A single center retrospective study: Comparison between centrifugal separation plasma exchange with ACD‐A and membrane separation plasma exchange with heparin on acute liver failure and acute on chronic liver failure

The purpose of this retrospective study is to compare the efficacy and safety of the centrifugal separation therapeutic plasma exchange (TPE) using citrate anticoagulant (cTPEc) with membrane separation TPE using heparin anticoagulant (mTPEh) in liver failure patients. The patients treated by cTPEc were defined as cTPEc group and those treated by mTPEh were defined as mTPEh group, respectively. Clinical characteristics were compared between the two groups. Survival analyses of two groups and subgroups classified by the model for end‐stage liver disease (MELD) score were performed by Kaplan–Meier method and were compared by the log‐rank test. In this study, there were 51 patients in cTPEc group and 18 patients in mTPEh group, respectively. The overall 28‐day survival rate was 76% (39/51) in cTPEc group and 61% (11/18) in mTPEh group (P > .05). The 90‐day survival rate was 69% (35/51) in cTPEc group and 50% (9/18) in mTPEh group (P > .05). MELD score = 30 was the best cut‐off value to predict the prognosis of patients with liver failure treated with TPE, in mTPEh group as well as cTPEc group. The median of total calcium/ionized calcium ratio (2.84, range from 2.20 to 3.71) after cTPEc was significantly higher than the ratio (1.97, range from 1.73 to 3.19) before cTPEc (P < .001). However, there was no significant difference between the mean concentrations of total calcium before cTPEc and at 48 h after cTPEc. Our study concludes that there was no statistically significant difference in survival rate and complications between cTPEc and mTPEh groups. The liver failure patients tolerated cTPEc treatment via peripheral vascular access with the prognosis similar to mTPEh. The prognosis in patients with MELD score < 30 was better than in patients with MELD score ≥ 30 in both groups. In this study, the patients with acute liver failure (ALF) and acute on chronic liver failure (ACLF) treated with cTPEc tolerated the TPE frequency of every other day without significant clinical adverse event of hypocalcemia with similar outcomes to the mTPEh treatment. For liver failure patients treated with cTPEc, close clinical observation and monitoring ionized calcium are necessary to ensure the patients' safety.


| INTRODUCTION
Liver failure in the context of either ALF or ACLF is associated with high mortality in the absence of a liver transplantation.The mechanisms of liver failure are thought to be driven by excessive systemic inflammation and dysregulated immune activation triggered by both microbial and non-microbial factors. 1 The liver failure patients require the extracorporeal removal of large compounds from the blood, including albumin-bound and water-soluble toxins and replacement with plasma and/or albumin.High volume therapeutic plasma exchange (TPE-HV), defined as exchange of 8-12 L of plasma volume per procedure, has been considered an effective bridging therapy in patients with liver failure, one procedure per day until they either undergo liver transplantation or spontaneous regeneration. 2According to American Society for Apheresis (ASFA) guidelines 2023, acute liver failure is category I indication of TPE-HV with grade 1A recommendation, while it is category III indication of TPE with grade 2B recommendation. 3Plasma exchange can be performed either by membrane separation or centrifugal separation.Membrane separation TPE using heparin anticoagulation (mTPEh) is a common TPE treatment method used in patients with liver failure. 4The mTPEh method needs central venous access, because it requires higher blood flow rate (100-150 mL/min).Patients with liver failure are prone to catheter-related infection and bleeding.In recent years, citrate anticoagulation has become a favorable alternative to heparin in critically ill patients at a high risk of bleeding.The required blood flow rate (50 mL/min) is lower for centrifugal separation plasma exchange using citrate anticoagulation (cTPEc) than that (150 mL/min) for mTPEh, which reduces the impact on hemodynamics and is especially suitable for patients with hemodynamic instability.Therefore, cTPEc could use superficial venous access, which reduces the catheter-related complications of central vein catheterization.But it is well-known that liver failure patients have increased risk to develop the citrate toxicity, because the citrate is mainly metabolized in the liver.A previous study showed that despite the occurrence of significant citrate accumulation after using regional citrate anticoagulation (RCA) in continuous renal replacement therapy (CRRT) in patients with liver failure, the side effects of citrate accumulation were not as severe as were expected. 5It may be time to reassess the use of RCA in patients with liver failure. 6However, more studies on the safety of citrate anticoagulation during TPEs in patients with liver failure are needed.
The purpose of this retrospective study is to compare the efficacy and safety of cTPEc with mTPEh in liver failure patients.

| Patients and study cohort
This is a retrospective study including all consecutive liver failure patients treated with TPE as an artificial bridging to liver transplantation or liver function recovery from a single large academic medical center, at the Department of Infectious Diseases, Huashan Hospital, Shanghai medical college, Fudan University, Shanghai, China from January 2014 to December 2021.The ALF was a well-defined medical emergency which was defined as a severe liver injury, leading to coagulation abnormality usually with an international normalized ratio (INR) ≥1.5, and any degree of mental status alteration (encephalopathy) in a patient without pre-existing liver disease and with an illness of up to 4 weeks duration.The ACLF was an acute hepatic insult manifesting as jaundice (serum bilirubin ≥5 mg/dL) and coagulopathy (INR ≥1.5) complicated within 4 weeks by clinical ascites and/or encephalopathy in a patient with underlying chronic liver disease or cirrhosis formalized by consensus recommendations of the Asian Pacific Association for the Study of the Liver. 2 However, the definition of ACLF, proposed by the European Association for the Study of the Liver-Chronic Liver Failure Consortium, includes extrahepatic organ failures, applies only to patients with acutely decompensated cirrhosis, with or without prior decompensation, and does not exclude extrahepatic precipitating events.Organ failures are identified with the use of a modified Sequential Organ Failure Assessment score, in which patients with serum total bilirubin ≥12 mg/dL receive a liver score of 3 points. 7Like the European definition, the definition of ACLF of the North American Consortium for the Study of End-Stage Liver Disease applies only to patients with acutely decompensated cirrhosis, does not exclude extrahepatic precipitating events, and considers organ failures as components of the syndrome. 8Thus, the total bilirubin of ACLF patients was defined as ≥10 mg/dL in this study, according to the liver failure guidelines of China, 9 which is closer to the definition of total bilirubin in the patients with liver failure proposed by the North American Consortium and European Association for the Study of the Liver-Chronic Liver Failure Consortium.Exclusion criteria were: patients aged less than 14 years old; patients infected with human immunodeficiency virus; patients during pregnancy or lactation; patients not treated with TPE.The patients treated by cTPEc were defined as cTPEc group and those treated by mTPEh were defined as mTPEh group, respectively.The patients were followed up for 3 months to 2 years.All the patients or their representatives provided the written informed consent.The study was performed in accordance with the Helsinki Declaration and was approved by the Ethical Committee of Huashan Hospital, Fudan University.

| Clinical characteristics and outcome parameters
The clinical characteristics including age, gender, total bilirubin, serum creatinine, INR, platelet count and etiology of liver failure, as well as the pre and post arterial blood gas analysis of cTPEc patients only, were obtained from patients' medical records.Post-TPE specimens were collected at 10 min after cTPEc.The critical value of the total calcium was <1.5 mmol/L and ionized calcium was <0.37 mmol/L at our institution.The blood gas analysis was not performed in the mTPEh group.All patients were evaluated by the model for end-stage liver disease (MELD) as a prognostic scoring system.MELD scores were calculated as follows: 9.6 Â ln[creatinine (mg/dL)] + 3.8 Â ln[bilirubin (mg/dL)] + 11.2 Â ln(INR) + 6.4 (etiology: 0 if cholestatic or alcoholic, 1 otherwise).Laboratory values less than 1.0 are set to 1.0 for the purposes of the MELD scores calculation.If serum creatinine is >4.0 mg/dL, the data is treated as 4.0 for calculation. 10he "overall survival rate" was defined as the number of patients survived (including the transplantation patients) of the total number of patients in each arm (cTPEc arm or mTPEh arm).The "survival rate of transplantation-free patients" was defined as the number of transplantation-free patients survived of the total number of transplantation-free patients in each arm.Each survival rate was calculated at 28 and 90 days after admission.

| TPE and anticoagulation
TPE was performed by centrifugal separation method on the hemocyte separator (FRESENIUS KABI COM.TEC, Germany) using blood preservation solution I (ACD-A, Shanghai Blood Biomedicine Co., Ltd., China) or membrane separation plasma exchange machine (KURURAY KM8900, Japan) using heparin anticoagulation.The plasma volume exchanged in each procedure was 1.0 total plasma volume of the patient calculated according to body weight and hematocrit.100% of replacement fluid was plasma for all procedures without albumin in this study.The blood flow rate during the procedure was around 50 mL/min in cTPEc and 150 mL/min in mTPEh and was adjusted according to each patient's conditions.ACD-A was infused immediately in draw line port automatically with the whole blood to the ACD-A ratio of 14:1-16:1 during cTPEc procedures.Calcium gluconate (each 3 g diluted with 5% glucose solution 50 mL) was infused with 40-60 mL/h via return line, about 1.5 g calcium gluconate per liter exchanged plasma, about 3-5 g calcium gluconate for each cTPEc procedure.For the cTPEc procedures, peripheral vascular access was used by two superficial vein indwelling 18G needles (BD Intima II, Beckton Dickinson Medical, Franklin Lakes, NJ, USA) without using a central venous catheter or port, with the median cubital vein as draw line, and the superficial vein as return line, since the blood flow rate is lower.Arterial access is atypical in most apheresis settings.For cTPEc procedures in the patients without an adequate superficial venous access, we used radial artery or dorsalis pedis artery access, to avoid using central venous access due to its related complications. 11The radial artery or dorsalis pedis artery was used as access for the draw line due to a higher blood flow rate, but all the patients used peripheral venous access for the return line.The needle was removed after each procedure.For mTPEh patients, the loading dose of heparin 0.3-0.5 mg/kg was given before the start of the procedure, then heparin was maintained at 0.1-0.2mg/kg/h via the return line of the blood.For all mTPEh procedures, the central venous access via femoral vein or internal jugular vein was obtained with ARROW two-lumen hemodialysis catheter 12F (Arrowg+ard Blue, Teleflex, Morrisville, USA).The advantages and disadvantages of each method are listed in Table 1.All patients were treated as inpatients.The number and frequency of TPE procedures were decided based on both clinical assessment and the availability of plasma since the plasma is not always available in our institution.No patients underwent daily TPE in this study.Each patient received one TPE treatment method, cTPEc or mTPEh.The type of TPE treatment (either cTPEc or mTPEh) for each patient was assigned by their attending physician who requested the apheresis treatment.

| Statistical analysis
Statistical analyses were performed with the Graphpad 9.3.1 (Graphpad Software, San Diego, CA).Variables were expressed as mean ± SD or median (range) unless otherwise specified.Survival probabilities were estimated by means of Kaplan-Meier method and were compared by the log-rank test.The performance of prognostic scores on the prediction of mortality was assessed by the receiver operating characteristic curve (ROC curve).Differences in the parameters were compared using the unpaired parametric t-test, paired parametric t-test or nonparametric Mann-Whitney U test as needed.A two-tailed P value of <.05 was considered statistically significant.collected at the admission day (day 0) and described in Table 2.In our study, the number of TPE procedures for each patient was a median of 3 procedures ranging from 1 procedure to 6 procedures; the frequency of the TPE procedures for each patient was a median of 2 procedures/week, ranging from 1 procedure/week to 3 procedures/week.The duration from admission to last TPE was a median of 2 weeks, ranging from 1 to 4 weeks.In cTPEc group, there were 15 ALF patients, of which 8 patients were caused by drug, 3 patients by acute hepatitis E, 1 patient by acute hepatitis B, and 3 patients by unknown etiology.There were 36 ACLF patients, of which 32 patients had chronic hepatitis B, 1 patient alcoholic liver disease, 2 patients autoimmune hepatitis, and 1 patient Wilson's disease as basic liver disease.

| The baseline characteristics of the liver failure patients treated with TPE
In mTPEh group, there were 6 ALF patients, of which 3 patients were caused by drug, in 1 patient by autoimmune hepatitis, and 2 patients by unknown etiology.There were 12 ACLF patients, of which 9 patients had chronic hepatitis B, 2 patients alcoholic liver disease, and 1 patient cirrhosis with unknown etiology as basic liver disease.
Table 2 shows the baseline characteristics of all patients, which included the transplantation-free and the transplantation patients.There was no statistical significance between the cTPEc group and the mTPEh group in age, gender, total bilirubin, serum creatinine, INR, platelet count, types of liver failure and MELD score.
Table 3 shows the baseline characteristics of the patients with liver failure who did not have a liver transplantation in this study.There were 37 patients in cTPEc group, and 18 patients in mTPEh group.The baseline INR of transplantation-free patients was significantly lower in cTPEc group than that in mTPEh group.In transplantation-free patients, there was no statistical significance in the age, gender, total bilirubin, serum creatinine, platelet count, types of liver failure and MELD score between the cTPEc group and the mTPEh group.
Accuracy of the MELD score in predicting 28-day survival of the transplantation-free patients with liver failure was calculated.MELD score best-fit value was 30, and area under the ROC curve was 0.7208 (P = .008)in the transplantation-free patients with liver failure in both cTPEc group and mTPEh group.MELD score best-fit value was 30, and area under the ROC curve was 0.7033 (P < .05) in the transplantation-free patients with liver failure in cTPEc group.Therefore, the transplantationfree patients with liver failure were divided into two subgroups: MELD score < 30 group and MELD score ≥ 30 group.The 28-day survival rate was 75% (33/44) in MELD score < 30 group and 27% (3/11) in MELD score ≥ 30 group (P < .001, Figure 3A).The 90-day survival rate was 68% (30/44) in MELD score < 30 group and 0 (0/11) in MELD score ≥ 30 group (P < .001, Figure 3B).

| Safety of TPE treatment
In cTPEc group, 134 TPE procedures were performed, with an average of 3 procedures per patient.There were Note: The survival rate was the number of patients survived/total number of patients, and the total number of patients refers to all patients enrolled in this study.The survival rate in the subgroup analysis was to compare the survival rate of all non-liver transplant patients, that was, the number of non-liver transplant patients survived/total number of non-liver transplant patients.
F I G U R E 1 A comparison of cumulative overall survival between the cTPEc group and the mTPEh group, using the Kaplan-Meier curve: (A) at day 28 (P = .15);(B) at day 90 (P = .11).
10 complications: 5 patients had rashes, one patient had chill, one patient had shock and 3 patients had aggravation of hepatic encephalopathy.In mTPEh group, 46 TPE procedures were performed, also with an average of 3 procedures per patient.There were 4 complications: 2 patients had shock and 2 patients had aggravation of hepatic encephalopathy.No patient had gastrointestinal bleeding or intracranial hemorrhage.Among the patients with adverse reaction, each patient only had one event, no patients had two or more adverse reactions.The rash was treated with antihistamines and disappeared within hours.The chills lasted for about half an hour and relieved without intervention.Aggravation of encephalopathy was considered to be due to the progression of the primary liver disease.Hepatic encephalopathy was treated with lactulose and ornithine aspartate.Hepatic encephalopathy could be corrected in a few days, lasting up to 1 week.The patients who suffered from shock were considered to have shock due to an allergic reaction to plasma, so those patients did not have further TPE treatments.Noradrenaline was used to treat shock.The shock was relieved quickly, and no patient died because of shock.There was no significant difference in the occurrence of complications between the two groups (P = .66).
In cTPEc group, 40 patients used radial artery or dorsalis pedis artery access for the draw line, but all the patients used peripheral venous access for the return line in cTPEc procedures.The use of the radial artery or the dorsalis pedis artery as vascular access for the draw line in cTPEc procedures was safe and no obvious adverse consequences arose from it in this study.
The arterial blood gas data of pre-cTPEc and post-cTPEc treatment were analyzed and presented in Table 5.The concentrations of calcium, chloride, hemoglobin, F I G U R E 3 A comparison of cumulative transplantation-free survival between MELD ≥30 group and the MELD <30 group, using the Kaplan-Meier curve: (A) at day 28 (P < .001);(B) at day 90 (P < .001).
F I G U R E 2 A comparison of cumulative transplantation-free survival between the cTPEc group and the mTPEh group, using the Kaplan-Meier curve: (A) at day 28 (P = .52);(B) at day 90 (P = .50).
hematocrit and ionized calcium before cTPEc were significantly higher than after cTPEc.The concentrations of glucose, lactate, actual HCO 3 , standard HCO 3 , pH, extracellular base excess, base excess, anion gap, total calcium, and total calcium/ionized calcium post-cTPEc were significantly higher than pre-cTPEc.The concentrations of serum magnesium (Mg++) after cTPEc (0.80, range from 0.72 to 1.26) was significantly lower than before cTPEc (0.85, range from 0.74 to 1.30, P = .002).Also, the concentrations of serum magnesium at 24 h after cTPEc (0.79, range from 0.69 to 1.33, P < .001) was significantly lower than before cTPEc.However, there was no significantly significant difference between the concentrations of serum magnesium after cTPEc and at 24 h after cTPEc (P > .05).
The median of total calcium/ionized calcium ratio (2.84, range from 2.20 to 3.71) after cTPEc was significantly higher than the ratio (1.97, range from 1.73 to 3.19) before cTPEc (P < .001, Figure 4A).The concentration of ionized calcium after cTPEc (0.89 mmol/L, range from 0.73 to 1.04 mmol/L) was significantly lower than before cTPEc (1.09 mmol/L, range from 0.84 to 1.21 mmol/L, P < .001).The concentration of total calcium at 10 min after cTPEc (2.60 ± 0.27 mmol/L) was significantly higher than before cTPEc (2.15 ± 0.23 mmol/L, P < .001).Also, at 24 h after cTPEc, the mean total calcium level (2.23 ± 0.17 mmol/L) was significantly higher than before cTPEc (P = .001).However, there was no significant difference between the concentrations of total calcium before cTPEc and at 48 h after cTPEc (2.18 ± 0.09 mmol/L, P = .68,Figure 4B).Although our data showed that the post-cTPEc ionized calcium after cTPEc was significantly lower than pre-cTPEc, no definite adverse event due to low calcium was identified in our study, suggesting that liver failure patients can tolerate the cTPEc in parameter settings in this study.However, for liver failure patients treated with cTPEc, close clinical observation and monitoring ionized calcium are necessary to ensure the patients' safety.
T A B L E 5 Arterial blood gas of pre-cTPEc and post-cTPEc treatment.

Characteristics
Pre The European Association for the Study of the Liver guidelines 2016 recommend high volume TPE as a grade 1 recommendation with level I evidence in ALF, but no recommendation has been made for ACLF. 12As per ASFA guidelines 2023, ALF is category I indication for TPE-HV with grade 1A recommendation (strong recommendation, high quality evidence), but category III indication for TPE with grade 2B recommendation (weak recommendation, moderate quality evidence). 3In ASFA guidelines 2023, the ALF is defined as that acute liver failure can develop in a normal liver (ALF) or in the setting of chronic liver disease (ACLF), therefore the term ALF used in ASFA guidelines 2023 includes ALF and ACLF discussed in our paper.Currently, there is no strong recommendation made for the general use of TPE in ALF and ACLF patients. 1In this single center retrospective study, we found that there was no difference in survival rate and complications between cTPEc and mTPEh in patients with liver failure.MELD = 30 was the best cut-off value to predict the prognosis of patients with liver failure treated with plasma exchange.This cut-point may help identify liver failure patients at high risk of mortality, prompting more aggressive management, such as liver transplantation.Our results are the same as the published data that decreasing the MELD score to <30 with artificial liver support system before liver transplantation, leads to reduced overall mortality. 13PE can be performed by either centrifugation or filtration-based mechanisms.Centrifugation separates the blood into its components based on density, whereas filtration separates the plasma from the cellular components based on a hollow fiber.Both centrifugation and filtration-based TPE systems are similar in safety, efficiency, therapeutic effects. 1 TPE performed by filtration-based method using heparin anticoagulation needs central venous access.However, the patients with liver failure are prone to catheter-related infection and bleeding.In this study, all TPE procedures in cTPEc group were performed via the peripheral vascular access to avoid complications from central venous catheter.The required flow rate is slow for cTPEc, which reduces the impact on hemodynamics.This is beneficial to TPE patients, especially hemodynamically unstable patients.
In adult patients with acute kidney injury, although there is no difference in mortality between the regional citrate and heparin treated groups, regional citrate is more efficacious in prolonging circuit lifespan and reducing the risk of bleeding and should be recommended as the priority anticoagulant for critically ill patients who require CRRT. 14In recently published surveys or large clinical trials, the use of RCA is still limited to 0%-20% of the patients/treatments.The Kidney Disease: Improving Global Outcomes (KDIGO) Work Group therefore only recommends the use of citrate for anticoagulation during CRRT in patients that do not have shock or severe liver failure. 15Therefore, it has been believed that one should be cautious when using citrate anticoagulation in the patients with liver failure, because of citrate accumulation and toxicity.But some studies have found that CRRT or intermittent renal replacement therapy using RCA might be safe and effective in patients with liver failure.The increased risk of citrate accumulation is likely to be well addressed by careful monitoring and timely adjustments in treatment. 16,17In one study about citrate accumulation, the mean dose of citrate was approximately 61.3 mmol and the volume of plasma exchanged was approximately 1500 mL.Citrate accumulation was well tolerated by ACLF patients who received plasma exchange therapy without filtration and dialysis. 18As expected, significant citrate accumulation did occur, predicted using the total calcium to ionized calcium ratio, but more interestingly, the adverse reaction of citrate accumulation was not as severe as expected.The cTPEc was tolerated well in terms of citrate accumulation.
At the end of the last century, some artificial liver support systems, including Molecular Adsorbent Recirculating System, Single-Pass Albumin Dialysis system, and F I G U R E 4 A comparison of calcium before cTPEc and after cTPEc.The total calcium/ionized calcium value (A) after cTPEc was significantly higher than before cTPEc (P < .001).Symbol "****" in panel (A) means P < .001.The concentration of total calcium (B) after cTPEc was significantly higher than before cTPEc (P < .001),as well as that at 24 h after cTPEc (P = .001)was also significantly higher than before cTPEc.However, there was no significant difference between the concentrations of total calcium before cTPEc and at 48 h after cTPEc (P = .68).Symbol "*" means P < .05 and abbreviation "ns" means no significant difference between two groups in panel (B).therapeutic plasma exchange alone or combined with plasma adsorption or high flux CRRT, began to be used as supportive therapy until liver transplantation. 19RCA, performed by infusion of 4% trisodium citrate solution (136 mmol/L citrate; Fresenius Kabi) before the roller pump of the hemodialysis machine with a starting dose of 2-3 mmol/L blood flow during the first treatment, can be a safe and effective method of anticoagulation during Molecular Adsorbent Recirculating System therapy. 20nother study found that the probability of filter survival in RCA group receiving Molecular Adsorbent Recirculating System therapy was 94% whereas the probability in heparin group was 82% (P = .20). 21The blood flow to Prismocitrate 18/0 (Na + 140 mmol/L, Cl 86 mmol/L, citrate 18 mmol/L, 244 mOsmol/L) flow was used in 6:1 ratio to achieve the theoretical concentration of 3 mmol/L of trisodium citrate in blood.RCA was found to be an effective and safe method of anticoagulation during albumin dialysis. 21A commercial citrate solution, used (anticoagulation citrate dextrose solution-A) for RCA, can also provide and effective anticoagulation for pediatric patients with liver failure during albumindialysis. 22A recent study found that the cTPEc is safe and well-tolerated, and it improves coagulation profile and liver function tests in critically ill liver disease patients, but the overall survival remains low (69%, 31/45). 23A 12 French femoral catheter was used for venous access in that study.No procedure-related complications were observed during the cTPEc, except for hypocalcemia. 23Another study found that cTPEc was a safe, tolerable therapy option in patients who had high total bilirubin level (>10 mg/dL) and early cTPEc may improve patient outcomes in severe acute toxic hepatitis. 24TPE was also performed using a central venous catheter in a centrifugal-based Fresenius COMTEC 204 cell separator.Here, we used peripheral vascular access by two superficial vein indwelling needles in cTPEc treatment to avoid complications from central venous catheter.RCA might be safe and effective in ACLF patients during plasma adsorption plus plasma exchange treatment. 25There are few studies of cTPEc treatment in both ALF and ACLF patients.A systematic review and metaanalysis showed that regional citrate anticoagulation might be effective and safe in liver failure patients receiving extracorporeal organ support. 26There were 19 eligible studies included, but only three studies were about TPE using RCA in the systematic review. 26Our research reveals that RCA, if the whole blood to the ACD-A ratio was 14:1 to 16:1, can be well tolerated by both ALF and ACLF patients receiving cTPEc, despite citrate accumulation, when compared with mTPEh group.
In this study, we found when the MELD score cut-off value was 30 for liver failure patients treated with TPE, area under the ROC curve was 0.7208.A retrospective study found that the hepatitis B virus infection related liver failure patients with MELD scores ≤26.6 had high negative predictive values of longer duration of citrate accumulation, receiving double plasma molecular adsorption system plus plasma exchange therapy with RCA. 27Another previous study demonstrated improved survival with TPE among patients with HBV-ACLF and MELD scores between 20 and 30 (50%) when compared to patients with MELD scores above 30 (31.7%). 28n this study, we used total calcium/ionized calcium ratio ≥ 2.5, instead of using citrate concentration to indirectly evaluate the occurrence of citrate accumulation. 29he median total calcium/ionized calcium ratio post-cTPEc was 2.84, which was significantly higher than pre-cTPEc (1.97, P < .001)(Table 5).Despite the presence of significant citrate accumulation post-cTPEc in patients with liver failure, there was no difference in survival rate and complications between cTPEc group and mTPEh group.Therefore, the adverse reactions of citrate accumulation are not as serious as expected.There was no significant difference between the concentrations of total calcium pre-cTPEc and at 48 h after cTPEc.Unfortunately, the concentrations of ionized calcium at 48 h were not detected in this study.Further research is required to confirm whether 2 days interval is adequate for the population of patients to metabolize and clear the citrated received from the previous cTPEc.However, the most published studies showed that the TPEs performed 2-3 procedures/ week with plasma as replacement fluid were well tolerated, until the patients clinically improved, liver -transplanted, or expired in the ACLF population. 1 Our study has several limitations.First, there may be bias in a single-center retrospective study and some data are incomplete.The sample size of the heparin arm is smaller, limiting the significance of the findings.We acknowledge the impact of possible confounding factors due to the absence of propensity score matching, such as a higher baseline serum creatinine of citrate arm than heparin arm, possibly contributing to hypocalcemia in the citrate arm.Pre-TPE and post-TPE serum electrolytes and arterial blood gas analysis of patients in the heparin arm were not able to be obtained, due to retrospective nature of the study.Second, the factor of transplantation is not matched between the two groups (cTPEc and mTPEh) because this is a retrospective study.Not all patients with liver failure need TPE treatment before liver transplantation.Due to the shortage of plasma supply, some patients directly underwent liver transplantation without TPE treatment.Therefore, patients without liver transplantation in the two groups (cTPEc and mTPEh) were selected for subgroup analysis.Since liver transplantation is a very important treatment and may affect the prognosis, there may be bias in the comparison of the prognosis of the all patients.Third, only the MELD scoring system were used in this study, since more often it is used to decide whether to have a liver transplant.The other prognostic scoring systems were not used in this study, such as MELD-sodium, Child-Turcotte-Pugh, sequential organ failure assessment score (SOFA), etc. [30][31][32][33] Lastly, the MELD score cut-off point to predict the prognosis for the patients with ALF or ACLF treated with TPEs was generated based on limited number of cases in this study, which should be further validated by more studies.

| CONCLUSIONS
Our study showed that there was no statistically significant difference in survival rate and complications for patient with ALF or ACLF between cTPEc and mTPEh groups.The liver failure patients tolerated cTPEc treatment using peripheral vascular access with the prognosis similar to mTPEh.The prognosis in patients with MELD score < 30 was better than in patients with MELD score ≥ 30 in both groups.In this study, the patients with ALF and ACLF treated with cTPEc tolerated the TPE frequency of every other day without significant clinical adverse event of hypocalcemia with similar outcomes to the mTPEh treatment.For liver failure patients treated with cTPEc, close clinical observation and monitoring ionized calcium are necessary to ensure the patients' safety.
There were 69 patients included in this study, 51 patients in cTPEc group and 18 patients in mTPEh group respectively.All patients were Chinese.Baseline characteristics The advantages and disadvantages of each method.
T A B L E 1Abbreviation: CRRT, continuous renal replacement therapy.
Baseline characteristics of all patients with liver failure.
T A B L E 2 a Continuous variables, presented as mean ± SD. b Non-normal distribution data, presented as median (range).T A B L E 3 Baseline characteristics of transplantation-free patients with liver failure.a Continuous variables, presented as mean ± SD. b Non-normal distribution data, presented as median (range).c Statistically significant difference.ZHU ET AL.
Outcome of liver failure patients treated with TPE.
T A B L E 4