Bleeding management in computed tomography-guided liver biopsies by biopsy tract plugging with gelatin sponge slurry

To evaluate the safety and impact of biopsy tract plugging with gelatin sponge slurry in percutaneous liver biopsy. 300 consecutive patients (158 females, 142 males; median age, 63 years) who underwent computed tomography-guided core biopsy of the liver in coaxial technique (16/18 Gauge) with and without biopsy tract plugging were retrospectively reviewed (January 2013 to May 2018). Complications were rated according to the common criteria for adverse events (NCI-CTCAE). The study cohort was dichotomized into a plugged (71%; n = 214) and an unplugged (29%; n = 86) biopsy tract group. Biopsy tract plugging with gelatin sponge slurry was technically successful in all cases. Major bleeding events were only observed in the unplugged group (0.7%; n = 2), whereas minor bleedings (4.3%) were observed in both groups (plugged, 3.6%, n = 11; unplugged, 0.7%, n = 2). Analysis of biopsies and adverse events showed a significant association between number of needle-passes and overall (P = 0.038; odds ratio: 1.395) as well as minor bleeding events (P = 0.020; odds ratio: 1.501). No complications associated with gelatin sponge slurry were observed. Biopsy tract plugging with gelatin sponge slurry is a technically easy and safe procedure that can prevent major bleeding events following liver biopsy.

www.nature.com/scientificreports/ using a 16-Gauge (G) core needle coaxial system in combination with an 18-G cutting needle (Cook Medical Europe Ltd., Ireland) were included in the study. Other needle diameters and non-coaxially obtained biopsies were excluded from the study cohort ( Fig. 1). Definitions. Procedures were dichotomized by application of gelatin sponge slurry into a plugged and an unplugged biopsy tract group. Postinterventional complications were assessed using the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE) Version 5.0. NCI-CTCAE standardizes five degrees of severity of complications following medical treatment or procedure. These complications are classified as minor (CTCAE 1), moderate (CTCAE 2), severe (CTCAE 3), life-threatening (CTCAE 4) or fatal (CTCAE 5). Depending on grade of severity, a therapeutical concept is derived 11 . Bleeding was divided into minor (Hgb loss < 2 mg/dl) and major (Hgb loss > 2 mg/dl) bleedings. In case of minor bleedings, observation only was sufficient, whereas patients with major bleedings were referred to ICU observation and/or surgical/ interventional treatment.

Data. Procedural parameters
Procedure technique. Procedures were performed by experienced interventional radiologists or interventional radiology residents under direct supervision. Pre-procedural sufficient coagulation status was obligate (PTT, INR, platelet count). Otherwise, sufficient correction of coagulation parameters by platelet transfusion or tranexamic acid was performed pre-interventionally by the referring ward. All patients received a pre-interventional non-enhanced CT at the time of intervention for biopsy planning. The biopsy path was planned on an individual patient basis. Tissue sampling was performed in sterile coaxial technique under CT-guidance as previously described 1 (Fig. 2). Specimens were preserved in formalin and sent to the department of pathology. Biopsy tract plugging was performed at the discretion of the attending interventional radiologist by using gelatin sponge slurry (CuraSpon ® , CuraMedical B.V., The Netherlands). A 5 ml suspension of gelatin sponge slurry was prepared immediately before the biopsy procedure by cutting the gelatin sponge into small squares (5 × 5 mm) and mixing 2 ml of compressed sponge with 4 ml of 0.9% NaCl solution. Mixing the components was performed by using two 5 ml Luer-Lock disposable syringes connected via a three-way stopcock, until a viscous embolizate was prepared (Fig. 3). After exclusion of puncture or penetration of critical vascular structures (hepatic vein, portal vein branch or artery) on interventional CT images, a volume between 2-3 ml of gelatin sponge slurry was continuously applied during withdrawal of the introducer needle until it exited the skin surface after completion of the biopsy.

Follow-up.
To detect early complications (2-4 h after the procedure 12 ), a post-interventional non enhanced CT was performed after completion of the procedure. Blood count control was performed within clinical course by the referring ward. Additional sonographic follow-up was ordered depending on the prior imaging and laboratory results. All biopsy procedures were performed on an in-patient basis with a post-interventional clinical follow-up of at least 24-48 h.    www.nature.com/scientificreports/ regression analysis was used to determine the odds ratio (OR). A P-value < 0.05 was considered statistically significant.

Parameter Valid, N All patients Plugged biopsy tract Unplugged biopsy tract P-value
No adverse events attributable to gelatin sponge slurry like anaphylaxis, vascular occlusion or local infections/ abscesses were recorded. No postinterventional fatality was reported.
Beside clinical observation, Hgb-control was performed by the referring ward. Average pre-interventional Hgb was 12.3, average post-interventional Hgb was 12.0. In case of minor bleeding, an average Hgb-drop of 0.6, in major bleedings an average Hgb-drop of 2.2 was observed.
One major bleeding (CTCAE 4) occurred in a 56-year-old patient with a metastatic gastrointestinal stromal tumor of the jejunum (liver lesion diameter, 30 mm; localization, right lobe; access route, lateral; number of biopsies, 2; biopsy tract length, 20 mm). Postinterventional non-enhanced CT showed no complications. Laboratory diagnostics in the referring ward revealed a significant Hgb drop of 1.1 (9.0 to 7.9). The patient was immediately transferred to the surgical intensive care unit by the medical emergency team and, due to circulatory instability, emergency laparotomy was performed identifying intraperitoneal bleeding from the liver puncture tract. Surgery led to successful hemostasis.
Another major bleeding (CTCAE 3) occurred in an 81-year-old patient with metastatic prostate cancer (liver lesion diameter, 33 mm; localization, right lobe; access route, ventral, number of biopsies, 4; biopsy tract length, 39 mm). Postinterventional non enhanced CT showed no complications. The patient presented with elevated liver enzymes and cholestasis parameters. Contrast enhanced CT revealed liver hematoma and hematoma in the gall bladder. The patient was referred to the ICU for blood transfusions. No further intervention was necessary.

Discussion
Personalized cancer therapies increase demand on tissue sampling. Besides establishing the nature of a lesion, staging, therapy-monitoring, research and clinical oncological studies are based on biopsies [1][2][3]13 . Larger tissue volume samples on the one hand provide more adequate information about tumor architecture/microenvironment and enable multiple analyses 2,14 , but may lead to increased bleeding risk and mortality on the other [15][16][17] . Therefore, complication management of percutaneous biopsies is important.
Overall major complication-rate in liver biopsy ranges between 0.3 and 3.3% 18 . Thereof, rate of major bleeding events range between 0.5-1.2%, though it is elevated in tumor patients (1.6%), the female sex, and advanced age 15,19 . Procurement of more than two biopsies was reported to elevate the risk for bleeding 16,17,20 . A number of > 4 biopsies is associated with a threefold higher risk of major bleeding events 16 .
Percutaneous needle biopsy can be performed in coaxial and non-coaxial technique. Coaxial technique, as chosen in this investigation, offers several procedural advantages. Besides collection of multiple samples in a single puncture, it has been reported to reduce bleeding risk (one puncture of the organ capsule) and prevent tumor cell seeding 1 . Even though this technique introduces a larger access route, it has not been associated with higher complication rates 1,21,22 . Additionally, coaxial technique allows for tract plugging, an effective tool in bleeding management/prevention. Compared to literature, the rate of major bleeding events in our cohort was within acceptable margins (0.7% vs. 0.5-1.2%), even though our patients demonstrated an elevated risk profile (76% tumor patients; median number of biopsies, 4 [range 1-11]; 53% females; median age, 63.0). www.nature.com/scientificreports/ Minor bleedings were reported in 4.3% (plugged, n = 11/214; unplugged, n = 2/86), which is also within acceptable margins (0-10.9%) 23 . Haage and colleagues described one subcapsular hematoma and one moderate bleeding from the biopsy tract in a study of 1999 without tract plugging 24 . Thus, the self-limiting hematoma rate in their study was notably lower, although the biopsy tract was not plugged with a gelatin sponge slurry. A possible explanation for this finding might be the higher number of samples per biopsy obtained in our study cohort in comparison to Haage et al., who in general obtained only one biopsy sample for pathological diagnosis.
The number of biopsies (plugged group, 4 [range 1-10]; unplugged group, 3 [range 1-11]) had a significant impact on bleeding events (P = 0.038; OR = 1.395, 95% CI lower = 1.018, 95% upper = 1.913), which is concordant with the pertinent literature, where significant influence on major hemorrhage was found 16,17,20 . In comparison to Boyum et al., our study cohort was rather small (n = 300 vs. n = 5011) and risk of major bleeding events in general is low. The number of samples per biopsy did not demonstrate a significant difference regarding major bleeding events in our cohort (P = 0.883), but did show a correlation for minor bleeding events (P = 0.020; OR = 1.501, 95% CI lower = 1.066, 95% upper = 2.113).
Beside gelatin sponge slurry, biopsy tract embolization can also be achieved with other materials, e.g. embolization coils, autologous blood clots, microfibrillary collagen, or gelatin sponge in rolled form 1,9,25,26 . One advantage of gelatin sponge slurry in contrast to e.g. coils is that it is absorbed completely within weeks. Risks of dislocation, delayed local infection, or imaging artefacts associated with permanent foreign bodies can therefore be avoided 27 .
Comparison of gelatin sponge and other absorbable embolizates has so far not been performed for liver biopsy tract plugging. Gelatin sponge was compared to microfibrillary collagen for tract embolization following percutaneous transhepatic islet cell transplantation and found to be inferior 9 . However, gelatin in this study was introduced into the biopsy tract in rolled form and not mixed as a suspension. It is possible, that gelatin suspension leads to a more adequate embolization due to a more sufficient distribution than the application of rolled gelatin sponges. An advantage of gelatin sponge over microfibrillary collagen, however, is the notably lower cost (< $ 4 at our institution vs. approx. $ 180.00 in 2016). Besides these advantages, preparation of gelatin sponge Table 4. Logistic regression analysis of the cohort given as N (%), Odds ratio (OR) with 95% confidence intervals (CI). www.nature.com/scientificreports/ slurry is fast and easy, especially in emergency cases. Severe adverse events induced by gelatin sponge slurry like anaphylaxis are rare (< 0.5%) 27,28 and did not occur in any of our patients. As described we performed non-contrast CT immediately after the intervention in order to detect postprocedural complications. It is arguable that post-interventional hemorrhage can also be detected sonographically forgoing the need for additional radiation exposure. However, there are inherent limitations of sonography in our patient cohort of mostly elderly, critically ill patients who were often unable to perform deep inspiration, adequate breath holds, or demonstrated meteorism. Therefore, we opted for postinterventional non-contrast enhanced CT. From a procedural point of view postinterventional non-contrast enhanced CT not only provides a basis for evaluation of complications without the need for repositioning the patient. It can also be followed by immediate contrast-enhanced CT angiography in case any complications are suspected without time loss and thus also provides the basis for adequate and immediate intervention (e.g., immediate transfer to angiography for further treatment).
This study is limited by its retrospective character. Furthermore, our sample size of 300 patients was rather small, in particular with regard to overall low rates of major bleedings following liver biopsy. Therefore, further validation of our data should be the subject of larger cohorts. In our institutional setting, a randomized trial of tract plugging in the presence of active bleeding from the biopsy site was not justifiable, though it would be an ideal study design to test the efficiency and safety of gelatin sponge slurry. Some post-procedural complications might not be reported to us due to follow-up treatment in another institution, especially in patients with a more distant place of residence. The amount of gelatin sponge slurry applied can only be estimated. Only the use of gelatin sponge slurry, not the amount was documented. To our knowledge, no comparative study between plugged and unplugged biopsy tract exists, therefore this study addresses an important gap of knowledge.

Conclusion
Biopsy tract plugging with gelatin sponge slurry is a technically easy and safe procedure that can prevent major bleeding events following biopsy procedures of the liver.