This is a randomized comparative trial between percutaneous liver biopsy and EUS FNB guided liver biopsy in patients with parenchymal liver disease.

Studies so far have been shown that EUS FNB guided liver biopsy is non inferior to percutaneous liver biopsy in terms of efficacy, provided several conditions are met (type of needle, wet suction, actuations).

În terms of safety, it may be that EUS FNB have an advantage, as needle diameter is slightly smaller, needle is inserted under better visualisation especially in patients with large subcutaneous fat tissue and the possibility of examining the needle tract and plugging it with needle content if needed ("the blood patch technique").

However, although severe, the incidence of liver bleeding is low, so a comparative trial with the hypothesis that EUS FNB has fewer liver bleedings than percutaneous techniques would have a very large sample size.

A recent study found out that abdominal pain at 2 hours after procedure is predictive for liver bleeding.

So we have design a randomized prospective trial assuming that EUS FNB guided liver biopsy has significantly less abdominal pain at 2 hours after procedure when compared to percutaneous route, using abdominal pain at 2 hours as a surrogate marker for risk of liver bleeding.

Why liver biopsy? Parenchymal liver diseases need sometimes diagnostic or staging assessment liver biopsy (eg. autoimmune hepatitis, primary biliary cholangitis, metabolic dysfunction associated fatty liver disease - MAFLD). Also, some focal liver lesions may need a diagnostic liver biopsy (eg. primitive liver tumors, liver metastases).

This can be performed percutaneously, trans-jugular and more recently trans-gastric or trans-duodenal under endoscopic ultrasound (EUS) guidance. The trans-jugular route was proposed for patients with coagulation disorders, as there is virtually no risk of bleeding, as the procedure is done within the supra-hepatic veins.

Currently, evidence exist that optimal results for EUS guided fine needle biopsy (FNB) are obtained by using a 19 Gauge Franseen FNB needle, by the heparin primed "wet suction" technique, with one or two passes (in both liver lobes), with at least 3 needle actuations [1,2,3,4]. Ensuring these conditions, the outcomes are non-inferior to percutaneous liver biopsy (length of the specimen, number of portal tracts, diagnostic accuracy).

So why choose EUS guided liver FNB? Two percutaneous liver biopsy adverse reactions are probably responsible for one's apprehension towards it - post procedure abdominal pain and the risk of liver bleeding. A recent meta-analysis [5] has found an incidence of mild abdominal pain at 12.9% and moderate-severe abdominal pain at 0.34%. The percentages for liver bleeding were 0.48% for major bleeding, 0.11% for major hematoma, 0.06% for hemoperitoneum and 0.01% for haemobilia (0.66% in total).

Bisonnette et al [6] have tried to find predictive factors for liver bleeding after percutaneous or trans-jugular liver biopsy. The only predictive factor was pain at 2h liver biopsy. This was more frequent following percutaneous than trans-jugular liver biopsy (33% vs. 15%; p < .001). Out of all patients with pain 2 h after liver biopsy, 15% had a liver biopsy-related bleeding, while out of those without pain 2 h after liver biopsy, 4% had a liver biopsy-related bleeding (p = .002). Pain 2 h after liver biopsy had a 55% (95% CI 0.35-0.73) sensitivity, 75% (95% CI 0.70-0.80) specificity, and 96% (95% CI 0.93-0.97) negative predictive value for liver biopsy -related bleeding.

EUS guided liver FNB seems to be safer than percutaneous liver biopsy. Eleven studies so far on 632 patients who had EUS FNB with 19G Franseen needle report on post procedure liver bleeding and abdominal pain [7-17]. Abdominal pain was present in 7.7%. of cases (28 of 362 reported patients), while liver bleeding was present in 0.16% (1 of 593, as "bile leak").

If we we're to design a study to prove that EUS guided liver FNB decreases bleeding rate from 0.66% for percutaneous route to 0.16%, with a type I error probability of 5% and a power of 80%, we would need to include 5126 patients, that is 2563 for each arm in a 1:1 allocation [7]. This is a near impossible study.

However, as Bisonnette et al [6] have proven, pain at 2h after liver biopsy is predictive for liver bleeding. So, we only need to prove that EUS guided liver FNB significantly decreases abdominal pain at 2h after procedure as compared to percutaneous route.

So, to prove that EUS guided liver FNB decreases abdominal pain rate at 2h after biopsy from 33% for percutaneous method to 7.7%, with a type I error probability of 5% and a power of 80%, we would need to include 78 patients, that is 39 for each arm in a 1:1 allocation [7].

Question of the research project We aim to answer the question whether EUS guided liver FNB has a significantly lower rate of abdominal pain at 2h after procedure than percutaneous liver biopsy.

Hypothesis The EUS guided liver FNB has a significantly lower rate of abdominal pain at 2h after procedure than percutaneous liver biopsy, which probably translates in a significantly lower liver bleeding rate.

Patients will be randomly allocated in one of the 2 study groups, following a randomization list (www.random.org), as it follows:

• Group one - percutaneous (P)
• Group two - EUS FNB (E)

Group one (P) - Percutaneous procedure will be performed under standard conditions, under transabdominal ultrasound guidance. After choosing the optimum needle trajectory, the skin site will be marked, standard skin sterilization with iodine and surgical draping of the site will be performed. Local anesthesia of the site with gradual lidocaine solution infiltration of the skin and underlying planes will be performed. Under ultrasound guidance, with collaboration of the patient for the respiratory movements, one or two passages with a 16G needle will be performed (BARD Max Core 16G, Becton Dickinson, USA). Specimens will be collected in formalin and send to pathology for processing.

Group two (E) - EUS guided liver FNB will be performed under deep sedation with propofol under anesthesiology monitoring. A 19G gauge Franseen needle (Acquire 19G, Boston Scientific, USA) will be used with the "wet suction" technique. Before inserting into the operating channel, the stylet will be removed, and the needle will be primed with heparin (500UI/5ml). After liver puncture, the 20ml aspiration syringe will be attached. Three needle actuations will be performed, one or two passes in the left liver lobe. At the end of the 3rd actuation, the syringe aspiration will be stopped, the needle will be left in place in the liver parenchyma for 2 to 3 minutes. The liver needle path will be examined with Doppler ultrasound to look for possible bleeding. If present, about 25% of the needle content will be pushed in the liver path, as previously described - the "blood patch" technique [19].

Inclusion Criteria:

• Diffuse parenchymal liver disease with indication for liver biopsy for diagnostic and/or staging purposes,
• Eligibility for deep sedation with propofol
• Age above 18 years old
• Informed consent

Exclusion Criteria:

• Coagulation disorders which cannot be corrected
• Anticoagulant and/or antiagregant medication which cannot be stopped temporarily.
• Contraindications for liver biopsy, to local lidocaine anesthesia and to deep propofol sedation
• Absence of written informed consent