Animals

Client-owned dogs with a diagnosis of IC were included in the study. Dogs with nonidiopathic chylothorax were excluded from the study. The study was approved by the IACUC at the UC-Davis School of Veterinary Medicine.

Study design

Three anesthetic events were planned for each dog. The first 2 events were aimed at diagnosis and surgical treatment; dogs were placed under general anesthesia the day prior to surgery for cardiac catheterization and CT lymphangiography (CTLA). On the second day, dogs underwent thoracoscopic TDL with or without thoracoscopic pericardiectomy followed by immediate postoperative CTLA. Dogs that had evidence of CPP on the basis of results of cardiac catheterization underwent thoracoscopic pericardiectomy in addition to thoracoscopic TDL (TDL/P group), whereas dogs that lacked evidence of CPP underwent TDL alone (TDL group). The third anesthetic event occurred approximately 3 months postoperatively to repeat the CTLA and evaluate the pattern of chyle flow redistribution postoperatively.

Instrumentation

For general anesthesia, dogs were instrumented with an 18- to 22-gauge cephalic catheter and a 20- to 22-gauge arterial catheter placed in the dorsal pedal artery. All dogs were administered lactated Ringer solution at a rate of 10 mL/kg/h, IV. Induction of anesthesia was performed using an induction protocol at the discretion of the attending anesthesiologist for all 3 anesthetic events.

Preoperative cardiac/pericardial evaluation

Each dog received a routine echocardiogram (EPIQ 7C; Koninklijke Philips NV) prior to cardiac catheterization to evaluate for echocardiographic signs of constrictive pericarditis. For the purposes of this study, all echocardiographic assessments were performed under the supervision of a board-certified cardiologist (JAS) at a single time point at a digital off-cart workstation (Syngo Dynamic Workplace; Siemens Medical Solutions Inc). Investigators were blinded to treatment received. Specifically, echocardiographic cine loops were evaluated for the presence of right atrial enlargement, septal bounce, pericardial thickening, and caval and hepatic vein enlargement. A score of 1 was given to each criteria fulfilled, and these were summated at the end. For the purposes of this study, an echo score ≥ 2 was considered to be consistent with constrictive physiology.

For cardiac catheterization, a vascular introducer was placed in the right jugular vein through a standard modified Seldinger approach. A balloon wedge catheter was directed to the level of the right atrium, and atrial pressure profiles were obtained via standard manometry techniques. The profile was recorded when stable for a minimum of 6 consecutive cardiac cycles. Atrial pressure profiles were evaluated and recorded. The balloon catheter was then advanced to the level of the right ventricle, and the right ventricular pressure profile was also recorded for a minimum of 6 consecutive cardiac cycles. After data recording was complete, the catheter was withdrawn, the introducer was removed, and direct manual pressure was applied to the jugular venipuncture site for a minimum of 15 minutes. The following invasive hemodynamic criteria were considered evidence of constrictive pericardial physiology¹³: mean right atrial pressure > 6 mm Hg and/or right ventricular end-diastolic pressure > 10 mm Hg. Previously reported normal reference values for mean right atrial pressure and right ventricular end-diastolic pressure are 3 mm Hg (± 2 mm Hg) and 0 to 5 mm Hg, respectively.¹³ Right ventricular peak pressure was also measured but not used as a criterion for CPP.

Computed tomography lymphangiography

The CTLA protocol used was similar to that described in previous publications.^14,15 Intra-abdominal lymph nodes were assessed ultrasonographically for their suitability to receive percutaneous injections of 1 to 2 mL of contrast (Isovue-370) using a 27-gauge needle, T-port, and 3-mL syringe. If this technique did not yield a diagnostic study or lymph node size or accessibility was deemed unworthy of attempting lymphangiography, then the popliteal and/or medial iliac lymph nodes were evaluated for potential injection.

The 3-month postoperative CTLA was evaluated to assess for the formation of collateral or missed ducts bypassing the ligation site. Ducts were classified as recanalized/incompletely attenuated if there was contrast flow through ducts that anatomically resembled ducts seen on the preoperative CTLA study and were at the site of previously applied hemoclips. Ducts were classified as collateral ducts if they were not seen on the preoperative study and took a substantially different anatomical course from those seen on the preoperative study. The nature of postoperative chyle flow redistribution at the 3-month postoperative time point was also documented.

Surgical technique

Thoracoscopic TDL was performed using a technique previously described.^2,3 After initial mediastinal dissection, 1 of the following 3 approaches for injection of methylene blue (MB) and/or indocyanine green (ICG) into the lymphatic system was performed: a 5-cm paracostal incision was made to expose a mesenteric lymph node for injection, injection was performed after a surgical cutdown to a popliteal lymph node, or injection was performed subcutaneously in a 4-quadrant fashion in the perineal area with subsequent massage of the injection sites, as has been previously described in cats.¹⁶ Either 1:1 diluted MB (American Reagent Inc) with sterile saline injected at a dose of 0.3 to 0.5 mL/kg and/or ICG (Akorn Inc) at a dose of 0.02 to 0.2 mg/kg diluted was injected. Thoracoscopic ICG/near-infrared fluorescence (NIRF) imaging was accomplished using a commercially available NIR D-light imaging system (Karl Storz Veterinary Endoscopy) using a previously reported technique.¹⁷ Thoracic ducts were isolated, and laparoscopic hemoclips (M/L-10; Microline Surgical Inc) were placed on the ducts.

For the thoracoscopic pericardiectomy procedure, either a subphrenic pericardiectomy^2,3,18 or creation of a large pericardial window with vertical pericardial fillets was performed.¹⁹ After all surgical procedures had been completed, a thoracic drain (MILA International Inc) was placed, the cannulae were removed, and the thoracic port incisions were closed routinely. If continued flow of contrast past the ligation site was evident on the immediate postoperative CTLA, most dogs were returned to the operating room and a second approach was made to locate and ligate any remaining patent TD branches. Anesthesia time was recorded and included time for the surgical procedures and postoperative CTLA. Surgical time from first incision to the placement of the last suture was also recorded.

Postoperative care

All dogs were recovered from anesthesia in the ICU. The thoracic drain was aspirated every 4 to 6 hours for 12 to 24 hours. Analgesia was provided using either intermittent oxymorphone (dose, 0.1 mg/kg, IV, q 4 to 6 h), methadone (dose, 0.1 to 0.2 mg/kg, IV, q 4 h), or fentanyl as a constant rate infusion (dose, 0.05 µg/kg/min, IV). The day after surgery, oral analgesics were initiated if the patient could tolerate them.

Follow-up and outcomes assessment

Complications were recorded, and conversion from thoracoscopic to open procedures was categorized on the basis of a previously reported standardized classification scheme.²⁰ Resolution of chylothorax postoperatively was defined as a lack of pleural effusion on postoperative imaging substantial enough to cause clinical signs or warrant thoracocentesis in the first 6 postoperative months. Late recurrence was defined as recurrence of chylous effusion > 6 months postoperatively in a dog that had previously shown resolution of chylothorax. Long-term follow-up was obtained by means of contact with the owner or referring veterinarian by telephone or email.

Statistical analysis

Due to the small sample sizes, differences between surgical groups were assessed using the Fisher exact test for categorical predictor variables, with statistical significance set at P < .05. Continuous predictor variables were assessed using binary logistic regression.

Sensitivity, specificity, positive predictive value, and negative predictive value were calculated to assess the feasibility of using a less invasive screening technique (echocardiography) to correctly identify dogs with CPP, compared to the more invasive and less accessible gold standard assessment (cardiac catheterization). All analyses were conducted using Stata analysis software (version 15; StataCorp LLC).

Animals

Twenty-six dogs were included in the study. Fifteen dogs were male castrated, 8 were female spayed, and 3 were male intact. Median weight was 25.4 kg (range, 5.8 to 71.7 kg). Median body condition score was 5/9 (range, 2/9 to 8/9). Median age was 59.5 months (range, 18 to 120 months). Breeds represented included 6 mixed-breed dogs, 4 Shiba Inus, 4 Bullmastiffs, 2 Australian Cattle Dogs, and 1 each of Alaskan Malamute, Dachshund, Doberman, Golden Retriever, Husky, Labrador Retriever, Newfoundland, Rhodesian Ridgeback, Shetland Sheepdog, and Tibetan Terrier.

Clinicopathological features

Clinical signs included dyspnea or increased respiratory rate (n = 15), exercise intolerance (5), lethargy (5), inappetence or anorexia (3), and weight loss (1). Clinical signs were absent in 3 dogs in which pleural effusion was diagnosed during evaluation for other conditions. Physical examination revealed decreased ventral heart and lung sounds on auscultation (n = 15), dyspnea, increased respiratory rate or persistent panting (6), poor body condition (3), and heart murmur (2). Five dogs had no significant abnormalities on physical examination. Median time from onset of clinical signs to surgery date was 43.5 days (range, 7 to 1,050 days). The median number of therapeutic thoracocentesis procedures performed preoperatively was 3 (range, 0 to 10). The estimated median total volume of chylous effusion withdrawn from 19 dogs in which reliable estimates could be made from the medical records was 3,000 mL (range, 200 to 11,250 mL).

The results of biochemical screening and CBCs were nonspecific and therefore are not reported in detail. Median pleural effusion triglyceride concentration was 1,197 mg/dL (range, 402 to 3,864 mg/dL), and median serum triglyceride concentration was 61.5 mg/dL (range, 33 to 221 mg/dL). Cytological analysis of pleural fluid was performed in 18 dogs and was consistent with a diagnosis of chylothorax in all cases. Three dogs had fluid culture and sensitivity testing from pleural effusion, none of which exhibited bacterial growth. All 18 dogs screened for heartworm antigen were negative. Three dogs were tested for coccidiomycosis, with 2 reporting negative results and 1 dog demonstrating a 1:1 titer, which was not considered to be a cause of the chylothorax.

Diagnostic imaging

Thoracic radiographs were available for review in 21 dogs, with all dogs demonstrating some degree of pleural effusion. One dog had a suspicious lung nodule that was not evaluated cytologically.

Preoperative CTLA was performed via mesenteric lymph node injection (n = 26) or subsequent popliteal lymph node injection (4; Table 1). Eight of 24 (33%) dogs demonstrated cranial mediastinal lymphangectasia, and 4 of 24 (17%) dogs demonstrated leakage of contrast agent in the cranial mediastinum (Figure 1).

View Full Size

Cranial mediastinal lymphangectasia (white arrow) is evident on this CT lymphangiography (CTLA) study from a dog with idiopathic chylothorax (A). Contrast leakage in the cranial mediastinum can be seen surrounding the cranial lung lobe and dissecting through tissue planes (arrows) on this transverse CTLA image of the cranial thorax (B). Spontaneous rupture of the cisterna chyli is seen in this dog on transverse CTLA study, evidenced by a blush of contrast leakage (arrow) at the level of the cisterna chyli (C). In this 3-month postoperative CTLA image, a collateral thoracic duct (arrow) can be clearly seen just dorsal to the azygous vein (D). This duct was not seen in either the preoperative or immediate postoperative CTLA study performed on this dog.

Citation: Journal of the American Veterinary Medical Association 261, 5; 10.2460/javma.22.08.0381

• Download Figure
• Download figure as PowerPoint slide

Twenty-four of 26 (92%) dogs underwent an immediate postoperative CTLA via mesenteric lymph node injection (n = 24) or subsequent popliteal lymph node (1). Four of 19 (21%) dogs that had diagnostic studies evaluated had persistent flow in the TD system. In all 4 dogs, ICG/NIRF had been used for intraoperative duct identification. Three of these 4 dogs were returned to surgery to have persistent ducts sealed, whereas in 1 of the 4 dogs, multiple ducts were visible ventral and lateral to the aorta that were not evident on the preoperative study and a return to surgery was not pursued, as it was considered unlikely that these ducts would all be amenable to complete ligation. In 1 of 3 dogs that were returned to surgery, a second postoperative CTLA was performed that confirmed complete cessation of contrast flow at the TDL site, whereas in the other 2 dogs a second postoperative CTLA was not pursued. In 13 dogs, retrograde flow of contrast to the area of the lumbar lymphatic plexus was seen. In 1 case, spontaneous rupture of the cisterna chyli with chyle leakage could be seen (Figure 1). In 4 dogs, no retrograde contrast flow could be appreciated, and in the remainder, the caudal abdomen was not included in the imaging study.

Nineteen of 26 (73%) dogs underwent a 3-month postoperative CTLA study (all via mesenteric lymph node injection), performed a median of 92.5 days (range, 60 to 161 days) postoperatively. Five of 17 (29%) dogs with diagnostic studies demonstrated ongoing passage of contrast through patent TDs. Of these, it appeared that 2 dogs had developed collateral TDs: 1 dog had a patent duct on the right side where no duct was seen preoperatively, and in a second dog, a collateral TD was seen dorsal to the azygous vein (Figure 1). In both dogs, resolution of IC occurred, and both dogs remained alive 37 and 50 months postoperatively with no recurrence. In the remaining 3 dogs, ongoing flow of contrast occurred due to recanalization/incomplete attenuation of ducts (Figure 2). These 3 dogs were alive at the time of last follow-up. Two of the 3 had IC resolve without recurrence (follow-up times of 10 and 25 months). In the third dog, resolution occurred, but recurrence was confirmed at 13 months postoperatively (triglyceride level in pleural effusion, 1,387 mg/dL). At the time of last follow-up, no further diagnostic or therapeutic interventions had been elected by the owners of this dog. In 15 of 17 dogs, the 3-month CTLA study showed retrograde flow of contrast to the lumbar lymphatic plexus (Figure 3).

View Full Size

Dorsal plane maximum intensity projection 3-month postoperative CTLA from 2 dogs with recanalization/missed ducts (shown in panels A and B). The most cranial site of ligation is indicated with black arrows, and the contrast column beyond the clips are indicated with white arrows. It is difficult to state with certainty whether ongoing flow represents recanalization of contrast through the applied hemoclips or new ducts are passing over the clips in the area.

Citation: Journal of the American Veterinary Medical Association 261, 5; 10.2460/javma.22.08.0381

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Sagittal (A) and dorsal plane (B) maximum intensity projections from a CTLA performed 3 months postoperatively in a dog that underwent thoracoscopic thoracic duct ligation. Retrograde flow of contrast to the lumbar lymphatic plexus and lymph nodes represented by far the most commonly seen pattern of postoperative redistribution of contrast flow in dogs after thoracic duct ligation.

Citation: Journal of the American Veterinary Medical Association 261, 5; 10.2460/javma.22.08.0381

• Download Figure
• Download figure as PowerPoint slide

Evaluation for constrictive pericardial physiology

Seven of 26 dogs had an echo score of 2 or greater (Table 2). Using a mean right atrial pressure of ≥ 6 mm Hg as the gold standard for diagnosis of CPP, sensitivity and specificity of an echocardiography score ≥ 2 to detect CPP were 94% and 66%, respectively. The positive predictive value of an echocardiography score ≥ 2 to detect CPP was 89% and negative predictive value was 80%.

Surgery and intraoperative imaging

Seventeen dogs underwent a TDL alone, and 9 dogs underwent TDL/P (Table 3). For the TDL procedure, a 3-port technique was used in 23 dogs and a 4-port technique was used in 3 dogs. Ports were inserted in variable combinations in intercostal spaces 8 through 11. In 25 dogs, a right-sided approach was initially performed, and in 1 dog, a bilateral approach was performed due to the presence of obvious TDs on both the right and left sides of the aorta in the CTLA study.

In 12 dogs operated on earlier in the study, direct injection of mesenteric lymph nodes (n = 10) or injection into the mesenteric root (2), for cases in which no mesenteric lymph nodes were identified, was performed through a 3- to 5-cm right-sided paracostal incision. Where mesenteric lymph nodes were directly injected, ICG/NIRF and MB were successfully visualized in the thoracic ducts in 8 of 8 (100%) and 2 of 4 (50%) dogs, respectively. For the 2 dogs in which the mesenteric root was injected, only ICG/NIRF was used and was successful in imaging the thoracic ducts in 1 of 2 (50%) dogs. In 13 dogs operated on later in the study, 4-quadrant subcutaneous perineal injections were performed. Successful visualization of ICG/NIRF and MB was achieved in 11 of 13 (85%) and 0 of 1 (0%) dogs, respectively. For 2 dogs in which no flow of contrast was seen after perineal subcutaneous injection, a paracostal approach was subsequently made and direct mesenteric lymph node injection was performed. In a single dog, a cutdown to a popliteal lymph node was performed. In this dog, ICG/NIRF and MB were injected, with only ICG/NIRF being successfully visualized.

Dissection into the caudal mediastinum was performed at a median of 2 spaces between vertebral arteries (range, 1 to 5). A median of 4.5 hemoclips (range, 2 to 12) were used for TDL. Median total anesthesia time for all dogs was 285 minutes (range, 165 to 540 minutes). Median total surgical time for all dogs was 135 minutes (range, 75 to 250 minutes). Median number of days in the hospital was 4 days (range, 3 to 9 days), and median number of postoperative hours in the hospital was 48 (range, 24 to 159 hours). Dogs in the TDL/P had significantly greater surgical (P = .007) and anesthesia (P = .046) times than dogs in the TDL group. No differences were detected in hours in hospital postoperatively between the TDL/P and TDL groups (P = .14).

Complications and conversion

Chyle leakage from the TDs during dissection occurred in 2 of 26 (8%) dogs (Table 3). In 1 dog, leakage was stopped by application of further clips on the abdominal side of the leak. In the second dog, complete cessation of leakage was not possible; this dog initially resolved but then experienced a recurrence of chylothorax at 77 days postoperatively and underwent a pericardiectomy and a subsequent en bloc TDL but was classified as a nonresolving case, as pleural effusion didn’t resolve within 6 months postoperatively. The dog eventually died 36 months postoperatively after developing pulmonary metastatic disease.

In 2 of 26 (8%) dogs, emergent Grade 4 conversion to an open approach was performed. In 1 dog, during the TDL, laceration of a vertebral artery occurred and was successfully treated by artery ligation. In a second dog, conversion was performed during the pericardiectomy to perform cardiopulmonary resuscitation after ventricular fibrillation had developed. Despite attempts at resuscitation, the dog died, resulting in an overall perioperative mortality rate of 1 in 26 (4%) dogs.

Long-term outcomes

Sixteen of 17 (94%) dogs in the TDL group and 7 of 8 (88%) dogs in the TDL/P group exhibited resolution of chylothorax, results of which were not statistically different (P = .55; Table 3). There were no statistically significant relationships between age (P = .68), weight (P = .22), time from onset of clinical signs to surgery (P = .49), number of preoperative thoracocentesis episodes performed (P = .62), or volume of pleural effusion (mL/kg) withdrawn preoperatively (P = .39) and the chances of resolution.

One of 25 (4%) dogs in the study (TDL group) developed late recurrence of chylothorax. At the time of last follow-up, 3 dogs in the TDL group had died from unrelated causes at a median of 33 months (range, 24 to 36 months) while the remaining 14 dogs were alive at a median of 16.5 months (range, 10 to 60 months). In the TDL/P group, of the 8 dogs that survived the perioperative period, 2 had died, one at 22 months postoperatively from an unrelated cause and the other at 4 months postoperatively because of nonresolution of chylothorax. The latter dog underwent a necropsy evaluation, which showed widespread pulmonary arterial embolic disease and chronic myocardial fibrosis and fatty infiltration but no obvious underlying cause of the chylothorax. Six of 8 dogs in the TDL/P group that survived the perioperative period remained alive at a median of 47 months (range, 13 to 56 months).