Hemodialysis-Related Portal-Systemic Encephalopathy Possibly Associated with Right Femoral Vein Thrombosis

Abstract

In hemodialysis-related portal systemic encephalopathy (HRPSE), transient negative pressure in the inferior vena cava (IVC) during dialysis increases the blood flow through a portal-systemic shunt, leading to encephalopathy. We report the case of a 74-year-old man with a gastrorenal shunt who developed HRPSE for the first time following venous occlusion due to thrombosis around a right femoral tunneled-cuffed hemodialysis catheter. Before the thrombosis dissolved, ammonia levels increased after dialysis. Conversely, after the thrombosis was dissolved, the ammonia levels decreased after dialysis. We hypothesized that venous stasis in the right lower limb due to thrombosis intensified the negative pressure in the IVC during dialysis, triggering HRPSE.

Introduction

Portosystemic shunts often form a compensatory mechanism for portal hypertension caused by liver cirrhosis. Portal-systemic shunt encephalopathy occurs when ammonia-rich portal blood bypasses the liver through the shunt (1). Encephalopathy can occur more frequently after dialysis in patients undergoing maintenance hemodialysis and portal-systemic shunts. Ubara et al. (2) defined this condition as hemodialysis-related portal systemic encephalopathy (HRPSE) and attributed it to increased shunt blood flow caused by transient negative pressure in the inferior vena cava (IVC) during ultrafiltration. We herein present a case of HRPSE that was potentially associated with deep vein thrombosis (DVT) around a tunnel-cuffed hemodialysis catheter in the right femoral vein.

Case Report

A 74-year-old man presented with a sudden onset of worsening edema in his right lower limb. He had a history of hypertension and chronic kidney disease due to benign nephrosclerosis, which had been present for a long time and resulted in the initiation of hemodialysis 4 years prior. At that time, computed tomography (CT) showed liver margin blunting and a 10-mm gastrorenal shunt (Fig. 1a-c; CT images during hospitalization), and blood tests revealed elevated liver fibrosis markers suggesting liver cirrhosis. The patient tested negative for hepatitis B and C, and did not have a history of excessive alcohol consumption, which led to suspicion of a nonalcoholic steatohepatitis. He had no ascites and his ammonia levels were normal at 32 μg/dL. Additionally, he did not exhibit any prior symptoms of hepatic encephalopathy, even after the initiation of hemodialysis. His medical history included coronary artery bypass grafting for ischemic heart disease, ablation for atrial fibrillation, transcatheter aortic valve implantation for aortic stenosis, carotid artery stenting for bilateral carotid artery stenosis, and right nephrectomy for renal cancer. The patient was treated with multiple medications, including clopidogrel sulfate. Six months earlier, he had consulted a respiratory physician due to right pleural effusion, of unclear origin, possibly resulting from cancer or infection. The patient received an arteriovenous fistula and a superficialized brachial artery as vascular access for hemodialysis, but unfortunately, they became occluded. Therefore, the medical team considered an arteriovenous graft as the next vascular access, while utilizing a tunneled-cuffed catheter as a temporary bridge. A catheter previously placed in the right internal jugular vein caused thrombosis and stenosis. As there was compensatory dilation of the left internal jugular vein, a right femoral vein tunneled-cuffed catheter was placed 3 months prior to avoid similar complications in the left internal jugular vein, which could increase the risk of cerebral congestion if occlusion occurred. Although an arteriovenous graft procedure was being considered at another hospital, the patient has been reluctant to avoid pain associated with injections. Despite no difficulty with walking or other daily activities, the patient continued to use the femoral vein catheter.

Figure 1.

CT images of the GR shunt. (a, b) CT image of the GR shunt observed during this hospitalization. (c) Three-dimensional rendering of the GR shunt from the current hospitalization. CT: computed tomography, GR: gastrorenal

The patient was 167.0 cm tall and had a dry weight 79.8 kg, which was appropriately managed at the dialysis clinic. A physical examination revealed pitting edema distal to the right hip with mild redness, no warmth, and minimal tenderness. The left lower limb showed no edema. Blood tests revealed mild inflammation (white blood cell count, 3,470/μL; C-reactive protein level, 1.91 mg/dL) and elevated D-dimer levels (13.0 μg/mL). Contrast-enhanced CT revealed thrombosis around the catheter in the right external iliac vein causing venous occlusion (Fig. 2a). Pulmonary emboli were not observed. Based on the catheter history, options for catheter replacement sites were limited, and blood flow issues at the dialysis clinic were absent. Therefore, the right femoral tunnel cuff catheter was not immediately removed. Anticoagulation therapy was initiated for DVT.

Figure 2.

CT images of DVT. (a) CT image on day 1 showing DVT around the tunneled-cuffed dialysis catheter in the right external iliac vein, causing venous occlusion. (b) CT image on day 13 showing thrombus dissolution and relieved occlusion. CT: computed tomography, DVT: deep vein thrombosis

The clinical course of the patient is shown in Fig. 3. On day 4, the patient underwent the first hemodialysis session of the week with an initial blood pressure of 147/72 mmHg and ultrafiltration of 2.9 kg over 4 h, ending at 137/58 mmHg without hypotension. After dialysis, the patient experienced fatigue and reduced appetite that persisted on day 5. A physical examination revealed the following: heart rate, 62 bpm; blood pressure, 163/64 mmHg; respiratory rate, 16 breaths per minute; SpO₂, 98%, and body temperature, 36.8°C. He was alert with a Glasgow Coma Scale score of E4V5M6; however, asterixis was noted. The head-to-pelvis CT findings were unremarkable. The blood test results showed an elevated ammonia level of 116 μg/dL (Table). Based on the physical findings and test results (high ammonia levels), we suspected grade II hepatic encephalopathy (3). Therefore, the patient was treated with intravenous branched-chain amino acids (BCAAs), resulting in an improvement of symptoms the following morning. Day 6 blood tests showed rising ammonia levels after dialysis (68-74 μg/dL). After receiving oral BCAAs, lactulose, and rifaximin, he did not experience any additional encephalopathy episodes. By days 8-10, the right lower limb edema had resolved, D-dimer levels had decreased, and CT confirmed thrombus dissolution (Fig. 2b). On day 13, the patient's ammonia level decreased from 113 to 88 μg/dL after dialysis. The patient was discharged on day 15 on continued anticoagulation therapy, with plans for arteriovenous graft creation at another hospital to eventually remove the tunneled-cuffed catheter.

Figure 3.

Clinical course of the patient. The clinical course, thrombus status, HRPSE, and ammonia level. Changes in ammonia levels before and after dialysis were reversed by thrombus dissolution. HRPSE: hemodialysis-related portal-systemic encephalopathy, BCAA: branched-chain amino acids

Table.

Laboratory Data on Day 5 of Hospitalization.

WBC	4,410	/μL	TP	6.2	g/dL	AST	16	IU/L
Seg	61	%	Alb	2.7	g/dL	ALT	10	IU/L
Lympho	26	%	BUN	39.6	mg/dL	T-Bil	0.7	mg/dL
Mono	10	%	Cr	8.88	mg/dL	γ-GTP	33	IU/L
Eosino	2	%	Na	135	mEq/L	ALP	261	IU/L
Baso	1	%	K	4.9	mEq/L	LDH	251	IU/L
RBC	339×104	/μL	CI	103	mEq/L	Amylase	91	IU/L
Hgb	11.4	g/dL	Ca	7.8	mg/dL	CK	79	U/L
Hct	33.50	%	IP	4.7	mg/dL	Ammonia	116	μg/dL
Plt	13.7×104	/μL	Glucose	68	mg/dL	CRP	1.60	mg/dL

WBC: white blood cell, Seg: segmented neutrophil, Lympho: lymphocyte, Mono: monocyte, Eosino: eosinophil, Baso: basophil, RBC: red blood cell, Hgb: hemoglobin, Hct: hematocrit, Plt: platelet, TP: total protein, Alb: albumin, BUN: blood urea nitrogen, Cr: creatinine, Na: sodium, K: potassium, Cl: chloride, Ca: calcium, IP: inorganic phosphorus, AST: aspartate aminotransferase, ALT: alanine aminotransferase, T-Bil: total bilirubin, γ-GTP: γ-glutamyltransferase, ALP: Alkaline phosphatase, LDH: Lactate dehydrogenase, CK: Creatine kinase, CRP: C-reactive protein

Discussion

Despite the removal of small molecules, such as ammonia during hemodialysis, patients with portal-systemic shunts may experience increased ammonia levels post-dialysis, for example, from 168 to 276 μmol/L, as reported by Oi et al. (4). This phenomenon is thought to occur because of rapid ultrafiltration during dialysis, leading to a transient reduction in the blood volume and venous pressure, which facilitates the flow of portal blood into the IVC through the shunt (2). Nishikawa et al. (5) reported a decrease in portal blood flow, as measured using Doppler, both before and after dialysis. Additionally, hypokalemia and metabolic alkalosis during dialysis may contribute to increased ammonia levels (6). Takashima et al. (7) proposed the following diagnostic criteria for HRPSE: the absence of severe liver dysfunction except hyperammonemia and either a shunt diameter exceeding 10 mm or improvement in hyperammonemia and neuropsychiatric symptoms following closure of the shunt flow by surgical ligation or balloon-occluded retrograde transvenous obliteration.

In this case, the gastrorenal shunt, which was attributed to portal hypertension due to cirrhosis, was identified 1 year before the initiation of dialysis. Despite this, although the patient's ammonia level was not measured after the initiation of dialysis, the patient did not experience symptoms after dialysis, indicating that the portal-systemic shunt was not significant enough to cause HRPSE. However, after the onset of DVT, the patient displayed symptoms of encephalopathy for the first time following the initial weekly dialysis sessions. We hypothesized that DVT resulted in blood stasis in the right lower limb. This, in turn, intensified the temporary negative pressure in the IVC during dialysis ultrafiltration, which triggered HRPSE (as shown in Fig. 4). The patient's baseline ammonia level likely decreased with treatment using BCAAs, lactulose, and rifaximin, as no encephalopathy symptoms were observed after day 6, making it challenging to definitively prove the hypothesis that the HRPSE was induced by DVT. In addition, the hypothesis was difficult to verify due to the absence of Doppler ultrasound assessment of blood flow in the IVC, portal vein, or renal artery since the patient was obese. The post-dialysis ammonia levels showed an increase on day 6 but a decrease on day 13 based on the blood test results. Thus, before the thrombosis dissolved, the ammonia levels increased post-dialysis, and after the thrombosis dissolved, the ammonia levels decreased post-dialysis. These results support our hypotheses. The patient did not experience HRPSE after dialysis at the clinic during the weekend immediately after noticing edema. We believe that this might be due to incomplete DVT formation and partial venous occlusion at that time, or because dialysis on day 4 post-admission was the first session of the week, during which the negative pressure in the IVC was strongest among the three weekly sessions due to solute removal and ultrafiltration. There are no data to confirm this hypothesis.

Figure 4.

Hypothesis explaining the contribution of DVT to the onset of HRPSE. As shown in (a), venous blood rich in ammonia from the intestine is normally metabolized in the liver before entering systemic circulation. When patients with portal-systemic shunts undergo dialysis, as shown in (b), ultrafiltration temporarily creates negative pressure in the IVC, increasing shunt blood flow and delivering ammonia-rich blood to the brain, leading to HRPSE. This is the general mechanism of the HRPSE. However, in this patient, the GR shunt was not significant enough to cause an HRPSE. We hypothesized that the thrombosis in the right external iliac vein, as shown in (c), caused blood stasis in the right lower limb, which intensified negative pressure in the IVC during dialysis and increased shunt blood flow. HRPSE: hemodialysis-related portal systemic encephalopathy, DVT: deep vein thrombosis, IVC: inferior vena cava, GR: gastrorenal

In comparison to previous reports of HRPSE, although the patient in the current case showed a delayed response in conversation relative to before the onset, his consciousness level was maintained at E4V5M6, which suggests that the severity was milder than that reported by Ubara et al. (2) and Kondo et al. (8). Furthermore, there are several reports in which shunt closure was performed to treat HRPSE (8-10), and there have also been cases where encephalopathy improved in non-dialysis patients following shunt closure (11). However, in the current case, no enlargement of the shunt was observed on CT relative to previous scans, and since the symptoms improved based on the above-mentioned hypothesis, shunt closure was not performed.

Risk factors for catheter-related venous thrombosis include a history of DVT, malignancy, and infection (12). This patient may have been at high risk because of an unexplained pleural effusion on the right side, which could be indicative of an infection or cancer. A previous report suggested that when possible, femoral tunneled-cuff catheter placement should be avoided as it was found to cause DVT in 15.7% of patients with tunneled-cuff dialysis catheters in the femoral vein (13). In a previous case, repositioning the catheter tip away from the shunt improved symptoms of HRPSE in a patient with a transjugular intrahepatic portosystemic shunt (14), suggesting that the proximity of the catheter tip to the left renal vein may have played a role in the onset of HRPSE (Fig. 1b, c). However, placing the catheter tip below the right atrium and above the bifurcation of the renal vein could potentially improve the clearance of ammonia. The relationship between blood access type and HRPSE remains unclear, as a number of case reports did not provide information about the access types that they used. If additional case reports are accumulated in the future and a clear relationship is established between blood access types and positions and HRPSE, it may be feasible to consider this when choosing blood access for patients who have portal-systemic shunts.

In conclusion, we encountered a case in which DVT around a right femoral tunnel-cuffed dialysis catheter possibly exacerbated the negative pressure in the IVC during dialysis ultrafiltration, contributing to the onset of HRPSE.

Informed consent was obtained from the patient described in this report.

The authors state that they have no Conflict of Interest (COI).