Unravelling risk factors for delayed bleeding following ultrasonography-guided liver biopsy: a retrospective analysis

Introduction

Background

Despite remarkable advances in medical imaging and artificial intelligence (AI) (1), percutaneous liver biopsy remains the gold standard for definitively diagnosing and evaluating a wide spectrum of liver diseases (2-4). Although noninvasive technologies such as ultrasonography or magnetic resonance imaging (MRI) offer valuable insights, they face challenges in pinpointing the exact cause of abnormalities. Similarly, although constantly evolving, AI algorithms have limitations in handling complex and rare liver diseases (5,6).

Percutaneous liver biopsy was first performed in 1883 and has been used in clinical practice for over 150 years. Although generally safe, potential complications, such as pain, bleeding, pneumothorax, and infection can occur (7). Notably, the use of image-guided liver punctures has significantly reduced the incidence of haemorrhage, a common and potentially life-threatening complication. Current reports on liver biopsies indicate a haemorrhage rate of less than 2%, while biopsy-related mortality has similarly decreased to <1% (8). However, although rare, post-biopsy haemorrhage can still occur and may lead to mortality (9,10).

Rationale and knowledge gap

Previous studies have shown that most post-biopsy bleeding manifests within the first 24 h (11-13). This has led some healthcare providers to advocate for a 24-h observation period following the procedure. However, our clinical experience has revealed a trend of delayed post-biopsy bleeding coupled with an unremarkable initial follow-up. Haemorrhagic shock may occur after hospital discharge or cessation of close monitoring, with patients often identified only when presenting with significant bleeding symptoms. This poses significant challenges for both patients and physicians because delayed diagnosis can lead to life-threatening complications (9,10).

Early identification of patients at risk of delayed haemorrhage is crucial for prompt intervention and improved patient outcomes. By recognizing these high-risk individuals, timely measures such as local compression, haemostatic drugs, or radiofrequency ablation can be implemented to effectively address minor bleeding and prevent potentially serious consequences (14). Therefore, identifying the risk factors for delayed haemorrhage and extending post-operative monitoring are critical steps in mitigating its severity.

Objective

This study aimed to investigate the factors that increase the risk of delayed haemorrhage following liver biopsy by meticulously reviewing cases in our hospital. By analysing past patient data, we hoped to gain valuable insights into the characteristics and conditions that predispose individuals to delayed bleeding complications following liver biopsy. This knowledge can be used to develop improved risk stratification strategies and optimise post-biopsy monitoring protocols, ultimately leading to safer and more effective management of liver diseases. This article is written following the STROBE reporting checklist (available at https://tgh.amegroups.com/article/view/10.21037/tgh-24-79/rc).

Methods

The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by the ethics committee of The First Affiliated Hospital of Zhejiang University School of Medicine (Reg. No. 20240695A) and the requirement for informed consent was waived owing to the retrospective nature of the study. Data were collected from the electronic medical records (EMR) system of the hospital for patients who underwent ultrasonography-guided liver puncture biopsy between January 2022 and December 2023. The inclusion criteria were as follows: (I) patients hospitalised within our medical facility and (II) patients undergoing biopsy for non-focal liver lesions. Conversely, the exclusion criteria were as follows: (I) insufficient preoperative or postoperative laboratory information and (II) concurrent procedures on the same day with potential haemorrhagic risks.

EMR-retrieved patient records encompassed demographic characteristics such as age and sex, radiological observations including hydrothorax and ascites, and laboratory assessments including haemoglobin levels, platelet counts, and prothrombin time (PT). Further, postoperative complications were scrutinised from medical records, along with the corresponding treatments administered and subsequent pathological findings.

The liver puncture technique

Prior to liver biopsy, the patients underwent a minimum of 8 hours of fasting. Preoperative blood tests were used to assess complete blood count and coagulation functions. If the platelet count was <50×10⁹/L or the PT was >18.5 s, the liver puncture biopsy was aborted. Additionally, the medication history was reviewed to confirm the absence of medications that could affect coagulation. Written informed consent was obtained from all patients, detailing the procedure and potential risks and benefits.

Liver biopsy was performed by a qualified surgeon with the title of attending or above, assisted by a trained medical professional. Preoperative imaging studies were reviewed to assess the presence of hydrothorax or ascites and to identify any abnormal vasculature near the targeted biopsy site. Ultrasonography guidance was employed to visualise the liver and identify a safe puncture route; if there was a large amount of hydrothorax or ascites that left a gap between the liver tissue around the puncture site and the chest wall, the puncture biopsy was aborted. Local anaesthesia with lidocaine was administered at the puncture site to minimise discomfort. A 17-gauge coaxial needle was inserted into the liver under continuous ultrasonographic visualisation. Colour Doppler flow imaging (CDFI) was conducted to identify potential blood vessels within the needle trajectory, ensuring a safe puncture path. Upon confirming a safe puncture path, the assistant triggered the biopsy needle. Adjustments to the puncture angle were made based on a real-time assessment of the specimen quality. Typically, two to four passes were performed to obtain sufficient tissue samples. The liver tissue specimens were preserved in a fixative solution and promptly transported to the pathology department for further analysis.

Post-biopsy monitoring

Following liver biopsy, the patient was transferred to the ward for recovery. To minimise the risk of bleeding, strict bed rest was enforced for at least 6 h. During this initial period, vital signs and any potential symptoms, such as pain or abdominal distention, were closely monitored. An ultrasonographic examination was typically performed within the first 24 h post-biopsy to check if the patient was experiencing signs or symptoms of bleeding, including significant pain, abdominal distension, or abnormal vital signs. In the absence of these acute presentations of bleeding, a blood test, ultrasonography, or both were scheduled within the first 24 h post-biopsy for routine evaluation.

Hemorrhage classification

The present report used a definition of haemorrhage that was based on specific criteria. Haemorrhage was confirmed by either ultrasonography or other imaging modalities that demonstrated a haematoma near the puncture site, or the presence of blood in the hydrothorax or ascites, as confirmed by diagnostic paracentesis. Additionally, a reduction in haemoglobin levels exceeding 2 g/dL post-biopsy was deemed to be indicative of haemorrhage, and bleeding events were categorised into minor and serious bleeding groups. Serious bleeding complications were defined as grade 3 or higher, according to the Common Terminology Criteria for Adverse Events version 5.0.

Adverse event timing

Adverse events were categorised based on their time of occurrence relative to the biopsy procedure. Acute adverse events were defined as those occurring within the first 24 h after biopsy. Conversely, delayed adverse events manifested more than 24 h after liver biopsy.

Statistical analysis

Statistical analyses were conducted using SPSS software (version 26.0). Descriptive statistics, including age, sex, and laboratory values, were used to characterise patient demographics. Quantitative data with a normal distribution were presented as mean ± standard deviation (SD), whereas non-normally distributed data were presented as median and interquartile range. Categorical variables were presented as frequencies and percentages. Chi-square tests or Fisher’s exact tests were used to assess differences in categorical variables, while independent samples t-tests or Mann-Whitney U-tests were used to compare quantitative variables. Statistical significance was set at P<0.05.

Propensity score matching (PSM) analyses were performed to mitigate potential bias arising from baseline confounding variables due to sample size variations. PSM was conducted using a 1:1 matching algorithm without replacement. Age and sex were chosen as matching variables, and the caliper was set to 0.2 SD of the propensity score. This approach aimed to create comparable groups with similar baseline characteristics, thereby reducing bias in subsequent analyses.

Results

Patient selection and characteristics

Our study commenced with an initial screening of 642 patients who underwent ultrasonography-guided liver biopsy at our institution between January 2022 and December 2023. Following a meticulous review of medical records, we excluded 15 individuals. Eleven patients were excluded because of incomplete pre- or post-operative data that may have compromised the accuracy of our analysis. Additionally, four patients underwent other procedures on the same day as the liver biopsy that could have potentially increased their bleeding risk. These patients were excluded from the present study to focus solely on the specific effects of the liver biopsy.

Our stringent data screening process resulted in a final study population of 627 patients. Among them, 37.16% were men and 62.84% were women. Table 1 presents a detailed breakdown of the primary reasons for liver puncture in this cohort. As the table illustrates, post-liver transplantation was the most prevalent reason for the biopsy (45.29%), followed by autoimmune liver disease (23.44%) and drug-induced liver injury (19.30%). This distribution indicates that liver biopsies play an important role in clinical practice.

Hemorrhage outcomes

Of the 627 patients finally included in the study, 11 (1.75%) experienced bleeding complications after ultrasonography-guided liver biopsy. Fortunately, the majority of these events (namely, eight cases that represented 72.73% of all bleeding complications) were classified as minor bleeding events that typically have a minimal clinical impact and can often be managed conservatively.

The remaining three cases (27.27% of all bleeding complications) were categorised as serious. Although less frequent than minor events, serious bleeding events can pose a significant threat to patient health and may require more intensive intervention. All three serious cases were classified as delayed bleeding events occurring after the initial 24-hour post-biopsy observation period. This finding highlights the importance of extended monitoring or implementation of strategies to identify patients at a higher risk of delayed bleeding complications.

Analysis of baseline characteristics and risk factors

We conducted a comprehensive statistical analysis to identify the potential risk factors associated with post-biopsy bleeding complications. This analysis compared the baseline characteristics of patients who experienced bleeding after liver biopsy (bleeding group) with those of patients who did not experience bleeding (non-bleeding group). However, we did not observe any statistically significant differences in baseline characteristics between these two groups.

To account for potential confounding variables that may have influenced the initial comparison, we employed a statistical technique termed PSM. Following PSM, we extracted 11 patient pairs (1:1 matching), ensuring a balanced representation in both the bleeding and non-bleeding groups. Similar to the results of the unadjusted analysis, no statistically significant differences were found in the baseline characteristics when comparing the propensity-matched groups. Detailed baseline information for all the patients with and without bleeding is presented in Table 2.

Although we were unable to identify definitive risk factors for overall post-biopsy bleeding events, our analysis did reveal interesting observations regarding delayed bleeding. Notably, patients who experienced delayed post-biopsy bleeding exhibited significantly lower platelet counts than those who experienced early bleeding (P=0.047). Additionally, patients in the delayed bleeding group were more likely to present with pre-operative hydrothorax or ascites than those in the early bleeding group (P=0.047). These findings are presented in Table 3, and the details of the patients with delayed haemorrhage are shown in Table 4. Notably, another patient with mild bleeding exhibited a low platelet count (52×10⁹/L) and the presence of both hydrothorax and ascites. However, because of close monitoring, bleeding was detected early in this patient.

Treatment for delayed bleeding

Fortunately, all three cases of delayed bleeding were successfully managed. Two patients required minimally invasive surgical procedures (laparoscopy and thoracoscopy) to control the bleeding. The remaining patient was treated with blood transfusion and intravenous haemostatic medications, while another patient with a low platelet count and the presence of both hydrothorax and ascites successfully overcame the critical period after receiving haemostatic medication.

Discussion

Key findings

Liver puncture biopsy, particularly image-guided biopsy, is considered safe. However, haemorrhage remains a potential complication that can have serious consequences, even posing a threat to life. This study aimed to investigate factors influencing post-puncture haemorrhage in patients undergoing liver biopsy at our hospital.

A total of 627 patients who underwent liver puncture biopsy and completed post-procedure follow-up were included in this study. Our analysis revealed that the most common reasons for biopsy were post-liver transplantation autoimmune liver disease and drug-induced liver injury. The high proportion of post-liver transplant recipients in our patient cohort likely reflects the fact that our hospital unit is the leading liver transplantation centre in the province. Notably, the prevalence of hepatitis B in our study was significantly lower than that reported by Chang (15). This difference might be attributed to the emergence of new diagnostic techniques, such as serum biomarkers (16,17) and elastography (18,19) that enable noninvasive assessment of hepatitis B severity (20). Our findings highlight limitations in the non-invasive diagnosis of autoimmune liver disease and drug-induced liver injury (21).

Not all patients in our cohort underwent ultrasonography or other imaging techniques on post-operative day-1. This may have resulted in an underestimation of the actual number of bleeding events. However, given that these undetected minor haematomas are typically not life threatening, they did not significantly affect our main findings. Consistent with previous studies by Caldwell (22), our analysis did not reveal statistically significant differences in baseline characteristics between patients who experienced post-biopsy bleeding and those who did not. We reason that this result may be explained by the predominantly mild nature of the observed bleeding, often stemming from minor vessel damage during puncture that is generally unpredictable. Additionally, such bleeding is usually identified using ultrasonography during the procedure, allowing the operator to take immediate action to control the bleeding, including local pressure, gelatin sponge insertion through the coaxial needle, and even radiofrequency ablation. Patients were not discharged from the operating room until active bleeding was confirmed. Consequently, early bleeding caused by intraoperative vessel damage is often relatively mild. In contrast, delayed hemorrhage typically does not originate from vital vessel damage during the puncture itself; rather, it is often associated with underlying coagulation issues in these patients.

In our study 34 patients underwent a third biopsy and only one patient underwent a fourth biopsy, however no bleeding was observed in any of these patients with more than two punctures. This may be due to the fact that regardless of whether 2 or 4 punctures were performed, we checked rigorously for active bleeding intraoperatively, and when bleeding occurred, we managed it aggressively, so the number of punctures had no effect on whether or not there was postoperative bleeding in our study.

Platelets play a crucial role in blood clotting and preventing bleeding. Lower platelet counts compromise clotting ability, potentially leading to uncontrolled bleeding, even from minor injuries (23,24). This aligns with the findings of Potretzke, who reported an increased risk of post-puncture bleeding in patients with platelet counts <70×10⁹/L (25).

Strengths and limitations

Our analysis revealed a concerning association between delayed post-biopsy bleeding and the presence of pre-operative hydrothorax or ascites. This association can be attributed to several factors. First, small volumes of hydrothorax or ascites often arise secondarily to underlying conditions like liver cirrhosis. Liver cirrhosis, characterised by scarring and distorted liver architecture, disrupts the production of various clotting factors that are essential components of the blood coagulation cascade. These clotting factors play a crucial role in the formation of blood clots and in preventing excessive bleeding. When the production of these factors is impaired, such as in liver cirrhosis, the ability of the body to control bleeding is significantly compromised. Even minor injuries, such as those caused by the biopsy needle, can lead to prolonged or uncontrolled bleeding. Second, post-operative compression is a standard technique used to control bleeding after liver biopsy. This technique involves applying external pressure at the biopsy site to promote clot formation and prevent further blood loss. However, the presence of hydrothorax or ascites can act as a physical barrier, hindering the effectiveness of this compression. Finally, imaging techniques such as ultrasonography or computed tomography may not easily distinguish between a pleural or abdominal effusion that was present preoperatively and one that increased postoperatively as a result of haemorrhage, especially if the change in volume was minimal. This may lead to a delay in the diagnosis of post-biopsy bleeding complications, particularly if the initial clinical symptoms were mild.

This study acknowledges the limitations arising from its single-centre design and relatively small sample size. Future studies involving data collection from multiple centres and larger patient cohorts may provide more robust and generalisable conclusions.

Conclusions

Our study showed that when patients presented with a lower-than-normal platelet count in conjunction with the presence of hydrothorax or ascites, there was an increased risk of delayed haemorrhage. Therefore, we recommend that patients should be adequately informed of the high risk of haemorrhage before liver puncture biopsy and the period of postoperative monitoring should be extended.

Acknowledgments

Funding: None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://tgh.amegroups.com/article/view/10.21037/tgh-24-79/rc

Data Sharing Statement: Available at https://tgh.amegroups.com/article/view/10.21037/tgh-24-79/dss

Peer Review File: Available at https://tgh.amegroups.com/article/view/10.21037/tgh-24-79/prf

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tgh.amegroups.com/article/view/10.21037/tgh-24-79/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by the ethics committee of The First Affiliated Hospital of Zhejiang University School of Medicine (Reg. No. 20240695A) and the requirement for informed consent was waived owing to the retrospective nature of the study.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Buzzetti E, Lombardi R, De Luca L, et al. Noninvasive Assessment of Fibrosis in Patients with Nonalcoholic Fatty Liver Disease. Int J Endocrinol 2015;2015:343828. [Crossref] [PubMed]
2. Ehsan NA, Elsabaawy MM, Sweed DM, et al. Role of liver biopsy in management of liver diseases without hepatic nodules following end of the interferon era: experience of a tertiary referral center. Clin Exp Med 2023;23:97-105. [PubMed]
3. Lim TS, Kim JK. Is liver biopsy still useful in the era of non-invasive tests? Clin Mol Hepatol 2020;26:302-4. [Crossref] [PubMed]
4. Tapper EB, Parikh ND. Diagnosis and Management of Cirrhosis and Its Complications: A Review. JAMA 2023;329:1589-602. [Crossref] [PubMed]
5. Nam D, Chapiro J, Paradis V, et al. Artificial intelligence in liver diseases: Improving diagnostics, prognostics and response prediction. JHEP Rep 2022;4:100443. [Crossref] [PubMed]
6. Lu F, Meng Y, Song X, et al. Artificial Intelligence in Liver Diseases: Recent Advances. Adv Ther 2024;41:967-90. [Crossref] [PubMed]
7. Tian G, Kong D, Jiang T, et al. Complications After Percutaneous Ultrasound-Guided Liver Biopsy: A Systematic Review and Meta-analysis of a Population of More Than 12,000 Patients From 51 Cohort Studies. J Ultrasound Med 2020;39:1355-65. [Crossref] [PubMed]
8. Bissonnette J, Riescher-Tuczkiewicz A, Gigante E, et al. Predicting bleeding after liver biopsy using comprehensive clinical and laboratory investigations: A prospective analysis of 302 procedures. J Thromb Haemost 2022;20:2786-96. [Crossref] [PubMed]
9. Huang JY, Lu Q, Liu JB. Delayed hepatic rupture post ultrasound-guided percutaneous liver biopsy: A case report. Medicine (Baltimore) 2018;97:e9955. [Crossref] [PubMed]
10. Kato H, Kanno S, Ohtaki J, et al. A lethal case of massive hemorrhage after percutaneous liver biopsy in a patient with thrombasthenia. Leg Med (Tokyo) 2023;65:102315. [Crossref] [PubMed]
11. Dotan Y, Carlebach M, Zuckerman E, et al. Delayed Bleeding after Percutaneous Liver Biopsy. Eur J Case Rep Intern Med 2016;3:000359. [Crossref] [PubMed]
12. Garrido I, Coelho R, Macedo G. Hospital discharge after percutaneous liver biopsy - less is more? Eur J Gastroenterol Hepatol 2023;35:1186-91. [Crossref] [PubMed]
13. Neuberger J, Patel J, Caldwell H, et al. Guidelines on the use of liver biopsy in clinical practice from the British Society of Gastroenterology, the Royal College of Radiologists and the Royal College of Pathology. Gut 2020;69:1382-403. [Crossref] [PubMed]
14. Zhang L, Wu Y, He X, et al. Post biopsy delayed liver hemorrhage successfully controlled by bedside CEUS-guided microwave ablation: A case report. Clin Case Rep 2023;11:e6916. [Crossref] [PubMed]
15. Chang Y, Kim JI, Lee B, et al. Clinical application of ultrasonography-guided percutaneous liver biopsy and its safety over 18 years. Clin Mol Hepatol 2020;26:318-27. [Crossref] [PubMed]
16. Kramvis A, Chang KM, Dandri M, et al. A roadmap for serum biomarkers for hepatitis B virus: current status and future outlook. Nat Rev Gastroenterol Hepatol 2022;19:727-45. [Crossref] [PubMed]
17. Zhan Y, Tao Q, Lang Z, et al. Serum ribonucleotide reductase M2 is a potential biomarker for diagnosing and monitoring liver fibrosis in chronic hepatitis B patients. J Med Virol 2023;95:e29157. [Crossref] [PubMed]
18. Ginès P, Castera L, Lammert F, et al. Population screening for liver fibrosis: Toward early diagnosis and intervention for chronic liver diseases. Hepatology 2022;75:219-28. [Crossref] [PubMed]
19. Ye J, Wang W, Feng S, et al. Precise fibrosis staging with shear wave elastography in chronic hepatitis B depends on liver inflammation and steatosis. Hepatol Int 2020;14:190-201. [Crossref] [PubMed]
20. Bera C, Hamdan-Perez N, Patel K. Non-Invasive Assessment of Liver Fibrosis in Hepatitis B Patients. J Clin Med 2024;13:1046. [Crossref] [PubMed]
21. Patel K, Sebastiani G. Limitations of non-invasive tests for assessment of liver fibrosis. JHEP Rep 2020;2:100067. [Crossref] [PubMed]
22. Caldwell SH, Intagliata NM. Bleeding after liver biopsy: Remarkably infrequent but potentially serious and unfortunately unpredictable by extensive clinical laboratory analyses-Take a closer look at the biopsy? J Thromb Haemost 2022;20:2756-7. [Crossref] [PubMed]
23. Scridon A. Platelets and Their Role in Hemostasis and Thrombosis-From Physiology to Pathophysiology and Therapeutic Implications. Int J Mol Sci 2022;23:12772. [Crossref] [PubMed]
24. Alvaro D, Caporaso N, Giannini EG, et al. Procedure-related bleeding risk in patients with cirrhosis and severe thrombocytopenia. Eur J Clin Invest 2021;51:e13508. [Crossref] [PubMed]
25. Potretzke TA, Saling LJ, Middleton WD, et al. Bleeding Complications After Percutaneous Liver Biopsy: Do Subcapsular Lesions Pose a Higher Risk? AJR Am J Roentgenol 2018;211:204-10. [Crossref] [PubMed]