Advertisement

Granulocyte-colony stimulating factor (G-CSF) to treat acute-on-chronic liver failure: A multicenter randomized trial (GRAFT study)

Open AccessPublished:August 05, 2021DOI:https://doi.org/10.1016/j.jhep.2021.07.033

      Highlights

      • G-CSF was expected to be a novel therapy for acute-on-chronic liver failure.
      • In this first multicenter, randomized phase II trial, G-CSF did not improve patient survival.
      • G-CSF was unable to reduce the rate of complications and did not improve organ function.

      Background & Aims

      Based on positive results from small single center studies, granulocyte-colony stimulating factor (G-CSF) is being widely used for the treatment of patients with acute-on-chronic liver failure (ACLF). Herein, we aimed to evaluate the safety and efficacy of G-CSF in patients with ACLF.

      Methods

      In this multicenter, prospective, controlled, open-label phase II study, 176 patients with ACLF (EASL-CLIF criteria) were randomized to receive G-CSF (5 μg/kg daily for the first 5 days and every third day thereafter until day 26) plus standard medical therapy (SMT) (n = 88) or SMT alone. The primary efficacy endpoint was 90-day transplant-free survival analyzed by Cox regression modeling. The key secondary endpoints were overall and transplant-free survival after 360 days, the development of ACLF-related complications, and the course of liver function scores during the entire observation period.

      Results

      Patients treated with G-CSF had a 90-day transplant-free survival rate of 34.1% compared to 37.5% in the SMT group (hazard ratio [HR] 1.05; 95% CI 0.711–1.551; p = 0.805). Transplant-free and overall survival at 360 days did not differ between the 2 arms (HR 0.998; 95% CI 0.697–1.430; p = 0.992 and HR 1.058; 95% CI 0.727–1.548; p = 0.768, respectively). G-CSF did not improve liver function scores, the occurrence of infections, or survival in subgroups of patients without infections, with alcohol-related ACLF, or with ACLF defined by the APASL criteria. Sixty-one serious adverse events were reported in the G-CSF+SMT group and 57 were reported in the SMT group. In total, 7 drug-related serious adverse reactions occurred in the G-CSF group. The study was prematurely terminated due to futility after conditional power calculation.

      Conclusions

      In contrast to previous findings, G-CSF had no significant beneficial effect on patients with ACLF in this multicenter controlled trial, which suggests that it should not be used as a standard treatment for ACLF.

      ClinicalTrials.gov number

      Lay summary

      Granulocyte-colony stimulating factor was considered as a novel treatment for acute-on-chronic liver failure (ACLF). We performed the first randomized, multicenter, controlled phase II trial, which showed that G-CSF did not improve survival or other clinical endpoints in patients with ACLF. Therefore, G-CSF should not be used to treat liver disease outside clinical studies.

      Graphical abstract

      Keywords

      Linked Article

      • G-CSF in acute-on-chronic liver failure – Art of ‘patient selection’ is paramount!
        Journal of HepatologyVol. 76Issue 2
        • Preview
          The study by Engelmann et al. failed to show a significant beneficial effect of granulocyte-colony stimulating factor (G-CSF) in treating patients with acute-on-chronic liver failure (ACLF).1 The use of G-CSF neither improved 3- and 12-month transplant-free survival nor led to an improvement in model for end-stage liver disease (MELD) score or new infections. This was independent of the nature of the precipitating event, severity of ACLF or type of organ failure. We believe the negative results in this study are mainly due to limitations in patient selection, as these cohorts did not match with the pathophysiological basis of the mechanisms of action of the growth factor used, G-CSF.
        • Full-Text
        • PDF
      • Reply to: “G-CSF in acute-on-chronic liver failure – Art of ‘patient selection’ is paramount!”
        Journal of HepatologyVol. 76Issue 2
        • Preview
          We read with great interest the Letter to the Editor by Jindal and Sarin hypothesizing that inadequate patient selection was the reason for the lack of efficacy of granulocyte-colony stimulating factor (G-CSF) in patients with acute-on-chronic liver failure (ACLF) reported in the European multicenter GRAFT study.1,2 The authors are experts in this field and have published several studies on G-CSF in decompensated cirrhosis, ACLF, and alcoholic hepatitis.3,4 Their main arguments are that the inclusion of patients with advanced extrahepatic organ failure, especially related to sepsis, together with depleted bone marrow capacity may have contributed to the negative outcome of the GRAFT trial.
        • Full-Text
        • PDF

      Introduction

      Acute-on-chronic liver failure (ACLF) develops in more than 30% of patients with acutely decompensated cirrhosis and is characterized by development of organ failures.
      • Arroyo V.
      • Moreau R.
      • Jalan R.
      Acute-on-Chronic liver failure.
      Cirrhosis accounts for about 1 million deaths worldwide and most of these patients die during an ACLF episode.
      • Pimpin L.
      • Cortez-Pinto H.
      • Negro F.
      • Corbould E.
      • Lazarus J.V.
      • Webber L.
      • et al.
      Burden of liver disease in Europe: epidemiology and analysis of risk factors to identify prevention policies.
      Its overall 90-day mortality is approximately 50%, which is substantially higher than the 14% seen in patients with acute decompensation without organ failures.
      • Moreau R.
      • Jalan R.
      • Gines P.
      • Pavesi M.
      • Angeli P.
      • Cordoba J.
      • et al.
      Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis.
      Liver transplantation is the only established treatment approach,
      • Artru F.
      • Louvet A.
      • Ruiz I.
      • Levesque E.
      • Labreuche J.
      • Ursic-Bedoya J.
      • et al.
      Liver transplantation in the most severely ill cirrhotic patients: a multicenter study in acute-on-chronic liver failure grade 3.
      ,
      • Engelmann C.
      • Thomsen K.L.
      • Zakeri N.
      • Sheikh M.
      • Agarwal B.
      • Jalan R.
      • et al.
      Validation of CLIF-C ACLF score to define a threshold for futility of intensive care support for patients with acute-on-chronic liver failure.
      but due to scarcity of donor organs, costs and contraindications it is only accessible to a minority of patients (approximately 5%).
      • Artru F.
      • Louvet A.
      • Ruiz I.
      • Levesque E.
      • Labreuche J.
      • Ursic-Bedoya J.
      • et al.
      Liver transplantation in the most severely ill cirrhotic patients: a multicenter study in acute-on-chronic liver failure grade 3.
      Novel treatment approaches aim to target ACLF-specific pathomechanisms, dominated by the co-existence of exaggerated inflammatory responses on the one hand, and immune paralysis on the other, leading to mitochondrial dysfunction, cell death and regenerative incapacity.
      • Albillos A.
      • Lario M.
      • Alvarez-Mon M.
      Cirrhosis-associated immune dysfunction: distinctive features and clinical relevance.
      • Moreau R.
      • Claria J.
      • Aguilar F.
      • Fenaille F.
      • Lozano J.J.
      • Junot C.
      • et al.
      Blood metabolomics uncovers inflammation-associated mitochondrial dysfunction as a potential mechanism underlying ACLF.
      • Engelmann C.
      • Becker C.
      • Boldt A.
      • Herta T.
      • Boehlig A.
      • Splith K.
      • et al.
      Ascites' neutrophil function is significantly impaired in patients with decompensated cirrhosis but can be restored by autologous plasma incubation.
      • Engelmann C.
      • Claria J.
      • Szabo G.
      • Bosch J.
      • Bernardi M.
      Pathophysiology of decompensated cirrhosis: portal hypertension, circulatory dysfunction, inflammation, metabolism and mitochondrial dysfunction.
      Bone marrow-derived stem cells have been shown to be potent in modulating immune functions and promoting regenerative capacities, thus mitigating various types of liver injury in experimental and clinical studies.
      • Li N.
      • Zhang L.
      • Li H.
      • Fang B.
      Human CD34+ cells mobilized by granulocyte colony-stimulating factor ameliorate radiation-induced liver damage in mice.
      • am Esch J.S.
      • Schmelzle M.
      • Furst G.
      • Robson S.C.
      • Krieg A.
      • Duhme C.
      • et al.
      Infusion of CD133+ bone marrow-derived stem cells after selective portal vein embolization enhances functional hepatic reserves after extended right hepatectomy: a retrospective single-center study.
      • Engelmann C.
      • Martino V.D.
      • Kerbert A.J.C.
      • Weil-Verhoeven D.
      • Aehling N.F.
      • Herber A.
      • et al.
      The current status of granulocyte-colony stimulating factor to treat acute-on-chronic liver failure.
      However, costs, high technical requirements and peri-interventional complications limit the general applicability of this approach.
      Granulocyte-colony stimulating factor (G-CSF) mobilizes hematopoietic stem and immune cells and represents an alternative for exogenous stem cell infusions.
      • Demetri G.D.
      • Griffin J.D.
      Granulocyte colony-stimulating factor and its receptor.
      ,
      • Engelmann C.
      • Splith K.
      • Berg T.
      • Schmelzle M.
      Effects of granulocyte-colony stimulating factor (G-CSF) on stem cell mobilization in patients with liver failure.
      Several randomized controlled trials in India and China using G-CSF in patients with ACLF reported improved survival and decreased rates of disease-related complications, such as bacterial infections.
      • Garg V.
      • Garg H.
      • Khan A.
      • Trehanpati N.
      • Kumar A.
      • Sharma B.C.
      • et al.
      Granulocyte colony-stimulating factor mobilizes CD34(+) cells and improves survival of patients with acute-on-chronic liver failure.
      ,
      • Duan X.Z.
      • Liu F.F.
      • Tong J.J.
      • Yang H.Z.
      • Chen J.
      • Liu X.Y.
      • et al.
      Granulocyte-colony stimulating factor therapy improves survival in patients with hepatitis B virus-associated acute-on-chronic liver failure.
      Consequently, G-CSF is being used as standard of care in many parts of the world. However, its role in treating ACLF is still under debate as data supporting its efficacy have been tested in relatively small clinical trials outside of Europe and the US, and in a limited number of centers. Therefore, the aim of our study was to explore the efficacy of G-CSF in patients with ACLF in a large multicenter, randomized and controlled trial.

      Patients and methods

      Trial design

      The GRAFT study was an investigator-initiated, multicenter, randomized, controlled trial, which recruited patients with ACLF in 18 German tertiary centers funded by the German research foundation (DFG). The trial protocol (supplementary trial protocol) was approved by the leading ethics committee (University Leipzig) and by the German competent authority (BfArM). The trial design followed the CONSORT guidelines and complied with the principals of the declaration of Helsinki from 1975 (Version Sommerset West 1996),
      • Shephard D.A.
      The 1975 declaration of Helsinki and consent.
      as well as all pertinent national laws and the ICH guidelines for Good Clinical Practice (GCP) issued in June 1996 and CPMP/ICH/135/95 from September 1997.

      Study population

      Patients with ACLF according to the criteria of the European Foundation for the Study of Chronic Liver Failure (EASL-CLIF) consortium
      • Jalan R.
      • Saliba F.
      • Pavesi M.
      • Amoros A.
      • Moreau R.
      • Gines P.
      • et al.
      Development and validation of a prognostic score to predict mortality in patients with acute-on-chronic liver failure.
      with underlying liver cirrhosis and aged ≥18 years were eligible for the trial. ACLF was defined as acutely decompensated cirrhosis (acute development or worsening of one of the following complications: ascites, hepatic encephalopathy, gastrointestinal hemorrhage, bacterial infection) combined with one or more organ failure(s) (definitions in supplementary information).
      • Moreau R.
      • Jalan R.
      • Gines P.
      • Pavesi M.
      • Angeli P.
      • Cordoba J.
      • et al.
      Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis.
      Patients with prior non-curatively treated or active malignancies, sickle cell disease, septic shock (defined by the presence of bacteriaemia, systemic inflammatory response syndrome, and vasopressor support) or white blood cell (WBC) count of >50 x 109/L were excluded (complete list of inclusion and exclusion criteria is included in the supplementary information). All patients gave their written consent before study inclusion.

      Intervention

      Patients were randomized in a 1:1 ratio to receive either standard medical therapy (SMT) (SMT group) or recombinant G-CSF (G-CSF, Ratiograstim®) plus SMT (G-CSF+SMT group). Patients were centrally randomized with a Web-based system by using a minimization algorithm with a factor for center and a probability of 0.2 for random allocation, to avoid potential loss of allocation concealment within centers. Patients randomized to the G-CSF+SMT group were scheduled to receive in total 12 doses of G-CSF subcutaneously injected on a daily basis for the first 5 days after randomization and every third day thereafter until day 26. The dosing scheme used corresponded to that published by Garg et al.
      • Garg V.
      • Garg H.
      • Khan A.
      • Trehanpati N.
      • Kumar A.
      • Sharma B.C.
      • et al.
      Granulocyte colony-stimulating factor mobilizes CD34(+) cells and improves survival of patients with acute-on-chronic liver failure.
      which was chosen based on its superior efficacy over a shorter treatment duration applied by Duan et al.
      • Duan X.Z.
      • Liu F.F.
      • Tong J.J.
      • Yang H.Z.
      • Chen J.
      • Liu X.Y.
      • et al.
      Granulocyte-colony stimulating factor therapy improves survival in patients with hepatitis B virus-associated acute-on-chronic liver failure.
      Further randomized efficacy trials were not available at the time of trial preparation. Injections were performed at 12 am (± 1 hour) on every treatment day. Using pre-filled syringes provided by the supplier (30 Mio IU or 48 Mio IU), as part of the clinical routine, G-CSF doses were guided by body weight using a cut-off value of 70 kg (≤70kg 30 Mio IU G-CSF, >70 kg 48 Mio IU G-CSF) (Fig. 1) (Supplementary CTAT Table).
      Figure thumbnail gr1
      Fig. 1Study design.
      After enrolment patients with G-CSF were randomized to receive either SMT or SMT+G-CSF for 26 days. The primary endpoint was assessed after 90 days. Patients were followed up for in total 360 days. There were 9 study visits (baseline, V1-V8) at which clinical data, adverse events and blood samples were collected. ACLF, acute-on-chronic liver failure; D, day of study visit; G-CSF, granulocyte-colony stimulating factor; SMT, standard medical therapy; V, study visit.
      A placebo controlled and blinded design was not considered feasible in the study population for 2 main reasons. First, patients with ACLF can have pre-existing systemic inflammatory responses and a high risk of infectious complications
      • Arroyo V.
      • Moreau R.
      • Jalan R.
      Acute-on-Chronic liver failure.
      ; thus, blinding could lead to misinterpretation of WBC count, as it will be unknown if an infection or G-CSF treatment causes the response in WBC count. Alternative diagnostic approaches are not reliable to detect infectious complications in cases of G-CSF-related WBC increases.
      • Pieri G.
      • Agarwal B.
      • Burroughs A.K.
      C-reactive protein and bacterial infection in cirrhosis.
      Second, according to the drug manufacturer’s instruction, G-CSF treatment was transiently stopped in cases where WBC rose to >70 x109/L and re-started at a WBC <50 x109/L, mainly because WBC exceeding 100 x109/L can be accompanied by a high risk of symptomatic hyperviscosity syndrome.
      • Malkan U.Y.
      • Ozcebe O.I.
      Leukapheresis do not improve early death rates in acute myeloid leukemia patients with hyperleukocytosis.
      Any subsequent new ACLF episode, which developed after the qualifying ACLF episode for inclusion into the trial, was treated with SMT in both treatment arms.

      Follow-up

      After randomization patients were followed-up for 360 days or until premature study termination defined as:
      • withdrawal of informed consent (drop-out)
      • complete loss of contact (drop-out)
      • death
      During follow up visits V1-V8 clinical data and samples were collected as outlined in Fig. 1.

      Endpoints

      The primary endpoint was defined as transplant-free survival 90 days after inclusion into the study, with death and orthotopic liver transplantation (OLT) counting as events. Patients were censored at the time point of last information still alive and without OLT. The key secondary endpoints were the overall and transplant-free survival after 360 days, the development of ACLF-related complications, bacterial infections and the course of liver function scores, such as the model of end-stage liver disease (MELD) score and Chronic Liver Failure-consortium (CLIF-C) organ failure (OF) score, during the entire observation period. A complete list of endpoints and definitions is presented in the supplementary information. We abstained from using the CLIF-C ACLF score
      • Jalan R.
      • Saliba F.
      • Pavesi M.
      • Amoros A.
      • Moreau R.
      • Gines P.
      • et al.
      Development and validation of a prognostic score to predict mortality in patients with acute-on-chronic liver failure.
      for further analysis, because it includes WBC count as part of the scoring system, making it susceptible to artificial WBC alteration by G-CSF-related cell mobilization. Safety was evaluated at each visit based on clinical data and physical examination. Severe adverse events (SAEs) were reported within 24 hours after awareness and monitored until resolution. The SAE reporting period started on day 0 and ended at day 28 except for newly developed malignancies. Annual safety analyses were performed. Safety reports and data analyses were reviewed by the data monitoring and safety board (DMSB) as requested or at least annually.

      Statistical analysis and sample size calculation

      Based on the CANONIC Study,
      • Moreau R.
      • Jalan R.
      • Gines P.
      • Pavesi M.
      • Angeli P.
      • Cordoba J.
      • et al.
      Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis.
      we expected a transplant-free survival rate of 42% at day 90 with SMT. Two small single-center, randomized G-CSF trials in patients with ACLF reported significant effects at improving survival from 26% to 66% after 60 days
      • Garg V.
      • Garg H.
      • Khan A.
      • Trehanpati N.
      • Kumar A.
      • Sharma B.C.
      • et al.
      Granulocyte colony-stimulating factor mobilizes CD34(+) cells and improves survival of patients with acute-on-chronic liver failure.
      and from 21% to 48% after 90 days.
      • Duan X.Z.
      • Liu F.F.
      • Tong J.J.
      • Yang H.Z.
      • Chen J.
      • Liu X.Y.
      • et al.
      Granulocyte-colony stimulating factor therapy improves survival in patients with hepatitis B virus-associated acute-on-chronic liver failure.
      We aimed to detect an absolute difference of 20% (from 42% to 62%, HR 1.815) in transplant-survival at day 90. Assuming a 10% loss to follow-up, we originally determined that 292 patients would be required to have 90% power, at a 2-sided significance level of 5%. One interim analysis was scheduled once the primary endpoint was available for 50% of patients, allowing for detection of marked superiority of G-CSF (of ≥33%) on the primary endpoint, with a power of at least 80%. Multiplicity adjustment followed O'Brien,
      • O'Brien P.C.
      • Fleming T.R.
      A multiple testing procedure for clinical trials.
      with nominal significance levels of p = 0.005 at interim analysis and p = 0.048 at final analysis. All patients enrolled up to this point were followed up until regular end of observation or an endpoint. Cox regression was used to estimate the treatment effect on hazard ratio (HR) scale with 2-sided 95% CIs provided in time-to-event endpoints. Two-sided chi-square or Fisher’s exact tests were used to compare qualitative variables, as appropriate. Courses of prognostic scores (MELD; CLIF-C OF score) were analyzed by repeated-measures ANCOVA, imputing single in-between values by linear interpolation and adding the last value observed in cases of death, OLT or dropout. We considered OLT competing with death and death/OLT competing with occurrence of disease-related complications in competing risk analyses (Fine and Gray proportional subdistribution hazard regression model, R version 3.6.3). Additional analyses were performed in the per protocol (PP) population. These patients fulfilled all inclusion and exclusion criteria and received at least 75% of the intended G-CSF total dose, if medical reasons did not result in premature end of treatment. Ninety-day transplant-free survival analysis was also calculated post hoc in various subpopulations. Detailed definitions are provided in the supplementary information. Analyses were performed with SSPS software, version 25 (IBM inc.) and R (Supplementary CTAT Table). Safety issues were carefully monitored by standardized SAE reporting and annual safety reports (according to legal requirements), and presented to the DMSB. All results shown here were calculated from the intention-to-treat (ITT) population unless otherwise specified.

      Results

      Patient and treatment characteristics

      After registering 180 patients in total, 176 were randomized between March 2016 and April 2019 to receive either G-CSF with SMT or SMT alone (n = 88 G-CSF+SMT group, n = 88 SMT group) (Fig. 2). Table 1 summarizes the baseline characteristics for both groups, which did not differ significantly, indicating good randomization. All patients had underlying cirrhosis. The mean ± SD age was 54.4 ± 10.2 years in the G-CSF+SMT group and 57.1 ± 9.6 years in the SMT group. The majority of patients enrolled were male (56.8% G-CSF+SMT group, 69.3% SMT group) and had an increased mean BMI >28 kg/m2. Most patients had ACLF grade 1 or 2, with a mean number of organ failures of 1.7 ± 0.7 in the G-CSF+SMT group vs. 1.4 ± 0.6 in the SMT group. The mean baseline CLIF-C OF score and MELD score were 10.4 ± 1.9 and 24.4 ± 6.3 in the G-CSF+SMT group and 10.3 ± 2 and 24.5 ± 6.1 in the SMT group, respectively. Liver and kidney failures were the predominant types of organ failure accounting for more than 60% of cases in both groups. Alcohol abuse, bacterial infections and gastrointestinal bleeding were the most frequent precipitating events (Table 1). At baseline, 96.6% of patients in both groups had ascites, and bacterial infections were present in 56.8% in the G-CSF+SMT group and 51.1% in the SMT group.
      Figure thumbnail gr2
      Fig. 2CONSORT flow chart of patients.
      After registering 180 patients, 176 patients fulfilled the inclusion and exclusion criteria, were randomized (n = 88 patients in each arm) and analyzed according to the ITT analysis. Patients were followed up for 360 days. ∗Other reasons for dropouts were: treatment of bronchial carcinoma n = 1 and medical decision n = 1; ACLF, acute-on-chronic liver failure; EF-CLIF, European Foundation for the Study of Chronic Liver Failure; ITT, intention-to-treat (all patients within the full analysis set were treated as randomized); OLT, orthotopic liver transplantation; SMT, standard medical care.
      Table 1Baseline characteristics for the 2 study arms.
      ParameterG-CSF+SMT group (n = 88)SMT group (n = 88)
      General
      Age (years)54.4 ± 10.257.1 ± 9.6
      Gender (w/m)38 (43.2%)/50 (56.8%)27 (30.7%)/61 (69.3%)
      BMI (kg/m2)28.9 ± 4.828.8 ± 5.6
      Mean arterial pressure (mmHg)79.9 ± 1282.7 ± 13.2
      Disease severity
      ACLF grade
      These values represent the number of patients in every respective ACLF grade. Values were calculated based on individual parameters. However, patients were included based on ACLF grade allocations provided by the study centers. Based on back calculations, 2 patients in the SMT group and 1 patient in the G-CSF+SMT group did not fulfil the criteria for ACLF. Percentages shown here were calculated against the overall number of patients included in the trial.
       Grade 139 (44.3%)45 (51.1%)
       Grade 237 (42.0%)28 (31.8%)
       Grade 312 (13.6%)15 (17%)
      Number of organ failures1.7 ± 0.71.4 ± 0.6
      CLIF-C OF score10.4 ± 1.910.3 ± 2
      MELD score24.4 ± 6.324.5 ± 6.1
      CLIF-C ACLF score51.9 ± 8.751.2 ± 7.4
      Organ failures
      Liver failure58 (65.9%)56 (63.6%)
      Kidney failure58 (65.9%)57 (64.8%)
      Brain failure5 (5.7%)5 (5.7%)
      Circulatory failure8 (9.1%)6 (6.8%)
      Coagulation failure22 (25%)15 (17%)
      Respiratory failure2 (2.3%)6 (6.8%)
      Precipitating events
      Precipitating events counted individually in patients with more than one precipitating event.
      Alcohol related49 (55.7%)46 (52.3%)
      Bacterial infection36 (40.9%)33 (37.5%)
      Bleeding13 (14.8%)7 (8%)
      Others7 (8%)8 (9.1%)
      Unknown16 (18.2%)18 (20.5%)
      Complications of cirrhosis
      Ascites85 (96.6%)85 (96.6%)
      Hepatorenal syndrome57 (64.8%)58 (65.9%)
      Hepatic encephalopathy55 (62.5%)61 (69.3%)
      Bacterial infection at baseline50 (56.8%)45 (51.1%)
      Laboratory values
      Bilirubin (mg/dl)18.0 ± 12.018.9 ± 14.5
      ALT (U/L)76.0 ± 144.781.7 ± 167.0
      AST (U/L)139.0 ± 217.0124.1 ± 136.5
      GGT (U/L)195.2 ± 287.5165.0 ± 224.0
      ALP (U/L)150.7 ± 80.6148.8 ± 116.4
      Creatinine (mg/dl)2.4 ± 1.62.4 ± 1.5
      Urea (mg/dl)101.5 ± 60.9115.3 ± 61.3
      INR2.2 ± 0.82.1 ± 1.0
      Albumin (g/dl)3.0 ± 0.73.0 ± 0.7
      C-reactive protein (mg/dl)39.9 ± 29.241.3 ± 39.6
      Procalcitonin (µg/L)1.7 ± 2.11.4 ± 1.4
      WBC count (109/L)14.8 ± 12.411.2 ± 7.1
      ACLF, acute-on-chronic liver failure; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CLIF-C, Chronic Liver Failure-consortium; G-CSF, granulocyte-colony stimulating factor; GGT, gamma glutamyltransferase; INR, international normalized ratio; MELD, model for end-stage liver disease; OF, organ failure; SMT, standard medical therapy.
      # Precipitating events counted individually in patients with more than one precipitating event.
      These values represent the number of patients in every respective ACLF grade. Values were calculated based on individual parameters. However, patients were included based on ACLF grade allocations provided by the study centers. Based on back calculations, 2 patients in the SMT group and 1 patient in the G-CSF+SMT group did not fulfil the criteria for ACLF. Percentages shown here were calculated against the overall number of patients included in the trial.
      In the G-CSF+SMT group, 83 patients (94.3%) received at least 1 G-CSF injection whereas 5 patients remained untreated due to early death (n = 4) and other medical reasons (n = 1). The mean number of G-CSF injections was 8.6 ± 3.9 and the mean duration of G-CSF treatment was 17.5 days ± 9.7. In total, 29 patients (33%) received all 12 injections and 51 patients ≥9 injections. Therefore, 80 patients in the G-CSF+SMT arm and 84 patients in the SMT arm fulfilled criteria for the PP cohort. Patients treated with G-CSF showed increasing WBC count during the treatment course, whereas it remained unchanged in the SMT group (p <0.001) (Fig. S1).

      Patient outcomes

      The primary endpoint, defined by death or OLT within 90 days after study enrolment, was reached in 54 patients (61.4%) in the G-CSF+SMT group vs. 51 patients (58.0%) in the SMT group from the ITT cohort. Five patients from the G-CSF+SMT group and 9 from the SMT group received an OLT. The HR for the risk of death or transplantation when treated with G-CSF was 1.05 (95% CI 0.711–1.551; p = 0.805) (Fig. 3, Fig. 4). The median transplant-free survival was 35 days (95% CI 19.8–50.2) in the G-CSF+SMT group and 34 days (95% CI 9.9–58.1) in the SMT group. Neither the overall survival (HR 1.058; 95% CI 0.727–1.548; p = 0.768), nor 360-day transplant-free survival (HR 0.998; 95% CI 0.697–1.430; p = 0.992) were statistically different between arms in the ITT cohort. Similar results were observed in the PP analysis. The HR for the primary endpoint was calculated as 1.034 (95% CI 0.698–1.531; p = 0.869] (Fig. 4). For overall survival, the HR was 1.059 (95% CI 0.724–1.548; p = 0.769) and for transplant-free survival 0.998 (95% CI 0.694–1.436; p = 0.992) when treated with G-CSF (Fig. 4).
      Figure thumbnail gr3
      Fig. 390-day transplant-free survival in ITT cohort.
      Out of 176 patients with ACLF, 88 were randomized to receive G-CSF+SMT (G-CSF+SMT group) or SMT only (SMT group). Death or liver transplantation were considered as events in a Cox Regression analysis calculating HR for G-CSF therapy. ACLF, acute-on-chronic liver failure; G-CSF, granulocyte-colony stimulating factor; HR, hazard ratio; ITT, intention-to-treat; SMT, standard medical therapy.
      Figure thumbnail gr4
      Fig. 4Forrest plot for risk of all sub-populations analyzed.
      (A) Cox regression analyses calculating the hazard ratio (HR) for G-CSF therapy in subgroups with different ACLF severity grades (ACLF grade 1 n = 83, ACLF grade 2 n = 64, ACLF grade 3 n = 27) and organ failures (liver failure n = 114, kidney failure n = 115) were performed to evaluate the impact of disease severity on treatment efficacy. G-CSF was unable to improve the 90-day transplant-free survival in any of the severity subgroups. (B) Cox regression analyses adjusted for ACLF grades were performed to calculate the hazard ratio of death or transplant until 90 days of observation in additional subgroups. Along with the ITT cohort and per protocol sets of patients, analyses were performed in a variety of subgroups of patients with predefined clinical conditions of potentially different prognosis. Best case (n = 65) – patients who survived regular end of treatment without transplant for at least 14 days and received at least 75% of intended dose if applicable; ACLF-APASL criteria (n = 114) – patients who fulfilled the APASL criteria for ACLF; alcohol-related ACLF (n = 95) – patients who had an alcohol-related precipitating event; 7-day survivors (n = 155) – patients who survived randomization for a minimum of 7 days; no infection (n = 73) – patients without detectable infection at inclusion; infection (n = 69) – patients with a bacterial infection as a precipitating event. ACLF, acute-on-chronic liver failure; APASL, Asian Pacific Association for the Study of the Liver; G-CSF, granulocyte-colony stimulating factor; ITT, intention-to-treat.
      During the study period most fatalities were caused by ACLF or other liver-related complications in both arms (G-CSF+SMT: 27/88 [30.7%]; SMT: 21/88 [22.7%]). Gastrointestinal bleeding leading to death was more frequent in the SMT arm (G-CSF+SMT: 2/88 [2.3%]; SMT 8/88 [9.1%]). In total 8 deaths, 3 in the G-CSF+SMT arm and 5 in the SMT arm, were related to bacterial infections (Table S2)
      As liver transplant and dropouts were competing with death, we performed a competing risk analysis by calculating the cumulative incidence of death, which confirmed that G-CSF treatment failed to show a significant benefit on survival in patients with ACLF (HR 1.11; 95% CI 1.61–0.77) (Fig. S2). For futility analysis, we calculated a conditional power of 1-β=0.0266 for transplant-free survival after 90-days and 0.0046 for overall survival, which led to the premature termination of patient recruitment after consultation with the study’s DMSB.

      Secondary endpoints

      Fatalities occured predominently in patients with increasing MELD score during the treatment period until day 28 in both groups (Fig. S5). The distribution of patients with increasing or decreasing MELD or CLIF-C OF score, did not differ between the 2 groups (ANCOVA analysis: significance between groups p = 0.884 for MELD and p = 0.757 for CLIF-C OF score; Fig. 5 and S5).
      Figure thumbnail gr5
      Fig. 5Development of CLIF-C OF score during the treatment course.
      The distribution of patients with increasing and decreasing (A) CLIF-C OF score and (B) MELD score (delta between baseline and day 28) was not different when comparing the G-CSF (G-CSF+SMT; blue bars) arm and the control arm (SMT; grey bars) (between groups effect: DeltaOF-score p = 0.757, DeltaMELD score p = 0.884). Difference between groups over time was calculated by ANCOVA. Missing values due to death or transplant were imputed by individuals’ last value. ACLF, acute-on-chronic liver failure; CLIF-C, Chronic Liver Failure-consortium; G-CSF, granulocyte-colony stimulating factor; MELD, model for end-stage liver disease; OF, organ failure; SMT, standard medical therapy.
      Patients who recovered from the first ACLF episode were followed up for development of subsequent ACLF episodes (ACLF recurrence). The risk of developing further ACLF episodes was not different between the 2 groups (HR 1.94; 95% CI 0.83–4.53; p = 0.13) (Fig. S4) although the number of patients with ACLF recurrence randomized into the G-CSF+SMT group was slightly higher (65.2% vs. 34.8 %, p = 0.19) (Table S3). Groups of patients with or without ACLF recurrence were further characterized at baseline and at the time point of remission from the first ACLF episode. Patients with ACLF recurrence more often had ACLF grade 1 at baseline (87% vs. 55%, p = 0.019). At the time point of remission the same patients had higher values for MELD score (17.8 vs. 13.1, p = 0.002), bilirubin (13.5 vs. 7, p = 0.024) and C-reactive protein (38.9 vs. 27.2; p = 0.004) and more often had high-grade ascites (60.9% vs. 22.5%, p = 0.002) (Table S3).
      Bacterial infections occurred in 71 (80.8%) patients in the G-CSF+SMT group and 69 (78.4%) in the SMT group (p = 0.709). Of those, 50 out of 88 (56.6%) in the G-CSF+SMT arm and 45 out of 88 (51.1%) had an infection at baseline and 31 out of 88 (35.2%) and 32 out of 88 (36.4%) in the respective treatment arms developed an infection during the observational period. Urinary tract infections, spontaneous bacterial peritonitis, pneumonia and blood stream infections were the most common sites of infection during the study. There was a predominance of spontaneous bacterial peritonitis in the G-CSF+SMT arm and blood stream infections in the SMT arm. There were no major differences in the rate of infections across the different ACLF grades in comparison between both arms (Table S4). There were 6 fungal infections in each cohort with Candida being the most frequent causative pathogen (6 out of 12). Two patients (1 in each group) had a fungal peritonitis (Table S5).
      There was no difference between both groups with respect to the development of other complications, such as gastrointestinal bleeding, hepatic encephalopathy, and hepatorenal syndrome during the observation period (Table 2).
      Table 2Development of cirrhosis-related complications during the observational period.
      Groups were balanced regarding history of complications.
      ComplicationTreatment group (G-CSF+SMT)Control group (SMT)p value
      ACLF remission observed
      Remission before an endpoint-related/competing event or end of individual observational period.
      33/87
      Only patients with ACLF grade of ≥1 at baseline were included.
      (37.9%)
      30/86
      Only patients with ACLF grade of ≥1 at baseline were included.
      (34.9%)
      0.68
      Competing risk analysis by cumulative incidence function and hazard ratio.
      ACLF re-episode
      ACLF re-episodes were considered only in case of a previous ACLF remission.
      15/87 (17.2%)8/86
      Only patients with ACLF grade of ≥1 at baseline were included.
      (9.3%)
      0.13
      Group comparisons were performed by Chi-Square test.
      Ascites87 (98.9%)87 (98.9%)1.000
      Group comparisons were performed by Chi-Square test.
      Hepatic encephalopathy64 (72.7%)67 (76.1%)0.604
      Group comparisons were performed by Chi-Square test.
      Hepatorenal syndrome71 (80.7%)69 (78.4%)0.709
      Group comparisons were performed by Chi-Square test.
      Bacterial infections71 (80.8%)69 (78.4%)0.709
      Group comparisons were performed by Chi-Square test.
      Variceal bleeding7 (8%)7 (8%)1.000
      Group comparisons were performed by Chi-Square test.
      ACLF, acute-on-chronic liver failure; G-CSF, recombinant granulocyte colony stimulating factor; SMT, standard medical therapy.
      Groups were balanced regarding history of complications.
      1 Remission before an endpoint-related/competing event or end of individual observational period.
      2 Only patients with ACLF grade of ≥1 at baseline were included.
      3 ACLF re-episodes were considered only in case of a previous ACLF remission.
      4 Group comparisons were performed by Chi-Square test.
      5 Competing risk analysis by cumulative incidence function and hazard ratio.

      Post hoc sensitivity analyses

      Sensitivity analyses focused on subgroups of patients with different disease severities (Fig. 4A) (ACLF grades and types of organ failures), on subgroups of patients who were included in previous G-CSF trials (Fig. 4B)
      • Garg V.
      • Garg H.
      • Khan A.
      • Trehanpati N.
      • Kumar A.
      • Sharma B.C.
      • et al.
      Granulocyte colony-stimulating factor mobilizes CD34(+) cells and improves survival of patients with acute-on-chronic liver failure.
      ,
      • Singh V.
      • Sharma A.K.
      • Narasimhan R.L.
      • Bhalla A.
      • Sharma N.
      • Sharma R.
      Granulocyte colony-stimulating factor in severe alcoholic hepatitis: a randomized pilot study.
      (ACLF defined according to the APASL criteria, ACLF due to an alcohol-related precipitating event) and on subgroups with a higher likelihood of treatment response (such as patients without infections, 7-day survivors and patients fulfilling criteria for best cases). In none of these cohorts did G-CSF show a significant effect on transplant-free survival after 90 days (Table S1).
      Multivariate Cox regression analysis identified risk factors for mortality such as age (HR 1.043; 95% CI 1.019–1.066; p <0.001), the ACLF severity grade at baseline (HR 2.249; 95% CI 1.703–2.970; p <0.001) and the WBC count at baseline (HR 1.044; 95% CI 1.02–1.066; p <0.001) whereas treatment with G-CSF, the presence of an alcohol-related precipitating event or bacterial infection at baseline (precipitating events) did not impact on transplant-free survival (Table 3).
      Table 3Multivariate Cox regression analysis for 90-day transplant-free survival.
      Baseline factorMultivariate analysis
      Hazard ratio95% CIp value
      Treatment arm (risk for G-CSF+SMT)0.9460.635–1.4080.785
      Age (years)1.0431.019–1.066<0.001
      ACLF grade2.2491.703–2.970<0.001
      Precipitating event bacterial infection1.0230.938–1.1170.604
      Precipitating event alcohol related1.0410.943–1.1490.426
      WBC count (109/L)1.0441.022–1.066<0.001
      ACLF, acute-on-chronic liver failure; G-CSF, recombinant granulocyte colony stimulating factor; SMT, standard medical therapy; WBC, white blood cell.

      Safety characteristics

      Overall, 403 adverse events were reported from 80 patients in the G-CSF+SMT group and 354 from 78 patients in the SMT group. Sixty-one SAEs were reported in 54 patients from the G-CSF+SMT group, as well as 57 SAEs in 47 patients from the SMT group, including 37 and 36 deaths during the SAE reporting period, respectively. ACLF was the most frequent cause of death in both groups (n = 17/37 [46%] and 16/36 [44%], respectively, p = 1.0), followed by hemorrhages (n = 6 [16%] and 8 [22%], respectively, p = 0.564), and infection-related deaths (n = 4 [11%] and n = 5 [14%], respectively, p = 0.736). In the G-CSF+SMT group, G-CSF-related adverse reactions were reported in 18 patients and 7 serious adverse reactions were reported in 4 patients (Table S6). The study treatment was paused in 3 out of 4 patients with serious adverse reactions, one of which was related to a WBC count above the threshold of 70 x109/L. In total, G-CSF therapy had to be transiently stopped due to a WBC increase in 8 patients (9.1%).

      Discussion

      This is the first large multicenter study evaluating G-CSF as a potential novel therapy for patients with bona fide ACLF. Despite G-CSF having immune-modulatory and pro-regenerative capacities,
      • Sarin S.K.
      • Kedarisetty C.K.
      • Abbas Z.
      • Amarapurkar D.
      • Bihari C.
      • Chan A.C.
      • et al.
      Acute-on-chronic liver failure: consensus recommendations of the asian pacific association for the study of the liver (APASL) 2014.
      our study confirmed the overall dismal prognosis of ACLF and failed to demonstrate superiority of G-CSF over SMT. Hence, the data provided here are in strong disagreement with 2 single-center trials that showed a marked improvement in survival and reduced infectious complications after G-CSF therapy
      • Demetri G.D.
      • Griffin J.D.
      Granulocyte colony-stimulating factor and its receptor.
      ,
      • Engelmann C.
      • Splith K.
      • Berg T.
      • Schmelzle M.
      Effects of granulocyte-colony stimulating factor (G-CSF) on stem cell mobilization in patients with liver failure.
      ; thus, our data do not support the use of G-CSF as a potential standard therapy for patients with ACLF.
      • Sarin S.K.
      • Kedarisetty C.K.
      • Abbas Z.
      • Amarapurkar D.
      • Bihari C.
      • Chan A.C.
      • et al.
      Acute-on-chronic liver failure: consensus recommendations of the asian pacific association for the study of the liver (APASL) 2014.
      We showed that 90-day transplant-free survival (HR 1.05, p = 0.805) and overall survival (HR 1.058, p = 0.768) were unaffected by G-CSF administration, and this observation was also confirmed in the subgroup of patients treated strictly according to the protocol (PP) with an HR of 1.034 (p = 0.869) for 90-day transplant-free survival and of 1.059 (p = 0.769) for overall survival. In smaller single-center trials, G-CSF reduced the number of infectious complications, improved immune dysfunction by increasing the number of dendritic cells whilst decreasing interferon-producing T-cells and enhanced the homing of CD34+ hematopoietic cells, which are believed to promote regenerative capacities.
      • Engelmann C.
      • Splith K.
      • Berg T.
      • Schmelzle M.
      Effects of granulocyte-colony stimulating factor (G-CSF) on stem cell mobilization in patients with liver failure.
      ,
      • Garg V.
      • Garg H.
      • Khan A.
      • Trehanpati N.
      • Kumar A.
      • Sharma B.C.
      • et al.
      Granulocyte colony-stimulating factor mobilizes CD34(+) cells and improves survival of patients with acute-on-chronic liver failure.
      ,
      • Khanam A.
      • Trehanpati N.
      • Garg V.
      • Kumar C.
      • Garg H.
      • Sharma B.C.
      • et al.
      Altered frequencies of dendritic cells and IFN-gamma-secreting T cells with granulocyte colony-stimulating factor (G-CSF) therapy in acute-on- chronic liver failure.
      However, in our trial, G-CSF was neither able to reduce the number of infections or any other complication nor improve severity scores, such as the MELD score and CLIF-C score significantly. This is in line with recently published data in children with ACLF,
      • Sharma S.
      • Lal S.B.
      • Sachdeva M.
      • Bhatia A.
      • Varma N.
      Role of granulocyte colony stimulating factor on the short-term outcome of children with acute on chronic liver failure.
      and supports other studies that did not find any beneficial effect of G-CSF.
      • Khanam A.
      • Trehanpati N.
      • Garg V.
      • Kumar C.
      • Garg H.
      • Sharma B.C.
      • et al.
      Altered frequencies of dendritic cells and IFN-gamma-secreting T cells with granulocyte colony-stimulating factor (G-CSF) therapy in acute-on- chronic liver failure.
      ,
      • Sharma S.
      • Lal S.B.
      • Sachdeva M.
      • Bhatia A.
      • Varma N.
      Role of granulocyte colony stimulating factor on the short-term outcome of children with acute on chronic liver failure.
      Notably, some pre-clinical studies have even shown a worsening of liver injury upon G-CSF use in ACLF.
      • Marot A.
      • Singal A.K.
      • Moreno C.
      • Deltenre P.
      Granulocyte colony-stimulating factor for alcoholic hepatitis: a systematic review and meta-analysis of randomised controlled trials.
      There are several reasons that may explain the discrepancies in G-CSF treatment outcome observed in our and previous single-center trials.
      • Garg V.
      • Garg H.
      • Khan A.
      • Trehanpati N.
      • Kumar A.
      • Sharma B.C.
      • et al.
      Granulocyte colony-stimulating factor mobilizes CD34(+) cells and improves survival of patients with acute-on-chronic liver failure.
      ,
      • Duan X.Z.
      • Liu F.F.
      • Tong J.J.
      • Yang H.Z.
      • Chen J.
      • Liu X.Y.
      • et al.
      Granulocyte-colony stimulating factor therapy improves survival in patients with hepatitis B virus-associated acute-on-chronic liver failure.
      A recent meta-analysis summarized all G-CSF studies which enrolled patients with severe alcoholic hepatitis and underlined the huge heterogeneity among studies and pointed towards a conflict between Asian and European results,
      • Marot A.
      • Singal A.K.
      • Moreno C.
      • Deltenre P.
      Granulocyte colony-stimulating factor for alcoholic hepatitis: a systematic review and meta-analysis of randomised controlled trials.
      suggesting that a selection bias might play a role. However, when we stratified our patients according to criteria that were used in the Asian studies, most importantly patients with an alcohol-related precipitating event and ACLF according to the APASL criteria, we were not able to demonstrate superiority of the G-CSF approach compared to SMT. In this context, it has to be mentioned that the definition of an alcohol-related precipitating event was at investigators discretion and not part of the study protocol. It has to be assumed that this cohort comprised predominantly patients with alcoholic hepatitis.
      A recent review has indicated already that disease severities and patient selections were comparable among Asian and European studies.
      • Engelmann C.
      • Martino V.D.
      • Kerbert A.J.C.
      • Weil-Verhoeven D.
      • Aehling N.F.
      • Herber A.
      • et al.
      The current status of granulocyte-colony stimulating factor to treat acute-on-chronic liver failure.
      Furthermore, G-CSF’s inefficacy was confirmed across the different ACLF severity grades and types of organ failures. Overall, these results indicate that selection criteria and the study design do not fully explain differences in efficacy data among previous studies and ours.
      It might be tempting to say that the high number of infections in our study has driven an overwhelming inflammatory response, which was aggravated by G-CSF thereby inhibiting its beneficial effect. And indeed, patients with ACLF recurrence after recovery from the first ACLF episode had higher C-reactive protein values and tend to be treated more often with G-CSF. The discrepancy between a high number of overall infections developing during the study and low number of infection-related deaths is intriguing but might be related to the fact that only primary causes of death were documented by investigators and data on a potential link between cause of death and underlying infection was not systematically collected.
      Stimulating the immune response may be an effect also related to G-CSF per se independent of infections as shown by several pre-clinical studies. The rational for treating ACLF with GCSF comes from animal models showing pro-regenerative and anti-inflammatory effects,
      • Piscaglia A.C.
      • Shupe T.D.
      • Oh S.H.
      • Gasbarrini A.
      • Petersen B.E.
      Granulocyte-colony stimulating factor promotes liver repair and induces oval cell migration and proliferation in rats.
      ,
      • Theocharis S.E.
      • Margeli A.P.
      • Kittas C.N.
      Effect of granulocyte colony-stimulating-factor administration on tissue regeneration due to thioacetamide-induced liver injury in rats.
      and the earlier Asian studies seemed to confirm these findings in humans.
      • Garg V.
      • Garg H.
      • Khan A.
      • Trehanpati N.
      • Kumar A.
      • Sharma B.C.
      • et al.
      Granulocyte colony-stimulating factor mobilizes CD34(+) cells and improves survival of patients with acute-on-chronic liver failure.
      ,
      • Duan X.Z.
      • Liu F.F.
      • Tong J.J.
      • Yang H.Z.
      • Chen J.
      • Liu X.Y.
      • et al.
      Granulocyte-colony stimulating factor therapy improves survival in patients with hepatitis B virus-associated acute-on-chronic liver failure.
      It has to be emphasized that G-CSF may act like a double-edged sword. Toll-like receptor 4 (TLR4)-activation was observed to be a major driver of disease progression in ACLF.
      • Engelmann C.
      • Sheikh M.
      • Sharma S.
      • Kondo T.
      • Loeffler-Wirth H.
      • Zheng Y.B.
      • et al.
      Toll-like receptor 4 is a therapeutic target for prevention and treatment of liver failure.
      ,
      • Engelmann C.
      • Adebayo D.
      • Oria M.
      • De Chiara F.
      • Novelli S.
      • Habtesion A.
      • et al.
      Recombinant alkaline phosphatase prevents acute on chronic liver failure.
      As shown in septic liver injury models, G-CSF might enhance sensitivity to endotoxins and increase hepatocyte uptake of lipopolysaccharide through the lipopolysaccharide-binding protein/TLR4 axis, thereby aggravating tissue injury, as shown by HMGB1 translocation and mortality in rat models of septic liver injury.
      • Fang H.
      • Jin H.
      • Hua C.
      • Liu A.
      • Song Z.
      • Chen X.
      • et al.
      The LPS responsiveness in BN and LEW rats and its severity are modulated by the liver.
      ,
      • Fang H.
      • Liu A.
      • Sun J.
      • Kitz A.
      • Dirsch O.
      • Dahmen U.
      Granulocyte colony stimulating factor induces lipopolysaccharide (LPS) sensitization via upregulation of LPS binding protein in rat.
      G-CSF also activates neutrophils that carry TLR4
      • Acorci-Valerio M.J.
      • Bordon-Graciani A.P.
      • Dias-Melicio L.A.
      • de Assis Golim M.
      • Nakaira-Takahagi E.
      • de Campos Soares A.M.
      Role of TLR2 and TLR4 in human neutrophil functions against Paracoccidioides brasiliensis.
      and migrate into injured tissues, potentially causing additional cellular damage.
      • Fiuza C.
      • Salcedo M.
      • Clemente G.
      • Tellado J.M.
      Granulocyte colony-stimulating factor improves deficient in vitro neutrophil transendothelial migration in patients with advanced liver disease.
      Therefore, the effect of G-CSF, either beneficial or detrimental, might depend on the disease stage. In our trial, the number of possibly drug-related SAEs and the trend that patients with ACLF recurrence were more often treated with G-CSF points towards a potential clinical risk when using G-CSF, although these data were purely descriptive. The comparison with previous trials is difficult as WBC-related or drug-related severe complications were not reported.
      • Garg V.
      • Garg H.
      • Khan A.
      • Trehanpati N.
      • Kumar A.
      • Sharma B.C.
      • et al.
      Granulocyte colony-stimulating factor mobilizes CD34(+) cells and improves survival of patients with acute-on-chronic liver failure.
      ,
      • Duan X.Z.
      • Liu F.F.
      • Tong J.J.
      • Yang H.Z.
      • Chen J.
      • Liu X.Y.
      • et al.
      Granulocyte-colony stimulating factor therapy improves survival in patients with hepatitis B virus-associated acute-on-chronic liver failure.
      The study presented here has several limitations. Firstly, patient recruitment had to be prematurely terminated due to futility after performance of conditional power analyses (in a variety of sub-populations) in parallel to the planned interim analysis. A calculated conditional power of 0.0266 for the primary endpoint and <0.005 for overall survival made it unethical to continue the publicly funded trial although there was no safety concern. We have also not provided any data about the mode of action of G-CSF in ACLF. We did not specifically address potential ethnic differences which could be a potential issue. However, so far there is no evidence to suggest that G-CSF itself acts differently or that the ethnic background of patients influences ACLF’s phenotype beyond the different diagnostic criteria applied in Asia and Europe. Open labeled trials are always at risk of bias. However, given the high number of positive previous study results one would expect an overestimation of a therapeutic effect rather than failing to proof the benefit of G-CSF. We therefore consider a bias of the effects associated with the study design to be very unlikely. Unfortunately, we were not collecting data on the etiology of the underlying chronic liver disease. That was related to the fact that our focus was on a clear diagnosis of ACLF according to the EASL-CLIF criteria and collection of acute precipitating events, which we considered as important determinants of the disease course.
      This multicenter randomized study confirms the dismal prognosis of patients with ACLF and failed to demonstrate any beneficial effect of treating this condition with G-CSF monotherapy. G-CSF was associated with drug-related serious adverse reactions and should not be used outside clinical trials for the treatment of ACLF.

      Abbreviations

      ACLF, acute-on-chronic liver failure; AE, adverse events; APASL, Asian Pacific Association for the Study of the Liver; CLIF-C, Chronic Liver Failure-consortium; DMSB, data monitoring and safety board; EASL-CLIF, European Foundation for the Study of Chronic Liver Failure; G-CSF, granulocyte-colony stimulating factor; HR, hazard ratio; ITT, intention-to-treat; MELD, model of end-stage liver disease score; OF, organ failure; OLT, orthotopic liver transplantation; PP, per protocol; SAEs, severe adverse events; SMT, standard medical therapy; TLR4, Toll-like receptor 4; WBC, white blood cell.

      Financial support

      The German Research Foundation (DFG) – EN 1100/1-1 , funded the study. Cornelius Engelmann is participant in the BIH-Charité Clinician Scientist Program funded by the Charité – Universitätsmedizin Berlin and the Berlin Institute of Health . Moritz Schmelzle and Katrin Splith received grants for ancillary studies funded by the German Research Foundation (DFG) ( SCHM2661/3-1 , SCHM2661/3-2 ). Jonel Trebicka is supported by grants from the German Research Foundation (DFG) ( SFB TRR57 to P18), European Union's Horizon 2020 Research and Innovation Programme (Galaxy, No. 668031 and MICROB-PREDICT, No. 825694), and Societal Challenges - Health, Demographic Change and Wellbeing (No. 731875 ), and Cellex Foundation (PREDICT).

      Authors’ contributions

      CE, TB, MS, KS, AF and AS were substantially involved the conception and design of the project. CE, TB and AF were involved in data analysis and interpretation. All authors were involved in data acquisition, revised the manuscript critically, approved the final version to be published and agreed on the accuracy and integrity of data presented here.

      Data availability statement

      Data are available upon request from the corresponding author.

      Conflict of interest

      CE: has on-going research collaboration with Merz Pharmaceutical and Novartis. He has received speaker fees from Novartis, Gilead and Merz Pharmaceuticals. TB: Receipt of honoraria or consultation fees or participation in a company sponsored speaker’s bureau: Abbvie, Alexion, Bayer, Gilead, Eisai, Falk Foundation, Intercept, Ipsen, Janssen, MSD/Merck, Novartis, and Sequana Medical. Receipt of grants/research supports: Abbvie, BMS, Gilead, MSD/Merck, Humedics, Intercept, Merz, Sequana Medical. MS: Receipt of honoraria or consultation fees or participation in a company sponsored speaker’s bureau: Merck Serono GmbH, Bayer AG, ERBE Elektromedizin GmbH, Amgen Inc., Johnson&Johnson Medical GmbH, ERBE Elektromedizin GmbH, Takeda Pharmaceutical Limited, Olympus K.K., Medtronic GmbH, Intuitive Surgical Inc.. JT: has received speaking and/or consulting fees from Gore, Bayer, Alexion, MSD, Gilead, Intercept, Norgine, Grifols, Versantis, and Martin Pharmaceutical. All other authors declared no conflict of interest.
      Please refer to the accompanying ICMJE disclosure forms for further details.

      Acknowledgement

      We would like to thank Laura A. Kehoe for her excellent English language editing. We also like to acknowledge Prof. Reiner Wiest, Prof. Tilmann Sauerbruch and Prof. Walter Lehmacher for their deliberate and critical contribution as members of the DMSB.

      References

        • Arroyo V.
        • Moreau R.
        • Jalan R.
        Acute-on-Chronic liver failure.
        N Engl J Med. 2020; 382: 2137-2145
        • Pimpin L.
        • Cortez-Pinto H.
        • Negro F.
        • Corbould E.
        • Lazarus J.V.
        • Webber L.
        • et al.
        Burden of liver disease in Europe: epidemiology and analysis of risk factors to identify prevention policies.
        J Hepatol. 2018; 69: 718-735
        • Moreau R.
        • Jalan R.
        • Gines P.
        • Pavesi M.
        • Angeli P.
        • Cordoba J.
        • et al.
        Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis.
        Gastroenterology. 2013; 144 (1426-1437, 1437 e1421-1429)
        • Artru F.
        • Louvet A.
        • Ruiz I.
        • Levesque E.
        • Labreuche J.
        • Ursic-Bedoya J.
        • et al.
        Liver transplantation in the most severely ill cirrhotic patients: a multicenter study in acute-on-chronic liver failure grade 3.
        J Hepatol. 2017; 67: 708-715
        • Engelmann C.
        • Thomsen K.L.
        • Zakeri N.
        • Sheikh M.
        • Agarwal B.
        • Jalan R.
        • et al.
        Validation of CLIF-C ACLF score to define a threshold for futility of intensive care support for patients with acute-on-chronic liver failure.
        Crit Care. 2018; 22: 254
        • Albillos A.
        • Lario M.
        • Alvarez-Mon M.
        Cirrhosis-associated immune dysfunction: distinctive features and clinical relevance.
        J Hepatol. 2014; 61: 1385-1396
        • Moreau R.
        • Claria J.
        • Aguilar F.
        • Fenaille F.
        • Lozano J.J.
        • Junot C.
        • et al.
        Blood metabolomics uncovers inflammation-associated mitochondrial dysfunction as a potential mechanism underlying ACLF.
        J Hepatol. 2020; 72: 688-701
        • Engelmann C.
        • Becker C.
        • Boldt A.
        • Herta T.
        • Boehlig A.
        • Splith K.
        • et al.
        Ascites' neutrophil function is significantly impaired in patients with decompensated cirrhosis but can be restored by autologous plasma incubation.
        Sci Rep. 2016; 6: 37926
        • Engelmann C.
        • Claria J.
        • Szabo G.
        • Bosch J.
        • Bernardi M.
        Pathophysiology of decompensated cirrhosis: portal hypertension, circulatory dysfunction, inflammation, metabolism and mitochondrial dysfunction.
        J Hepatol. 2021; 75: S49-S66
        • Li N.
        • Zhang L.
        • Li H.
        • Fang B.
        Human CD34+ cells mobilized by granulocyte colony-stimulating factor ameliorate radiation-induced liver damage in mice.
        Stem Cell Res Ther. 2010; 1: 22
        • am Esch J.S.
        • Schmelzle M.
        • Furst G.
        • Robson S.C.
        • Krieg A.
        • Duhme C.
        • et al.
        Infusion of CD133+ bone marrow-derived stem cells after selective portal vein embolization enhances functional hepatic reserves after extended right hepatectomy: a retrospective single-center study.
        Ann Surg. 2012; 255: 79-85
        • Engelmann C.
        • Martino V.D.
        • Kerbert A.J.C.
        • Weil-Verhoeven D.
        • Aehling N.F.
        • Herber A.
        • et al.
        The current status of granulocyte-colony stimulating factor to treat acute-on-chronic liver failure.
        Semin Liver Dis. 2021;
        • Demetri G.D.
        • Griffin J.D.
        Granulocyte colony-stimulating factor and its receptor.
        Blood. 1991; 78: 2791-2808
        • Engelmann C.
        • Splith K.
        • Berg T.
        • Schmelzle M.
        Effects of granulocyte-colony stimulating factor (G-CSF) on stem cell mobilization in patients with liver failure.
        Eur J Intern Med. 2016; 36: e37-e39
        • Garg V.
        • Garg H.
        • Khan A.
        • Trehanpati N.
        • Kumar A.
        • Sharma B.C.
        • et al.
        Granulocyte colony-stimulating factor mobilizes CD34(+) cells and improves survival of patients with acute-on-chronic liver failure.
        Gastroenterology. 2012; 142: 505-512 e501
        • Duan X.Z.
        • Liu F.F.
        • Tong J.J.
        • Yang H.Z.
        • Chen J.
        • Liu X.Y.
        • et al.
        Granulocyte-colony stimulating factor therapy improves survival in patients with hepatitis B virus-associated acute-on-chronic liver failure.
        World J Gastroenterol. 2013; 19: 1104-1110
        • Shephard D.A.
        The 1975 declaration of Helsinki and consent.
        Can Med Assoc J. 1976; 115: 1191-1192
        • Jalan R.
        • Saliba F.
        • Pavesi M.
        • Amoros A.
        • Moreau R.
        • Gines P.
        • et al.
        Development and validation of a prognostic score to predict mortality in patients with acute-on-chronic liver failure.
        J Hepatol. 2014; 61: 1038-1047
        • Pieri G.
        • Agarwal B.
        • Burroughs A.K.
        C-reactive protein and bacterial infection in cirrhosis.
        Ann Gastroenterol. 2014; 27: 113-120
        • Malkan U.Y.
        • Ozcebe O.I.
        Leukapheresis do not improve early death rates in acute myeloid leukemia patients with hyperleukocytosis.
        Transfus Apher Sci. 2017; 56: 880-882
        • O'Brien P.C.
        • Fleming T.R.
        A multiple testing procedure for clinical trials.
        Biometrics. 1979; 35: 549-556
        • Singh V.
        • Sharma A.K.
        • Narasimhan R.L.
        • Bhalla A.
        • Sharma N.
        • Sharma R.
        Granulocyte colony-stimulating factor in severe alcoholic hepatitis: a randomized pilot study.
        Am J Gastroenterol. 2014; 109: 1417-1423
        • Sarin S.K.
        • Kedarisetty C.K.
        • Abbas Z.
        • Amarapurkar D.
        • Bihari C.
        • Chan A.C.
        • et al.
        Acute-on-chronic liver failure: consensus recommendations of the asian pacific association for the study of the liver (APASL) 2014.
        Hepatol Int. 2014; 8: 453-471
        • Khanam A.
        • Trehanpati N.
        • Garg V.
        • Kumar C.
        • Garg H.
        • Sharma B.C.
        • et al.
        Altered frequencies of dendritic cells and IFN-gamma-secreting T cells with granulocyte colony-stimulating factor (G-CSF) therapy in acute-on- chronic liver failure.
        Liver Int. 2014; 34: 505-513
        • Sharma S.
        • Lal S.B.
        • Sachdeva M.
        • Bhatia A.
        • Varma N.
        Role of granulocyte colony stimulating factor on the short-term outcome of children with acute on chronic liver failure.
        J Clin Exp Hepatol. 2020; 10: 201-210
        • Marot A.
        • Singal A.K.
        • Moreno C.
        • Deltenre P.
        Granulocyte colony-stimulating factor for alcoholic hepatitis: a systematic review and meta-analysis of randomised controlled trials.
        JHEP Rep. 2020; 2: 100139
        • Piscaglia A.C.
        • Shupe T.D.
        • Oh S.H.
        • Gasbarrini A.
        • Petersen B.E.
        Granulocyte-colony stimulating factor promotes liver repair and induces oval cell migration and proliferation in rats.
        Gastroenterology. 2007; 133: 619-631
        • Theocharis S.E.
        • Margeli A.P.
        • Kittas C.N.
        Effect of granulocyte colony-stimulating-factor administration on tissue regeneration due to thioacetamide-induced liver injury in rats.
        Dig Dis Sci. 1999; 44: 1990-1996
        • Engelmann C.
        • Sheikh M.
        • Sharma S.
        • Kondo T.
        • Loeffler-Wirth H.
        • Zheng Y.B.
        • et al.
        Toll-like receptor 4 is a therapeutic target for prevention and treatment of liver failure.
        J Hepatol. 2020; 73: 102-112
        • Engelmann C.
        • Adebayo D.
        • Oria M.
        • De Chiara F.
        • Novelli S.
        • Habtesion A.
        • et al.
        Recombinant alkaline phosphatase prevents acute on chronic liver failure.
        Sci Rep. 2020; 10: 389
        • Fang H.
        • Jin H.
        • Hua C.
        • Liu A.
        • Song Z.
        • Chen X.
        • et al.
        The LPS responsiveness in BN and LEW rats and its severity are modulated by the liver.
        J Immunol Res. 2018; 2018: 6328713
        • Fang H.
        • Liu A.
        • Sun J.
        • Kitz A.
        • Dirsch O.
        • Dahmen U.
        Granulocyte colony stimulating factor induces lipopolysaccharide (LPS) sensitization via upregulation of LPS binding protein in rat.
        PloS One. 2013; 8e56654
        • Acorci-Valerio M.J.
        • Bordon-Graciani A.P.
        • Dias-Melicio L.A.
        • de Assis Golim M.
        • Nakaira-Takahagi E.
        • de Campos Soares A.M.
        Role of TLR2 and TLR4 in human neutrophil functions against Paracoccidioides brasiliensis.
        Scand J Immunol. 2010; 71: 99-108
        • Fiuza C.
        • Salcedo M.
        • Clemente G.
        • Tellado J.M.
        Granulocyte colony-stimulating factor improves deficient in vitro neutrophil transendothelial migration in patients with advanced liver disease.
        Clin Diagn Lab Immunol. 2002; 9: 433-439