Haemoperitoneum after an endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) of a pancreatic lesion in a peritoneal dialysis patient

Abstract

Haemoperitoneum was observed in a peritoneal dialysis (PD) patient after undergoing endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA). EUS-FNA was performed to evaluate a pancreatic cyst seen on a prekidney transplant evaluation abdominal CT scan. Haemoperitoneum cleared with a PD exchange. In this case report, we discuss aetiologies for bleeding risks in patients with chronic kidney disease and focus on haemoperitoneum in patients receiving PD. We will also explore treatment options to minimise bleeding associated with an abdominal procedure such as EUS-FNA.

Background

A literature search reveals only one case in which endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) was used in a peritoneal dialysis (PD) patient to evaluate a pancreatic cyst.¹ EUS-FNA did not result in any gastrointestinal (GI) or PD-related complications and was thus deemed a successful and safe procedure to perform in PD patients.¹ The patient was able to resume PD after a 48 hours hiatus to minimise infection, bleeding and peritoneal fluid leakage and without having to transition to haemodialysis (HD).

In 2019, President Trump signed an executive order to increase the number of patients with end-stage renal disease (ESRD) being treated with PD or a kidney transplant.² Depending on the patient’s age and comorbidities, a prekidney transplant abdominal CT may be required as part of the evaluation process. These incidental pancreatic lesions may require further evaluation with EUS-FNA. We believe the number of EUS-FNA procedures in patients receiving PD will increase.

We report the first case of haemoperitoneum in a PD patient after EUS-FNA which was used to evaluate suspicious pancreatic lesions seen on a prekidney transplant evaluation CT scan (figure 1). A discussion regarding aetiologies, evaluation strategies and management of haemoperitoneum follows the case presentation.

Figure 1.

CT scan with pancreatic lesions. Key: (1) multiple cystic lesions in the pancreas, with the largest being amulti-lobulated 5 cm lesion in the tail (numbered); (2) atrophic kidneys; (3) abdominal aorta, with extensive atherosclerosis; (4) small to moderate amount of pneumoperitoneum, likely related topatient’s peritoneal dialysis; (5) normal liver.

Case presentation

A 66-year-old woman receiving PD for type 1 diabetes mellitus-induced ESRD carries diagnoses of hypertension, hyperlipidaemia, hypothyroidism, depression, breast cancer in remission, anaemia and secondary hyperparathyroidism. The PD prescription includes 2 L of PD fluid per exchange, with 3 cycles over 8 hours during the night using a PD cycler, 0.2 L last fill and a 2 L day exchange with less than 4 hour dwell. The reason for the low day fill volume is that she gradually absorbs the PD fluid once it dwells longer than 4 hours. She started her kidney transplant evaluation shortly after starting dialysis. During the first evaluation, she was diagnosed with breast cancer which was treated with lumpectomy and radiation. After clearance from the breast surgeon and oncologist 2 years later, she resumes her kidney transplant evaluation. Due to her age and breast cancer diagnosis and treatment, she had abdominal and chest CT scans performed. The abdominal CT scan showed multiple cystic lesions in the pancreas, including a multilobulated 5 cm lesion in the tail, a 1.7×1.3 cm cystic lesion in the uncinated process, another 1.9 cm lesion in the uncinated process and a 1.2 cm cystic lesion in the head of the pancreas. The structure labelled #1 in figure 1 shows the largest multilobulated lesion in the tail. These incidental findings did not cause any symptoms such as fever, weight loss, pain or changes in PD fluid characteristics. In fact, with a combination of medications and PD, she was feeling well and employed full time. Nevertheless, these lesions require further evaluation before she could obtain medical clearance for a kidney transplant.

Investigations

The gastroenterologist suggested EUS-FNA to obtain fluid for malignancy and infection assessment. The nephrologist expressed concerns of possible bleeding, infection and PD fluid leakage any time the peritoneal membrane is violated.³ With the cooperation of the patient, PD nurses, nephrologist and gastroenterologist, the following actions were instituted to minimise these complications. She stopped any medications that may interfere with coagulation such as aspirin. Her haemoglobin was optimised for a patient with ESRD to 97 g/L with the use of an erythropoietic stimulating agent. The patient optimised her fluid and electrolyte status and improved the uremic state by performing extra dialysis for 2 days prior to the procedure. She drained the PD fluid from the peritoneum before reporting to the procedure in order to decrease leakage and peritonitis risks (by avoiding glucose in the peritoneum as a source of nutrition for bacteria).⁴ She received prophylactic antibiotics, namely, ampicillin 1 g intravenously and gentamicin 1 mg/kg intravenously in the endoscopy suite preprocedure to decrease infection risk.⁵

EUS revealed a 39×25 mm anechoic cyst with internal septations. No additional mural nodules or lesions with thickened walls or calcifications were seen in the pancreatic head or neck. The examination was limited due to a large amount of food in the stomach. Communication of the cyst with the pancreatic duct was unable to be assessed. FNA was performed with a 22-gauge Acquire FNA Needle. Approximately 1.5 mL of serosanguinous fluid was initially aspirated, which was either from the cystic wall or gastric body. Subsequently, 5 mL of clear, thick fluid was aspirated from the cyst. There were no complications, and the procedure was well tolerated.

PD was suspended for 48 hours postprocedure to allow the puncture wound to heal in order to prevent PD fluid leakage and to avoid disrupting the blood clot by PD fluid in order to maintain haemostasis.⁶ When she resumed PD, she reported asymptomatic bloody PD fluid with the first exchange (figure 2), which subsequently cleared with another exchange. She did not report cloudy peritoneal fluid, PD fluid leakage or GI symptoms such as nausea, vomiting, diarrhoea, changes in stool consistency or abdominal pain.

Figure 2.

Blood in peritoneal dialysis fluid.

The PD catheter represents a window to the peritoneum. A drop of blood will render the PD fluid bloody. Haemoperitoneum has been reported in PD patients.⁷

Differential diagnosis

The aetiology for haemoperitoneum in a patient receiving peritoneal diagnosis is the same as the general population plus complications related to the peritoneal catheter (during its placement or due its presence in the peritoneum) and PD. Any structure in or contiguous to the peritoneum or peritoneal lining may be a source for haemorrhage. Thus, structures related to the GI tract, urological, gynaecological or vascular systems may be involved. In addition, trauma, infection (peritonitis) and coagulation disorders may result in haemoperitoneum.

Common causes for haemoperitoneum include infection such as peritonitis and gynaecological-related matters such as menstruation and ovulation.^{7 8} Reported unusual causes for haemoperitoneum include other gynaecological issues such as haemorrhagic luteal cyst and ovarian cyst rupture; intra-abdominal lesions such as kidney or liver cyst rupture, kidney or liver cancers, liver or splenic rupture; procedure-related causes such as pericardiocentesis, radiation, and colonoscopy; bleeding diatheses such as uremic platelet dysfunction and anaemia and vascular issues such as aneurysm rupture and erosion of mesenteric vessel by a Tenckhoff catheter.⁷

The cause for haemoperitoneum in this patient was not due to peritonitis as she did not complain of abdominal pain or cloudy fluid. Moreover, peritonitis would be unlikely since she received prophylactic antibiotics and had no glucose-based fluid in the peritoneum. Gynecology-related issues was unlikely as she is postmenopausal. According to her CT scan, she did not have liver or kidney cysts or cancerous lesions. She did not have a catastrophic bleed from a major vessel as she was asymptomic (she was not hypotensive or weak). Bleeding diathesis from platelet dysfunction and anaemia could have contributed to haemoperitoneum. Haemoperitoneum has not been associated with an upper endoscopic procedure. If haemoperitoneum was caused by the EUS-FNA procedure, it would be the first to be reported. EUS-FNA associated haemoperitoneum would be categorised as procedure-related.

Treatment

Patients with chronic kidney disease, especially those with ESRD have a higher risk for bleeding. Uraemic platelet dysfunction and anaemia contribute to bleeding risk.⁹ In a PD patient undergoing endoscopy, haemoperitoneum may be prevented by stopping any medication that interferes with coagulation, optimising platelet function,^{10 11} correcting anaemia to a target haemoglobin above 80 g/L,¹² avoiding clot disruption (eg, a clot formed after a procedure, or a surgical incision) by delaying PD for 48 hours⁶ and administering cryoprecipitate¹³ or desmopressin or1-deamino-8-D-arginine vasopressin¹⁰ to individuals at high risk of bleeding or those actively bleeding.

A PD patient undergoing EUS-FNA should have their coagulation status optimised and be forewarned that haemoperitoneum may be observed with the first exchange post procedure. A subsequent rapid exchange producing a lighter pink colour or clear effluent would suggest that bleeding was limited at the time of the procedure. However, persistent bloody effluent would suggest ongoing bleeding. The patient should then seek medical attention for further evaluation and management of haemoperitoneum.

Evaluation for common causes for haemoperitoneum such as peritonitis and gynaecological-related issues should be conducted initially. PD related peritonitis can be treated with antibiotics.⁵ The initiation of empiric antibiotics with gram positive and negative coverage may be used until culture identifies the organism and its sensitivity. Gynaecological-related haemoperitoneum (ie, ovulation or retrograde menstrual flow) occasionally clears with a few rapid exchanges. It has been recommended that the patient use room temperature dialysis fluid to perform a PD exchange in an attempt to abate idiopathic bleeding.¹⁴ Ongoing haemoperitoneum not responding to a few rapid exchanges with cool dialysis fluid will require further evaluation with imaging studies such as CT or MRI to define the source of bleeding. Additional approaches to haemostasis will depend on its source. Some examples include endovascular intervention, laparoscopy and exploratory laparotomy. Exploratory laparotomy would be the last diagnostic/treatment option because surgical invasion of the peritoneum may preclude or limit PD as a treatment option.

Outcome and follow-up

Fluid analysis revealed elevated carcinoembryonic antigen (12 870 ng/mL) and low amylase (5.1 U/L) consistent with a mucinous cyst, either side-branch intraductal papillary mucinous neoplasm or mucinous cystic neoplasm. She was cleared to proceed with kidney transplant evaluation.

With respect to haemoperitoneum, bloody fluid was observed in only the first exchange. Luckily for our patient, asymptomatic bleeding appeared to have occurred at the time of the procedure and not ongoing. Since the subsequent PD exchanges resulted in clear fluid, no additional action was needed. She did not experience any short-term consequences from haemoperitoneum. However, 11 months after the procedure, she transitioned to HD due to PD catheter malfunction. At the time of PD catheter removal, another catheter was not inserted due to observed peritoneal membrane inflammation by the surgeon.

Discussion

A well-established minimally invasive GI procedure, EUS-FNA has been recognised as the standard approach to diagnose and stage cancer of the pancreas, upper GI tract and mediastinum.¹⁵ The EUS-FNA procedure allows the endoscopist to guide a small gauged-needle through the intestinal wall to aspirate contents from a pancreatic lesion by using an endoscope with ultrasound capability. EUS has become the primary tool for diagnosing pancreatic cancer and tissue obtained via FNA has significantly improved diagnostic accuracy.¹⁶

Morbidity and mortality rates from EUS-FNA have been reported to be <1%.^{17 18} Complications include perforation, infection, iatrogenic pancreatitis, haemorrhage, bile peritonitis and malignant seeding.¹⁸

Risk of significant bleeding occurs in about 1% of EUS-FNA of pancreatic lesions with <1% risk of death. Mild intraluminal bleeding (4%)¹⁹ and acute extraluminal haemorrhage associated with EUS-FNA have been reported.²⁰ Our patient represents another example of extraluminal haemorrhage after EUS-FNA and the first in an individual treated with PD. Extraluminal haemorrhage was observed only because she has a PD catheter which provides a ‘window into the peritoneum’. Such a finding suggests that extraluminal haemorrhage may occur more frequently than reported. Symptomatic extraluminal haemorrhage would warrant additional imaging and intervention to stop haemorrhage. Although she has a PD catheter that allows a ‘window into the peritoneum’, we opted to hold PD for 48 hours to avoid disturbing the blood clot with PD fluid rather than to diagnose haemoperitoneum. The risk for haemorrhage associated with EUS-FNA remains low and insignificant as our case demonstrated.

Learning points.

• Patients with end-stage renal disease have a higher risk for bleeding due to platelet dysfunction and anaemia.

• Patients with a peritoneal dialysis catheter will observe blood-tinged peritoneal dialysis effluent when bleeding occurs in the peritoneum.

• Bleeding complications can be minimised by optimising coagulation and correcting uraemia and anaemia prior to an invasive procedure.

• Despite haemoperitoneum, endoscopic ultrasound-guided fine-needle aspiration is a relatively safe procedure for evaluating a pancreatic cyst.