Reoperation for post hepatectomy complications

Reoperation for post hepatectomy complications is associated with high rates of morbidity and mortality. We aim to describe the frequency, indications, and risk factors for reoperation after liver resection in a single centre.


Introduction
Hepatectomy is the gold standard treatment for liver malignancies, including primary liver cancer and metastasis from other sites, most commonly colon.Historically, hepatectomy has carried a high operative mortality rate of 20%. 1 Due to advances in surgical technique, imaging tools, and perioperative care, the mortality rate is near to 0% in high volume centres. 2,3However, post-operative complications that require reoperation are associated with high mortality and morbidity. 4,5The risk of perioperative mortality within 90 days increases seven-fold (23.0%versus 3.4%) in patients who required reoperation compared to those who did not require reoperation. 6he most common reasons for reoperation include haemorrhage, intra-abdominal collection, wound dehiscence, and bile leak. 6,7here is paucity of published data exploring the risk factors associated with reoperation post-hepatectomy.Currently, known risk factors associated with reoperation include male sex, American Society of Anaesthesiologists (ASA) class 4, concomitant major procedures, and intraoperative blood loss. 6,8is study aims to describe common indications and frequency of reoperation after a liver resection, independent risk factors for reoperation, and the outcomes of these patients, during the 20-year experience of a single institution.

Methods
A retrospective analysis of outcomes from a prospectively maintained database of consecutive patients who underwent hepatectomy at The Queen Elizabeth Hospital (Adelaide, South Australia) from January 2001 to December 2020 was performed.
Data analysed included: preoperative data including age at the time of operation, sex, body mass index (BMI), ASA score, and indication for operation were obtained.Intraoperative data included duration of operation, duration of Pringle manoeuvre, number of segments resected, mass of liver resected, maximum lesion size, performance of additional major non-hepatic procedure, histopathology, intraoperative blood loss, and intraoperative blood transfusion.Minor resection was defined by resection of three or less segments and major resection was defined by four or more segments.Performance of an additional major non-hepatic procedure included non-hepatobiliary organ resection and hernia repair, but excluded cholecystectomy and adhesiolysis.Postoperative data included length of stay (LOS) in hospital, length of stay in intensive care unit (ICU), occurrence and severity of post-operative complications graded according to the Clavien Dindo classification.Thirty days and 90 day mortality was calculated.
Reoperation, or return to theatre, was determined to have occurred when unplanned procedures requiring general anaesthetic (Clavien Dindo grade 3B or higher) were needed prior to hospital discharge.Procedures that did not require general anaesthetic (Clavien Dindo grade 3A) or procedures that were planned such as removal of packs for intra-operative bleeding were not included in the reoperation group.

Standard hepatectomy procedure
Preoperatively, patients receive a workup consisting of history, physical exam, blood tests, and staging investigations.Definitive management is determined at a multidisciplinary meeting consisting of surgeons, radiologists, oncologists, and pathologists.Ceftriaxone and Metronidazole are used for induction surgical prophylaxis.Most open hepatectomies are performed via a subcostal incision.Intraoperatively, the liver is palpated, and ultrasound is used to confirm lesion location and margins.Haemostasis is achieved using compression, pack, argon, and fibrillar.A pringle manoeuvre is employed if necessary and the timing is recorded on the operation note.Parenchyma is dissected using bipolar electrocautery or harmonic scalpel.Fascia is closed using PDS loop and skin is closed with staples.Patients receive abdominal drains and ICU admission if required.DVT prophylaxis (compression stocking, calf stimulators and Clexane) is prescribed post operatively.There were no laparoscopic or robotic resections performed at our centre during the period of this study.

Statistical analysis
Comparison of the group requiring reoperation to the group which did not require reoperation was performed using Wilcoxon Rank Sum Test and independent t-test.Univariate binary logistic regression models were performed for the binary outcome: reoperation versus various a priori predictors.The statistical software used was SAS On Demand for Academics (SAS Institute Inc. 2021).A P-value of <0.05 was considered statistically significant.

Characteristics and perioperative data
A total of 464 hepatectomy patients were included in these analyses.Baseline characteristics and perioperative details of all patients are listed in Table 1.271 (58.4%) were male and 193 (41.6%) were female.The average age was 63.8 years (range 20-89).184 (41.4%) patients had an ASA ≥3.Malignancy was confirmed by histopathology in 427 (92.4%) patients, which consisted of 67 primary malignancies and 360 metastatic disease cases.Of the liver primary malignancies there were 47 hepatocellular carcinoma, 18 cholangiocarcinoma, 1 combined hepatocellular carcinoma cholangiocarcinoma, and 1 follicular dendritic cell sarcoma.Metastatic colorectal cancer (n = 309) was the most common malignant diagnosis.Overall, 142 (30.9%) patients underwent major hepatectomy, nine (6.3%) patients had extended hepatectomy, and 62 (13.4%) patients had one or more additional major non-hepatic procedure.The mean number of segments resected was 2.6 (range 1-7) and the mean liver mass resected was 438.9 g (range 3-2156 g).The mean operating time was 204 min (range 56-805 min).Pringle manoeuvre was used in 188 procedures, with an overall mean Pringle time of 19.8 minutes.The mean intraoperative blood loss was 887.8 mL (range 20-10 000 mL) and 119 (25.6%) patients required intraoperative blood transfusion.The mean and median length of stay was 10 and 8 days respectively (range 1-55 days).288 (62.1%) patients experienced complications, with a mean Clavien Dindo grade of 2.1.The number of patients who had a severe complication with Clavien Dindo grade 3 or more was 65 (14%).Overall, there were 9 deaths within 90 days of procedure.Thirty days and 90 day mortality rates were 0.9% and 1.9% respectively.

Re-operative patients
Only seven (1.5%) patients required reoperation post hepatectomy.Four (57.1%) were male and the mean age was 71 years (range 47-81 years).Overall, five patients (71.4%) had an ASA ≥3, and mean ASA score was 2.8.Five patients had malignant colorectal cancer and two had primary liver cancer (cholangiocarcinoma).The mean number of segments resected was 3.1 (range 1-5).Five patients underwent major hepatectomy and three had additional major non hepatic procedures, including extrahepatic bile duct resection, hepaticojejunostomy, jejuno-jejunostomy (n = 1), wedge resection of splenic flexure (n = 1), right hemicolectomy, debulking of mass and peritoneal disease, bilateral oophorectomy, hyperthermic intraperitoneal chemotherapy (n = 1).The mean duration of operation was 247 minutes (range 140-465 min).Pringle manoeuvre was used in 3 (42.3%)procedures and overall mean pringle time was 28 minutes (range 20-34 min).The mean liver mass resected was 733.6 g (range 70-1544 g), and the mean size was 42.5 mm (range 22-120 mm).The mean intraoperative blood loss was 814 mL (range 100-1500 mL) and three (42.3%)patients required intraoperative blood transfusion.The mean and median length of stay was 19 and 16 days respectively (range 4-39 days).The mean number of days until return to theatre was 8.4 days (range 0-20 days).There were two patients who required a reoperation procedure involving removal of packs for intra-operative bleeding which have been excluded from the reoperation group because they were planned procedures.This is summarized in Table 2.

Reoperation for post hepatectomy complications
Indications for reoperation were three (42.9%)intra-abdominal abscess, two (28.6%)post-operative haemorrhage, one (14.3%)ischaemic bowel, and one (14.3%)bile leak (Table 2).Three patients who had intra-abdominal abscess required incision and drainage, washout, and wound closure.Two of these patients returned to theatre three times for repeat incision and drainage.The other patient required one incision and drainage procedure.
Two patients had post-operative haemorrhage, one had bleeding from the wound edge/muscle (n = 1), and one from the staple line of right portal pedicle, with the venous bleeding point at liver parenchyma (n = 1).One patient had ischaemic bowel due to superior mesenteric artery thrombosis complicated by small bowel obstruction, requiring three separate reoperations on day 7, 8 and 15 post operation.The procedures performed during the reoperation include laparotomy and small bowel resection on the first reoperation, relook laparotomy, small bowel resection, and washout on the second reoperation, relook laparotomy, washout for a leak in duodenal jejunal flexure on the third reoperation.There was one patient who had bile leak requiring a laparotomy and drain insertion.

Outcomes after re-operation for post hepatectomy complications
Three patients (43%) died following reoperation.One patient underwent re-operation within 24 h after the initial operation due to post-operative haemorrhage from liver parenchyma.Haemostasis was achieved during the re-laparotomy and the patient's haemodynamic status improved, however, the patient died on day 4 due to multi-organ failure from hypovolemic shock.One patient had three separate reoperations for ischaemic bowel, small bowel obstruction, and duodenal jejunal leak.The patient died 37 days after the index operation due to hypovolemic shock secondary to uncontrollable intra-abdominal bleeding.One patient had an intraabdominal abscess and sepsis secondary to an anastomotic leak, requiring three re-operations for washout.This was complicated by a large pleural effusion, vocal cord injury secondary to intubation.Pathology revealed cholangiocarcinoma and extensive peritoneal disease.A palliative approach was taken, and the patient died 39 days after index operation.

Risk factors for reoperation
Patients who had an additional major non hepatic procedure during index operation had significantly higher odds of requiring reoperation (Odds Ratio = 5.06 (95% CI: 1.10, 23.17, P = 0.04).Patients who died within 90 days were more likely to have had a  reoperation compared to those did not (Odds Ratio = 56.38 (95% CI: 10.30, 308.56,P < 0.0001).No statistically significant associations were found between reoperation and sex, ASA (3/4 versus 1/2), major versus minor hepatectomy, intraoperative transfusion, or intraoperative blood loss.This data is summarized in Table 3.

Discussion
In this study, indications and outcomes of reoperation in patients after initial hepatectomy at a single institution were assessed.Despite significant improvements in operative mortality in the last 20 years, hepatectomy is still a major operation with a risk of morbidity and mortality.There is a lack of data in the literature for the frequency, risk factors, and indications for reoperation post hepatectomy.][8] The most clinically significant finding of this study was that patients who required reoperation post hepatectomy were 56 times more likely to experience mortality within 90 days of index hospitalization.For patients who did not have reoperation, the 90 day mortality rate was 1.3%.
Binary logistic regression analysis identified additional major non-hepatic procedure as a risk factor associated with need for reoperation.This finding was consistent with the study by Barbas et al. which found that performance of a major concomitant nonhepatic procedure was independently associated with need for reoperation. 6In our study, three out of seven patients who required early reoperation had additional major non-hepatic procedures.The additional procedures performed included extra-hepatic bile duct resection, wedge resection of splenic flexure, and right hemicolectomy.The extrahepatic bile duct resection, hepaticojejunostomy, jejunojejunostomy was performed as part of standard treatment for primary cholangiocarcinoma.Resection of splenic flexure was performed for resection of colorectal cancer.The right hemicolectomy was performed for large bowel obstruction and bilateral oophorectomy was performed for significant peritoneal disease.Additional procedures are associated with increased blood loss, longer operation time, and more surgical stress, which increase the risk of complications post operatively that may necessitate reoperation.Therefore, it may be appropriate for some patients to undergo staged procedure to minimize risks and optimize outcomes.
Two of the three patients with intra-abdominal abscess had additional major non hepatic procedures performed with a range of operation time from 219 to 465 min.The long duration of operation may increase risk of exposure of abdominal cavity to environment and increase surgical stress to the immune system, which increase the risk of infection.This is in agreement with the literature which shows that the risk of post-operative infection doubles with each additional hour of surgery. 9The nature of the additional procedures may also play a role in formation of intra-abdominal abscess because both patients had bowel anastomosis which could result in anastomotic leaks.Therefore, the duration of operation and performance of additional major non-hepatic procedures should be minimized when possible.The use of intravenous or oral antibiotic prophylaxis and intraoperative leak test are necessary to prevent an anastomotic leak resulting in abscess formation or sepsis.It is important to note that both patients with post-operative haemorrhage had high amounts of intraoperative blood loss (1100 and 1500 mL respectively) and a significant mass of liver resected (804-1053 g, 4-5 segments respectively).This was likely to result in loss of synthetic function of the liver and increase the risk of haemorrhage.Thus, the use of haemostatic agents intra-operatively and timely replacement of clotting factors are vital to prevent postoperative haemorrhage.One patient was hemodynamically unstable and actively bleeding post index hepatectomy and one patient had bleeding from the wound site.Both returned to theatre within 24 h after index hepatectomy.The patient who had bleeding at liver parenchyma died later due to hypovolemic shock.The other patient who had bleeding at the wound edge/ muscle recovered well.The in-hospital mortality rate for our two patients with post hepatectomy haemorrhage was 50% which was consistent with the literature. 10Therefore, it is critical to reduce the risk of post hepatectomy haemorrhage by careful manipulation and meticulous haemostasis at the end of the operation.Post operatively, close monitoring of the excision site/drains and vital signs are crucial to detect intra-abdominal haemorrhage early, and correction of coagulopathy is important to reduce bleeding risk.
Bile leak is a serious complication post hepatectomy.While reoperation for bile leak is rarely required, major bile leaks could lead to intra-abdominal sepsis and post-operative liver failure, with a mortality rate as high as 40%-50%. 11,12Until the proposal of a standardized definition of bile leak by the International Study Group of Liver Surgery, 13 little was known about the risk factors, frequency, outcomes of bile leak.In our study, the rate of bile leak was found to be 1.9%, which was lower than previously reported in the literature. 14,15One patient with bile leak and peritonitis required reoperation but ultimately recovered well.Eight other patients with a bile leak did not require reoperation and were managed conservatively or with radiological drainage.The use of intraoperative drains and increasing intraoperative blood loss have been identified as risk factors for bile leak. 14However, it has been shown that patients with a drain placed intraoperatively showed no change in clinical course, indicating that intraoperative drainage may not necessarily prevent bile leak or intraabdominal collection. 14The best way to prevent bile leak involves intraoperative detection and repair of injury through methods such the 'air leak test', 16 indocyanine green or methylene blue dyes.Postoperatively, routine intraoperative drainage may not be necessary.Selective drain placement for patients with significant intraoperative blood loss may have the most benefit.While minimally invasive hepatectomy remain challenging especially for major resections, it has less risk of bile leak when compared with open hepatectomy. 17iven our centre is experienced in open hepatectomy and has previously shown excellent peri-operative outcomes, we have not adopted laparoscopic hepatectomy. 18Previous meta-analysis demonstrated laparoscopic hepatectomy of small hepatocellular carcinoma was associated with shorter hospital stay, lower morbidity, and lower 30-day mortality rates. 19However, their oncological outcomes in terms of overall survival and disease-free survival were equivocal.Furthermore, there was no significant difference in reoperation and 90-day mortality rates. 20Our centre does not have access to robotic surgery.
It is important to note that our study is a retrospective analysis of a prospective database, and as such, there are some limitations.Important peri-operative patient factors such as BMI, patient comorbidities, smoking history, pre-op liver function and albumin are not available, and thus cannot be incorporated in reoperation risk assessment.The analyses are also limited to post-operative complications and does not include recurrence or survival in the long term.A small number of patients in our study underwent reoperation, and further studies involving larger sample sizes are necessary to fully appreciate the risk factors associated with reoperation after liver resection and expand upon our findings.
One of the strengths of this analyses is that the operation data is complete, with the operation time being tracked from initial incision until final skin closure.Additionally, the operations in our study were performed by at least one experienced consultant, which helps minimize potential differences in outcomes due to variations in surgical technique or expertise.Further research that employs larger sample sizes and prospective data collection methods may be needed to enhance the accuracy and completeness of postoperative data.Nonetheless, our study provides valuable insights into the frequency and risk factors for reoperation after liver resection and can help guide the development of strategies to reduce the need for reoperation and improve patient outcomes.

Conclusion
Reoperation for post hepatectomy complications is associated with high morbidity and mortality.Hepatectomy patients who undergo simultaneous major non-hepatic procedures such as bowel resection are more likely to require reoperation and more likely to experience 90 day mortality.Clinicians should ensure multi procedure patients are well monitored post procedure, and where possible, plan staged procedure to minimize risk and optimize outcomes.

Table 1
Baseline characteristics and perioperative details of patients Abbreviations: ASA, American Society of Anaestheologists; LOS, length of stay; IQR, interquartile range; SD, standard deviation.*Wilcoxon Rank Sum Test P-value.**Independent t-test P-value.

Table 3
Univariate binary logistic regressions for reoperation (Yes/No) versus various predictors American Society of Anaestheologists; LOS, length of stay; IQR, interquartile range; SD, standard deviation.
© 2023 The Authors.ANZ Journal of Surgery published by John Wiley & Sons Australia, Ltd on behalf of Royal Australasian College of Surgeons.