Table 2

Studies with mortality and morbidity data in relation to non-hepatic surgery in patients with cirrhosis and portal hypertension, categorised by type of surgery

Type of surgeryReference and dateType and details of studySalient findings
Colorectal surgeryGervaz et al46 2003Retrospective analysis of 72 colorectal adenocarcinoma operations between 1976 and 2001 on cirrhosis patients.
  • Post-operative death—13%

  • Risk factors predictive of postoperative mortality include elevated bilirubin (p=0.01) and prolonged prothrombin time (p=0.009).

  • CTP-A patients had significantly better survival rates than combined group of CTP-B and C patients (p=0.008)

Csikesz et al10 2009Retrospective analysis of national database between 1998 and 2005; 22 569 patients with cirrhosis (of whom 4214 had portal hypertension)
  • Higher mortality for patients undergoing colectomy compared with controls (HR 3.7, 95% CI 2.6 to 5.2)

  • Presence of portal hypertension confers even higher risk (HR 14.3, 95% CI 9.7 to 21.0)

Nguyen et al39 2009Population based study (1998–2005) identifying 4042 patients
  • Patients with cirrhosis and cirrhosis with portal hypertension had higher in-hospital mortality in comparison to patients without cirrhosis (14% and 29% vs 5%, respectively, p<0.0001)

  • Approximately four-fold higher rate of in-hospital mortality for emergency and urgent procedures in comparison to elective procedures in liver cirrhosis patients

  • Higher mortality rate of cirrhotic patients with portal hypertension undergoing surgery (HR: 5.8; 95% CI 4.9 to 7.6).

  • Independent risk factors for mortality were cirrhosis, portal hypertension, old age, colectomy and comorbidities which included cardiovascular disease, chronic kidney disease, paraplegia and malnutrition.

Ghaferi et al47 2010Prospective, clinical database between 2005 and 2007 identifying 1565 patients with chronic liver disease undergoing colorectal resections
  • Chronic liver disease patients were identified using clinical characteristic: ascites, oesophageal varices or total bilirubin >34 umol/L.

  • 6.5-fold increased risk of mortality following colorectal operations in chronic liver disease patients.

  • MELD score >15 associated with higher rates of mortality and complications

(Ascites, infection, bleeding anastomotic leaks and stoma issues including leaks, difficulty closing and peristomal variceal bleeding).
Montomoli et al87 2013Population-based study between 1996 and 2009 identifying 39 840 patients undergoing colorectal surgery with 158 (0.4%) having cirrhosis
  • Higher 30-day mortality (24.1%) corresponding to adjusted RR of 2.59 (95% CI 1.86 to 3.61)

Lee et al48 2017Retrospective analysis of 161 cirrhosis patient undergoing surgery for colorectal cancer (CRC)
  • MELD score >8 influenced overall survival (p<0.001)

Lee et al88 2018Retrospective, observational, population-based study between 2005 and 2014 identifying 7463 patients with CRC who underwent colorectal surgery.
  • Increased risk of in-hospital mortality (adjusted OR 2.05, p<0.001)

  • No significant increase in postoperative complications (adjusted OR 0.91, p=0.192)

Cheng et al45 2021Meta-analysis identifying 2485 patients from 5 studies
  • Cirrhotic group experienced more major complications (OR 5.15; p=0.005), higher rates of return to theatre (OR: 2.04; p=0.03), higher short-term mortality (OR: 2.8; p<0.00001) and shorter survival (HR 2.96, p<0.00001)

Peptic ulcer diseaseLehnert and Herfarth64 1993Retrospective analysis (1972–1991), 69 patients undergoing surgery (90%—emergency) for gastroduodenal disease
  • Overall mortality—54% with bleeding and multiorgan failure being leading causes of death (70%).

Gastric cancer surgeryIsozaki et al65 1997Retrospective analysis of 39 patients with liver cirrhosis undergoing curative surgery for gastric cancer between 1978 and 1994
  • Postoperative complications were observed in 10 (25.6%) of patients with 4 (10.3%) hospital deaths.

Guo et al66 2014Retrospective analysis of 58 patients with cirrhosis undergoing radical gastrectomy between 2001 and 2012.
  • Forty-four patients received subtotal gastrectomy and 14 received total gastrectomy accompanied by D1 (26 patients) or D2 (32 patients) lymphadenectomy.

  • Severe postoperative complications occurred in 58.6% of patients and occurred more frequently in CTP-B (p=0.03) or if underwent D2 lymphadenectomy (p=0.015).

  • Postoperative mortality occurred more frequently in CTP-B patients (p=0.033).

  • 100% mortality was experienced in CTP-C.

Oesophageal surgeryValmasoni et al89 2017Retrospective analysis of oesophageal cancer database identifying 3445 oesophageal cancer patients, 73 with cirrhosis undergoing surgery.
  • Cirrhosis patients experienced more respiratory events (p=0.013), infections (p=0.005) and severe anastomotic complications (p=0.046)

  • MELD >9 associated with decreased 5 year survival (p=0.004) and MELD score or nine or lower showed outcomes similar to that of non-cirrhotic patients.

Cheng et al90 2020Retrospective, propensity-matched study (cirrhotic (n=50) and non-cirrhotic patients (n=100))
  • Patients with cirrhosis experience higher rates of postoperative complications including postoperative pneumonia (22 vs 9%, p=0.027), pleural effusion (38 vs 20%, p=0.018), chylothorax (10 vs 1%, p=0.016) and had longer intensive care unit (ICU) stay (mean: 6.10 vs 2.58 days, p=0.002) compared with controls

Schizas et al68 2020Meta-analysis of 12 observational studies including 1938 patients (238 with cirrhosis)
  • 30-day mortality higher in cirrhosis (OR 3.04, 95% CI 1.71 to 5.39) although this was not observed at 90 days (OR 2.84, 95% CI 0.94 to 8.93) or long term (mean follow up - 24-months) (OR 1.70, 95% CI 0.53 to 5.51).

  • Anastomotic leak occurred at a higher rate in patients with cirrhosis (OR 2.81, 95% CI 1.05 to 7.49).

  • CTP-A patients were associated with a significantly lower 30-day mortality compared with CTP-B (OR 0.14, 95% CI 0.04 to 0.54).

Bariatric surgeryLee et al91 2021Retrospective, 1:1 propensity-matched case control study of 957 patients with and without cirrhosis undergoing bariatric surgery.
  • No difference in mortality (OR 1.73; p=0.33)

  • Comparing decompensated (n=117) and compensated (n=957) cirrhosis, increased mortality was observed (7.69 vs 0.94%, p<0.001).

Mavilia et al92 2020Retrospective study of 20 096 chronic liver disease (using ICD coding) patients undergoing bariatric surgeryChronic liver disease patients experience high inpatient mortality (adjusted OR 1.47, 95% CI 1.24 to 1.73) but significantly less surgical revision, improper wound healing and postoperative infection following bariatric surgery
Agarwal et al71 2021Systematic review and meta-analysis including 18 studies and 471 patients with obesity and liver cirrhosis undergoing bariatric surgery (mainly LSG and RYGB)
  • 397/423 patients with defined CTP class were CTP-A

  • The weighted pooled proportion of overall complications was 22.14% and all-cause 90-day mortality was 0%.

  • Significant increase in postoperative complications (p<0.001) but no difference in all-cause 90-day mortality (p=0.1165) compared with controls.

AppendectomyRashid et al73 2022Meta-analysis of 923 patients with cirrhosis undergoing appendectomy
  • 30-day mortality 9% vs 0.3% in cirrhosis and non-cirrhosis patients, respectively.

  • Laparoscopic appendectomy appeared safer with a mortality of 0.5% in comparison to open appendectomy with mortality of 3.2%)

Umbilical herniaSnitkjær et al50 2022Systematic review of 13 prospective, 10 retrospective studies including 3229 patients
  • Evidence was graded as very low quality for all outcomes.

  • Mortality quoted to be 6% (n=191).

  • Patients with cirrhosis are eight times more like to die after surgery compared with patients without cirrhosis

De Goede et al52 2021Randomised controlled trial (CRUCIAL trial) with 2 years of follow-up recruiting 34 patients with umbilical hernia in liver cirrhosis and ascites
  • Randomised controlled trial (16—elective repair, 18—conservative management)

  • After 24 months, 8 (50%) assigned to elective repair, in comparison to 14 (77.8%) assigned to conservative treatment had no significant difference in morbidity.

Grey et al93 2008Retrospective analysis of 1421 cases, 127 (8.9%) cirrhotics
  • Elective repair in cirrhosis is associated with similar outcomes in patients without cirrhosis.

  • Cirrhotic patients were more likely to undergo emergency repair (26% vs 4.8%; p<0.0001), concomitant bowel resection (8.8% vs 0.8%; p<0.0001), return to theatre (7.9% vs 2.5%, p=0.0006) and increased length of stay (4 vs 2 days; p=0.01)

Eker et al94 2011Prospective study
  • In total, 30 patients (6 CTP-A, 19 CTP-B and 5 CTP-C) with a median MELD score was 12 (IQR 8–16).

  • Elective umbilical hernia repair is safe with no post-operative intensive care admissions and only 2 of 30 patients died; neither of deaths were attributable to umbilical hernia repair.

Carbonell et al95 2005Nationwide retrospective cohort study 32 033 patients (30, 836 non-cirrhotic and 1197 cirrhotics)
  • Cirrhotics had a higher age distribution (p<0.0001) underwent ICU admission more commonly (15.9% vs 6%; p<0.0001), had a longer length of stay (5.4 vs 3.7 days), and higher morbidity (16.5% vs 13.8%; p=0.008), and mortality (2.5% vs 0.2%; p<0.0001) compared with non-cirrhotics.

  • Mortality was seven-fold higher in patients undergoing emergency repair (3.8% vs 0.5%; p<0.0001)

Marsman et al96 2007Retrospective study between 1990 and 2004
34 cirrhosis patients
  • Elective hernia repair was successful in 12 out of 17 patients without complications and recurrence. 3/17 wound related problems and 4/17 had recurrence.

  • Conservative management was only successful in 23%; 10/13 attended hospitals for incarceration of which 6 required emergency hernia repairs. Two patients managed conservatively died from complications of umbilical hernia.

Pinheiro et al97 2020Prospective cohort study
246 patients with cirrhosis
  • A total of 246 patients (57 underwent elective hernia repair and 189 who opted for ‘wait and see’ approach) were included in the study. Of the latter, 43 (22.7%) patients required emergency hernia repair due to complications such as ascites leakage due to skin rupture in hernia site (n=28), incarceration (n=7), small bowel strangulation (n=5), and extensive skin necrosis or ulceration (n=3).

  • MELD score >11 (HR 7.8; p=0.011) and emergency hernia repair (HR 5.35; p=0.005) were identified as risk factors for 30-day mortality.

Inguinal herniaOh et al98 2011Retrospective study over 10-year period
780 patients having inguinal hernia repair. 129 patients with cirrhosis
  • Morbidity (9.1–16.7%) is not significantly higher than patients without cirrhosis

  • Overall mortality ranging from <1% to 2.7%

  • Cumulative recurrence rates were not significantly different between cirrhosis and non-cirrhosis group (p=0.87)

Patti et al99 2008Prospective evaluation of QOL by questionnaire including 32 patients.
  • Inguinal hernioplasty in patients with cirrhosis is a safe procedure and improves quality of life.

CardiovascularChou et al54 2017Nationwide, population-based study from Taiwan between 1997 and 2001 including 1030 Liver patients and 1040 matched controls without cirrhosis.
  • 1 year survival was 68 vs 81% (p<0.001) in cirrhosis for CABG and valve surgery

Hayashida et al55 2004Retrospective study between 1989 and 2003 of 18 patients with cirrhosis undergoing cardiac operations
  • Overall postoperative mortality—17%

  • CTP-A—no increased mortality when undergoing elective cardiac surgery

  • CTP-B and C—mortality rate 50%–100% after cardiopulmonary bypass

Jacob et al100 2015Systematic review (19 studies) of short-term and overall mortality in patients with liver cirrhosis classified by CTP score undergoing cardiac surgery.
  • CTP score reporting 30-day mortality noted to be 9%, 37% and 52% for CTP class A, B and C, respectively.

  • One-year mortality was reported to be 27.2%, 66.2% and 78.9%, respectively, for CTP class A, B and C, respectively.

Araujo et al56 2017Retrospective, propensity-matched, case–control study of 1197 patients with liver dysfunction undergoing cardiac surgery (n=755 CABG, n=442 valve surgery)
  • Increased mortality was observed for both CABG (OR 5.19, p<0.0001) and valve surgery (OR 7.49, p<0.0001) in comparison to controls.

  • Higher rates of complications (bleeding, respiratory, renal, infections) in patients with liver dysfunction and CABG.

Hseih et al57 2015Meta-analysis (22 studies, 939 patients—CABG, valve surgery and cardiopulmonary bypass)
  • Nineteen of the studies evaluated mortality with 354 patients in CTP-A, 205 in CTP-B and 33 in CTP-C.

  • Mean mortality rates were 20.58%, 43.58% and 56.48% for patients in class A, B and C, respectively (p<0.01 for comparisons between each class).

  • Major postoperative morbidity with rates up to 60%, 100% and 100% for CTP-A, B and C, respectively.

Steffen et al53 2017Retrospective, propensity-matched study between 1998 and 2011 identifying 2769 patients with cirrhosis undergoing surgical aortic valve replacement
  • Aortic valve surgery in-hospital mortality was 16 vs 5% in controls (OR 3.6, p<0.0001) and greater rate of complications (55% vs 45% for controls).

  • Risk factors of mortality included congestive cardiac failure, fluid and electrolyte disturbances, pulmonary circulation disorder and weight loss.

AAA repairMarrocco-Trischitta et al63 2011Retrospective, single centre study between 2001 and 2006 identifying 24 patients with Liver cirrhosis undergoing elective open repair of infrarenal AAA.
  • CTP-B and MELD>10—associated with reduced survival

  • Significant difference in 2-year survival (77.4% vs 97.8%; p=0.03)

Elective Hip and knee arthroplastyCohen et al74 2005Retrospective analysis of outcomes of primary total hip arthroplasty and total knee arthroplasty in cirrhotic patients.
  • Complication rates, decompensation and/or death in up to 80% of cirrhosis patients after emergency THA due to a fracture.

  • Primary THA or TKA can be safely performed electively in CTP-A and CTP-B patients

Bell et al101 2020Retrospective study identifying 18 129 cirrhotic patients undergoing TKA and compared with control of 1716439 TKA patients.
  • Cirrhosis was associated with increased rate of major complications (3.7% vs 2.3%; OR 1.23, 95% CI 1.13 to 1.33; p<0.001) Higher risk of periprosthetic joint infection compared with controls, minor medical complications (13.5% vs 7.4%; OR 1.52, 95% CI 1.45 to 1.59, p<0.001), transfusion (2.8% vs 1.4%; OR 1.66, 95% CI 1.51 to 1.81, p<0.001), encephalopathy (1.0% vs 0.2%; OR 3.00, 95% CI 2.55 to 3.51, p<0.001), DIC (<0.001) within 90 days

  • Alcohol and viral aetiologies were associated with increased rate of major complications.

Onochie et al78 2019Systematic review identifying eight studies on 28514 THA’s
  • Increased postoperative infection rates of 0.5% (p<0.001) and perioperative mortality of 4.1% (p<0.001).

  • Frequent need for revision surgery at 4% (p<0.001).

  • Aetiology of need for revision surgery included periprosthetic infection (70%), aseptic loosening (13%), instability (13%), periprosthetic fracture (2%) and linear wear (2%).

  • AAA, abdominal aortic aneurysm; CABG, coronary artery bypass grafting; CTP, child-turcotte-pugh; ICD, international classification of diseases; LC, laparoscopic cholecystectomy; LSG, laparoscopic sleeve gastrectomy; MELD, model for end-stage liver disease; OC, open cholecystectomy; RR, relative risk; RYGB, roux-en-y gastric bypass; THA, total hip arthroplasty; TKA, total knee arthroplasty.