Introduction

A 19-year-old 225-kg male bottlenose dolphin (Tursiops truncatus) under managed care presented with acute inappetence and avoidant behavior. The animal was alert and responsive, with normal respiratory rate and character. Vaccination against Erysipelothrix rhusiopathiae was current. Complete blood count revealed mildly decreased hemoglobin (12.7 g/dL; reference range [RR], 13.5 to 15.5 g/dL¹), 9% decrease in Hct (47% to 38%; RR, 38% to 44%; Cell Dyn 3700 hematology analyzer; Abbott Laboratories), and 7% drop in spun PCV (42% to 35%) compared to a routine sample 10 days prior. Serum biochemistry analytes were within RR except for alkaline phosphatase (120 U/L; RR, 300 to 1300 U/L). There was a decrease in serum iron (376 μg/dL to 260 μg/dL; RR, 120 to 340 μg/dL) and increase in erythrocyte sedimentation rate (13 to 77 mm/h; RR, not available; changes in alkaline phosphatase, iron, and erythrocyte sedimentation rate are typical for systemic inflammation in marine mammals).^2,3 Levofloxacin (5 mg/kg, PO, q 12 h), metronidazole (7 mg/kg, PO, q 12 h), vitamin K (0.4 mg/kg, PO, q 24 h), nystatin (7,000 IU/kg, PO, q 8 h), and bismuth subsalicylate (525 mg total dose, PO, q 12 h) were prescribed empirically. Ultrasound revealed a large-volume left-sided pleural effusion (approx fluid depth, > 20 cm), lung consolidation, and a small amount of peritoneal effusion. Large echogenic flocculent particles were present within the pleural fluid (Figure 1). The animal was sedated with 50 mg of diazepam PO. A diagnostic thoracocentesis was performed out of water following a surgical scrub of the left side of the thorax, by use of a 20-gauge 6-inch spinal needle attached to a syringe, and 35 mL of fluid was collected for cytologic evaluation and culture. The fluid was red and cloudy with a PCV of 5%, total solids of 6.3 g/dL, and specific gravity of 1.036. Microscopically, the fluid had abundant RBCs and mononuclear phagocytes (ie, mesothelial cells and macrophages; Table 1). An unidentified gram-positive and catalase-positive coccobacillus and Staphylococcus capitis (possibly a contaminant; VITEK 2 Systems Analyzer version 07.01; BioMérieux) were identified on culture of the first fluid sample, and blood culture was negative. Doxycycline (1.5 mg/kg, PO, q 12 h) was added to the treatment regimen on the basis of the sensitivities of both organisms. One week later, repeat CBC was within RR and serum biochemistry analytes revealed icteric serum with elevated bilirubin (1.02 mg/dL; RR, 0.1 to 0.2 mg/dL), iron (461 μg/dL; RR, 120 to 340 μg/dL), and lactate dehydrogenase (1122 U/L; RR, 350 to 500 U/L).

View Full Size

Left thoracic ultrasonographic image demonstrating a large-volume echogenic pleural effusion with hyperechoic flocculent particles in a 19-year-old 225-kg male bottlenose dolphin (Tursiops truncatus) with acute inappetence and avoidant behavior. The image was acquired by use of a 3 MHz curvilinear transducer with the animal resting in water in right lateral recumbency. The marker (GE) on the left denotes cranial, and the scale on the right denotes depth in centimeters.

Citation: Journal of the American Veterinary Medical Association 260, 9; 10.2460/javma.21.06.0311

• Download Figure
• Download figure as PowerPoint slide

Initial differential diagnoses included hemorrhage, pleuropneumonia, thoracic or abdominal abscess, coagulopathy, toxin exposure, neoplasia, diaphragmatic hernia, lung lobe torsion, cholangiohepatitis, or pancreatitis +/− pancreatic pseudocyst.⁴ A coagulation profile (activated partial thromboplastin time, 159.3 seconds; prothrombin time, 21.6 seconds; thrombin clotting time, 6 seconds) was not significantly different when compared to a healthy conspecific (activated partial thromboplastin time, > 180 seconds; prothrombin time, 21.2 seconds; thrombin clotting time, 6 seconds).⁵ Fecal cytology, culture, and evaluation for parasites were unremarkable. Shortly after initiating medical management, the animal appeared comfortable and eupneic, with normal behavior and appetite and only mild trepopnea during in-water thoracic ultrasound in right lateral recumbency (not repeatable in left lateral recumbency). Hematological and serum biochemical abnormalities normalized; however, the effusion persisted.

In terrestrial animals, massive pleural effusion warrants chest tube placement⁶; however, in an aquatic animal, this would be difficult to maintain and a nidus for infection or water infiltration. Repeated pleural drainage was planned instead for therapeutic and diagnostic purposes. Drainage was performed 7 times over a 24-month period, by use of the technique described herein, with a total volume of 21.52 L removed. Thoracic ultrasounds were performed regularly to monitor fluid depth, and drainage procedures were staged to minimize handling stress for the animal. Once fluid was no longer appreciable, thoracic ultrasounds, CBC, serum biochemistry analysis, and serum amyloid A (SAA) concentration measurements were performed monthly for monitoring. Drainage was reinitiated once pleural fluid depth exceeded 5 cm and SAA levels increased above the RR (0.15 to 5.98 mg/L; University of Miami Avian and Wildlife Laboratory).

Throughout treatment, SAA concentration was the only abnormal serum marker^7,8 and concentrations were positively associated with amount of pleural fluid seen ultrasonographically (Table 1). The fluid became less hemorrhagic and more straw-colored with each procedure. At no time did the fluid appear milky, opalescent, or typical of a chylous effusion. Following the first drainage procedure, nystatin was discontinued and voriconazole initiated (2.5 mg/kg, PO, q 12 h for 3 days, followed by 2.5 mg/kg, PO, q 7 d) empirically for possible primary or secondary fungal infection. Serum voriconazole levels were monitored regularly by high performance liquid chromatography assay (UT Health South Texas Reference Laboratory; lowest limit of quantification is 0.2 µg/mL) and ranged from 1.35 to 3.31 µg/mL during treatment. In humans, there is no established RR; however, concentrations of < 1.0 µg/mL may be suboptimal while levels > 6.0 µg/mL may be associated with toxicity.⁹

In addition to pleural drainage, levofloxacin, voriconazole, s-adenosylmethionine and silybin (1275 mg/360 mg total dose, PO, q 24 h), and tramadol (0.1 mg/kg, PO, q 12 h) were the mainstay of medical management until normalization of SAA and absence of pleural fluid was confirmed ultrasonographically, 14 months after presentation.

Pleural fluid drainage

Diazepam (0.2 mg/kg, PO, once) was administered 2 hours prior to each procedure (each procedure was performed similarly each time). The animal was lifted in a stretcher and placed in sternal recumbency on a bed of thick foam. A bolus of midazolam (0.05 mg/kg, IV, once via the ventral fluke periarterial venous rete) was administered on arrival in hospital. Respiratory rate, heart rate, and ECG complexes were continuously monitored, and flow-by oxygen was administered.

Drainage site selection was based on maximum depth of fluid on ultrasound (approx 20 cm), approximately between the 7th to 9th rib spaces, as for domestic mammals.¹⁰ The skin was surgically prepared and a local anesthetic block (1:1 lidocaine and sodium bicarbonate solution) administered. Following a second surgical scrub, a stab incision was made into the skin with a size 11 surgical blade. A closed system was used consisting of a 14-gauge 120-mm insufflation needle (Surgineedle; Covidien) attached to an extension line, 3-way tap, and 60-mL Luer-lock syringe (Figure 2). With the tap in the off position, the needle was inserted into the pleural space under ultrasound guidance, through the skin incision and intercostal muscle at the cranial aspect of the rib.

View Full Size

Equipment used for pleural drainage in the bottlenose dolphin of Figure 1. A—A 14-gauge 120-mm insufflation needle is attached to a 3-way tap and 60-mL Luer-lock syringe. An extension line connects the 3-way tap to surgical suction tubing via a tubing adapter. B—Pleural drainage procedure. The animal is sedated on a foam bed in the hospital treatment room. Following aseptic collection of diagnostic samples into a 60-mL syringe, the 3-way tap was turned on to a surgical suction unit.

Citation: Journal of the American Veterinary Medical Association 260, 9; 10.2460/javma.21.06.0311

• Download Figure
• Download figure as PowerPoint slide

The tap was turned on to the syringe, and a sample was aspirated for cytology, aerobic culture and susceptibility testing, generic fungal PCR, generic bacterial PCR, Nocardia, Brucella, Streptococcus phocae PCR, mycobacterial culture and PCR, coronavirus PCR, and virus isolation (Cornell University Animal Health Diagnostic Center; fluid sample for virus isolation was inoculated in Vero DogSLAM cells and bottlenose dolphin epidermal cells). All were negative.

The tap was then turned on to a 1/4” × 12’ sterile suction hose (Jorgensen Labs) connected to a surgical suction unit, and fluid was aspirated by means of suction (Figure 2). Continuous suction at a pressure of approximately 15 cm H₂O was used, and insufflation needle tip location relative to lung tissue and level of fluid were monitored by means of ultrasound. A total volume of 5 L of fluid was removed before the animal became mildly tachypneic, prompting discontinuation of the procedure. Sedation was antagonized with flumazenil (0.02 mg/kg, once, administered half IM into the epaxial muscles and half IV via the ventral fluke periarterial venous rete).

Fluid analysis from all samples is summarized (Table 1). A blood sample collected during the first procedure was submitted for culture (no growth), CBC and serum biochemistry analysis (unremarkable), erythrocyte sedimentation rate (20 mm/h), fibrinogen (311 mg/dL), fungal serology by immunodiffusion (negative for Candida, Cryptococcus, Coccidioides, Histoplasma, Aspergillus, and Blastomyces antibodies), serum amylase (< 10 U/L) and lipase (19 U/L), SAA (30.7 mg/L), Brucella PCR (negative), Apophysomyces spp antibody ELISA (negative), and anticoagulant rodenticide screen (none detected).

Right lateral and dorsoventral radiographs following drainage revealed a right mediastinal shift and possible rib fracture in the cranial left thorax on the dorsoventral view. This was not confirmed on lateral projections.

Upon return to the water, the animal was bright, alert, responsive, and respiration rate returned to normal limits (1 to 4 respirations/min). Follow-up thoracic ultrasound revealed a marked reduction in pleural fluid volume with no signs of alveolar interstitial syndrome¹¹ or pleural surface disease. Tachypnea was not observed during subsequent drainage procedures.

Fluid cytology

Direct pleural fluid smear preparation, collected following the first drainage procedure, was highly cellular and adequately preserved on a pale eosinophilic background with frequent erythrocytes (Table 1). A highly cellular inflammatory cell population was present with a predominance of mononuclear phagocytes (including macrophages and mesothelial cells) that frequently contained abundant small, distinct, clear vacuoles. These cells were often binucleated or rarely multinucleated. Infrequent mononuclear phagocytes contained phagocytized erythrocytes or a small amount of greenish basophilic hemosiderin pigment, which was confirmed to contain hemosiderin by Prussian blue reaction. Fewer small, well-differentiated lymphocytes and mature nondegenerate neutrophils were also noted. No infectious agents or overt neoplastic cells were identified.

In a follow-up sample 3 months later, the oil red O stain (for visualizing lipid¹²) was mildly focally positive in the background of the preparation; frequent mononuclear phagocytes exhibited mild diffuse or strong focal positive staining within vacuoles, consistent with phagocytized lipid material (Figure 3).

View Full Size

Cytologic examination of a pleural fluid sample from the bottlenose dolphin of Figure 1. Pleural fluid sediment smear preparation with Wright Giemsa stain (panel A) and oil red O stain (panel B). Abundant mononuclear phagocytes with intracytoplasmic clear vacuoles are recognized as lipid material by oil red O stain. Small well-differentiated lymphocytes are also noted. Objective X100; bar = 100 µm.

Citation: Journal of the American Veterinary Medical Association 260, 9; 10.2460/javma.21.06.0311

• Download Figure
• Download figure as PowerPoint slide

By the fourth drainage procedure, the fluid was straw colored and cloudy (Table 1). The cytologic and biochemical findings (eg, low triglycerides and high cholesterol) were similar to previously submitted samples and were consistent with persistent chronic nonchylous lymphatic effusion.

Lipidomics

Nontargeted semiquantitative lipidomics of serum and thoracic fluid samples was performed on both patient (n = 5 serum and 5 fluid; same individual at different collection points in time) and control serum samples (3; different healthy adult bottlenose dolphins). Lipid extracts were obtained using chloroform/methanol containing known amounts of labeled internal standards (EquiSplash mix; Avanti Lipids) for each lipid class. Lipid extracts were analyzed using a nontargeted lipidomics method by high-performance liquid chromatography and tandem mass spectrometry (Vanquish HPLC system and Q-Exactive Orbitrap mass spectrometer; Thermo Scientific). Spectra were acquired in full-scan positive and negative ion modes, and tandem mass spectrometry was performed by use of data-dependent acquisition for top 10 ions. Lipid identifications and peak areas were determined by use of LipidMatch software.¹³ Totals of 711 and 719 lipids were identified in serum and fluid samples, respectively (Figure 4).

View Full Size

Distribution of lipids in serum and pleural fluid for the bottlenose dolphin of this report. Panel A—Number of lipid identifications. Panel B—Semiquantitative amounts within total lipids. AcCar = Acylcarnitine. CE = Cholesteryl ester. Cer-NS = Ceramide nonhydroxy fatty acid–sphingosine. DG = Diacylglycerol. DMPE = Dimethyl-phosphatidylethanolamine. LPC = Lysophosphatidylcholine. PC = Phosphatidylcholine. PE = Phosphatidylethanolamine. PI = Phosphatidylinositol. PS = Phosphatidylserine. SM = Sphingomyelin. TG = Triacylglycerol.

Citation: Journal of the American Veterinary Medical Association 260, 9; 10.2460/javma.21.06.0311

• Download Figure
• Download figure as PowerPoint slide

There was overlap between serum and fluid samples following principal component analysis (Supplementary Figure S1), but there was no separation between patient and control serum samples. Fifteen lipids were found to be significantly increased in patient serum samples, compared with the control samples (Supplementary Figure S2). This included 3 ceramides, a lysophosphatidylserine, and 11 ether phospholipids (P < 0.05 for all; Figure 4), which have been implicated in various conditions, including inflammation, cell death, and neoplasia. Interestingly, the same ether phospholipids that were significantly higher in the serum from the patient, when compared with serum from healthy individuals, were also detected in the fluid.

Advanced diagnostic imaging

A noncontrast thoracic CT was obtained following the first drainage procedure, with the animal in sternal recumbency, under sedation with 12 mg midazolam IV.

A large-volume pleural effusion occupied the left hemithorax (Figure 5). It was evenly distributed with a maximum peripheral depth of 9 cm (perpendicular to the lung surface). The fluid was homogeneous in appearance, with an attenuation of 22 to 35 HU (water = 0 HU). Irregular, variably sized, soft tissue pleural nodules were noted in a scattered distribution in the left hemithorax. A hyperattenuating, ovoid soft tissue mass (50 HU) measuring 2.7 × 5.5 cm in cross section was also apparent, suspended in the pleural fluid.

View Full Size

Cranial to caudal (panels A to D) transverse CT images revealing a large-volume left-sided hypoattenuating pleural effusion (white arrow) in the bottlenose dolphin of this report. Images were obtained (Revolution GSI CT unit; GE Healthcare) by use of a standard kernel, 140 kV, auto mA, and 1.25-mm slice thickness and are displayed in a soft tissue window (window width, 350 HU; window length, 50 HU). The right side of the animal is on the left side of each image. Image A is the most cranial and D is the most caudal.

Citation: Journal of the American Veterinary Medical Association 260, 9; 10.2460/javma.21.06.0311

• Download Figure
• Download figure as PowerPoint slide

A severe right-sided mediastinal shift was present with the heart markedly displaced away from the fluid. There was complete loss of conspicuity of the thyroid and common brachiocephalic vein, structures typically seen on thoracic CT of Tursiops sp.¹⁴ The left lung exhibited severe loss of volume (approx 50%; Supplementary Figure S3) and a corresponding diffuse increase in attenuation and was retracted cranially and caudally.¹⁵ The ventral left lung narrowed to a 2- to 3-mm-wide ribbon-like strip that was highly folded within the fluid. Small foci of parenchymal infiltrates were noted in the left caudal lung, near the diaphragm. There was no CT evidence of lung lobe torsion. There was a single 2 × 2-cm well-defined focus of irregular parenchymal consolidation in the right midventral lung. The remainder of the right lung was within normal limits.

The left dorsal mediastinum was abnormally expanded with a large, tubular structure displacing the aorta to the right. It contained hypoattenuating fluid (5 to 11 HU) and extended from the level of the aortic arch to the caudal thorax. Within the caudal mediastinum, the structure extended ventral to the aorta and became more dilated. The esophagus, distinct from the aorta and this structure, was also displaced to the right. Primary consideration was given to a dilated thoracic duct (Figure 6). Another smaller dilated tubular structure 2 × 3 cm in cross section was present in the retrosternal region, running craniocaudally through the effusion, dorsal to the sternum, and to the left of the heart.

View Full Size

Transverse CT images of thoracic structures in the bottlenose dolphin of this report. A—CT image following first pleural fluid drainage procedure with relevant structures labeled. B—CT image obtained 6 months later following 4 thoracic drainage procedures. The esophagus, aorta, and suspected dilated thoracic duct (white arrow) are pictured. The right side of the animal is on the left side of each image.

Citation: Journal of the American Veterinary Medical Association 260, 9; 10.2460/javma.21.06.0311

• Download Figure
• Download figure as PowerPoint slide

A left-sided cranial abdominal effusion was present with notable displacement of cranial abdominal organs to the right. A rough estimate of the area of fluid was > 20 cm in depth × 20 cm in height × > 33 cm in length. Differentials included chyloabdomen (rare but has been reported in conjunction with chylothorax),¹⁶ transudate secondary to intrathoracic disease and potential obstruction of lymphatics, or exudate secondary to abdominal abscess. The caudal vena cava was abnormally dilated to 5 to 6 cm at the level of the diaphragm, consistent with passive congestion.

The left first sternal rib was fractured, resulting in a 2.2-cm length of free sternal rib between its junction with the vertebral rib dorsally and the fracture ventrally. There was also a healed fracture of the left fourth rib.

Following repeated drainage procedures, a 6-month follow-up noncontrast thoracic CT in sternal recumbency was obtained under sedation with 12.3 mg of midazolam IV. Contrast CT was attempted, but motion artifact hindered interpretation despite sedation. Images demonstrated significant reduction in the pleural effusion, with only a small amount of fluid remaining in the left cranial thorax. The previously observed fractures of the left ribs, as well as the pulmonary parenchyma, could not be reassessed due to the degree of motion present during image acquisition. There was, however, considerable expansion of the left lung relative to the initial study. The re-expanded left lung had a much more normal, hypoattenuating appearance; however, the ribbon-like ventral margin persisted.

Thickened tissue (76 HU) had developed along the left ventral pleural surface where there was previously fluid, and a large soft tissue mass was again appreciable to the left of the heart, rounded cranially and tapering caudally. It was separated from the heart by a faintly hypoattenuating margin. The tissue of origin could not be determined in the noncontrast study, but differentials included pulmonary, lymph node, pleural, or mediastinal origin.

The dilated fluid-filled structure seen in the left dorsal mediastinum on the first CT, suspected to be dilated thoracic duct, was difficult to assess completely due to motion. It was identified in select caudal thoracic images to the left of the descending aorta (Figure 6).

Outcome

Drainage was performed 7 times over a 24-month period, by use of the technique described, with a total volume of 21.52 L removed. Serial welfare evaluations performed by professional animal care staff determined that the animal continued to have a good quality of life during this period, and the decision to euthanize was made as soon as assessment indicated otherwise.

Recrudescence of the left-sided effusion was diagnosed 19 months after presentation, during a routine ultrasound. No clinical signs were present. Effusion persisted despite reinitiation of drainage procedures and medical management. The animal was humanely euthanized 27 months after presentation, following acute clinical decline, development of pericardial effusion diagnosed by ultrasound, and indication by animal welfare assessment.

Necropsy revealed healed mid-diaphyseal fractures of the fourth and fifth ribs, with the previously reported first rib fracture fragment not appreciated. There was a stable callus present at both sites with fracture fragments well aligned. The left hemithorax had severe diffuse parietal pleural thickening and contained approximately 1.5 to 2 L of reddish tan fluid with cream flocculent precipitate. The caudal aspect of the left lung was congested with rounded caudal margins and a 14-cm-long area of markedly thickened pleura. The mediastinum had an area approximately 7 cm long by 5 cm wide of nodular soft tissue associated with the esophagus at the level of the heart base that was light tan with discrete black and gray areas 3 to 5 mm in diameter. The heart was grossly and microscopically unremarkable except for 2 incidental blood cysts on the mitral valve. The heart weighed 925 g, and relative (to body) heart weight was similar to other adult collection T truncatus (N Stedman, DVM, PhD, DACVP, Busch Gardens, email communication, 08/13/2020).

Microscopically, pleura was up to 8 mm thick and consisted of vascular dense collagenous connective tissue with infiltrates of siderophages. Pleura was covered by hypertrophied mesothelium with mild multifocal hyperplasia. Bronchioles in the pulmonary parenchyma subjacent to thickened pleura and at the lung hilus were partially filled with macrophages and neutrophils, but infectious agents were not identified. The lungs were otherwise microscopically unremarkable. The nodular mediastinal tissue consisted of dense collagenous connective tissue with frequent small aggregates of siderophages and multifocal entrapped adipose. Multiple lymph nodes that were either unremarkable or contained small numbers of sinusoidal or medullary cord pigment-laden macrophages were embedded within adipose within or adjacent to the fibrosis. Remnants of thymic tissue were also identified within the fibrosis. The fibrosis extended into surrounding skeletal muscle of the cranial thoracic body wall with associated lymphocytic and histiocytic myositis and myofiber necrosis and atrophy.

The liver had increased numbers of hemosiderin-laden Kupffer cells and mild acute hepatocellular necrosis around most centrilobular veins consistent with passive congestion.

Final diagnoses were cranial mediastinal and pleural fibrosis with chronic hemorrhage and left pleural effusion, slight multifocal neutrophilic and histiocytic pneumonia of the left lung, chronic passive hepatic congestion with mild multifocal acute centrilobular necrosis, and incidental mitral valve blood cysts.

Discussion

This is the first report of chronic nonchylous lymphatic pleural effusion, likely secondary to rib fractures, in a bottlenose dolphin. The clinical history and prominent mononuclear phagocyte component of the fluid in this case were most consistent with chronicity of effusion. The hemorrhagic component of the fluid indicated mild recent hemorrhage into the thoracic cavity; an additional consideration included diapedesis, possibly from increased hydraulic pressure within the thoracic cavity. The discontiguous rib segment identified on CT may also have contributed to ongoing focal chest wall disruption and exacerbated the effusion.

In humans, lymphorrhagic effusions are differentiated into chylous (ie, chylothorax) or nonchylous (ie, cholesterol, chyliform, or pseudochylous) effusions.¹⁷ A chylous pleural effusion is characterized by triglycerides > 110 mg/dL and cholesterol < 200 mg/dL. This condition often results from thoracic duct rupture and leakage from the lymphatic system into the pleural space.¹⁸ A disruption of the thoracic duct above the fifth thoracic vertebra leads to a left-sided chylothorax in humans, whereas damage below this level results in a right-sided effusion.¹⁸ Nonchylous effusions are rare in human medicine and have been sparsely reported in veterinary medicine.¹⁹ They are characterized by increased cholesterol (> 200 mg/dL) relative to triglycerides (typically < 100 mg/dL) with a ratio of cholesterol to triglycerides > 1.²⁰ Unlike chylothorax, nonchylous effusions do not result from thoracic duct leakage and are often unilateral.²¹ They are associated with chronic pleurisy, with primary human differential diagnoses being infectious (eg, tuberculosis) or autoimmune (eg, rheumatoid arthritis). Empyema, hemothorax, and trauma have been reported but are far less common.^20,22 Although leakage of lymph is often secondary to damaged lymphatic vessels, it has also been associated with occlusion or increased hydrostatic pressure in the vena cava.²³

Nonchylous lymphatic pleural effusions have been associated with cardiac disease in cats,¹⁹ but no evidence of this was observed on necropsy or histopathology in this case. The pathogenesis of nonchylous effusions is not fully understood; however, it is generally accepted that the cholesterol originates from cellular degradation, becomes trapped in the pleural space due to poor reabsorption as a result of a chronically thickened pleura,²⁴ and with chronicity undergoes a change in lipoprotein binding characteristics.²² The observed pleural thickening in this case thus may have been secondary to the rib fractures, thoracocenteses, or chronic effusion itself, as has been previously reported in human patients.²² Nonchylous lymphatic effusions may result from damaged lymphatic vessels that are not located in the drainage path from intestines to the thoracic duct and therefore do not contain chyle or chylomicrons. An additional consideration for lack of triglycerides in these types of effusions includes lack of dietary lipids, which may have contributed here at initial presentation, but anorexia resolved shortly after initiation of supportive care. The thickened pleura may contain cholesterol deposits histologically,²² although this was not observed in this case.

The peritoneal effusion and dilated thoracic structures may have occurred secondary to increased hydrostatic pressure from the pleural effusion; likewise, the space-occupying effect of fluid within the relatively small thoracic and abdominal cavities of a bottlenose dolphin¹⁵ may have accounted for the unusual contralateral shift in viscera away from the fluid observed on CT. While this displacement is typically suggestive of a space-occupying lesion, no abdominal masses were identified in the initial CT study (although only a limited portion of the abdomen was included) or at necropsy.

Thoracocentesis is often performed in cetaceans to characterize effusions for diagnosis¹¹; however, pleural drainage for effusion management is rarely reported.²⁵ Main considerations for the pleural drainage procedure described here include the use of an insufflation needle, needle placement, and care regarding fluid suction. In a previous report describing a similar procedure, the nylon IV catheter used was prone to crimping²⁵; use of an insufflation needle in this case provided a reliably patent bore with an atraumatic tip, which was readily visualized with ultrasound. The needle was placed on the cranial aspect of the rib. This is also recommended for thoracocentesis in domestic animals, as intercostal vessels and nerves are located on the caudal border of each rib.¹⁰ In cetaceans, an extensive network of vessels, known as a rete mirabile, covers the entire thoracic surface¹⁵; thus, placement on the cranial or caudal aspect of the ribs in this species may be equally likely to result in iatrogenic hemorrhage. The cranial aspect was nevertheless chosen in this case to avoid iatrogenic nerve damage.

The observed mild tachypnea after aspiration of 5 L of fluid during the first procedure prompted termination. In human medicine, re-expansion pulmonary edema is an uncommon complication of drainage of large pleural fluid volumes.²⁶ If this was a factor in the present case, it is speculated to have been mild and self-limiting. Edema was not appreciable on radiographs, and the patient became eupneic upon return to water. Tachypnea was not observed during subsequent drainage procedures. In humans with large-volume thoracic effusions, it is recommended that small volumes be drained over consecutive days by use of low negative suction pressure (< 20 cm H₂O).²⁶ Pressures of approximately 15 cm H₂O were used in this case, and procedures were staged over longer intervals to reduce stress and potential adverse effects associated with multiple out-of-water procedures. The use of pleural manometry is also recommended for monitoring, although it was not available for this case.^26,27

The nontargeted lipidomics analyses performed herein resulted in identification of various lipid compounds (ceramides, lysophosphatidylserine, and ether phospholipids), which could possibly serve as biomarkers if based on proper validation studies. It was limited in its diagnostic utility for this case due to lack of available control thoracic fluid, which is of negligible volume in healthy dolphins. With further validation that these could be true biomarkers for this disease process, targeted mass spectrometric assays that are quantitative and affordable may have potential. In human medicine, lipidomic profiling of pleural effusions has resulted in discovery of biomarkers for neoplasia and tuberculosis with excellent diagnostic performance.^28,29

SAA was a valuable diagnostic in this case.^30–32 Although a recent study found that SAA concentrations were not significantly higher in sick dolphins compared to healthy dolphins, increased SAA did correlate with increased inflammatory markers, such as erythrocyte sedimentation rate, WBC, and fibrinogen.³³ Following rapid normalization of initial clinical signs, CBC, and serum biochemistry analysis, SAA remained the only abnormal marker in this case throughout the treatment period. SAA decreased shortly after each drainage procedure, and increases in SAA preceded increases in pleural fluid depth seen on ultrasound. SAA thus served as a valuable noninvasive measure to evaluate clinical status, despite no identifiable underlying infection. The interventions utilized resulted in a survival time of 834 days from the date of initial presentation.

The most common pulmonary pathological condition in wild and managed cetaceans is pneumonia^11,34 and treatment of possible primary infection was a vital part of clinical management in this case, given the initial fluid culture and CT results. The nature and chronicity of the disease also put the animal at significant risk of immunosuppression and thus prevention of secondary infection was essential. Mild multifocal pneumonia was apparent at necropsy despite the absence of infectious agents.

Rib fractures are commonly found in wild dolphins at necropsy,³⁵ and in 1 report,³⁶ half of all cetaceans diagnosed with intra-interspecific trauma presented bone fractures, with the thorax being the most affected region. Conservative management is generally recommended in human medicine for cases with large-volume pleural effusions¹⁸ and most human rib fracture cases heal without surgical intervention.³⁷ The described pleural drainage procedures resulted in resolution of effusion for a 5-month period. It is unclear why recrudescence occurred, but it is speculated that the severity of pleural fibrosis resulted in impaired lymphatic and venous return.

While postcontrast CT attempts were unsuccessful in this case, it is highly recommended and would have provided further insight into the integrity of the lymphatic structures. Considerations for similar cases include lymphangiography, thoracoscopy, and pleurodesis. In chylous effusions, lymphangiography can be useful to identify leakage sites and has been shown to result in clinical resolution in some cases.³⁸ Transcutaneous biopsy of the observed mass-like structures under sedation was considered but was ultimately not attempted to prevent iatrogenic trauma or other complications.

This report described chronic nonchylous lymphatic pleural effusion in a bottlenose dolphin under managed care. While clinical signs were nonspecific, repeated thoracocentesis, diagnostic testing, and imaging were essential for diagnosis, treatment, and clinical management. Good quality of life was maintained for over 2 years after initial diagnosis.

Supplementary Materials

Supplementary materials are posted online at the journal website: avmajournals.avma.org