Operative management for refractory chyle leakage after pancreatectomy: A case report and literature review

Abstract

Post-pancreatectomy chyle leak is a common complication typically managed conservatively. However, rare ISGPS grade C leaks that produce high output can lead to significant morbidity and often resist non-operative treatments. We report two cases of such refractory leaks successfully addressed through surgical intervention. We describe two complex cases of postoperative chyle leakage. The first case involves a 66-year-old female who underwent pylorus-preserving pancreaticoduodenectomy with portal vein resection. The second case is a 67-year-old female who underwent completion total pancreatectomy. Both patients experienced persistent, high-output chyle leakage that did not respond to conservative management and several interventional procedures. Ultimately, both were successfully treated with exploratory laparotomy and multiple lymphatic ligations after undergoing preoperative fat loading. Postoperative chyle leak is initially managed conservatively, with interventional options reserved for refractory cases, though these are often ineffective. When conservative measures fail, surgical management with fat loading, targeted or prophylactic lymphatic ligation, and adjunct sealants can provide lasting control. Surgical management presents a viable salvage strategy for refractory postoperative chyle leaks when nonoperative measures are ineffective. Preoperative fat loading aids in intraoperative localization. Timely surgical intervention can prevent malnutrition and preserve opportunities for adjuvant therapy.

INTRODUCTION

Chyle leak is defined as the postoperative drainage of chylous fluid resulting from injury to the cisterna chyli or its tributary lymphatic channels [1]. It is frequently encountered after pancreatic surgery, where dissection of the hepatoduodenal ligament, celiac axis, and the root of the mesentery around the superior mesenteric vein (SMV) and superior mesenteric artery (SMA) brings the operative field into proximity with major abdominal lymphatic pathways [2-7].

The incidence of chyle leak following pancreatic resection has been reported to range from 1.3% to 16%, with distribution categorized by the International Study Group of Pancreatic Surgery grades A–C based on clinical impact and required management [1-5,8-11]. Grade A indicates a clinically irrelevant chyle leak that does not require further treatment, while grade B necessitates dietary restrictions, medication such as octreotide, or drainage through intervention or surgery [1]. Grade C requires more invasive management, including embolization, sclerosis, and reoperation, which may lead to readmission, intensive care unit admission, or mortality [1]. Although grade C events are uncommon (0%–4.6%), they are clinically significant [6,9]. While most grade A or B cases resolve with conservative therapy, persistent high-output leakage can prolong hospital stays, cause protein-calorie malnutrition, and require extended drainage, increasing the risk of intra-abdominal infection [2-4,9-11]. For intractable chyle leakage, interventional approaches—including lymphangiography with embolization or sclerotherapy, and in select reports, radiotherapy—have been described [12]. However, operative management has been rarely reported and remains insufficiently characterized. Here, we present two cases of refractory postoperative chyle leak successfully treated through operative intervention. This case report adheres to the Surgical CAse Report (SCARE) guidelines [13].

CASES

Case 1

A 66-year-old female patient presented to the outpatient clinic with elevated liver function tests and an unremarkable past medical history. Work-up revealed borderline-resectable pancreatic head adenocarcinoma 2.3 cm with SMV involvement. She received neoadjuvant modified FOLFIRINOX (mFOLFIRINOX) for 12 cycles over 6 months, achieving stable disease.

Following chemotherapy, the patient underwent pylorus-preserving pancreaticoduodenectomy (PPPD) with resection of the portal vein (PV) and divestment of the common hepatic artery (CHA). Tumor infiltration was found to extend from the celiac axis to the CHA, necessitating adventitial divestment of major vascular structures. The tumor also demonstrated involvement of the PV and the root of SMA. An anatomical variant of the right posterior hepatic artery arising from the SMA was identified and was sacrificed due to tumor invasion. The PV was resected for a length of 6 cm, including the portomesenteric junction, and was subsequently reconstructed using an end-to-end anastomosis with 6-0 polypropylene sutures (continuous). The splenic vein and artery were ligated as part of the procedure. A regional lymphadenectomy was performed, encompassing stations 7, 8, 9, 12, 13, 14, and 17. The reconstruction involved duct-to-mucosa pancreaticojejunostomy, hepaticojejunostomy, and hand-sewn duodenojejunostomy. Two Jackson-Pratt (JP) drains were placed—one near the caudate lobe and another at the pancreatojejunostomy site. The total operative time was 413 minutes, during which 2 units of red blood cells were transfused.

Diet was done routinely according to our center’s enhanced recovery after surgery protocol: postoperative day (POD) 1 sips of water, POD 2 liquid diet, and POD 3 soft diet. During PODs 1–3, combined drain output totaled 800–900 mL/day, and the appearance changed from serous to chylous. On POD 3, drain fluid amylase/lipase levels were 9/5 U/L each, excluding postoperative pancreatic fistula, while triglycerides measured 185 mg/dL, exceeding the diagnostic threshold for chylous effusion (> 110 mg/dL). A low-fat, medium-chain triglyceride (MCT) diet was initiated; the drain fluid reverted to serous, but the volume increased to 1,500 mL/day on PODs 4–5. Computed tomography (CT) on POD 5 demonstrated perivascular fluid causing PV luminal narrowing; one pancreaticojejunostomy-adjacent JP had migrated into the pelvis and was removed. On POD 6–8, drainage remained as high as 1,200 mL/day. By POD 9, the remaining drain output was serous without clinical signs of inflammation, and the drain was removed. She was discharged on POD 12 after a 3-day observation period.

On POD 26, she was readmitted with abdominal distension and bilateral leg edema. CT revealed massive ascites, and a percutaneous catheter drainage (PCD) was inserted (Fig. 1), yielding 700 mL of chylous ascites at placement; daily drainage was limited to 1,500 mL. Ascitic triglycerides were 1,196 mg/dL. Conservative therapy with nil per os (NPO) and total parenteral nutrition (TPN) was provided. After 11 days (POD 38), the PCD appearance became serous with volumes up to 1 L/day. On POD 39, bipedal lymphangiography via bilateral inguinal lymph node access achieved adequate opacification (Fig. 2) but did not localize a discrete leak along the iliac/retroperitoneal channels or thoracic duct; a mesenteric lymphatic source was suspected. Mesenteric lymphangiography was not accessible due to limited medical resources at that time.

Fig. 1.

Computed tomography scan on postoperative day 26 (case 1) showing a large amount of ascites throughout the abdomen (arrow).

Fig. 2.

Intraoperative finding of case 1 demostrating pulsatile lymphatic channels, presumed to represent branches of intestinal trunks or cisterna chyli (circled).

Given ongoing malnutrition, oral intake on a low-fat/MCT diet was cautiously resumed on POD 40, after which drainage increased to 2,000 mL/day. Over one month of conservative care, she lost 7 kg. Considering the refractory high-output drainage and malnutrition, operative management was deemed necessary.

At 2 months after PPPD, she underwent exploratory laparotomy. She was kept NPO from midnight, and 4 hours before the incision, she ingested 50 mL of milk mixed with 50 g of melted butter to enhance lymph flow. Intraoperatively, severe adhesions were observed between the abdominal wall and small bowel, and 2,500 mL of chylous ascites was drained. Pulsatile lymphatic channels, thought to be branches of the intestinal trunks or cisterna chyli, were identified beneath the PJ. Four stitches of 5-0 Prolene suture (Ethicon, Inc.) were used to achieve immediate cessation of chylous leakage (Fig. 3). Surgicel (absorbable hemostat; Ethicon, Inc.) and Greenplast Q (fibrin sealant; GC Pharma) were applied at the suture site.

Fig. 3.

Magnetic resonance lymphangiography on postoperative day 46 (case 2) demonstrating suspicious contrast leakage at the right paraaortic area, just below the renal artery (L1-L2) (arrows).

Within seven days of lymphatic ligation, pelvic PCD output decreased to 400 mL/day and remained serous. On POD 9, the PCD output further decreased to 10 mL/day, and the PCD was removed on POD 13. After discharge, she received adjuvant mFOLFIRINOX (6 cycles over 4 months). She continues outpatient follow-up and remains without evidence of disease at 26 months after the initial operation.

Case 2

A 67-year-old female presented with pancreatitis in the remnant pancreas one year after undergoing laparoscopic distal pancreatectomy for pT2N0 pancreatic cancer. Her medical history included hypertension, diabetes mellitus, and heart failure with left ventricular dysfunction. Evaluation revealed metachronous pancreatic cancer involving most of the pancreatic head and body, with abutment to the SMV and SMA. She received 9 cycles of mFOLFIRINOX.

Following neoadjuvant chemotherapy, a completion total pancreatectomy (TP) was performed. Intraoperatively, a 7-cm mass with inflammatory changes was observed around the CHA and PV. A thrombus at the previously transected splenic artery stump was identified, removed, and ligated with Prolene. Lymph node dissection included stations 7, 8, 9, 12, and 13. Reconstruction was performed via a retrocolic Roux limb with hepaticojejunostomy and duodenojejunostomy. The operative time was 326 minutes, with an estimated blood loss of 300 mL. Two JP drains were placed.

On POD 4, a liquid diet was initiated. The JP output changed from serous to chylous, reaching 1,500 mL/day on POD 5–7. A low-fat MCT diet was started, but the output did not decrease. From POD 10–12, the patient was kept NPO and received TPN, which reduced drainage to 600 mL/day. The drain was removed on POD 13, and she was discharged on POD 15.

On POD 23, she was readmitted with abdominal distension. A CT scan revealed massive ascites, and a PCD was inserted. The ascitic fluid triglyceride level was 1,359 mg/dL, confirming chylous ascites, with daily drainage reaching up to 2,500 mL. She was managed with NPO, TPN, and intravenous furosemide for 7 days, during which drainage decreased to 1,000 mL/day and became serous. After resuming a low-fat MCT diet, drainage again became chylous (2,000 mL/day). From POD 34 to POD 41, she was maintained on NPO, TPN, and diuretics, after which drainage decreased to 600 mL/day, and the PCD was removed on POD 40. However, five days later, she developed a 7-kg weight gain and an increase in abdominal circumference (from 72 to 89 cm). On POD 45, PCD reinsertion showed chylous ascites.

On POD 46, magnetic resonanc lymphangiography with ultrasound-guided inguinal lymph node contrast injection revealed suspicious leakage in the right para-aortic region below the renal artery, while the cisterna chyli and thoracic duct remained intact (Fig. 3). An attempt at lymphangiography and embolization via the left brachial vein failed, as did liver lymphatic cannulation. On POD 53, right femoral vein puncture revealed complete occlusion of the left brachiocephalic vein. Multiple attempts at left hepatic lymphangiography were unsuccessful, but right hepatic lymphangiography succeeded, demonstrating a leakage site. Embolization with NBCA (N-butyl2-Cyanoacrylate) was performed (Fig. 4); however, daily drainage remained 1,000–2,000 mL.

Fig. 4.

Lymphangiography and N-butyl2-Cyanoacrylate embolization on postoperative day 54 (case 2) demonstrating lymphatic leakage around the surgical clips (rectangles).

On POD 68, an exploratory laparotomy was performed. The patient was kept NPO from midnight and ingested 50 mL of milk mixed with 50 g of butter 4 hours before the incision to enhance lymph flow. Intraoperatively, 800 mL of chylous ascites was drained, revealing dense adhesions around the liver, colon, afferent loop, and stomach. After adhesiolysis, a visible leak was identified in the retroperitoneum between the lesser curvature of the stomach and the left liver. Due to severe adhesions, direct dissection of the lymphatic vessels was not feasible; therefore, multiple Prolene ligations were performed instead. No further leakage was observed, and adjunctive hemostatic and sealant materials (TachoSil [Corza Medical], Surgicel, Greenplast Q) were applied. A JP drain was placed in the pelvis, and the PCD was removed.

On POD 4 after the exploration, a low-fat MCT diet was reintroduced. The JP drain output remained serous at 400 mL/day and was removed on POD 7. The patient was discharged on POD 13 (POD 81 from completion TP). After a multidisciplinary discussion, adjuvant chemotherapy was determined to have limited benefit beyond 3 months postoperatively, and she was followed with surveillance. At 18 months, a positron emission tomography-CT revealed hepatic and bone metastases with peritoneal seeding. She died 39 months after the completion TP.

DISCUSSION

Both of our cases highlight the clinical challenge of refractory postoperative chyle leak following pancreatic resection. Despite initial conservative management with dietary modification, bowel rest, and TPN, both patients developed persistent high-output drainage and progressive malnutrition. Lymphangiography, performed to localize and embolize the leakage site, failed to provide definitive control in either case. Surgical intervention, following preoperative fat loading, enabled direct identification of the lymphatic leakage and successful ligation. These experiences underscore that while conservative management and interventional radiology represent first-line strategies, operative management remains an essential salvage option for patients with refractory chyle leak after pancreatic surgery.

Contemporary management of postoperative chyle leak generally follows a step-up strategy [5,11]. Leaks are pragmatically stratified into low-output and high-output categories based on daily drainage volumes and clinical trajectory. For low-output leaks, conservative measures are first-line: dietary fat restriction with a MCT diet, TPN as needed, and maintenance of adequate external drainage. Adjuncts such as somatostatin analogues have also been reported [3-5,10,11].

When chyle output persists despite optimized conservative therapy, minimally invasive interventions can be considered. The criteria reported in the literature include: no improvement within two weeks, drainage output exceeding 1 L per day or 500 mL per day for five consecutive days, or the presence of metabolic or immunologic complications [5]. Intervention should be attempted before surgery, as operative management carries inherent risks of surgical complications. Intranodal or transabdominal lymphangiography may identify the leakage site and allow for embolization or sclerotherapy, with success rates approaching 90% when cisterna chyli cannulation is achieved; however, the overall rate is approximately 70% due to the technical demands of the procedure [12,14]. At our institution, five cases underwent lymphangiography for intractable chyle leak following hepatobiliary-pancreatic surgery between 2008 and 2025. Successful embolization in a single session was achieved in only one case. One patient died in the intensive care unit following failed lymphangiography complicated by intractable ascites and acute kidney injury. Another patient underwent two lymphangiographic procedures—only one of which was successful—and required nearly one year of repeated hospitalizations before eventual resolution. The remaining two cases are reported in the present study: in Case 1, the leakage point from a mesenteric lymph node could not be identified; in Case 2, the first attempt failed due to unsuccessful cannulation, and although the second attempt achieved embolization, the chyle leak persisted and ultimately required surgical intervention.

There are no established guidelines or criteria for transitioning from interventional to operative management. However, based on our institutional experience, surgery is considered when chyle output exceeds 1 L per day for more than two weeks despite intervention, or when malnutrition develops. No studies have directly compared interventional and surgical approaches; surgery is typically pursued as salvage therapy following failed intervention.

Reports of surgical intervention are very rare. In Japan, several case reports have described the use of Denver shunt (BD) placement for intractable chyle leak [15]. The Denver shunt is a peritoneovenous bypass device that diverts refractory ascites into the central venous system; however, it does not directly resolve the underlying chyle leak. Although not following pancreatic surgery, a case reported after laparoscopic cholecystectomy demonstrated that suture ligation successfully controlled the leak [16]. During laparoscopic exploration, an active chyle leak was identified from the gallbladder fossa and was treated with figure-of-eight suture placement and fibrin glue application, resulting in complete resolution (Table 1).

Table 1.

Details of cases treated with surgical management for intractable chyle leak

First author (yr)	Sex	Age (yr)	Index surgery	Duration of chyle leak after surgery (POD)	Prior treatment	Surgical management	Outcome
Tahara et al. (2022) [15]	M	74	Total pancreatectomy	120 (60–180)	Low-fat MCT diet, TPN, lymphangiography	Denver peritoneovenous shunt	Resolved; discharged at POD 6
Tahara et al. (2022) [15]	F	71	Pylorus-preserving pancreaticoduodenectomy	59 (1–60)	Low-fat MCT diet, TPN	Denver peritoneovenous shunt	Resolved; discharged at POD 8
Yao et al. (2018) [16]	M	40	Laparoscopic cholecystectomy	8 (2–10)	Low-fat MCT diet, TPN	Laparoscopic exploration with suture ligation	Resolved; drain removed after 8 months
Malik et al. (2007) [8]	N/A	N/A	Pancreaticoduodenectomy	N/A	Low-fat MCT diet, TPN, diagnostic tapping, somatostatin	Peritoneo-venous shunt	N/A
Present case 1	F	66	PPPD + PV resection	55 (5–60)	Low-fat MCT diet, TPN, lymphangiography	Exploratory laparotomy with fat loading, suture ligation	Resolved; drain removed at POD 13
Present case 2	F	67	Completion total pancreatectomy	65 (3–68)	Low-fat MCT diet, TPN, lymphangiography	Exploratory laparotomy with fat loading, suture ligation	Resolved; drain removed at POD 7

M, male; F, female; MCT, medium-chain triglyceride; TPN, total parenteral nutrition; POD, postoperative day; N/A, not available; PPPD, pylorus-preserving pancreaticoduodenectomy; PV, portal vein.

Operative treatment aims to (1) localize the leaking lymphatic channel and (2) achieve durable control through ligation and/or reinforcement. To facilitate intraoperative identification, a fat-loading challenge can physiologically augment lymph flow, with lymphatic drainage peaking 4 to 6 hours after fat ingestion. Thus, administration 4 hours before surgery is optimal for leak localization [17]. Several approaches have been described, including the oral or nasogastric administration of milk or butter before surgery [18,19], and direct intraluminal injection of 10% lipid intraoperatively [8]. Regarding safety, current anesthesia guidelines generally recommend a minimum fasting period of 2 hours for clear liquids and 6 hours for non-clear liquids or light meals, such as milk-based products [20]. However, to capture the peak of chyle flow for optimal intraoperative localization, our institutional protocol utilizes a 4-hour fasting window following the ingestion of a specific fat load (50 mL of milk mixed with 50 g of melted butter). To mitigate potential risks, this approach is limited to patients without underlying gastric motility disorders, and anesthetic induction is performed under close monitoring by experienced anesthesiologists. This protocol strikes an effective balance between surgical necessity and patient safety.

We utilize a two-step approach for managing chyle leaks. First, we systematically explore the expected lymphatic routes around the hepatoduodenal ligament, celiac axis, and mesenteric root. Following preoperative fat loading, chyle may present as pulsatile, opalescent fluid from a pinpoint channel, prompting the application of targeted clip/tie or fine suture ligation to the identified vessels. If no distinct leak is observed, we proceed with prophylactic suture ligation of anatomically suspect channels, guided by established patterns of abdominal lymphatic drainage near the pancreas and great vessels [7]. Commonly examined areas include the cisterna chyli, para-aortic lymphatics, and retroperitoneal channels along the superior mesenteric vessels and celiac axis, which require careful inspection and stitch reinforcement. Additionally, hemostatic and adhesive oxidized regenerated cellulose sheets or fibrin sealant, may be utilized to manage minor oozing from microscopic channels.

Collectively, these cases support surgical intervention as a viable rescue option for patients experiencing persistent and clinically significant chyle leaks when conservative and interventional measures are ineffective or unavailable.

However, surgical management of postoperative chyle leaks has inherent limitations. First, re-operation necessitates additional recovery time and exposes patients to further surgical stress, which can compromise their overall condition, particularly in individuals already weakened by prior pancreatectomy and nutritional decline. Second, operative procedures carry risks of complications such as bleeding, infection, and delayed wound healing. Dense adhesions following major pancreatic surgery often require extensive adhesiolysis, increasing the risk of inadvertent bowel or vascular injury. Third, the benefits of re-operation must be carefully weighed against potential morbidity, especially in patients with advanced malignancies who may have limited tolerance for prolonged treatment or delays in resuming systemic therapy. Finally, the leakage site intraoperatively is not always assured, even with preoperative fat loading. Failure to localize the leak may result in incomplete control and recurrence. These considerations emphasize that while surgical management is an essential option for refractory cases, its application must be judiciously balanced with the patient's condition, technical feasibility, and the potential impact on the overall oncologic treatment plan.

In conclusion, in cases of refractory postoperative chyle leak following pancreatic surgery, a structured step-up algorithm is essential. Operative management can act as a salvage strategy when conservative and interventional approaches fail or are not available, although robust evidence is still limited. Preoperative fat loading can improve intraoperative localization, allowing for precise suture or clip ligation of lymphatic channels. If no focal source is identified, anatomically guided prophylactic ligations combined with adjunct packing may provide lasting control.