Autologous hematopoietic stem cell therapy in liver cirrhosis 목요세미나 Autologous hematopoietic stem cell therapy in liver cirrhosis 2019. 07. 18. R3. 김보경 / Pf. 이윤빈

Introduction Chronic liver disease Common cause of death and the incidence is rising globally (Alcohol, obesity, viral) Increasing clinical burden and absence of effective therapies Consideration of innovative approaches 만성 간질환은 사망의 주요한 원인이며, 전 세계적으로는 알코올 섭취 증가와 obesity, virus 등의 영향으로 그 유병률이 조금씩 증가하고 있습니다. 하지만 그 Clinical burden 에 비해 간 이식 외에 효과적인 치료법이 아직은 없는게 현실입니다. 그래서 새로운 치료법들에 대한 관심이 대두되고 있습니다. Lancet. 2017;389:1043–105

Introduction Autologous stem cell therapy in liver cirrhosis Reduce hepatic fibrosis, stimulate liver regeneration and improve biochemical indexes Transdifferentiation, oval cell replacement, paracrine effects, immunomodulation, remodeling fibrosis 이러한 관점에서 최근 수년 사이 LC 환자에서 자가조혈모세포 이식이 새롭게 조명되고 있습니다. 여러 전임상연구나 임상연구를 통해서 조혈모세포가 liver fibrosis 나 cirrhosis 가 있는 상황에서 간 섬유화를 줄이고, 간의 재생을 촉진시키며 간기능을 호전시킨다는 연구가 보고되었습니다. 첫번째 논문 : 2010 년 Eur J Gastroenterol Hepatol. 에 개재된 RCT로, advanced LC 환자를 대상으로, 자가조혈모세포를 hepatic artery 로 주입하였을 때 초기 90일에 Child pugh score, MELD score 에서 효과가 있었다는 연구입니다. 2) 두번째 논문 : 2013 liver international 에 개재된 논문으로, A-LC 환자 12명을 대상으로 자가조혈모세포이식을 하기전과 한 후의 liver biopsy 결과를 비교하였을 때 치료한 군에서 섬유화의 호전이 있었다고 보고하였습니다. 조혈모세포가 어떤 기전을 통해 간기능을 호전시키는지에 대해서 다양한 가설이 있는데, 3) 세번재 논문 : 이 그림은 2013 stem cells 에 개재된 review 에 나온 그림이고 조혈모세포가 hepatocyte 로 Transdifferenciation 되거나 endogenous hepatocyte 과 fusion 혹은 oval cell 로 변화하는 기전, extracellular matrix 를 제공하고 anti-inflammatory effect 를 나타내는 paracrine effect, 그리고 kupffer cell modification 을 통한 immunomodulation, fibrosis remodeling 등으로 나누어 설명하였습니다. Eur J Gastroenterol Hepatol. 2010;22:33–42. Liver Int, 34 (2014), pp. 33-41. Stem Cells, 31 (2013), pp. 2313-2329

Introduction Autologous stem cell therapy in liver cirrhosis Mostly small cohorts, short follow-up, not universally observed and not always durable 그러나 현재까지 임상연구는 규모가 작은 코호트 연구가 대부분이며, f/u 기간이 짧고, 어떤 연구에서는 효과가 있다고 나오고 다른 연구에서는 효과가 없다고 나오는 등 일관성이 부족하고, 효과가 있는 경우에도 시간이 지나면 차이가 없어지는 등 지속성이 짧다고 보고된 경우가 있었습니다. The size and nature of the trial designs of many of the previous studies make it impossible to draw meaningful conclusions on clinical outcomes, and thus the efficacy of bone-marrow-derived stem cells for liver cirrhosis has yet to be proven However, these effects are not universally detected between studies and when seen are not always durable The size and nature of the trial designs of many of the previous studies make it impossible to draw meaningful conclusions on clinical outcomes, and thus the efficacy of bone-marrow-derived stem cells for liver cirrhosis has yet to be proven. Systematic review – 2015, AP &T Background As morbidity and mortality from liver disease continues to rise, new strategies are necessary. Liver transplantation is not only an expensive resource committing the patient to lifelong immunosuppression but also suitable donor organs are in short supply. Against this background, autologous stem cell therapy has emerged as a potential treatment option. Aim To evaluate if it is possible to make a judgement on the safety, feasibility and effect of autologous stem cell therapy for patients with liver disease. Methods MEDLINE and EMBASE were searched up until July 2013 to identify studies where autologous stem cell therapy was administered to patients with liver disease. Results Of 1668 studies identified, 33 were eligible for inclusion evaluating a median sample size of 10 patients for a median follow‐up of 6 months. Although there was marked heterogeneity between studies with regards to type, dose and route of delivery of stem cell, the treatment was shown to be safe and feasible largely when a peripheral route of administration was used. Of the studies which also looked at biochemical outcome, statistically significant improvement in liver function tests was seen in 16 studies post‐treatment. Conclusion Although autologous stem cell therapy is a much needed possibility in the treatment of liver disease, further robust clinical trials and collaborative protocols are required. Aliment Pharmacol Ther. 2014;39:673–685 Stem Cells, 31 (2013), pp. 2313-2329

Granulocyte colony-stimulating factor and autologous CD133-positive stem-cell therapy in liver cirrhosis (REALISTIC): an open-label, randomized, controlled phase 2 trial P.N. Newsome, et al. Lancet Gastroenterol Hepatol 2018;3(1):25-36. Long-term Outcomes of Autologous Peripheral Blood Stem Cell Transplantation in Patients With Cirrhosis Changcun Guo, et al. Clin Gastroenterol and Hepatol 2019;17(6):1175-1182. 오늘 말씀드리고자 하는 두 논문은 간경변에서 조혈모세포 이식에 대한 논문으로, 주논문은 2018년 lancet 에 발표되었고 compensated LC 환자를 대상으로 G-CSF 투여 및 자가조혈모세포치료의 효과를 비교한 open label randomized controlled phase 2 trial 이고, 부논문은 2019 년 clinical gastroenterology hepatology 에 발표된 논문으로, decompensated LC 를 대상으로 자가조혈모세포 이식을 했을때의 long term outcome 에 대한 연구입니다.

먼저 주논문입니다.

Background & Aims Autologous stem cell therapy in liver disease No adequately powered RCTs Multiple administrations of stem cells might have a greater antifibrotic effect Assess the safety and efficacy of G-CSF and hematopoietic stem-cell infusions in patients with compensated liver cirrhosis 앞서 말씀드린 것처럼 간경변에서의 조혈모세포이식에 대해서는 잘 설계된 RCT 데이터가 부족하여 치료의 근거가 아직은 미약한 상태입니다. 이 연구에서는 간경변이 chronic 한 컨디션이므로 자가조혈모세포를 single dose 로 투약하는 것보다 여러 번 나누어 투약하는 것이 antifibrotic effect 에 더 효과적일 것이라고 가정하였고, Compensated LC 환자를 대상으로 G-CSF 와 자가조혈모세포 이식이 간기능에 어떠한 효과를 갖는지와 그 안전성에 대해 평가하는 것을 목적으로 하였습니다. 여하는 것은 효과가 불충분The size and nature of the trial designs of many of the previous studies make it impossible to draw meaningful conclusions on clinical outcomes, and thus the efficacy of bone-marrow-derived stem cells for liver cirrhosis has yet to be proven.8 In one adequately powered randomised controlled study, whole bone marrow mononuclear cells were infused into patients with acute inflammatory disease (alcoholic hepatitis), but had no effect.12 However, alcoholic hepatitis is a rarer and entirely different clinicopathological entity to that seen in patients with compensated liver cirrhosis, in which there is minimal inflammation and fibrosis is the major feature. Furthermore the mononuclear cells infused contained a mixed cell population. Because liver cirrhosis is a chronic condition and a single dose of cells might have little effect, as suggested in a previous study0 we reasoned that multiple administrations of stem cells might have a greater antifibrotic effect, which was supported by our preclinical findings.6 We therefore designed a sufficiently powered, randomised controlled trial of granulocyte colony-stimulating factor (G-CSF) with or without repeated haemopoietic stem-cell (HSC) therapy in patients with compensated liver cirrhosis to detect clinically important effects on liver function and liver fibrosis and to assess safety. Nat Rev Gastroenterol Hepatol. 2016;13:473–485.

Methods Multicenter, open-label, randomized, controlled phase 2 trial Three hospitals in UK May 18, 2010 ~ Feb 6, 2015 이 연구는 Multicenter, open-label, randomized, controlled phase 2 trial 이고, 영국의 세개 병원에서 2010년 5월부터 2015년 2월까지 enroll 하였습니다. The REpeated AutoLogous Infusions of STem cells In Cirrhosis (REALISTIC) trial was a multicentre, open-label, randomised, controlled phase 2 trial done at three hospitals in Birmingham, Edinburgh, and Nottingham, UK. The National Research Ethics Service (NRES) Oxfordshire REC A committee (REC reference 09/H0604/64) and the Medicines and Healthcare Products Regulatory Agency (MHRA) approved all versions of the study protocol. Additionally, all recruitment sites obtained approval from their local hospital research and development departments. The University of Birmingham (Birmingham, UK) acted as the sponsor of the trial. A detailed version of the REALISTIC protocol has been published.13

Methods Inclusion Criteria Exclusion Criteria 18 < Age ≤ 75 Compensated cirrhosis (liver biopsy, fibroscan, APRI or clinical features) 11 < MELD <15.5 Etiology : Alcohol related, Hepatitis B/C, Primary biliary cirrhosis, Hemochromatosis, NAFLD, Alpha-1 antitrypsin deficiency, Cryptogenic Exclusion Criteria Average alcohol ingestion >21 units/week (male) / >14units/week (female) Ascites (unless minimal and well controlled) Encephalopathy (requiring hospitalization, in the last 3 months) Portal hypertensive bleeding (requiring treatment or hospitalization in the last 3 months) HCC Previous or listed for Liver Transplantation Recent history of pneumonia Other clinical problems potentially compromise patient safety Cancer in last 5 years Pregnancy or breastfeeding. Inclusion criteria 는 18세 이상, 75세 이상이며 Liver biopsy 나 fibroscan, APRI 혹은 임상적으로 cirrhosis 가 확인된 compendsated cirrhosis 중 MELD 가 11 에서 15.5 사이인 환자를 대상으로 하였고, autoimmune hepatitis 를 재외한 대부분의 etiology 를 포함하였습니다. Exclusion criteria 는 음주를 남자의 경우 주 21unit, 여자의 경우 주 14unit 이상 하거나 잘 조절되지 않는 복수, 최근 3달 내 간성혼수나 bleeding 으로 입원치료를 요한 경우, HCC 나 TPL 병력, 최근 폐렴이 있는 경우, 5년 이내 발생한 다른 암이 있는 경우, 임신이나 수유, 그리고 trial 이 환자의 안전에 영향을 미칠만한 다른 comorbidity 가 있을때로 설정하였습니다. The trial entry criteria were based on the presence of liver cirrhosis identified on biopsy or on clinical grounds as detailed in the appendix (p 3). Participants had to be aged 18–75 years with predominantly compensated cirrhosis (most causes were allowed except for autoimmune hepatitis) and a MELD score of 11·0–15·5. Exclusion criteria were average alcohol ingestion greater than 21 units per week for men or greater than 14 units for women, any alcohol consumption within 6 months for patients with alcohol-related liver disease, ascites unless deemed by the investigator to be minimal and well controlled with no changes to diuretic therapy in the past 3 months, encephalopathy that was either current or required hospital admission for treatment in the past 3 months, portal hypertensive bleeding with an episode requiring treatment or hospital admission in the past 3 months, previous or current history of hepatocellular carcinoma, presence of dysplastic or indeterminate nodules, previous liver transplant (or being on a list for a transplant), and recent history of pulmonary infiltrates or pneumonia. All patients with hepatitis C virus (HCV) infection had to have received previous antiviral treatment before being considered for the trial, but none had cleared the virus after treatment and thus all had ongoing viral injury. We recruited participants from among patients attending the liver clinics. All patients gave written informed consent, the study was done by site investigators, and data were gathered by specifically trained personnel.

Methods: Study Interventions There are 2 major types of adult stem cell: HSCs and mesenchymal stem cells(MSCs), both of which are found predominantly in the bone marrow, although the latter can be found in adipose tissue in reasonable numbers.58 HSCs have been well characterized with respect to their recapitulation of the major hematopoietic lineages, and cell surface markers that can be used to identify, and isolate, them. In human beings, expression of CD34 on the cell surface is largely used in clinical practice, although CD133 expression is believed to represent a more stem cell–enriched subpopulation of the CD34+ cells.59 The expression of these markers does not denote that these cells are stem cells; rather it implies that the population is enriched for stem cells. True pluripotent stem cells would constitute about 0.1% of a CD133+ population Randomized (1:1:1) Standard care G-CSF G-CSF + stem-cell infusion Minimization algorithm (trial site and cause of liver disease) Lenograstim 15 μg/kg x 5 days G-CSF plus stem cells Leukapheresis, CD133+ HSCs (0·2 × 106 cells/kg/infusion) on days 5,30,60 Pheripheral vein Flow cytometry (CD 133+, CD34+, CD45+) Randomization 은 antiviral agent 등 disease-specific medications 과 LC complication 등의 기본적인 치료를 standard care 그룹과, standard care 에 더불어 G-CSF 를 추가로 투여한 G-CSF group , 그리고 standard care 를 하면서 G-CSF 를 투여하고 이후 자가조혈모세포까지 투여한 G-CSF + stem cell infusion group 으로 1:1:1 로 나누었습니다. Randomization 은 Minimization 알고리즘을 사용하였습니다. 이는 임상시험에서 각 치료군 간 균형을 맞추기 위한 방법인데, 중요한 공변량에서 치료군간에 불균형이 생기지 않도록 이전에 배정된 피험자의 공변량을 고려하여 새로운 피험자를 특정한 치료군에 순차적으로 배정하는 방법이고, 이 연구의 경우에는 trial site 와 etiology 에 따라 나누었습니다. G-CSF 는 15mcg/kg 로 5일간 투여하였고 자가조혈모세포까지 이식하는 군에서는 이후 leukapheresis 를 하고, flow cytometry 를 통해 stem cell marker 인 CD 34 그리고 CD 133 을 expression 하는 cell 을 분리하여 peripheral vein 을 통하여 5일, 30일, 60일째에 3회에 걸쳐 나누어 투여하였습니다. ** hematologic malignancy 에서는 4~6 x 10^6/kg target (층화로 인한 문제를 극복할 수 있으나, 공변량이 많을수록 수행이 복잡하고 선택편향의 가능성이 여전히 있음. 분석을 정확히 하려면 시뮬레이션을 통해 유의수준을 파악하고 공분산 분석을 진행해야 함) Randomisation Centre-delegated staff telephoned randomisation officers at the Cancer Research UK Clinical Trials Unit (CRCTU; Birmingham, UK), who used a computer-generated, centrally administered procedure to randomly assign eligible patients (1:1:1) to one of three treatments: standard care alone, G-CSF, or G-CSF plus stem-cell infusion. Randomisation was based on a minimisation algorithm (Randomisation was done by Cancer Research UK Clinical Trials Unit staff with a minimisation algorithm that stratified by trial site and cause of liver disease. prepared and validated by the CRCTU programming and statistical team.  The 81 patients recruited were not equally allocated to the treatment groups, because the minimisation procedure allocated patients to ensure that groups were balanced with respect to the specified minimisation factors. ) and patients were also stratified by trial site and cause of disease (alcohol-related liver disease, HCV, and other). Each patient was then allocated a unique patient trial number and scheduled for treatment and follow-up visits. The local site staff could not predetermine treatment allocation, the treatment allocation was open, and all clinicians and local site staff, as well as the patients, were aware of which treatment had been allocated (open-label therapy) following randomisation. Procedures All patients received standard management for compensated cirrhosis, which could include disease-specific medications and treatments for the complications of cirrhosis. Patients assigned to the control group received standard of care management alone. Patients assigned to the G-CSF group received subcutaneous injections of G-CSF (lenograstim; Chugai Pharmaceuticals, London, UK) at 15 μg/kg bodyweight daily for 5 consecutive days. Patients assigned to receive G-CSF plus stem cells received subcutaneous injections of G-CSF 15 μg/kg bodyweight per day for 5 consecutive days and underwent leukapheresis and intravenous infusion of three doses of CD133-positive haemopoietic stem cells (HSCs; 0·2 × 106 cells per kg) on days 5, 30, and 60 after randomisation. Isolation of CD133-positive HSCs from the harvested peripheral blood mononuclear cells was done under aseptic conditions within clean room facilities in accordance with Good Manufacturing Practice regulations (MHRA/HTA, UK). Cell sorting was done in a closed, sterile system to provide clinical grade enrichment (CliniMACS Plus, Miltenyi Biotec, Bergisch Gladbach, Germany). Cells were tested with flow cytometry (CD45-positive, CD34-positive, and CD133-positive) on the day of isolation to establish positive and negative fractions. The gating strategy was set up to exclude debris, beads, and non-viable cells, before gating on CD45 and then CD133 expression such that total cells included were viable and positive for CD45 and CD133. This information was used to aliquot three doses 0·2 × 106 cells per kg for each patient. Cells had to have greater than 50% viability, which is in line with clinical practice. Fresh CD133-positive HSCs were administered intravenously for immediate reinfusion and a further two aliquots were cryopreserved according to site standard protocols for later reinfusion at day 30 and day 60 post randomisation, thus requiring the administration of a minimum total number of cells of 0·6 × 106 cells per kg. Cells for days 30 and 60 were supplied to the ward frozen and thawed before being administered; results for sterility testing from BacTalert bottles were provided at the same time, in line with standard clinical practice for the administration of HSCs to patients receiving stem cell transplants. If there were insufficient cells for three doses, cells were allocated preferentially to the first, then second, and then third dose. The procurement, processing, storage and distribution of the autologous CD133-positive HSCs was done in accordance with Tissue Quality and Safety Regulations by establishments holding Human Tissue Authority licences. After randomisation, all patients returned for study visits at days 30, 60, 90, 180, and 360 (end of study). The coprimary outcomes used data captured from baseline to day 90. The schedule for the study visits and data collection is summarised in the appendix (p 5).

Methods: Outcome measures Co-primary outcome Change in MELD score at 90 days The trend of treatment activity established by incorporating MELD score measured at baseline and days 30, 60, and 90 Secondary outcome Liver stiffness (Fibroscan) Enhanced liver fibrosis (ELF) test Chronic Liver Disease Questionnaire (CLDQ) scores Individual components of liver function (bilirubin, albumin, INR, creatinine) UK End-Stage Liver Disease (UKELD) score Circulating peripheral blood HSCs Long-term MELD and UKELD (to day 360) Clinical events Transplant-free survival Safety and adverse events Co-primary outcome 으로 2개의 outcome 을 설정하였는데 하나는 90일째의 MELD 변화이고 두번째는 30,60,90 일째의 MELD 의 변화 양상에 따라 치료 효과의 trend 를 파악하는 것이었습니다. Secondary outcome 은, liver stiffness, Enhanced liver fibrosis test, CLDQ score, serum bilirubin, albumin, INR, Cr 수치, UK end stage liver disease score, 1년째 MELD 및 UK end stage liver disease score, clinical events, transplant free survival 과, adverse event 를 평가하였습니다. Outcomes The coprimary outcome measures were change in MELD score at 90 days from baseline and the trend of treatment activity established by incorporating MELD score measured at baseline and days 30, 60, and 90. The protocol was updated in March, 2015, to include the second coprimary outcome to make better use of data collected and detect differences before 90 days. MELD score is calculated from objective variables that are readily obtained: serum bilirubin, serum creatinine, and International Normalised Ratio (INR).14Secondary outcome measures were liver stiffness (Fibroscan), enhanced liver fibrosis (ELF) test, Chronic Liver Disease Questionnaire (CLDQ) scores, individual components of liver function (bilirubin, albumin, INR, and creatinine values [creatinine was unplanned]), UK End-Stage Liver Disease (UKELD) score, circulating peripheral blood HSCs, long-term MELD and UKELD (to day 360), clinical events, and transplant-free survival. Safety and adverse events were assessed with standard reporting forms by trained investigators. The reporting period for adverse events started from the date of patient consent and continued throughout the study until visit 7 (day 360). Serious adverse events were reported from date of consent until 30 days after the last possible stem-cell infusion (day 90) for all treatment groups, therefore ensuring that the reporting period stayed the same for all treatment groups. All serious adverse events and adverse events were assessed by the local investigators and recorded. We used the National Cancer Institute's common terminology criteria for adverse events (CTCAE version 4.02–2010) to grade each adverse event. The CRCTU kept detailed records of all adverse events reported (nature, onset, duration, severity, and outcome) and did an assessment with respect to seriousness, causality, and expectedness.

Methods: Statistical analysis Three-armed study with one control group Powered to compare each treatment to control Not powered to detect differences between the two treatment groups 27 patients into each group Pooled SD assumed to be 1·25, Standardized effect size of at least 0·8 -> 1 point reduction in MELD Two-sided α 0·05, 80% power First co-primary outcome : Non-parametric two-sample Wilcoxon test, exploratory linear regression models Second co-primary outcome : Linear mixed-effects models Statistical analysis 입니다. 한 개의 control group 을 포함한 3 arm study 로, G-CSF group 과 자가조혈모세포이식 group 이 각각 control 과 치료효과를 비교하도록 power 되었으나, G-CSF group 과 자가조혈모세포이식 group 끼리의 차이는 전제로 하지 않고 설계하였습니다. 또한 합동 표준편차를 1.25, 표준효과크기를 0.8로 가정하였고 각 치료군에 27명씩 필요한 것으로 확인되었습니다. 90일째 MELD 의 변화인 첫번재 Co-primary outcome 을 확인하기 위해, Non-parametric two-sample Wilcoxon test 와 exploratory linear regression models 를 사용하였고, 각 치료군에 따라 30,60,90일째의 MELD 의 변화 trend 인 두번째 co-primary outcome 을 확인하기 위해 Linear mixed-effects model 을 사용하였습니다. The sample size calculation was based on the change in MELD score from baseline to 90 days after randomisation. With pooled standard deviation assumed to be 1·25 in this controlled setting, the trial aimed to detect a standardised effect size of at least 0·8 between treatment groups and the control group, which would equate to a 1 point reduction in MELD. To have error rates based on two-sided α of 0·05 (α=0·10 split equally between the two hypotheses) and 80% power, we needed to recruit 27 patients into each group.13 The trial was designed as a three-armed study with one control group and was powered to compare each treatment to control with respect to the first coprimary outcome, but was not powered to detect differences between the two treatment groups. The hypothesis of the primary analysis was designed to assess activity and, as such, all analyses were done in the modified intention-to-treat (mITT) population. The mITT population included all participants who received at least one day of treatment (one day of G-CSF at 15 μg/kg bodyweight in the treatment groups, plus one infusion of at least 0·17 × 106 cells per kg for the G-CSF plus stem-cell infusion group), with patients retained in their randomly assigned treatment groups, including those who violated the protocol or were ineligible. The per-protocol population was defined as any patients who received 5 days of G-CSF at an average daily dose of at least 12 μg/kg and any patients who received 5 days of G-CSF plus three infusions at a minimum of 0·17 × 106 cells per kg each. All patients in the control group were included in the mITT and per-protocol populations. We defined the safety population on the basis of the actual treatment received. We calculated the first coprimary outcome, change in MELD score from baseline to day 90, for each participant and we compared treatment groups with the control using the non-parametric two-sample Wilcoxon test. We fitted exploratory linear regression models, with transformed MELD scores where required, to enable adjustment of the estimates of the first coprimary outcome. For the second coprimary outcome, we explored the effect of treatment on the change in MELD score through linear mixed-effects models (details of the modelling approach used to find the most appropriate parsimonious model are shown in the appendix [p 7]), which incorporated measurements taken at baseline and at days 30, 60, and 90. We assigned random effects at the patient level. The first coprimary outcome was conditional on the availability of day 90 MELD measures, whereas the second coprimary outcome was not. Sensitivity analyses included adjustment of model-based analyses for known prognosticators, such as alcohol intake and cause of liver disease, minimisation factors, and differences at baseline. We analysed UKELD in the same manner as the coprimary outcomes, and we assessed CLDQ responses through area-under-curve analyses with a set of varied assumptions applied to address censoring, and, as a sensitivity analysis, removing participants who had major adverse events. We also assessed long-term MELD and UKELD, measured to day 360, in the same way as the coprimary outcomes. We compared average change in liver parameters between treatment groups and the control group from baseline using a test appropriate to the data (t tests or Wilcoxon rank sum). We report other outcome measures descriptively. We used Stata version 14 for all analyses. Data were collected and presented to the independent data management committee regularly throughout the recruitment, treatment, and follow-up periods of the study (appendix p 2). The committee advised the trial management team throughout the study in relation to trial conduct and specifically whether extra data monitoring was needed throughout the trial. The data monitoring committee operated in accordance with a trial-specific charter based on the template created by the Damocles Group. This trial was registered at Current Controlled Trials on Nov 18, 2009; ISRCTN, number 91288089; and the European Clinical Trials Database, number 2009-010335-41.

Results Multicenter, open-label, randomized, controlled phase 2 trial May 18, 2010 ~ Feb 6, 2015 총 153명의 환자가 스크리닝되었고 exclusion criteria 에 해당하는 72 명이 제외되었으며, minimization 알고리즘에 따라 각 군에 27, 26, 28 명씩 assign 되었습니다. 자가조혈모세포 이식 군에서 1명은 치료 전 사망하고 한명은 충분한 cell 이 모이지 않아 enroll 을 철회하였습니다. 또한 이식군중 5명은 목표에 비해 suboptimal dose 의 cell 이 투여되었습니다. 그리서 90일째에 평가하였을 때 Intention to treat 군에는 각각 27, 26, 26명 그리고 per protocol analysis 에서는 27 26 21명이 있는것으로 확인되었습니다. Between May 18, 2010, and Feb 6, 2015, 153 patients with liver cirrhosis were screened, and 81 underwent randomisation (figure 1). We recruited 58 patients from the Birmingham centre, 19 from the Edinburgh centre, and four from the Nottingham centre. The 81 patients recruited were not equally allocated to the treatment groups, because the minimisation procedure allocated patients to ensure that groups were balanced with respect to the specified minimisation factors. The final patient's characteristics resulted in a balanced state and hence the allocation was made randomly to the G-CSF plus stem-cell infusion group. The mITT population included 27 patients in group 1, 26 patients in group 2, and 26 patients in group 3, whereas the per-protocol population included 27 patients in group 1, 26 patients in group 2, and 21 patients in group 3. In group 3, one patient received no G-CSF nor any cells (this patient later died after 2·3 months), one patient received G-CSF only (withdrew from study), and a further five patientsreceived G-CSF but did not receive sufficient cells to complete three full doses (appendix p 13).

Results : Characteristic of patients randomized Baseline characteristic 을 비교한 표입니다, standard care, G-CSF, G-CSF 와 자가조혈모세포 투여 군의 나이, 성별, BMI, 알코올 섭취, LC 의 etiology, MELD, UK ELD, child pugh score, 복수나 varix bleeding, 간성혼수 등의 liver comorbidity 와 medication, LFT 포함한 lab, 그리고 fibroscan 등의 non invasive hepatic biomarker, QOL score 등을 비교한 표입니다. 대부분의 군에서 baseline characteristic 의 큰 차이는 없었으나, standard 에 비해 두 치료군에서 남성이 더 많았고, 조혈모세포 이식 군에서 serum urea 와 creatinine 이 약간 더 높았습니다. Median MELD score 는 각각 13 내외로 비슷하였습니다. 또한 일부 환자들이 복수나 varix bleeding, encephalopathy 가 있었으나 mild degree 였고 standard medical treatment 로 잘 조절되는 정도였습니다. At baseline, we noted differences between the treatment groups in terms of gender, urea, and creatinine (table 1). Patients had features of liver decompensation (ascites, encephalopathy, and prior variceal bleeding), although these were mild and responsive to standard medical treatment. The median MELD score was 13·1 (IQR 12·4–13·8) in the standard treatment group, 12·7 (12·0–13·1) in the G-CSF group, and 13·2 (12·1–13·9) in the G-CSF plus stem-cell infusion group. 22 patients had previous histological confirmation of liver cirrhosis.

Results : Co-primary outcome : Change in MELD and UKELD scores No evidence of a trend in MELD over time, nor of a difference between the groups 다음은 co primary outcome 입니다. 각 군별 시간에 따른 MELD 의 변화를 simple mixed effect model 로 확인한 그래프이고, 파란색이 standard care, 빨간색이 G-CSF only, 초록색이 G-CSF 와 조혈모세포 투여를 같이 한 군입니다. 위쪽 그래프는 각 치료군에 따른 MELD 의 변화를 Simple mixed-effects models 로 나타낸 것이고, 왼쪽이 MELD, 오른쪽이 UK ELD score 입니다. 시간에 따른 각 군별 MELD 의 유의한 변화는 없었으며, 치료군 별로도 MELD 의 유의한 변화는 없었습니다. 아래 그래프는 두번째 co-primary outcome 의 적합도(fit) 을 증대하기 위하여, 30일째를 기준으로 change point 를 두고, 치료군별 그리고 치료 시기간의 차이를 추가로 염두하여 restricted cubic splines 을 토대로 linear mixed-effects model 을 만든것인데, 그렇게 하여 적합도를 올림에도 불구하고 시간에 다른 유의미한 변화나, 치료군간의 유의미한 차이는 관찰되지 않았습니다. Despite an improved fit, we found no evidence of a difference in trends in either model A 는 median MELD 를 Simple mixed-effects models showed no evidence of a trend in MELD over time (p=0·33 for all groups), nor of a difference between the treatment and control groups (p=0·55 for the G-CSF group vs the standard care group and p=0·75 for the G-CSF plus stem-cell infusion group vs the standard care group). The final, more flexible model for the second coprimary outcome included a change-point at day 30 and an interaction between treatment group and time to allow for different trends in each period (appendix p 7). Despite an improved fit, we found no evidence of a difference in trends in either model (figure 2). (A) (B) : Observed median change (C) (D) : Model fit based on predictions from a linear mixed-effects model incorporating restricted cubic splines

Results : Change in liver parameters from baseline to day 30 and day 90 No difference in change in MELD scores between days 0 and 90 in either of the treatment groups compared with the standard care group No differences in either composite or individual markers of liver dysfunction, fibrosis or QOL scores across any of the groups 다음은 각 군별로 치료 시작 후 30일, 90일째의 MELD, UK end stage liver disease score, child pugh score 등 간기능 지표와, lab, fibroscan 등의 non invasive hepatic biomarker 그리고 QoL score 을 나타낸 표입니다. 첫번재 primary outcome 인 90일째의 MELD score 변화는 모두 -0.5 점으로 세 군에서 동일하였으며, 이외 다른 검사들에서 각 치료군별 그리고 시기별 유의미한 변화는 없었습니다. We detected no differences in either composite or individual markers of liver dysfunction across any of the groups Median change in MELD from day 0 to 90 was −0·5 (IQR −1·5 to 1·1) in the standard care group, −0·5 (−1·7 to 0·5) in the G-CSF group, and −0·5 (−1·3 to 1·0) in the G-CSF plus stem-cell infusion group. We found no evidence of a difference in change in MELD scores between days 0 and 90 in either of the treatment groups compared with the standard care group (p=0·72 for G-CSF vs standard care and p=0·90 for the G-CSF plus stem-cell infusion group vsstandard care). Waterfall plots of day 90 change in MELD score and best change in MELD score are presented in the appendix (pp 15, 16).

Results Co-primary outcomes Unchanged when adjusted for participant characteristics (cause of liver disease, alcohol consumption history) MELD and UKELD scores at days 180 and 360 showed no difference between groups Neither treatment group had a reduction in mortality or admissions to hospital compared with the control Co-primary outcome 을 LC etiology 나 alcohol 섭취력에 따라 다시 adjust 하여도 치료군별 유의한 변화는 없었습니다. 또한 180일째와 360일째의 각 군별 MELD, UK ELD score 도 유의한 변화는 없었으며, 사망이나 입원의 감소에 있어서도 두 치료균에서 standart 치료와 비교하여 유의한 변화는 없었습니다. Change in MELD score was not affected by either previous consumption of alcohol (appendix p 10) or length of abstinence from alcohol (appendix p 11). We also evaluated MELD and UKELD scores at days 180 and 360 (appendix p 17), we detected no differences between groups.

Results : Adverse events 3 [11%] of 28 patients 3 [11%] of 28 patients 12 [43%] of 28 patients 치료 후 1년동안 Adverse event 를 비교하였을 때, 왼쪽이 standard care, 가운데가 G-CSF only, 오른쪽이 G-CSF + 조혈모세포이식 군인데, sepsis 나 encephalopathy, bleeding, hepatic decompensation 등의 serious adverse event 를 비교하였을 때, 각각 11%, 11%, 43% 로 조혈모세포이식 군에서 유의하게 심각한 부작용이 많이 발생하였습니다. HCC 가 발생한 사람은 없었습니다. Three patients died during the study, including one in the standard care group (variceal bleed) and two in the G-CSF plus stem-cell infusion group (one myocardial infarction and one progressive liver disease). Neither treatment group had a reduction in mortality or admissions to hospital compared with the control. Patients in both treatment groups had to be considered for liver transplantation, and patients in the G-CSF plus stem-cell infusion group had more serious adverse events (12 [43%] of 28 patients) than patients in the G-CSF group (three [11%] of 28 patients) and the standard care group (three [12%] of 26 patients; table 4). At the time of adjudication, none of the serious adverse events were judged to be related to treatment. The frequency of adverse events reported by CTCAE category and grade was similar between groups, except for nervous system disorders and musculoskeletal disorders, which were more common in the treatment groups than in the standard care group (appendix p 18). Patients were monitored for a year, including standard screening for hepatocellular carcinoma every 6 months and no malignancies were detected. We also followed up patients for outcomes such as death, assessment for liver transplantation, and serious adverse events up to 1 year after randomisation (table 4). Three patients died during the study, including one in the standard care group (variceal bleed) and two in the G-CSF plus stem-cell infusion group (one myocardial infarction and one progressive liver disease). Patients in the G-CSF plus stem-cell infusion group had more serious adverse events

Discussion G-CSF with or without hematopoietic stem-cell infusion did not improve liver dysfunction or fibrosis and might be associated with increased frequency of adverse events. MELD Objective prognostic scoring system (mortality, decompensation) Magnitude and direction of change is a significant independent predictor of mortality and complications Compensated cirrhosis May have greater potential for regeneration and regression of fibrosis Fibrosis in decompensated cirrhosis would be more advanced and less likely to be reversible Discussion 입니다. Compensated LC 환자를 대상으로 standard care 를 시행한 것에 비하여 G-CSF 혹은 G-CSF 에 이어 자가조혈모세포 이식을 한 군에서 유의미한 간기능 호전은 관찰할 수 없었고, 오히라 심각한 부작용이 자가조혈모세포 이식 군에서 더 많이 발생했습니다. 연구에서 MELD 를 primary outcome 으로 설정한 이유는, compensated cirrhosis 환자를 대상으로 decompensation 을 예측할 수 있고, mortality 의 객관적인 prognostic scoring system 이며, 또한 MELD score 변화의 정도와 변화의 방향이 사망률과 LC 합병증의 predictor 로 작용하기 때문이었습니다. 이 연구의 특징이 환자 대상을 compensated cirrhosis 로 둔 것인데, decompensated LC 의 경우에 fibrosis 가 더 진행하고 reversibility 가 떨어질 것이라고 생각해서 liver regeneration 이나 fibrosis regression 등이 compensated LC 군에서 더 효과적으로 나타날 것이라고 가정했습니다. In this study, we found that neither G-CSF nor G-CSF plus stem-cell infusion showed improvement in MELD score in patients with compensated liver cirrhosis. These results contrast with the findings of previous smaller studies.8 We chose MELD score as the primary endpoint because it is an objective prognostic scoring system validated in large independent cohorts worldwide for patients with chronic liver disease. It independently predicts clinical decompensation in patients with compensated cirrhosis15 and is highly accurate in the prediction of 1-week, 3-month, and 1-year mortality. For any given MELD score, the magnitude and direction of change in score during the previous 30 days is a significant independent predictor of mortality, making change in score a more important determinant than initial MELD score alone. The magnitude of the change in MELD score can also predict the development of complications of cirrhosis such as ascites and variceal bleeding. We supplemented the use of MELD as a primary endpoint with many secondary endpoints of liver dysfunction, as well as markers of liver fibrosis (serum ELF and Fibroscan), none of which improved with therapy. We specifically chose to recruit patients with predominantly compensated cirrhosis, as opposed to those with more advanced decompensation, in the belief that the included patients had a greater potential for regeneration and regression of fibrosis. In this respect, a tangible effect of stem-cell therapy could have been to prevent disease progression that would require consideration for liver transplantation. It seems highly unlikely that the interventions in this study would be more effective in a cohort of patients with more advanced decompensated disease given that the fibrosis in such patients would be more advanced and less likely to be reversible.16  Gastroenterology. 2007;133:481–488. Semin Liver Dis. 2016;36:87–98

Discussion Strengths Largest and most rigorous RCT on stem-cell therapy for LC Inclusion of a stand-alone G-CSF group (G-CSF improved liver regeneration and function in preclinical studies) Defined stem-cell population Representative spread of causes of liver disease Well characterized patients within a narrow range of MELD scores 연구의 강점은, 현재까지 LC 환자를 대상으로 한 자가조혈모세포치료에 대한 연구중에 가장 대규모의 RCT 라는 점이고, 조혈모세포 이식 군 뿐만 아니라 G-CSF 단독 치료군을 설정한 것인데, 전임상연구를 통하여 조혈모세포를 따로 추출하여 이식하지 않아도 G-CSF 많으로도 간기능에 호전이 있다는 데이터가 있어서 임상에서도 그 효과가 있는지를 추가로 확인할 수가 있었습니다. 또한 정제된 stem cell population 을 규정하여 사용하였는데, 동물 모형에서 mixed bone marrow infusion 의 경우 오히려 liver fibrosis 를 일으키고, purified stem cell 의 경우 fibrosis 를 감소시킨다는 연구가 있어, 이점이 중요하다고 할 수 있겠습니다. 또한 환자들이 alcohol, viral, non alcoholic fatty liver disease 등 다양한 기저간질환을 갖고 있어 대표성이 있었고, 비교적 좁은 범위의 MELD score 에 드는 잘 characterize 된 환자군이었습니다. viral This study has several strengths. It is, to our knowledge, the largest and most rigorous randomised controlled trial on stem-cell therapy for cirrhosis of the liver thus far, with robust clinical trial governance and sufficient power to detect a clinically meaningful difference. The trial also included a representative spread of causes of liver disease, between which we saw no differences in response to the interventions (appendix p 12). Moreover, the patients were well characterised within a narrow range of MELD scores (11·50–15·50) and with respect to potentially confounding effects such as the duration of abstinence from alcohol and concomitant use of antiviral therapy. The study also used a defined stem-cell population, which is important for regulatory and scientific reasons; in animal models, mixed bone marrow infusions have been reported to contribute to liver fibrosis,4 whereas purified HSCs reduce fibrosis.6 CD133-positive cells represent a more enriched subpopulation of the CD34-positive cells, with true pluripotent stem cells constituting about 0·1% of a CD133-positive population.17 Moreover CD133-selected cells have previously been infused in patients with liver9 and cardiovascular disease,18 producing benefits and with no safety concerns. We chose the dose used in this study (0·2 × 106 cells per kg per infusion) because it was higher than that used in previous studies and we felt it to be achievable. Notably, in several patients, this amount of cells could not be mobilised, suggesting that our chosen number of cells might have been too high. CD133-positive doses of up to 2·5 × 106 cells per kg have been used during bone marrow transplantation, although this was in the setting of bone marrow malignancy and chemotherapy. Another strength of this study was the inclusion of a stand-alone G-CSF group. Results from a dose-finding study have showed that G-CSF at a dose of 15 μg/kg per day for 5 days was safe in patients with cirrhosis and was deemed to be the optimum dose for effective stem-cell mobilisation without adverse effects.19 This dose is higher than the routinely used dose of 10 μg/kg per day because patients with cirrhosis have a poorer mobilisation capacity than other patients, which is possibly related to splenic sequestration and poor bone marrow function. This dose was well tolerated by patients and no adverse effects were recorded. Another previous study that used the standard dosage showed a reversible increase in spleen size.20 While G-CSF is used to mobilise cells for leukapheresis, it has also been reported to improve liver regeneration and function in preclinical studies21 and to improve mortality in a clinical trial of acute-on-chronic liver failure in India, possibly through a reduction in deaths from sepsis.22 Although differing in dose and clinical target from the study by Garg and colleagues,22 our study found no evidence of any benefit for the use of G-CSF in terms of liver fibrosis or function, or indeed sepsis, as has previously been reported. Gastroenterology. 2017;153:233–248. Gastroenterology. 2012;142:505–512. Gastroenterology. 2007;133:619–631.

Discussion Limitation Absence of a true placebo and masking of treatment allocation (overt side-effects of G-CSF and ethical challenge) Absence of a histological endpoint (macrohistological nature, clinical risk) This study challenges the findings of other reports that show bone marrow cell therapy plus G-CSF to be effective for improving liver function in LC. Such therapies might even increase patient morbidity. 연구의 한계로는, G-CSF 투여와 leukapheresis 등의 과정이 있어서 placebo 의 설정과 allocation masking 이 어려웠다는 점이 있고, 이 연구에서는 biopsy 를 하지 않아 조직학적 평가가 어려웠습니다. 결론적으로 이 연구를 통해 compensated LC 에서 G-CSF 나 조혈모세포 이식이 간기능 호전에는 효과가 없으며, 오히려 부작용을 증가시킨다는 것을 알 수 있었고, 만성 간질환에서 자가조혈모세포 이식이 간기능을 호전시킨다는 이전 여러 연구 결과들을 반박하는 결과였습니다. A potential limitation of the study could be the absence of a true placebo and masking of treatment allocation, although in the absence of an effect this point is less pertinent. Given the overt side-effects of G-CSF (eg, bone pain), a placebo would be readily unmasked, whereas sham leukapheresis would pose ethical challenges. Another limitation is the absence of a histological endpoint. Although histological analysis to assess the effect on fibrosis and hepatocyte proliferation might have been revealing, we considered this to be unlikely. Microscopic analysis of needle biopsy material in patients with cirrhosis can be misleading because of the macrohistological nature of the disease. For example, biopsy of regenerative nodules surrounded by bands of fibrosis can have the appearance of normal liver, and different stages of fibrosis can be seen when multiple biopsies are taken from the same liver at transplantation.23 Furthermore, in the setting of cirrhosis and coagulopathy, two biopsies within a short timeframe would constitute a substantial undertaking not without clinical risk; even transjugular liver biopsy in patients with cirrhosis has major complications (eg, perforation of the hepatic capsule, cholangitis, and intraperitoneal bleeding) in 1–3% of cases. Mortality related to the procedure varies from 0·2% to 0·3%.24 Indeed, it is not clear that regulatory approval would even be granted for such a study in western Europe or the USA. Finally, this trial was powered on a clinically relevant marker of liver dysfunction, for which a liver biopsy would provide no relevant information. It is possible that the absence of histological confirmation of cirrhosis resulted in patients without cirrhosis being included, although this is very unlikely given that impaired liver synthetic function was a key inclusion criteria. We decided not to track cells after infusion to confirm homing to the liver. Although these results would have been informative, the technical and regulatory barriers to non-invasively labelling and tracking such cells are substantial and there is a substantial risk that labelling of CD133-positive cells before their infusion could alter their viability and efficacy. While our published preclinical data6suggested that HSCs efficiently engraft the recipient liver at first, there is minimal medium-term engraftment, and, notably, the HSCs induce longer-lasting changes in endogenous inflammatory cells such as macrophages.6 Clinical data from the setting of alcoholic hepatitis also support this possibility.25 The absence of any effect of HSC infusions on liver fibrosis and function in this study differs from the positive effects reported by us and others in rodent models of liver fibrosis.6 This result comes despite the dose of cells and G-CSF being higher than that reported in other studies.8 Furthermore, given the failure to mobilise sufficient HSCs in all patients in this study, it would be difficult to aim for an even higher infused cell dose. Thus, this result is likely to be a true reflection of the lack of action of HSCs in the setting of advanced liver fibrosis and suggests caution when extrapolating the efficacy of any putative antifibrotic or proregenerative therapy in rodent models of fibrosis to the clinical setting. Fibrosis is qualitatively and quantitatively different in rodent models, with rodent fibrosis having less collagen cross-linking and more spontaneous remodelling than is typically seen in humans.26, 27 Studies in rodents also often use freshly isolated cells derived from healthy animals, thus in this study, the use of autologous frozen cells might be relevant, although frozen cells are commonly used in bone marrow transplantation to good effect. It is possible that the peripheral venous administration of the cells resulted in few cells reaching the liver, and in theory, direct administration of cells into the vessels supplying the liver (the portal vein and hepatic artery) might be more effective. However, this approach is invasive and carries substantial risks, including bleeding and thrombosis. Although such approaches were initially reported as being used without complications, a later study28 showed that direct infusions can lead to portal hypertensive bleeding following cell injection and another study29 was terminated early following complications in two of four patients after hepatic artery injection.29 Moreover, the scientific literature does not support any advantage of directed infusion of cells to the liver over peripheral infusion.30 Our trial found no evidence to support any benefit from G-CSF alone or G-CSF plus stem-cell infusion in liver cirrhosis. Given the often intense enthusiasm surrounding stem-cell therapy, despite potential safety concerns,31 this important negative result reinforces the need for rigorous clinical trials to test efficacy in the field of stem cell-based therapy. This study also provides a template for future studies of novel anti-fibrotic therapies in patients with liver cirrhosis. J Clin Pathol. 1996;49:556–559. Gastroenterology. 2017;153:233–248.

Granulocyte colony-stimulating factor and autologous CD133-positive stem-cell therapy in liver cirrhosis (REALISTIC): an open-label, randomized, controlled phase 2 trial P.N. Newsome, et al. Lancet Gastroenterol Hepatol 2018;3(1):25-36. Long-term Outcomes of Autologous Peripheral Blood Stem Cell Transplantation in Patients With Cirrhosis Changcun Guo, et al. Clin Gastroenterol and Hepatol 2019;17(6):1175-1182. 다음으로 부논문입니다. 2019 년 clinical gastroenterology hepatology 에 발표된 논문으로, decompensated LC 를 대상으로 자가조혈모세포 이식을 했을때의 long term outcome 에 대한 연구입니다.

Background & Aims Long-term effects of stem cell treatments on survival and risk of hepatocellular carcinoma (HCC) in patients with cirrhosis have not been determined. Risk of HCC even at compensated or early stage in cirrhotic patients. Investigate the long-term effects of autologous stem cell  transplantation and risk of HCC in patients with cirrhosis. 간경변 환자를 대상으로 한 자가 조혈모세포 이식에 대한 앞선 연구는 대부분 관찰 기간이 길지 않으며 long term survival benefit 에 대한 연구 및 HCC 발생과 관련된 연구는 아직 제대로 밝혀지지 않았습니다. 또한 대상성 및 초기 LC 에서도 HCC 가 발생할 수 있기 때문에 HCC risk 에 대한 평가도 필요할 것으로 생각되며, 그래서 이 연구에서는 간경변 환자를 대상으로 조혈모세포 이식을 한 뒤 장기간 f/u 하는 대규모의 cohort study 를 통해 long term effect 과 HCC risk 와의 관련성을 밝히는 것을 목표로 하였습니다. HCC 와의 연관성에 대해서도 PBSC transplantation is safe and well tolerated.21 However, few studies have reported patient survival as their primary end point, especially in cirrhotic patients. Data on long-term benefits and adverse effects of stem cell treatment in cirrhosis still are lacking.28 Given the risk of hepatocellular carcinoma (HCC) even at compensated or early stage in cirrhotic patients,29–31 it is important to address the long-term benefit as well as risk of stem cell treatment. We analyzed the outcomes of 282 cirrhosis patients treated with autologous PBSC transplantation and 286 controls over 10 years. This was a large cohort study with a long follow-up period of autologous PBSC transplantation in cirrhosis

Methods Retrospective cohort study January 1, 2006 ~ December 31, 2016 Military medical university in China Control group : Hospitalized and treated with standard medical care (antiviral therapy, complication management) Treatment group : Autologous PBSC transplantation 이 연구는 retrospective cohort study 이고, 2006년 1월부터 2016년 12월까지 10년동안 f/u 을 하였고, 중국의 한 국군병원에서 진행되었습니다. 환자를 control cohort은 입원하여 standard medical care 를 받는 환자를 대상으로 하였고, 앞선 연구와 같이 antiviral therapy 나 LC complication management 등 기본적인 치료를 하였습니다. Treatment corhort 은 standard treatment 와 더불어 자가조혈모세포 이식을 시행한 군이었습니다. . Study Population The study was approved by the Institutional Ethics Committee of Xijing Hospital, The Fourth Military Medical University. Two cohorts of decompensated cirrhosis patients were followed up from January 1, 2006, to December 31, 2016. The treatment group consisted of patients treated with autologous PBSC transplantation, and the control cohort comprised patients hospitalized and treated with standard medical care. All of the patients in the treatment group received only 1 PBSC transplantation and were treated in our Department of Gastroenterology at Xijing Hospital. The procedure protocol and reporting was standardized and strictly followed. Other clinical examinations and tests were performed according to the good clinical practice regulations of the hospital. Control patients were selected from the follow-up database. They were patients hospitalized during the same period, and only those who met the inclusion and exclusion criterial requirements of PBSC transplantation at baseline were selected as controls. Clinical data were retrieved from the case records or were collected during the follow-up visits

Methods Inclusion criteria Exclusion Criteria Age 20 – 70 Decompensated LC, Child-Pugh score of 7 or greater At least 6 months of antiviral treatment Exclusion Criteria HCC or other malignancy History of hepatorenal syndrome or recurrent variceal bleeding Hepatic encephalopathy, spontaneous peritonitis or variceal bleeding in the past 30 days Alcohol abuse Inclusion criteria 는 20세에서 70세 성인 중 decompensated LC 로 child pugh score 7점 이상이며 적어도 6개월 이상의 항바이러스제를 복용하는 경우로 설정하였습니다. 이 경우 HBV 는 nucleoside analogues 를 사용하였고, HCV 의 경우 중국에서는 direct-acting antivirals 을 아직 사용할 수 없기 때문에 interferon plus ribavirin 을 사용하였습니다. Exclusion criteria 는 HCC 나 다른 malignancy 가 있거나, hepatorenal syndrome 이나 recurrent varix bleeding 의 병력이 있는 경우 최근 1달 사이 hepatic encephalopathy, SBP, varix bleeding 이 발생한 경우 및 alcohol abuse 로 설정하였습니다. Diagnoses and Interventions We only included patients with decompensated cirrhosis in the analysis. Decompensated cirrhosis was diagnosed based on ultrasound or computerized tomography, upper digestive tract endoscopy, and patient history. Chronic hepatitis B or C was diagnosed according to patient history, serologic tests, and biochemical tests. Patients in both groups received standard medical treatment (SMT), which consisted of antiviral therapy (nucleoside analogues for hepatitis B virus or interferon plus ribavirin for hepatitis C virus) and management of complications such as ascites, variceal bleeding, hepatorenal syndromes, and hepatoencephalopathy according to updated guidelines.32–37 Because direct-acting antivirals are not available in China, only an interferon plus ribavirin regimen was used to treat hepatitis C virus. Anti–hepatitis B virus medications included entecavir, lamivudine, adefovir, telbivudine, tenofovir, or interferon. Autologous PBSC transplantation was approved by the Institutional Ethics Committee of Xijing Hospital, The Fourth Military Medical University. Written informed consent was obtained. The inclusion criteria included the following: age between 20 and 70 years, decompensated cirrhosis (hepatitis B or C), CTP score of 7 or greater, and at least 6 months of antiviral treatment before PBSC transplantation. The exclusion criteria included the following: hepatocellular carcinoma or other malignancy, history of hepatorenal syndrome or recurrent variceal bleeding, and a history of hepatic encephalopathy, spontaneous peritonitis or variceal bleeding in the past 30 days, and alcohol abuse. Patients in the autologous PBSC treatment group received G-CSF for 4 consecutive days (5–10 mg/kg/d, administered subcutaneously daily). Patients underwent ultrasound surveillance of the spleen during G-CSF mobilization. PBSCs were collected by leukapheresis using the COBER Spectra Apheresis System (Gambro BCT, Inc, Stockholm, Sweden) with a target of 107 to 108 /kg peripheral blood mononuclear cells (PBMCs). A total of 50 mL leukapheresis products (containing 107 –108 cells/kg body weight) were administered to patients via the hepatic artery. Flow cytometry was used to identify the percentages of CD34þ and CD14þ cells in the cell preparations after leukapheresis. Approximately 90% of the freshly isolated PBMCs were CD14þ, and 20% (18.9%–21.4%) of the cells were CD34þ. There was a minimum of 50 mL of freshly isolated PBMCs that contained 6 108 PBMC What You Need to Know Background Stem cells have been used to treat cirrhosis. However, long-term benefit of stem cell treatment and risk of hepatocyte carcinoma have not been reported. Findings Our retrospective cohort study lasted more than 10 years and included a large sample number. It suggested that stem cell transplantation improved the long-term survival of patients with decompensated cirrhosis without increasing risk of liver cancer. Implications for patient care Stem cell transplantation may be used as potential treatment for end-stage liver disease such as decompensated cirrhosis. 1176 Guo et al Clinical Gastroenterology and Hepatology Vol. 17, No. 6 cells. In total, more than 1 108 CD34þ cells were transferred via hepatic artery in each procedure. Patients were discharged after 5 days.

Methods Interventions G-CSF 5–10 μg/kg/day x 4 days Leukapheresis, 107 - 108 /kg peripheral blood mononuclear cells (PBMCs) 1 PBSC transplantation Hepatic artery Flow cytometry (CD34+ and CD14+ detection) 자가조혈모세포 치료를 받은 환자의 경우에는 G-CSF 를 5-10mcg/kg 로 4일동안 투여하고, leukapheresis 및 flow cytometry 를 통해 peripheral blood mononuclear cells (PBMCs) 를 채취하여 hepatic artery 로 1회 주입하는 방식으로 이루어졌습니다. 20% (18.9%–21.4%) of the cells were CD34þ 5×106 CD34+ cells/kg

Methods : Outcome measures Primary outcome Liver transplantation free survival (time interval between first hospitalization and death or liver transplantation) Secondary outcome HCC Effects on liver function (MELD, Child-Pugh score, Individual components (bilirubin, albumin, INR, Na, creatinine, PLT) Primary outcome 은 liver transplantation free survival 로, 첫 입원 이후 사망 혹은 liver transplantation 까지의 시간으로 설정하였으며 Secondary outcome 으로 HCC 의 발생, MELD나 child pugh score 등 간기능의 변화, serum bilirubin, albumin, INR, Na, Cr, PLT 등의 변화로 설정하였습니다. HCC Outcomes and Statistical Analysis All patients were followed up until death or liver transplantation. Mortality data were obtained by case records and confirmed by telephone calls. Survival time was calculated as the time interval between the first hospitalization and death or liver transplantation. was confirmed by computerized tomography or ultrasound. The chi-square test was used for comparing categoric variables. Means and medians were compared using the independent sample t test, the Fisher exact probability test, or the Wilcoxon rank-sum test. One-to-one matching were achieved by propensity score matching using the Matchit R package (http://www.jstatsoft.org/v42/i08/). Thirteen factors were included in the propensity score matching analysis including sex, age, etiology, alanine aminotransferase, aspartate aminotransferase, total bilirubin, albumin, international normalized ratio, serum sodium, creatinine, platelets, CTP score, and MELD score. Univariate and multivariate Cox’s proportional hazard regression analyses were performed to identify the prognostic significance of the demographic and clinical characteristics. Continuous variables were transformed into dichotomous categoric variables for regression analysis. Variables that were significant on univariate analysis were entered into multivariate regression. For survival analysis, survival curves were plotted using the Kaplan–Meier method and analyzed by log-rank test. All statistical analyses were performed using SPSS software (version 13.0; SPSS, Inc, Chicago, IL). A 2-sided P value less than .05 was considered significant.

Results 2 cohorts (Decompensated LC) January 1, 2006, through December 31, 2016 Propensity score matching to adjust the differences between the 2 groups 2006년부터 2016년까지 입원한 비대상성 간경변 환자를 대상으로 retrospective cohort study 를 하였고, 두 군간의 비교를 위하여 propensity score matching 을 하였습니다. 그래서 자가조혈모세포 이식을 한 288명과, standard treatment 를 한 290명을 비교하였습니다. 이중 각각 6명 4명이 LT 를 받았고 나머지 282명과 286명을 대상으로 166명씩 1:1 propensity score matching 을 하여 분석하였습니다. This study included a total of 282 patients in the PBSC group and 286 patients in the control group (Supplementary Figure 1).  Supplementary Figure 1. Flow chart of patient enrollment. Patients from the PBSC group and the SMT group who received a liver transplantation (LT) were excluded. PS, propensity score.

Results : Baseline characteristics Baseline Characteristics 입니다. 성별, 나이, HBV, HCV 등 etiology, LFT 와 PT INR 등의 lab, 그리고 child pugh score, MELD 를 비교하였고, 전체 코호트를 비교하였을 때는 치료군이 더 젊고, ALB 이 낮으며 PT INR prolongation 되어있고, MELD score 도 3점 가량 더 높았습니다. 이를 바탕으로 Propensity score matched cohort 를 만들었고, matched cohort 에서는 유의한 차이가 없었습니다. This study included a total of 282 patients in the PBSC group and 286 patients in the control group (Supplementary Figure 1). The baseline characteristics of the 2 cohorts are listed in Table 1. The control patients were older than the PBSC group (50.64 vs 48.24 y; P = .005). Patients in the PBSC group had poorer liver function as indicated by albumin levels, international normalized ratio, and MELD scores. The bias was addressed by propensity score matching, which resulted in 166 patients in each group with comparable baseline characteristics (Table 1).

Results : Liver Transplantation-Free Survival Adjusted 5-year survival rate : 71.2% (PBSC) v.s. 52.1% (control) Estimated mean survival : 94.2 ± 6.01 months (PBSC) v.s. 71.2 ± 6.30 months (control) Overall mortality : 28.3% (PBSC) v.s. 30.7% (control) Propensity score–matched cohort. Entire cohort. Primary outcome 입니다. Propensity score matched cohort 에서 adjusted 5 year survival rate 은 자가조혈모세포 이식 군에서 71.2%, control 에서 52.1%로 유의하게 치료군이 우월하였습니다. Estimated Mean survival 은 치료군과 대조군이 각각 94.1개월, 71.2개월이고 overall mortality 도 각각 28.3%, 30.7% 로 치료군이 우월하였습니다. Kaplan-meier 생존곡선을 보면, 1,2 년때는 유의미한 차이가 없으나 3년째에 control group 에서 급격히 떨어지며 전체 코호트를 대상으로 한 분석에서도 비슷한 결과였습니다. Liver Transplantation-Free Survival Adjusted and unadjusted survival curves are shown in Figure 1. The Kaplan–Meier survival curve showed a steep decline after 36 months for the control group, while the curve for the PBSC group showed a linear pattern (log-rank test, P = .001) (Figure 1A). There was no significant difference in adjusted survival between the groups at 1 year (94.8% in the control group, compared with 95.1% for PBSC group) and at 2 years (86.5% vs 86.2%). The adjusted 5-year survival rate was 71.2% in the PBSC group and 52.1% in the control group. Patients receiving PBSC transplantation had significantly better long-term survival with an estimated mean survival of 94.2 ± 6.01 months for the PBSC group and 71.2 ± 6.30 months for SMT. In the matched cohorts, the overall mortality rate was 28.3% (47 of 166) for PBSC patients and 30.7% (51 of 166) for control patients (P = .718). In the whole study population, PBSC treatment also showed benefit on long-term survival, as shown by the log-rank test (P = .007) (Figure 1B).

Results : Prognostic factors for transplantation-free survival PBSC treatment was found to be an independent beneficial factor for patient survival Predictive factors for poor survival in propensity score–matched patients : age 50 years and older, CTP class C, and hyponatremia 이와같은 결과를 바탕으로, survival 에 영향을 미치는 Prognostic Factors 를 분석하였습니다. Entire cohort, 그리고 propensity score matched cohort 를 대상으로, 각각 univariate, multivariate regression analysis 를 하였고, 성별, 나이, Etiology, child pugh class, MELD score, lab 결과, 마지막으로 조혈모세포 이식 여부에 따라 분석을 하였습니다. 먼저 조혈모세포이식은 생존에 영향을 주는 변수로 확인되었고, propensity score matched cohort 에서 multivariate analysis 를 하였을 때 HR 0.576 으로 유의하게 생존에 영향을 미쳤습니다. 한편 50세 이상의 나이, child pugh class C 이상, hyponatremia 137 이하도 poor survival 의 독립적 예측인자였습니다. The mean time from the first hospitalization to December 31, 2016, for the PBSC patients was 68.7 ± 26.7 months, and for the control patients was 57.2 ± 24.1 months. Factors associated with patient survival are listed in Table 2. By multivariate regression analysis, PBSC treatment was found to be an independent beneficial factor for patient survival (hazard ratio, 0.576; 95% CI, 0.374–0.887; P < .012). Risk factors including age 50 years and older, CTP class C, and hyponatremia were independent predictive factors for poor survival in propensity score–matched patients. In the whole study population, PBSC also was associated with longer survival. To identify factors that affect survival of PBSC patients, multivariate regression also was performed. Age older than 50 years, CTP class C, and hyponatremia were independent factors associated with poor survival (Supplementary Table 1).

Results : Hepatocellular Carcinoma Prevalence and Causes of Death HCC incidence : PBSC (21.1%) ≈ Control (20.4%) Causes of death in the study cohorts HCC : PBSC ≈ Control Cirrhosis complication : PBSC > Control Liver failure : PBSC < Control PBSC Control Control PBSC Control PBSC HCC 의 발생과, 환자의 사망 원인에 대해 분석하였을 때 HCC prevelance 는 propensity score matched cohort 에서 치료군이 21.1%, 대조군이 20.4% 이며 유의한 차이는 없었습니다. 또한 오른쪽 그래프는 HCC 발생까지의 시간을 하얀색으로 보이는 대조군과 회색으로 보이는 치료군에서 비교한 표인데, 대조군이 HCC 발생까지의 평균 시간이 조금 더 짧은 것으로 보이나 유의미한 차이는 아니었습니다. 왼쪽 그래프는 사망 원인에 대해 분석한 표로, 왼쪽이 전체코호트, 오른쪽이 propensity matched cohort 이고, 그중에서도 왼쪽이 control, 오른쪽이 자가조혈모세포 이식 군입니다. 제일 아래는 HCC, 회색은 LC complication, 흰색은 liver fialure 에 의한 사망인데, 전체적으로 HCC 에 의한 사망은 두 군이 비슷하였고, control 에서는 liver failure 에 의한 사망이, 조혈모세포 이식 군에서는 LC complication 에 의한 사망이 더 많았습니다. Hepatocellular Carcinoma Prevalence and Causes of Death The prevalence of HCC between the treatment and control groups was comparable in the entire cohort (18.1% vs 19.2%; P = .748) and in the matched cohort (21.1% vs 20.4%; P = .999). Three patients in the control group developed cancer other than HCC: 1 patient had bladder cancer, 1 patient had lymphoma, and another patient had breast cancer. No malignancies other than HCC were documented in the PBSC treatment group. Causes of deaths are shown in Figure 2 and Supplementary Table 2. HCC remained a major cause of death in both cohorts, and there was no difference in deaths resulting from HCC between the 2 groups. More patients in the PBSC group died of cirrhosis complications than controls (48% vs 30% for the entire cohort; P = .007; 45% vs 12% for the matched cohort; P < .001, respectively). However, more control patients died of liver failure (20% vs 10% for the entire cohort; P = .012; 41% vs 11% for the matched cohort; P = .001, respectively). Other causes of death in the control patients were as follows: 5 patients died from non–liver-related incidences such as myocardial infarction (1 patient), a fall injury (1 patient), bladder cancer (1 patient), lymphoma (1 patient), and breast cancer (1 patient). In the PBSC patients, 3 patients died from non–liver-related incidences (myocardial infarction, cerebral hemorrhage, and car accident). Causes of death in the study cohorts Time to develop HCC

Results : Effects on Liver function Changes in MELD score in the entire study population over time compared with baseline 간기능에 미치는 변화를 MELD score 변화를 통해 확인한 결과입니다. 검정색이 조혈모세포 이식군, 회색이 대조군이며, 가장 왼쪽이 1달때, 그리고 가장 오른쪽이 60개월째의 MELD 의 변화입니다. 결과를 보면, 1,3,6 개월에는 유의마한 차이는 없으나, 1년째부터 1,2,3,5년째에 두 군간에 MELD 변화가 유의하게 차이가 납니다. Effects on Liver function Overall, patients receiving SMT had only moderate MELD score improvement when compared with patients in the PBSC treatment group (Figure 3). A significantly greater reduction in MELD score was observed at 1 year, 2 years, 3 years, and 5 years after PBSC transplantation. A significantly greater reduction in MELD score was observed at 1 year, 2 years, 3 years, and 5 years after PBSC transplantation

During the recent years stem cells have been used to treat various disorders. Studies in early 2000s suggested that multipotent stem cells may exist among bone marrow (BM) cells; and bone marrow stem cells contribute to liver regeneration after injury.1,2 Subsequent studies demonstrated that cell fusion rather than differentiation was the principal source of bone marrow-derived hepatocytes.3,4 These observations led investigators to examine the therapeutic potentials of BM-stem cells in liver disorders. Several animal studies suggest that transplantation of BM-stem cells reduces liver fibrosis.5,6 Subsequent uncontrolled clinical trials suggested that infusion of hematopietic CD34+ stem cells,7-10 BM-mononuclear cells,11,12 or BM-mesenchymal stem cells13,14 may transiently improve liver function in cirrhosis in human. Malignant transformation is a potential concern associated with cell based therapies. Teratoma has been reported after transplantation of pleuripotent embryonic stem cells in nude or immunocompetent mice.15,16Teratoma has also been observed after transplantation of induced pleuripotent stem (iPS) cells.17 iPS cells can also lead to malignant tumors in transplanted mice. This has been attributed to induction of c-Myc gene in iPS cells.17 Furthermore, there is some evidence suggesting that MSCs could immortalize and transform spontaneously after long term in-vitro expansion.18,19 Theoretically, malignancy is more likely to develop after cell therapy at the site of cellular engraftment. For instance, there is a report of donor-derived brain tumor development after intracerebellar injection of fetal stem cells.20 Therefore, there is a need to assess the potential risk of HCC after cell-based therapies in patients with cirrhosis. Discussion Stem cell transplantation improved the long-term survival of patients with decompensated cirrhosis without increasing risk of liver cancer. Long term follow-up, Large cohort Survival rates were higher than reported previously Stem cell transplantation and HCC Potential concern for malignant transformation Stem cell transplantation may improve the immunologic status and microenvironment of target tissues potentially may delay HCC development.   Discussion 입니다. Decompensated LC 에서 자가조혈모세포이식은 long term survival 을 늘렸고, HCC 의 risk 를 올리지 않았습니다. 이 연구는 대규모의 cohort 로 long term f/u 을 통해 survival gain 을 분석한 연구로 그 의미가 있습니다. 생존율의 경우 초기 2년까지는 유의한 차이가 없었으나 3년 이후에 차이를 보였는데, decompensated LC 의 경우 예전 연구들에서 초기 1,2 년째 생존율을 각각 60, 50 % 로 보고하였는데, 이 연구에서는 이전 다른 연구들에 비해서는 초기 생존율이 좋았고, 아마도 항바이러스제 등에 이해 이전에 비해 생존율이 높게 확인되었다고 평가하였습니다. 한편 이 연구는 HCC 발생에 관해 분석하였는데, 간경변의 경우 초기 단계에서도 HCC 가 발생할 수 있어서 HCC 발생에 예방적 효과가 있는지에 대한 정보가 중요할 것으로 생각되고, 이와 별개로 stem cell 에 대한 다양한 연구들에서 stem cell infusion 이 teratoma 를 발생시키거나, induced pleuripotent stem cell 이 mouse 에서 malignant tumor 로 변하는 경우가 보고되어서 자가조혈모세포 이식의 malignant potential 에 대한 평가도 필요할 것으로 생각됩니다. 이 연구에서는 HCC 의 발생에 차이는 없었는데, HCC 의 발생까지의 시간이, 비록 유의하지는 않지만, 치료군에서 조금 더 긴 시간이 소요된 것으로 보아서, 자가조혈모세포 이식이 target tissue 의 immunologic status 나 microinvironment 를 증진시켜 상대적으로 HCC발생을 delay 하는 효과가 있을 가능성을 생각해 볼 수 있겠습니다. 이에 대해서는 추가 연구가 필요하겠습니다. decompensated LC 의 경우 각각 60, 50% 연은 Studies in animal models have shown that stem cell treatment can improve survival from liver injuries.38, 39 This benefit also was reported in PBSC studies that included patients with liver failure or severe alcoholic hepatitis.1, 9, 10, 13 Few studies on patients with cirrhosis reported survival outcome. Our data indicate that patients treated with PBSC transplantation had a significantly better long-term outcome, although 1-year and 2-year survival showed no difference. The 1-year and 2-year survival of decompensated cirrhosis patients was reported in a previous study to be 60% and 50%, respectively.40 With advancements in antiviral drugs, patients with cirrhosis may have better survival than in the past.41 As was the case with our control cohort, whose survival rates were higher than reported previously. Stem cell treatment has been shown to improve liver function. However, the improvement of liver function may not translate into immediate benefit on survival. In this study, we showed sustained improvement of liver function after PBSC treatment and both the survival curve and Cox hazard regression confirmed the survival benefit of PBSC. Only patients with decompensated cirrhosis were included in the study. A recent randomized controlled study showed that stem cell infusiondid not improve liver dysfunction or fibrosis.42 This study included only compensated cirrhosis patients. Another double-blind, randomized, controlled study found that bone marrow–derived stem cells improved MELD score in decompensated cirrhosis patients.43 The inconsistency of outcomes of stem cell treatment on cirrhosis may be explained by the difference in study population, type of stem cell, or administration methods. Stem cell treatment has been shown to be quite safe for liver disease. The most predominant adverse event of the G-CSF mobilization and intra-arterial PBSC transplantation was fever, bleeding and pain at the puncture site, malaise, and back pain/body ache (data not shown). These adverse events were only mild to moderate and did not require medical intervention. Patients with cirrhosis are at risk of developing HCC even at a compensated or early stage.29, 30, 31 Our data did not find a difference in HCC prevalence between SMT and PBSC treatment, suggesting that stem cell treatment may not increase HCC risk. Stem cell transplantation may improve the immunologic status and microenvironment of target tissues44 and therefore potentially may delay HCC development. Further studies are necessary to define the effects of stem cell treatment on HCC development. Other than HCC, complications of cirrhosis were the main cause of mortality in PBSC transplantation patients, whereas liver failure was the major cause of mortality in patients who received SMT. Hence, HCC and cirrhosis complications should be monitored closely and treated actively in PBSC transplantation patients. Expert Rev Clin Immunol, 10 (2014), pp. 1185-1196 Middle East J Dig Dis. 2012 Jul;4(3):145-9

Discussion Limitations Retrospective study, lack of randomization, and missing clinical data (antiviral treatments, virologic response, serologic response rate) PBSC transplantation is a safe treatment regimen for patients with cirrhosis and can improve long-term survival without increasing the risks of HCC 연구의 한계는 retrospective study 였고 randomizaiton 이 되지 않았다는 점을 들 수 있고, 항바이러스 치료를 구체적으로 어떻게 진행하였는지, virologic respose 나 serologic response rate 가 어땠는지에 대한 정보가 부족한데 이러한 차이가 환자 치료 반응에 차이를 일읔였을 수 있습니다. 그리고 비록 propensity score matching 을 하였지만 이러한 이휴로 bias 가 발생할 수 있어 결국 long term survival 에 관해서도 RCT 가 필요할 것으로 생각됩니다. 이 연구를 통해서 decompensated LC 에서 자가조혈모세포 이식이 long term survival 을 늘리고, HCC 발생에는 영향을 주지 않는 치료임을 확인할 수 있었습니다. The limitations of our study were its retrospective nature, lack of randomization, and missing clinical data such as antiviral treatments, virologic response, and serologic response rate, which also may have affected patient survival. Although we addressed the patient bias by propensity score matching, further prospective randomized controlled studies are needed to confirm the findings in our study. In conclusion, this study reported long-term follow-up findings in a large cohort of decompensated cirrhosis patients. Based on the available evidence from the literature and the findings of our study, PBSC transplantation is a safe treatment regimen for patients with cirrhosis and can improve long-term survival without increasing the risks of HCC.

Critical Appraisal Inconsistency of outcomes Studying the Fate of Infused BMSCs Do BMSCs transdifferentiate into functional hepatocytes, fuse with pre-existing hepatocytes, or remain unchanged and release angiogenic factors and/or antifibrotic factors in the cirrhotic liver? It is imperative that future studies attempt to map the fate of the infused cells and distinguish them from host cells so that we can clearly define the mechanism of action. In none of the clinical trials has colonization or engraftment of the transplanted cells been looked for in the recipient liver. There are several cell tracing techniques by which the fate of infused stem cells can be followed, each with its own advantages and disadvantages.80 Briefly, there are a number of different stem cell labeling techniques available, including bromodeoxyuridine, fluorescent dye, green fluorescent protein, and magnetic resonance imaging and isotope labeling techniques. Bromodeoxyuridine is a synthetic nucleoside that is readily incorporated into newly synthesized progeny of proliferating cells. Immunohistochemical markers that identify bromodeoxyuridine mark out daughter stem cells and their differentiating offspring. The technique has been criticized, however, because of poor sensitivity and specificity.81 There is a series of fluorescent dyes that can be used for cell tracking including CM-Dil, CFSE, Hoechst 33342, DAPI, and PKH26. Many of these dyes carry the caveat of producing significant cellular toxicity after cell exposure. PKH26, however, is a fluorescent marker that stains the stem cell membrane and carries the lowest levels of cellular toxicity. This marker has been used to track hematopoietic stem cell migration, engraftment, and transdifferentiation in murine models.23, 82, 83 An additional limitation of fluorescent dyes (eg, CM-Dil, PKH26) is the gradual loss of the dye by leakage from the cell and by reduction during cell division. This method of tracking cells is therefore only useful for very short studies. Green fluorescent protein is a fluorophore used in the tracking of stem cells. Labeled cells absorb blue light and emit green fluorescence, easily facilitating their identification after transplantation. Magnetic resonance imaging has been used to trace the precise location of transplanted stem cells labeled with tracers such as gadolinium and ferric oxidewith a high degree of success.84, 85 Although single-photon emission computed tomography has been used to trace isotope-labeled stem cells, this technique is still in evolution and is not used commonly.86 The dynamic in vivo distribution of bone marrow–derived mesenchymal stem cells has been shown in a rodent model after cell labeling with 111In-oxine. After intra-arterial, intravenous, and intraperitoneal infusions, cell migration was mapped using a gamma camera.87 A detailed understanding of the mechanisms that regulate the migration of bone marrow–derived stem cells to injured liver will be required if we are to optimally manipulate the contribution of these cells to liver repair. This area is beyond the scope of this review, and is covered elsewhere88 (Figure 2). Critical Appraisal Inconsistency of outcomes Difference in study population, type of stem cell, or administration methods Tracking cells to confirm homing to liver None of the clinical trials has colonization or engraftment of the transplanted cells been looked for in the recipient liver. Stem cell labeling techniques (bromodeoxyuridine, fluorescent dye, green fluorescent protein, MRI and isotope labeling techniques.) Concerns: poor sensitivity and specificity, cellular toxicity, loss of dye (leakage and cell division) Histological analysis Fibrosis in decompensated cirrhosis would be more advanced and less likely to be reversible (?) Difficulty of tracking cell homing Critical appraisal 입니다. 지금까지 간경변에 대한 자가조혈모세포 이식에 대한 연구들에서 연구 결과에 일관성이 없었는데, 앞의 두 논문에서도 한 연구에서는 효과가 없는 것으로 확인되었고, 한 연구에서는 survival gain 이 있는 것으로 확인되었습니다. 주논문은 compensated LC 를 대상으로 하였고 MELD 의 범위도 좁았으며, peripheral line 으로 여러 번 나눠서 주입한 반면 부논문은 decompensated LC 를 대상으로 hepatic artery 로 1회에 주입하는 등 study population 과 stem cell 추출 과정 및 주입 방법등에 차이가 나서 발생한 결과일 수 있습니다. Stem cell labeling techniques 그리고 이전 두 논문 모두, 자가조혈모세포를 주입하는 과정은 있으나, 조혈모세포가 간에 잘 homing 이 되었는지에 대한 평가는 이루어지지 않았습니다. 주 논문의 경우 치료 효과가 없는 것으로 확인되었는데,간에 잘 homing 을 했음에도 불구하고 효과가 없었던 것인지, 혹은 간으로 cell 이 잘 homing 되지 않아 효과가 없었던 것인지를 평가할 수 없었습니다. 현재까지 간경변에서 자가조혈모세포 이식에 대한 연구에서 cell homing 을 구체적으로 평가한 연구는 없었습니다. Stem cell labelling 테크닉에는 bromodeoxyuridine, fluorescent dye, green fluorescent protein, MRI 및 isotope labeling technique 등이 있는데, 각각의 검사들이 민감도/특이도가 낮거나 약제가 cellular toxicity 가 있고, 시간이 지나고 세포가 분열함에 따라 dye 가 loss 되는 경우 등이 있어서 현실적으로 적용하기에 어려운 경우가 많은 것이 문제입니다 결국 자가조혈모세포가 liver 에 잘 도착해서 그 안에서 어떠한 작용을 하는지를 확인하기 위해서는 Cell homing 평가가 필수적일 것으로 생각합니다. 또한 두 연구에서 모두 조직학적 평가는 이루이지지 않았는데요, 조혈모세포 이식이 liver regeneration 및 섬유화에 어떠한 영향을 미치는지 확인하기 위해서는 조직검사가 golden standard 이고, 특히 첫번째 논문에서는 환자 대상을 compensated LC 로 잡은 이유가 초기 단계일수록 regeneration and regression of fibrosis 에 효과가 있을 것으로 생각해서인데, 그렇다면 조직검사를 통해 확인했어야 하는게 아닌지 의문이 듭니다. 또한 조직검사를 통해 homing 에 대한 평가도 할 수 있을것으로 생각됩니다. 종합해보면, 두 연구를 통해 compensated LC 및 decompensated LC 를 대상으로 각각 간기능 변화에 대한 RCT 와, long term survival 에 대한 대규모 retrospective cohort study 를 하였는데 상반된 결과가 확인되었습니다. 향후에는 조금더 대규모의 RCT 를 장기적으로 시행하고, 가능하다면 cell homing 에 대한 평가와 liver biopsy 를 시행하여 자가조혈모세포가 간에 어떠한 변화를 유발하는지 및 간기능과 생존율에 어떠한 변화를 가져오는지를 구체적으로 확인하여 간경변 환자의 자가조혈모세포 이식에 대한 명확한 근거가 마련될 필요가 있을 것으로 생각됩니다. Gastroenterology. 2008;135:438–450.

 Thank you