the caval foramen. Intrathoracic location of the right medial and lateral lobes of the liver, a loop of duodenum, and the greater omentum was confirmed. The herniated liver lobes appeared congested. The right medial liver lobe was covered with numerous linear to globoid cystic structures (2 to 4 mm in size), which were fluid-filled and would disappear with digital pressure and then reappear when the pressure was released [Figure 3]. This lobe was firm and dark brown in color. A biopsy of the right medial liver lobe was performed and submitted for histopathological evaluation. The lymphatics were dilated along the intestines and were very tortuous near the biliary tree. The heart was gently palpated and visualized by extending a hand through the diaphragmatic rent. The heart and pericardium were normal on visual inspection and palpation. No pericardial fluid was present. The viscera were returned to the abdomen, and the rent in the diaphragm was repaired. A chest tube was placed through a subcutaneous tunnel prior to closure of the diaphragm.

The dog recovered from anesthesia without complication and was maintained on hetastarch^b for 1 day and IV fluids^d for 4 days after surgery. Over the first 2 days after surgery, the chest tube recovered 285 mL of serosanguineous fluid. The chest tube was removed on the 4th day, after 48 hours without aspiration and no clinical signs of respiratory distress. By the 3rd day after surgery, the serum total protein had returned to normal (6.0 g/dL). Histopathological evaluation of the right middle liver lobe revealed severe vascular congestion with hepatic atrophy, fibrosis, and telangiectatic sinusoidal dilatation. No recognizable normal liver was present.

A repeat echocardiogram done on the 5th day after surgery demonstrated complete resolution of the right atrial collapse and minimal pleural fluid [Figure 4; Video 2, http://www.jaaha.org/cgi/content/full/43/3/157/DC2]. The dog was released from the hospital on the 5th postoperative day, by which time the ventral subcutaneous edema had resolved. A telephone conversation with the owners 4 months later revealed the dog was clinically normal.

Discussion

In cases of diaphragmatic hernia with liver lobe entrapment within the thoracic cavity, it is not uncommon to have bicavitary effusions develop.¹⁸ The dog in this case report had severe pleural and peritoneal effusions associated with organ herniation into the thoracic cavity. On echocardiographic evaluation, this dog had right atrial collapse, a finding usually indicating cardiac tamponade. Resolution of the right atrial collapse was achieved by removal (via thoracocentesis and surgery) of the pleural effusion. Echocardiographic signs of cardiac tamponade without pericardial effusion have not been previously reported clinically in dogs. Experimentally, pleural effusions have been shown to cause cardiac chamber collapse in dogs, and several case reports exist of large-volume pleural effusions causing similar cardiac chamber collapse in humans.^{7–9,11–17}

This dog’s pleural and abdominal effusions were deemed secondary to the herniation of the right liver lobes through the diaphragm. Herniation of the liver can cause increased intrahepatic sinusoidal pressure from compression of the caudal vena cava and hepatic veins.¹⁸ Consequently, hepatic congestion and dilatation of hepatic lymph vessels result in lymph extravasation.¹⁸ The end result is a panhypoproteinemia with reduction in serum globulin and albumin, and a transudative cavitary effusion, as seen in this dog.¹⁹ Gross appearance and histopathology of the right medial liver lobe were consistent with chronic vascular congestion and subsequent hepatic atrophy and fibrosis. The cystic appearance of the liver lobe and the dilated lymphatics seen in the area of the biliary tree and along the intestines were consistent with exudation of albumin and globulin via the mechanism described above.^18,19

Other potential causes of transudative effusions are hypoproteinemia, right heart failure, and chronic hepatic disease.¹⁹ Hypoproteinemia as a primary cause of the observed cavitary effusions was considered unlikely, as those effusions do not generally occur until serum albumin is <1.0 g/dL.¹⁹ The serum albumin in this dog was 2.9 g/dL. Chronic hepatic disease was ruled out because no clinicopathological (e.g., low serum glucose, BUN, or cholesterol) or histopathological evidence of primary liver disease was present. Echocardiography ruled out primary right-sided heart failure, although right atrial collapse could have contributed or exacerbated the effusion via systemic hypertension.¹ It was impossible to know the exact contribution of the right atrial collapse to the continuing effusion.

The resolution of the pleural effusion, abdominal effusion, and return of serum total protein to normal after surgery suggested resolution of the hepatic congestion and lymphatic extravasation after correct repositioning of the liver and other abdominal contents. A urinalysis was not performed to rule out protein-losing nephropathy, but antithrombin III blood levels were normal, which suggested that hypoproteinemia was secondary to extravasation and third-space sequestering of fluid.¹⁹ Serial measurements of colloid oncotic pressure would have been helpful to better quantify the importance of the increased total protein after surgery and to rule out hypoalbuminemia alone as a cause of the effusion. Such measurements were not done based on the clinical improvement of the animal, the cessation of effusion, increased total protein, and resolution of the subcutaneous edema.

At surgery, the diaphragmatic hernia was thought to be traumatic in origin because of scarring and rents in the mesentery, as well as chronic in nature because of the observed peritoneal adhesions and fibrosis of the edges of the tear in the diaphragm. The precise timing of this hernia was impossible to decipher from the history. The laboratory abnormalities were consistent with a stress leukogram and the effects of third-space fluid sequestration on serum electrolytes and protein values.¹⁹

The sensitivity and specificity of right atrial diastolic collapse in predicting cardiac tamponade are nearly 100%.^1,4–10 The right atrium is thin-walled, and diastolic atrial collapse is seen with small elevations of pericardial pressure before right ventricular diastolic collapse occurs.⁵ Thus, right atrial collapse is observed before significant hemodynamic deterioration develops.¹⁰ Experimentally, intrapleural infusions of saline in dogs with insignificant pericardial effusions have caused tamponade-like effects with right ventricular diastolic collapse.^16,17 These studies found that increased intrapleural pressure, created by intrapleural fluid infusion, also increased intrapericardial pressure.^16,17 As intrapleural pressure increased, intrapericardial pressure also increased at a uniform rate.^16,17 The onset of right ventricular collapse occurred at the same intrapericardial pressure whether the pressure elevation was created by intrapleural or intrapericardial fluid infusion.^16,17

The hemodynamic consequences of cardiac chamber collapse are graded and increase as intrapericardial pressure increases; the higher the intrapericardial pressure, the greater the hemodynamic consequence.¹ The onset of right ventricular diastolic collapse occurs when intrapericardial pressure exceeds ventricular diastolic pressure. Right ventricular collapse has been associated with a 20% decrease in cardiac output.^4,17 Interestingly, the hemodynamic consequences of right ventricular diastolic collapse, seen experimentally with intrapleural infusion of saline, were not as severe as intrapericardial infusion when the intrapericardial pressure was the same.^16,17 Right ventricular diastolic collapse occurred at statistically the same intrapericardial pressure, regardless of whether it was caused by intrapericardial or intrapleural infusion (8.17 mm Hg with pericardial infusion versus 9.47 mm Hg with intrapleural infusion).¹⁷ However, intrapleural infusion produced a smaller decrease in stroke volume at the same heart rate, no change in mean aortic blood pressure, less pulsus paradoxus, and higher pulmonary artery blood pressure than did the same intrapericardial pressure created by intrapericardial infusion.^16,17 The mechanism for these observed changes was not known.

Some humans that had echocardiographic signs of cardiac tamponade with right atrial, right ventricular, or left ventricular collapse and significant pleural effusion exhibited no pericardial effusion, while others had small amounts of pericardial effusion.^7,9,11–15 The hemodynamic consequences in these patients varied from none to extremely significant changes. In all reported cases, drainage of the pleural effusion resulted in resolution of echocardiographic abnormalities and clinical signs.^7,9,11–15 In cases with mild pericardial fluid accumulation and significant pleural effusion, echocardiographic signs of cardiac tamponade resolved with thoracocentesis, not pericardiocentesis.^11,12,15

It was impossible to know the full extent of hemodynamic effects in this dog, because blood pressure measurements and cardiac evaluation were performed only after removal of 1.5 L of fluid from the pleural cavity; however, all blood pressure measurements were within the normal range. Femoral pulses ranged from weak to strong, but neither jugular distension nor pulsus paradoxus were identified. Only right atrial collapse was identified on echocardiography; the right ventricle was not collapsed. It appeared this dog did not have severe hemodynamic consequences based on physical examination findings and test results. Right atrial collapse occurs before severe hemodynamic effects develop, further supporting limited hemodynamic compromise in this dog.¹⁰ Measurements of blood pressure at presentation and before and after thoracocentesis would have provided useful information. Monitoring other measurements, such as central venous pressure or right atrial pressure, would also have given valuable information about the cardiac effects of the right atrial collapse. Ideally, a Swan-Ganz catheter^d would have been inserted into the right heart, as this would have provided the most information on cardiac function. Such monitoring was not performed, partly because of the initial lack of recognition of the potential effects of pleural effusion on cardiac function, the improvement of the dog in response to thoracocentesis, and expenses related to this level of monitoring.

Conclusion

Right atrial collapse, a sign typically associated with cardiac tamponade, was diagnosed in a 16-month-old German shepherd dog. The dog had severe pleural effusion associated with a diaphragmatic hernia and liver lobe entrapment; no pericardial abnormalities were identified. Resolution of the right atrial collapse occurred when the pleural effusion subsided after surgical repair of the hernia. Based on the results of this case, thoracocentesis is indicated for significant pleural effusion, and the effect of pleural fluid on the hemodynamic state of the animal should be evaluated by echocardiography. Future studies are needed to determine the type and volume of pleural effusion necessary to cause cardiac chamber collapse and changes typical of cardiac tamponade in the absence of pericardial abnormalities.

Acknowledgment

The authors thank the radiology department at Cornell University Hospital for Animals for providing and interpreting the radiographs.