Materials and Methods

Case selection—Medical records of dogs examined between 2007 and 2010 were searched, and dogs meeting the following criteria were included in the study: a diagnosis of IC on the basis of ruling out other known underlying etiologies; treatment via thoracoscopic TDL and SPP performed either by the first author (PDM) or under his direct supervision; and availability of outcome data, including postoperative radiographs at 1 or more time points ≥ 4 weeks after surgery and clinical follow-up by direct examination or telephone interview ≥ 6 months after surgery. Dogs were excluded if they did not have a diagnosis of IC, did not undergo thoracoscopic TDL and SPP, had undergone a previous surgical intervention for IC, or were lost to follow-up prior to the time points noted.

Anesthesia and patient preparation—A variety of anesthetic protocols were used on the basis of the personal preferences of the treating anesthesiologists. In all patients, routine monitoring of end-tidal capnography or blood gas analysis was performed. In all dogs, direct or indirect blood pressure, Spo₂, and ECGs were monitored intraoperatively. In some patients, alternating OLV performed with double-lumen endobronchial intubation, as has been described,¹⁷ was used during the SPP procedure only. All patients were prepared for aseptic surgery in a similar fashion by clipping of hair and aseptic preparation of the entire abdomen and thorax. For patients in which intraoperative lymphangiography was performed, a C-arm–compatible motorized float table was used. For TDL performed with the patient in lateral recumbency, the endoscopic tower was placed directly opposite the surgeon on the other side of the dog. For SPP, the dog was placed in dorsal recumbency and the endoscopic tower was moved to the head of the patient.

TDL—Dogs were positioned in left lateral recumbency for TDL. The first thoracic portal for the placement of the telescope was placed at the midthoracic to dorsal third thoracic level in the eighth or ninth intercostal space, allowing visualization of the caudal mediastinal area. Two other instrument portals were placed at the seventh or eighth and ninth or tenth intercostal spaces in a slightly more dorsal location and in a triangulating pattern around the caudal mediastinum. Portals were established with a combination of either 5-mm cannulae^c or 11.5-mm cannulae,^d but in each case, at least one 11.5-mm cannula was used to accommodate the 10-mm clip applier used.

Once port placement was complete, a 3- to 4-cm laparotomy incision was made in the patients' right cranial abdominal quadrant at the midabdominal level, and a wound retractore was inserted through the incision. Mesenteric lymph nodes in the area of the ileocecocolic junction were exteriorized through this incision. For dogs in which an intraoperative lymphangiogram was attempted, 1 to 3 mL of iohexol diluted 1:1 with sterile saline (0.9% NaCl) solution, a nonionic iodinated contrast agent, was slowly infused into the lymph node with a 3-mL syringe and 25-gauge needle. Images were collected fluoroscopically by use of a mobile C-arm unit to assess the anatomic arrangement of the thoracic duct and its branches. Methylene blue dye diluted 1:1 with sterile saline solution was then slowly infused into the same or an adjacent lymph node to a dose ≤ 0.5 mg/kg (0.23 mg/lb).

As soon as the methylene blue was seen in the thoracic duct branches, blunt dissection around the duct branches in the mediastinal root dorsal to the aorta was initiated with a combination of laparoscopic instruments and a vessel-sealing device.^f A space between vertebral arterial branches was selected as far caudal within the thoracic cavity as possible while still allowing dissection and clip placement on the thoracic duct and its branches. As much dissection across the mediastinal base as possible was performed, and all visualized thoracic duct branches were clipped by means of a multiuse 10-mm medium to large clip applier.^g At the completion of the surgery, another injection of contrast agent was attempted in some patients to confirm complete duct ligation. Thereafter, the thoracic port incisions and the flank approach incision were closed routinely.

SPP—Patients were then repositioned in dorsal recumbency, and a second aseptic surgical scrub of the entire thoracic area was performed. A subxiphoid camera portal was established with a 5-mm cannula,^c and on entry into the thoracic cavity, an instrument portal was established at the fourth to sixth intercostal space on the left-hand side in the ventral third of the thoracic cavity. A vessel-sealing device^f was placed into the instrument portal and used to section the ventral mediastinal attachments to the sternum, thereby providing access to both hemithoraces. A second instrument portal was placed at the fourth to sixth intercostal space on the right side and also in the ventral third of the thoracic cavity. The SPP was then initiated and performed in a manner identical to that described in a previous report.¹⁷ A Babcock or Kelley forceps was used for retraction of the pericardium or overlying pericardial fat, and endoscopic scissors were used to incise the pericardium over the apex of the heart. At this point, the patient was rotated over toward the right side either manually or by lateral tilting of a motorized float table. Through the established pericardial incision, a vessel-sealing device^f was used to continue sectioning the pericardium in a caudodorsal direction down to the level of the phrenic nerve close to its diaphragmatic insertion. For cases in which OLV was used, it was initiated at the point where dissection close to the phrenic nerve was performed to maximize visualization of the nerve. To establish OLV, fresh gas inflow to the left hemithorax was then stopped, resulting in left-sided lung atalectasis. Once dissection down to the level of the caudodorsal pericardium close to the phrenic nerve was complete, the thoracoscope was removed from the subxiphoid port and placed into the left-sided instrument port. The vessel-sealing device was placed into the subxiphoid port, and a Babcock forceps was placed into the contralateral instrument port to allow the cut edge of the pericardium to be retracted medially, thus improving visualization of the phrenic nerve. The pericardium was sectioned along a line parallel to the phrenic nerve in a caudal to cranial direction. For patients in which OLV was used, the nonventilated lung was reinflated, and fresh gas inflow to the contralateral side was halted. The patient was tilted to the opposite side, and the procedure was repeated on that side until only a small cranial attachment of the pericardium remained. The thoracoscope was replaced into the subxiphoid port, and a Babcock forceps was replaced into 1 instrument port with the vessel-sealing device placed into the other port so that final sectioning of the remaining cranial attachment of the pericardium could be accomplished. Pericardial tissue with associated fat was then withdrawn in either 1 piece or piecemeal through one of the 11.5-mm instrument ports if it was small or placed within a specimen retrieval bag if the sample was larger. Reinflation of both lungs was visually verified after completion of the procedure, and a thoracic tube was placed to evacuate the pneumothorax prior to routine closure of the thoracoscopic port incisions.

Outcome—Clinical follow-up was obtained by examination of the patient or by telephone interview of the owner. In each case, owners were questioned as to whether their dog was still alive and, if so, whether the dog had any obvious recurrence of clinical signs associated with chylothorax. If the dog had died, they were questioned to attempt to establish whether the cause of death was known or was related to IC or respiratory disease and whether clinical signs or a diagnosis of IC recurrence had ever been made while the dog was still alive. Radiographic follow-up at various time points (but ≥ 4 weeks) after surgery was obtained, and any evidence of persistence or recurrence of pleural effusion was recorded.

Results

Six dogs met all inclusion criteria and were included in the study. This represented all cases of IC referred to the first author during the study period in which the owners elected surgical management and no previous surgical interventions had been attempted.

Three dogs were castrated males, and 3 were spayed females. Breeds represented included 2 mixed-breed dogs and 1 each of German Shepherd Dog, Whippet, Coton de Tulear, and Shetland Sheepdog. Dogs had a median body weight of 18 kg (39.6 lb; range, 5.1 to 29.2 kg [11.2 to 64.2 lb]). Median age was 3 years (range, 2 to 8 years). All dogs had a history of increased respiratory rate, increased panting, or dyspnea. Other historical abnormalities included lethargy (n = 5), decreased appetite or anorexia (3), and vomiting (1). Physical examination findings included an increase in respiratory rate in 3 dogs (median respiratory rate, 36 breaths/min; range, 28 to 40 breaths/min). The 3 remaining dogs were described as having persistent panting during examination. On thoracic auscultation, 4 of 6 dogs were described as having dull respiratory sounds over the ventral thoracic wall when evaluated in a standing position.

Thoracic radiography was performed in all cases and showed different amounts of pleural effusion in all dogs. All dogs had an echocardiogram performed by a board-certified cardiologist.

In 5 dogs, findings were normal, and 1 dog had evidence of mild mitral valve endocardiosis. Five dogs underwent an abdominal ultrasonographic examination; no major abnormalities were noted in any dog. Two dogs had thoracic CT performed. In 1 dog, there was a mild bilateral pneumothorax that was presumed to be secondary to a previous thoracocentesis. Otherwise, no important abnormalities were detected on CT in those 2 patients.

Thoracocentesis was performed in all 6 dogs and revealed variable amounts of opaque milky fluid. Pleural fluid triglyceride concentrations were greatly elevated in all dogs (median, 833 mg/dL; range, 479 to 1,840 mg/dL; reference range, 29 to 166 mg/dL). Five dogs had heartworm screening tests performed, the results of which were negative in all dogs. All 6 patients also underwent CBCs and serum biochemical analyses prior to surgery. No clinically important abnormalities were noted in any dog.

In 4 patients, double-lumen endobronchial intubation to allow alternating OLV was used during the SPP procedure.¹⁷ In the 2 remaining dogs, an SPP was able to be performed without OLV. One-lung ventilation was not used during any of the TDL procedures.

In all patients, thoracoscopic TDL and SPP were successfully completed in a median surgical time of 177 minutes (range, 140 to 210 minutes). Pre- and post-TDL lymphangiography was attempted in 2 patients by direct infusion of contrast into the same or an adjacent lymph node. In 2 other patients, only post-TDL lymphangiography was attempted. In both cases for which pre-TDL lymphangiograms were performed, the images were considered of good quality, but in only 1 of the 4 cases that underwent post-TDL was ductal opacification considered of adequate quality to completely rule out persistent flow in the ductal system. Mesenteric lymph node infusion with methylene blue resulted in good coloration of the thoracic duct and its branches in all dogs. However, persistent flow was not obvious in any of the patients.

No major intraoperative complications occurred, and in no dogs was conversion to an open surgical approach necessary. Immediately postoperatively, an injectable opioid drug was administered for 12 to 24 hours (hydromorphone hydrochloride, 0.1 mg/kg [0.045 mg/lb], IV, q 4 h, or buprenorphine hydrochloride, 0.01 mg/kg [0.0045 mg/lb], IV, q 6 h). On discharge, tramadol hydrochloride was prescribed to 3 dogs (2 to 4 mg/kg [0.91 to 1.82 mg/lb], PO, q 8 h for 3 to 5 days), and deracoxib was prescribed to 3 dogs (1 mg/kg [0.45 mg/lb], PO, q 24 h for 3 to 5 days). Thoracostomy tubes were maintained for 1 to 5 days after surgery, and all dogs were discharged from the hospital 2 to 6 days after surgery.

In all cases, resected pericardial tissue was submitted for histologic analysis. In all dogs, a varying degree of chronic fibrosing pericarditis was evident. Neoplastic disease of the pericardium was not diagnosed in any dog. In 1 dog, biopsy of a small pleural nodule revealed mild plasmacytic pleuritis. No dogs had bacterial or fungal culture and antimicrobial susceptibility testing performed.

Outcome—There were no major long-term complications in any of the 6 patients. Follow-up thoracic radiographs were obtained in all patients at least once after surgery. The most recent radiographic follow-up was at a median of 14.5 months (range, 7 to 25 months) after surgery. In 5 of 6 dogs, no pleural effusion was detected in the final set of thoracic radiographs obtained after surgery. In the 5.1-kg Coton de Tulear, radiographs obtained at 6 weeks and 7 months revealed a very small amount of pleural effusion, which was not considered a sufficiently large volume to aspirate and was not associated with any clinical signs. This dog did not require any further thoracocentesis and did not develop clinical signs within a 25-month follow-up period. Final clinical follow-up for all 6 dogs occurred at a median of 39 months (range, 19 to 60 months) after surgery. The German Shepherd Dog had no further clinical signs of respiratory disease during a follow-up period of 28 months and died of chronic progressive ambulation problems at the age of 14 years. The Whippet was lost to follow-up 31 months postoperatively but had no recurrence of clinical signs and was clinically normal at that time. The 4 other dogs were all alive with no recurrence of clinical signs.

Discussion

The results of the present small series of patients suggested that a minimally invasive TDL-SPP combined surgical technique for management of IC in dogs may be associated with a similarly successful outcome as has been reported for open surgical TDL-SPP. In 1 dog in the present report, a small-volume pleural effusion persisted at a 7-month radiographic examination, but this was not associated with clinical signs and therefore did not require drainage via thoracocentesis at any time after surgery. This dog had remained free of clinical signs for 25 months after surgery.

Many modalities have been used for the management of IC in dogs. The combination of TDL and SPP has been associated with some of the highest success rates reported in dogs with the idiopathic form of chylothorax, although only in very small series of patients.^5,7 Prior to this report, the ability to perform a minimally invasive approach for the management of IC was evaluated in 7 dogs in a previous study⁸ with successful results. In the prior study by Allman et al,⁸ the procedures were not performed in an identical manner to those in the present study; nonetheless, long-term resolution of clinical signs was seen in 6 of 6 patients in our series and 7 of 8 cases in the previous study.⁸ In 1 dog of the present case series, a very small volume of fluid was seen on postoperative thoracic radiographs obtained at 6 weeks and 7 months after surgery. Because this dog did not have any clinical signs and has not required thoracocentesis at any point after surgery, we classified this as a positive outcome; however, it remains debatable whether this case should be categorized as a treatment success or failure. Generally, a complete resolution of pleural effusion is the end goal of surgery when treating patients with IC, because chronic pleural effusion is thought to lead to progressive constrictive pleuritis, which can be associated with serious clinical signs even in the absence of ongoing effusion.¹⁸ It is important to note that the success rates quoted for minimally invasive TDL-SPP treatment of IC are those reported for newly diagnosed and previously untreated cases of IC. Patients that have previously undergone unsuccessful surgery or have identifiable underlying disease processes may not have as high a resolution rate with minimally invasive TDL-SPP.⁸ This fact underlines the need for thorough history, diagnostic evaluation, and treatment of any underlying conditions in dogs with chylothorax.

The thoracoscopic approach to the thoracic duct was first described in an experimental study of dogs by Radlinsky et al,³ and several modifications have been more recently reported in clinical cases by the same group.⁸ We performed the minimally invasive approach to the thoracic duct with dogs in lateral recumbency and similarly without OLV. Port placement for access to the thoracic duct was similar to that reported⁸ previously, and generally, 3 ports located in a triangulating pattern at the seventh; eighth and ninth; or eighth, ninth, and tenth intercostal spaces provided good access to the base of the mediastinum caudally.^3,8 In all patients, dissection across the base of the mediastinum to the contralateral hemithorax was attempted to try and visualize and clip all visible thoracic duct branches. This was challenging in the smallest patient with a weight of only 5.1 kg, making the available space between vertebral arterial branches for thorough dissection limited. This is the smallest dog reported in the literature to have undergone thoracoscopic TDL and SPP, and despite the principal thoracic duct branch being clearly identified and easily clipped, dissection through to the other side and around the aorta was probably not as thorough as performed in some of the other dogs because of anatomic limitations. It remains unknown whether this limitation contributed to the small-volume pleural effusion that remained in this patient, because lymphangiography was not used after surgery to confirm the absence of persistent ductal flow.

In all patients described in the present report, methylene blue was injected into the mesenteric lymph nodes, and the dye effectively highlighted the thoracic duct and its branches. This simple technique greatly facilitated ductal identification and was especially helpful for visualization of the smaller ductal branches.

Pre- and post-TDL lymphangiography was attempted in 2 patients described in the present report, and post-TDL only was attempted in another 2 dogs. In only 1 case were data obtained that showed a lack of flow in the thoracic ductal system after ligation. Ideally, this technique is performed by lymphatic catheterization of an afferent lymphatic vessel. This provides the ability to provide high flow and pressure within the lymphatic system during injection of contrast. These larger lymphatic vessels at the root of the mesentery are difficult to access by making a small paracostal incision as was performed in dogs of this report. Others have found a similarly low rate of success with this technique.^5,8 Improvements in lymphangiography technique have recently been reported by use of CT,^19,20 and these techniques are likely to be superior to the technique used in this study.

Because postoperative lymphangiography was not performed or was not successful in all dogs in this study, we cannot confirm that all branches of the TD were ligated. However, others have also found that not performing lymphangiography after TDL does not preclude obtaining high resolution rates of IC after surgery.^5,7,8 Two possible explanations may exist for this observation: first, it may be that successful ligation of all TD branches is occurring in most cases despite our inability to confirm this; second, it may be possible that it is not necessary to ligate all branches, given that other pathophysiologic mechanisms are occurring after TDL-SPP that lead to resolution and are unrelated to complete TD ligation.

Ideally, the underlying pathophysiology of IC needs to be evaluated more thoroughly. The therapeutic rationale for TDL is that occlusion of all ductal branches in the caudal thorax will lead to the development of abdominal lymphaticovenous anastomoses, thereby creating a conduit for the return of lymphatic fluid to the venous system within the abdomen. The available evidence for this proposed pathophysiologic response to TDL is somewhat conflicting. In a study by Fossum and Birchard,²¹ 3 dogs underwent ligation of the thoracic duct at the lymphaticovenous anastomosis in the cranial thorax. In all 3 dogs, lymphaticovenous anastomoses were formed within the abdominal cavity, and at necropsy, it appeared that all flow of chyle through the intrathoracic portions of the ductal system had ceased.²¹ Subsequently, in an experimental study of dogs²² that evaluated the lymphatic response to TDL in isolation as well as in combination with cisterna chyli ablation, 3 clinically normal dogs that underwent TDL alone did not form lymphaticovenous anastomoses in the abdominal cavity but formed draining vessels to the azygous vein within the thoracic cavity. In theory, this is a suboptimal response to treatment because diversion of drainage pathways away from the thoracic cavity, where chyle accumulation is so poorly tolerated, is desirable. It remains unknown, however, whether development of this type of collateral circulation within the caudal thoracic cavity would always lead to chyle accumulation within the pleural space in dogs with naturally occurring disease. When cisterna chyli ablation was performed in addition to TDL, it was found that 5 of 6 dogs exclusively developed intra-abdominal lymphaticovenous anastomoses. The authors postulated that this may represent a more desirable response than when TDL alone was performed.²² The limited data from canine patients suggest that response rates to TDL alone are not encouraging. Thoracic duct ligation in dogs with naturally occurring IC was associated with a 59% resolution rate when isolation of the thoracic duct with subsequent ligation² was performed and a 50% response rate when the surgery was performed in an en bloc fashion.⁴

The effect of pericardiectomy has been evaluated to an even lesser degree than TDL. It has been proposed that the pericardium in patients with chylothorax may be somewhat thickened, compared with results in clinically normal dogs, and that this may limit diastolic filling of the right side of the heart, which in turn can elevate right atrial and central venous pressure.⁵ This may lead to increased flow of chyle within the thoracic duct system and may decrease the tendency for chyle to drain into the venous system.⁵ Pericardiectomy has been advocated by some authors⁵ as a useful adjunct in the management of chylothorax, and small numbers of cases have been described in which pericardiectomy alone was successful in achieving resolution of chylothorax when other modalities had failed.⁵ No large case series describes the success rate of pericardiectomy alone as a treatment modality for chylothorax. Furthermore, evidence of venous or lymphatic hypertension has never been reported in clinical cases with chylothorax, although an experimental model of chylothorax in dogs has been created by ligation of the cranial vena cava with a subsequent substantial increase in central venous pressure.²¹ The extent of pericardial resection performed may or may not have an effect on the likelihood of success. Most small animal surgeons consider that a pericardial window may be inadequate to decompress any constriction placed on the right side of the heart by the pericardium.^5,7,8 Variations in the technique used include resection of all pericardial tissue ventral to the heart accessible through a lateral thoracotomy^5,7 and thoracoscopic formation of a pericardial window with creation of 2 to 3 additional fenestrations that run up to phrenic nerves.⁸ In the present study, all patients were treated with a true bilateral SPP in which all tissues ventral to the phrenic nerve were resected.¹⁷ In 2 cases, this was achievable without OLV, but in 4 dogs, a technique involving the use of alternating OLV was used that has been described in research animals.¹⁷ If there is a decompressive effect on the heart from pericardiectomy, it seems logical to remove as much of the pericardium as possible. However, currently, the advantages of this approach in dogs with IC remain speculative because of a lack of experimental or clinical data. Nonetheless, we suggest that response to treatment appears to be higher in cases with IC for which TDL and pericardiectomy are performed in combination. Resolution of clinical signs has been reported for 10 of 10 dogs⁵ and 13 of 14 dogs⁷ when TDL and SPP were performed with an open technique and in 6 of 7 dogs⁸ and 6 of 6 dogs (patients of the present study) when performed in a minimally invasive fashion.

There are several limitations to this study. Unfortunately, all clinical studies of IC are limited by having small case numbers because this challenging condition is rare and, to our knowledge, no single institution has evaluated a large number of cases. The retrospective nature of the present study is limited by the inherent challenges of gathering information retrospectively; therefore, not all data were available for all dogs, although these cases were followed closely in the postoperative period. Additionally, although the present study may provide information that can help to document clinical success of the currently recommended surgical techniques, such studies can provide little additional information on the underlying pathophysiology of the condition and the response to treatment. Future studies are necessary to confirm the similar outcome reported for minimally invasive TDL and SPP versus the open surgical approach for the treatment of IC in dogs.