Bilirubin Removal with Therapeutic Plasma Exchange or Molecular Adsorbent Recirculating System as Treatment for Cholemic Nephropathy in Patients with Cirrhosis and Acute-on-Chronic Liver Failure: A Case Series

Bile cast or cholemic nephropathy (CN) is a known cause of acute kidney injury (AKI) in patients with high bilirubin levels (generally >20 mg/dL) and advanced liver disease [1]. Despite being described many decades ago, it remains an underdiagnosed cause of AKI in cirrhosis, with potentially high clinical impact [2, 3]. CN pathophysiology involves the formation of kidney tubular bile casts, composed of bilirubin, bile acids, and other bile components [4]. The diagnostic cornerstone is the presence of tubular bile casts in the kidney biopsy, though it is rarely performed in the presence of cirrhosis due to increased bleeding risk associated with the disease [5]. Alternatively, the identification of bile casts in urinary sediment is a safer surrogate diagnostic tool in this setting, although more studies are needed to standardize the technique [1, 6‒8].

Extracorporeal bilirubin and bile acid removal by albumin-coupled dialysis adsorption systems (Molecular Adsorbent Recirculating System [MARS^®]), plasma adsorption systems, or therapeutic plasma exchange (TPE) have been reported as potential treatments for CN associated AKI [9‒13]. Whether their effects solely rely on bilirubin clearance or other bile components remains unclear. We present a contemporary cohort of 15 cirrhotic patients with severe cholemic AKI (KDIGO 3) [14] in whom bilirubin removal treatments (TPE or MARS) were applied to avoid further kidney derangement.

We retrospectively identified all patients between 2016 and 2023 admitted to our unit with diagnosis of cirrhosis and cholemic-associated AKI in whom a bilirubin removal therapy was applied (MARS or TPE). Cirrhosis was diagnosed based on clinical, ultrasound, and laboratory data. All included patients accomplished acute-on-chronic liver failure criteria since presenting at least two organ dysfunctions (liver and kidney). Diagnosis of CN was reached in the setting of kidney dysfunction and high bilirubin levels (all patients above 30 mg/dL) after excluding and treating other causes of AKI (volume depletion, hepatorenal syndrome, nephrotoxicity). The diagnosis of CN was finally established by direct observation (no staining) at optical microscope of biliary casts after centrifugation of a urine sample (10 mL, 10 min, 1,000 rpm), shown in Figure 1. Kidney biopsies were not considered due to bleeding risks in this setting. These patients, with grade III AKI (KDIGO) attributable to CN, were considered for bilirubin and bile component removal techniques to prevent further kidney derangement; most of them were candidates or already listed for liver transplantation (LT). The first 3 patients (2016–18) were treated with MARS system (usual configuration, 8 h minimum per session) [15] and the following 12 with TPE (2019–23). TPE exchanged 1.2–1.4 plasma volumes in each session with a median composition of 40% fresh frozen plasma and 60% Albumin 5% (Spectra Optia cell separator; TerumoBCT, Lakewood, CO, USA) [13]. A descriptive report of liver and kidney variables (hemodialysis, urine output, creatinine) and clinical outcomes up to 1 year is provided. Data were retrieved from electronic medical registries. The study was approved by the Local Ethics Committee (2022-0985 and 2023-0196).

Fifteen patients with cirrhosis and AKI stage III (KDIGO) secondary to CN underwent MARS therapy or TPE between 2016 and 2023. Clinical characteristics are reported in Table 1; all of them accomplished acute-on-chronic liver failure criteria due to liver and kidney dysfunction. Cholemic triggers were superimposed hepatitis in 5 patients (2 immune hepatitis, 2 viral, 1 post-embolization), 4 alcohol-associated hepatitis, and baseline etiology progression in others (cholestatic, viral, alcohol). Median bilirubin and creatinine before treatment were 40.7 and 3.04 mg/dL, respectively. Urine output was relatively preserved (aquaporin 2 damage has been reported in this setting [4, 5]) with a variable degree of proteinuria in most patients (median of 552 mg/24 h). Normalized median neutrophil gelatinase-associated lipocalin (n = 13) levels were 666 μg/g; 4 patients also had urinary β2-microglobulin (median 5,848 μg/L). MARS sessions were performed on a daily regimen while TPE sessions were applied on days 1, 2, and every 48 h later on, targeting bilirubin <20 mg/dL. Ten patients received 3–4 sessions of TPE/MARS and 5 patients underwent extended TPE (7–12 in total) as a bridge to LT due to bilirubin increase and AKI recurrence after initial sessions. Treatments were well tolerated without major adverse events.

After treatment, most patients improved or normalized kidney function with preserved and improved diuresis in parallel to bilirubin decrease (Fig. 2). Five patients required intermittent hemodialysis (2–3 sessions) concomitantly to TPE/MARS treatment due to severe baseline AKI, recovering kidney function thereafter in 4 of the 5 patients. Only 1 of these patients worsened their kidney function after the first sessions of TPE (new bilirubin rise) that required chronic dialysis and combined kidney-liver transplantation a few months later. Creatinine and bilirubin values significantly decreased from intervention to hospital discharge or before LT (if still admitted) as shown in Figure 2. Ten patients underwent LT during the first year (median time 46 days), 7 of whom while on extended TPE sessions. Four patients died: 3 by day 30 without LT and another at 1 year after being transplanted.

Since the initial description of CN in 1899 from autopsies of patients with hyperbilirubinemia, work has been done to identify and clarify the underlying mechanisms for the development of kidney injury in this setting, but they remain still unknown. In the presence of jaundice, bilirubin and bile acids are excreted via the urine in a higher amount than usual, which may be nephrotoxic by leading to mitochondrial damage and inhibition of oxidative phosphorylation [1]. However, there is no solid evidence to blame bilirubin as the main cause of tubular injury as it may be just a bystander of parallel pathological mechanisms or even play a nephroprotective role. Bile acids are pathogenic suspicions in CN, the deposition of which has been demonstrated as causing tubular damage in animal models [4, 16].

The typical finding in kidney biopsies of patients with CN is the presence of intratubular casts positive for Hall’s stain, indicative of bilirubin content, associated with signs of tubular damage [2]. There is evidence that these cholemic casts can precipitate into urine being mirrored in urinary sediment [17, 18], but no standardized protocols for their interpretation are available currently. Indeed, we used this surrogate diagnostic approach for practical and safety reasons after excluding other causes of AKI. The role of tubular damage biomarkers in this clinical scenario remains unexplored though they would be expected to increase as opposed to a pure functional dysfunction like hepatorenal syndrome-AKI. In fact, in our series, most patients had increased normalized neutrophil gelatinase-associated lipocalin (or β2-microglobulin) with only 4 of them below the cutoff 220 µg/g. The role of other promising biomarkers (TIMP-2, IGFBP7) remains to be clarified in this setting.

Therefore, CN remains a widely largely overlooked cause of AKI in cirrhosis with important implications since it may have an effective treatment through bilirubin and other bile component removal. Reluctance for kidney biopsies in cirrhosis and the lack of validated noninvasive tools (casts visualization) further make its diagnosis difficult. Individual reports on bilirubin removal to treat CN have been published, though the present series is the largest one reporting these therapeutic strategies. We applied them as a rescue measure to avoid worsening kidney function and the need of renal replacement therapy, which markedly darkens prognosis in these patients [19]. In fact, 5 patients required transient dialysis due to mandatory indications (hypervolemia, uremia, pH-electrolytes). All patients initially stabilized and improved kidney function while bilirubin decreased and stabilized around or below 20 mg/dL, allowing to reach LT (n = 6) in a few weeks. Only 1 patient worsened kidney function, finally needing chronic hemodialysis and combined transplantation. Probably this approach avoided the need of further kidney replacement and transplantation in many of them, though a control group is lacking to confirm this hypothesis.

The use of MARS or TPE obeyed temporal and clinical trends: the first 3 patients received MARS (2016–18) while the next 6 patients received plasma exchange after the fully adoption in the unit of these strategy for ALF [20]. The use of combined albumin and plasma for TPE was based on costs and safety issues [21]. TPE may have a benefit over MARS in terms of duration (3 h vs. 8 h), costs, tolerance, easier settings and greater bilirubin decreases with each session. It must be noted that other adsorption systems (double plasma molecular adsorption and plasma and coupled plasma fraction adsorption systems) have shown efficacy to diminish bilirubin and bile acids in similar settings, so they also could be alternative strategies for CN [12, 13]. The final decision will probably depend on local preferences, availability, costs and the best option will be hard to identify in a controlled way.

Limitations of the series rely on its observational and retrospective nature without control group, the use of different techniques to remove bilirubin, and the use of a surrogate diagnostic tool (urinary bile casts) to diagnose CN. However, all these factors point at important unmet needs for CN: the improvement and standardization of noninvasive diagnosis by visualization and quantification of urinary bile casts, and the design of controlled studies to confirm bilirubin and bile component removal to manage cholemic AKI. Our series, the largest reporting bilirubin removal strategies to treat CN, shows promising results with relatively safe and simple interventions (TPE easier than MARS), which may preserve kidney and avoid dialysis or even a combined liver-kidney transplantation in these very sick patients.

This study protocol was approved by the Local Ethics Committee “Comité de ética de la investigación con medicamentos, Hospital Clinic de Barcelona,” Approval Nos. HCB/2022/0985 and HCB/2023/0196. Due to the retrospective nature of the study, the need for informed consent was waived by “Comité de ética de la investigación con medicamentos, Hospital Clinic de Barcelona.”

Authors disclose no conflicts of interest regarding the present manuscript.

N.J.-E. is supported by a research contract from Hospital Clínic (“Emili Letang-Josep Font” end of residency awards 2023).

E.R., N.J.-E., G.P., and J.F. were responsible for study concept and design. E.R., N.J.-E., A.C., O.O., G.P., L.R.-C., M.C.S., M.L., D.T., J.C., O.B., M.S., E.C.-P., and P.C. participated in the management and screening of patients and acquisition of data. E.R. and N.J.-E. were responsible for analysis and interpretation of data and drafting of the manuscript. J.F., E.P., and M.L. provided an expert critical revision of the manuscript. E.R. is the guarantor of the article. All authors read and approved the final version of the paper.

The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants but are available from E.R. ([email protected]) upon reasonable request.