Cirrhosis is not a contraindication to laparoscopic cholecystectomy: results and practical recommendations

Abstract

Background

Gallstones appear more frequently in patients with cirrhosis and open cholecystectomy in this patient population is associated with higher morbidity and mortality. The aim of the present study was to evaluate experience with laparoscopic cholecystectomy in patients with cirrhosis and to provide recommendations for management.

Methods

Retrospective review of laparoscopic cholecystectomy in patients with cirrhosis from March 1999 to May 2008 was performed. Peri-operative characteristics and subgroup analysis were performed in patients with Child–Pugh's classes A, B and C cirrhosis.

Results

A total of 68 patients were reviewed in this study. In all, 69% of the patients were Child's class A. The most common indication for cholecystectomy was chronic/symptomatic cholelithiasis (68%). Compared with patients with Child's class B and C, laparoscopic cholecystectomy in patients with Child's class A was associated with significantly decreased operative time (P= 0.01), blood loss (P= 0.001), conversion to open cholecystectomy (P= 0.001) and length of hospital stay (P= 0.001).

Conclusions

Laparoscopic cholecystectomy in patients with cirrhosis is feasible with no mortality and low morbidity, especially in patients with Child's class A cirrhosis.

Introduction

Gallstones appear to occur more frequently in patients with cirrhosis compared with the non-cirrhotic population (23–79% vs. 7–23%, respectively).^1–5 Open cholecystectomies in the cirrhotic patient population are reported to have high mortality (from 8% to as high as 83% in patients with a prothrombin time >2.5 s over control) and morbidity (6.6–57%).^6–9 In the late 1980s laparoscopy was introduced, which eventually became the standard approach for most cholecystectomies; however, the NIH consensus statement in 1992 stated that end-stage cirrhosis was a contraindication to laparoscopic cholecystectomy.¹⁰ Several series have since suggested that laparoscopic cholecystectomy in patients with cirrhosis is safe and associated with low mortality (0–4%) and acceptable morbidity rates (0–52%).^11–19

The present study is a retrospective review (1999–2008) of the largest single institutional series reported in the United States. The aim of the study was to evaluate the experience with laparoscopic cholecystectomy in patients with cirrhosis and to provide practical recommendations for management.

Patients and methods

Between March 1999 and May 2008, a retrospective review of all cholecystectomies performed by two surgeons (T.C.G. and D.A.G.) at the UPMC Liver Cancer Center was performed. Patients who had underlying cirrhosis were included for further analysis. All surgical procedures were performed at a single tertiary level hospital.

Pre-operative characteristics gathered from the medical records included age, gender, pre-operative laboratory values, Child–Pugh's classification at the time of surgery and indication for cholecystectomy. Based on the documented laboratory values at the time of surgery [albumin, total bilirubin and international normalized ratio (INR)] and clinical presentation (ascites and encephalopathy), the Child–Pugh's score was calculated. Gallstone or biliary disease was diagnosed with ultrasonography (US), computed tomography (CT), Endoscopic retrograde cholangiopancreatography (ERCP), magnetic resonance imaging/magnetic resonance cholangiopancreatography (MRI/MRCP) or hepatobiliary iminodiacetic acid (HIDA) scan. Pre-operative imaging with CT or MRI of the abdomen was performed in 76% of patients. This was done most often at the request of the surgical team as part of the pre-operative work-up for laparoscopic cholecystectomy in order to evaluate for the presence of abdominal wall varices, a recanalized umbilical vein or concurrent hepatoma. Screening endoscopy for evidence of portal hypertension is a valuable tool however was not pursued pre-operatively.

Intra-operative data recorded from the operative and anaesthesia records included the following: operative time, estimated blood loss, peri-operative transfusion requirement [within 1 week of operative date including packed red blood cells (pRBCs), platelets and fresh frozen plasma (FFP)] and the use of haemostatic devices employed. Furthermore, conversion to open cholecystectomy and reasons for doing so were also noted. Information regarding placement of additional laparoscopic ports including location, size and indication was obtained from the operative record.

Post-operative features that were gathered from the medical record included: intensive care unit (ICU) admission, length of hospital stay, presence of acute hepatic decompensation peri-operatively, 30-day mortality and hospital re-admission, and post-operative infectious complications.

Anaesthesia was induced with Diprivan, fentanyl, muscle relaxant and maintained with inhalation anaesthetics supplemented with muscle relaxant and fentanyl. Patients had routine monitoring, which included electrocardiogram, non-invasive systemic blood pressure, arterial oxygen saturation and end-tidal carbon dioxide concentration. Invasive monitoring of central venous pressure and systemic arterial pressure was determined based on the comorbidity factors associated with cirrhosis.

An open Hasson technique via an infra-umbilical incision was used to gain access to the peritoneal cavity after which other ports were placed under direct visualization with transabdominal illumination. Placement of the umbilical port was preferentially in the infraumbilical position to avoid the prevalent recanalized umbilical vein. The subxiphoid port was placed to the right of the midline in order to avoid any potential recanalized collaterals in the falciform ligament. If the left lateral segment appeared hypertrophic limiting visualization of the gallbladder neck, a fifth 5-mm port was placed in the left upper quadrant and an assistant lifted the left side of the liver with an instrument, such as a laparoscopic Kittner dissector (Ethicon EndoSurgery, Cincinnati, OH, USA). The gallbladder was decompressed as needed if dilated. Care was taken to avoid excessive traction during cephalad retraction of the gallbladder fundus. The triangle of Calot was dissected to identify the cystic duct and artery. If the triangle was too inflamed, then a prograde (dome down) dissection of the gallbladder was considered. A cholangiogram was performed as needed, but not uniformly. The cystic duct and artery were doubly clipped and ligated in the usual fashion. The gallbladder was removed with an endocatch bag though the umbilical port. Haemostasis of the gallbladder bed was achieved by TissueLink™ or Argon beam coagulation. All port site fascial defects were closed using an 0-Vicryl. Skin opening was closed with skin glue (e.g. Endermil).

Statistical analysis was used to calculate the mean, median and percentages. Analysis of variance (anova) was used to evaluate the difference in operative time. The Kruskal–Wallis test was used to evaluate differences in blood loss and length of hospital stay. Chi-square analysis was used to evaluate the differences in ascites and drain placement.

Results

Of a total 282 laparoscopic cholecystectomies, 68 were attempted in patients with liver cirrhosis. Demographics, pre-operative biochemistry and Child–Pugh score are provided in Table 1. The pre-operative indication for cholecystectomy is shown in Table 2. Several patients were seen on multiple occasions by the surgical team for refractory symptoms before undergoing cholecystectomy. Differential diagnosis of these patients should include spontaneous peritonitis if ascites is present. The retrospective nature of our report limits our ability to report the working differential diagnosis in this series of patients. This was usually in an effort to optimize their liver disease in conjunction with the hepatology medicine team.

Table 1.

Pre-operative characteristics of patients with cirrhosis undergoing laparoscopic cholecystectomy

Total no. of patients	68
Age, median and range in years
Gender	56 (range, 23–87)
Pre-operative lab values, median and range	35 male, 33 female
Albumin	3.8 g/dl (1.8–4.8 g/dl)
Total bilirubin	0.9 mg/dl (0.1–13.4 mg/dl)
AST	40 U/l (15–240 U/l)
ALT	40 U/l (15–486 U/l)
Alkaline phosphatase	124 IU/l (40–537 IU/l)
INR	1.1 (0.9–2.4)
Platelet	152 K/ml (51 K–416 K/ml)
AFP	5 ng/ml (2–39 ng/ml)
Child–Pugh's classification
Class A	47 (69%)
Class B	19 (28%)
Class C	2 (3%)
Pre-op abdominal CT/MRI	54 (76.1%)

INR, international normalized ratio; ALT, alanine transaminase.

Table 2.

Indications for laparoscopic cholecystectomy in patients with cirrhosis

Chronic cholecystitis	31 (46%)
Symptomatic cholecystitis	15 (22%)
Gallstone pancreatitis	7 (10%)
Gallbladder mass	5 (7%)
Biliary dyskinesia	3 (4%)
Acute cholecystitis	2 (3%)
Prior acute cholecystitis	2 (3%)
Pre-operative kidney transplant, cholelithiasis	2 (3%)
Prior common bile duct stone	1 (2%)

Intra-operative factors are shown in Table 3. Of the four patients (6%) that were converted to an open cholecystectomy, two of these patients had undergone prior failed attempts at cholecystectomy.

Table 3.

Intra-operative characteristics of laparoscopic cholecystectomy for patients with cirrhosis

Operative time, median and range in minutes	120 min (35–229 min)
Estimated blood loss, median and range in ml	25 ml (25–600 ml)
5th additional port placement	8/68 patients (11%)
TissueLink/argon beam coagulator	61/68 patients (90%)
Product transfusions (pRBCs, Plts, FFP)	4/68 patients (6%)
pRBC transfusion	1/68 patient (1%)
Intra-peritoneal drain placement	24/68 patients (35%)
Conversion to open	4/68 patients (6%)

pRBC, packed red blood cells.

There was no 30-day mortality or acute hepatic decompensation post-operatively in this series (Table 4). One patient developed a cystic stump bile leak that resolved with ERCP and stent placement. Other complications included urinary tract infection in two patients, pneumonia in one patient and one wound infection that occurred in a patient who had been converted to an open cholecystectomy and had had a concurrent urinary tract infection. Four patients were admitted to the ICU for post-operative monitoring. Median length of overall hospital stay was 1 day (range, 1–13 days). Eleven patients (16%) were re-admitted within 30 days of discharge, with the primary reasons for re-admissions being non-surgical-related morbidity.

Table 4.

Post-operative course after laparoscopic cholecystecomy for patients with cirrhosis patients

30-day mortality	0
Acute hepatic decompensation	0
Post-op bile leak	1/68 (1.5%)
Post-operative infectious complications	2 UTI, 1 pneumonia, 1 wound infection
ICU admission	4/68 (5.6%)
Length of hospital stay	1 day (1–13)
30-day re-admissions	11/68 (15.5%)

Based on laboratory and clinical factors, the total group of patients were classified as Child's class A (47 patients, 69%), class B (19 patients, 28%), or class C (2 patients, 3%). Using anova, significant differences according to Child–Pugh's classification (A, B, C) were found in operating room (OR) time (P= 0.01). Patients with Child's C had the longest OR time (mean = 163 min, SD = 17 min) when compared to patients with Child's B (mean 145 min, SD = 43 min) and Child's A (mean = 112 min, SD = 44 min). If one examines patients with Child's B and C collectively, the results remain significant when compared to Child's A patients (P= 0.004). Significant differences were also observed between the Child-Pugh's classifications and intra-operative blood loss (EBL) (Chi-Square = 14.5, P= 0.001). Child's A, B and C patients had a median blood loss of 39 ml, 158 ml, and 250 ml, respectively. Conversion to open cholecystectomy was significantly more common with increasing Child-Pugh's classification (P= 0.001). No Child's class A patient required conversion, while Child's B and C patients were converted to open 16% and 50% of the time, respectively.

Increasing Child-Pugh's classification was significantly associated with increasing length of hospital stay (LOS) (Chi-Square = 17.6, P= 0.001). Median length of stay for patients with Child's A, Child's B, and Child's C were 1.6, 3.4, and 7 days, respectively. If Child's B and C patients were combined, the results remain significant for EBL (P= 0.001) and LOS (P= 0.001).

Discussion

Patients with cirrhosis have a higher rate of developing gallstones compared with the general population.^1–5 At the advent of laparoscopy for cholecystectomies, cirrhosis was considered a contraindication. However, several series have since reported laparoscopic cholecystectomy in the setting of cirrhosis as safe and associated with acceptable morbidity and mortality.^11–19 Most of the series to date have originated from Asia or Europe. The present study represents the experience at a large tertiary centre in the United States.

In the present study, there was no mortality and only minimal morbidity associated with laparoscopic cholecystectomy in patients with cirrhosis. No patients developed acute liver failure and only a small percentage (7%) required post-operative ICU admission. Median length of stay was 1 day with all complications and 30-day readmission non-surgical related. As one would expect, with worsening liver disease, based on increasing Child–Pugh's classification, laparoscopic cholecystectomy was associated with longer operative time, more blood loss, higher conversion rate, more ascites, longer length of stay and a high rate of 30-day re-admission. However, in the present study, the complication rates were not significantly different with increasing Child–Pugh's classification.

Puggioni and Wong (2003) performed a meta-analysis from 17 studies with a total of 351 patients, of whom 75.5% were Child's class A, 21.4% were Child's class B and 1.7% were Child's class C, which similar ratios as seen in this series.²⁰ Laparoscopic cholecystectomy was performed electively in 83% of the patients and for emergent, acute cholecystitis in 17% of patients. The overall morbidity after a laparoscopic cholecystectomy in cirrhotic patients was 18.2% of patients and mortality occurred in 0.28% of patients. Compared with non-cirrhotic patients who underwent laparoscopic cholecystectomy, cirrhotic patients tended to have a higher rate of acute cholecystitis (47% vs. 15%, P < 0.001), longer operative time (98.2 min vs. 70 min, P= 0.005), higher conversion to open rate (7.1% vs. 3.6%, P= 0.02), higher complication rate (20.9% vs. 8%, P < 0.001) and more intra-operative bleeding (26.4% vs. 3.1%, P < 0.001). There was no significant difference in wound infection and mortality between cirrhotic and non-cirrhotic patients. In patients with cirrhosis who underwent laparoscopic versus open cholecystectomy, the laparoscopic group had a shorter operative time (123.3 vs. 150.4 min, P= 0.04), less blood loss (113 vs. 425.5 ml, P= 0.015) and shorter length of hospital stay (6 vs. 12.2 days, P < 0.001). There was no difference in morbidity (9.25%–15%), mortality (0–4.76%) or wound infection rates (0–0.13%). The authors concluded that although laparoscopic cholecystectomy in patients with cirrhosis was associated with higher morbidity compared with non-cirrhotic patients, the laparoscopic approach was also associated with shorter operative time, less blood loss and shorter length of hospital stay in patients with cirrhosis with symptomatic biliary disease compared with the open approach. Given the small number of patients with Child's class C cirrhosis, no conclusions could be made in this patient population.

In Puggioni and Wong's meta-analysis, laparoscopic cholecystectomies were performed for emergencies in 17% of patients with cirrhosis who presented with biliary disease.²⁰ This study series differs significantly as acute cholecystitis was present in only 3% of patients. For patients with advanced cirrhosis based on the Child–Pugh's score and acute cholecystitis, ICU admission for monitoring and observation and conservative management with fluid rescuscitation and intravenous antibiotics should be the initial step of management. If their clinical status worsens, coagulopathy should be corrected immediately with fresh frozen plasma and vitamin K and a cholecystectomy performed. One should also consider the selective use of a decompression cholecystostomy tube in patients who may not tolerate general anaesthesia.²¹ If there are concerns regarding collateral vessels in the gallbladder bed or in the hilum, then a laparoscopic subtotal cholecystectomy may be considered, although this was not utilized in our series.^22–24

Pre-operative imaging, such as a CT scan or MRI, was obtained in 76% of patients. This provided invaluable information regarding the presence or absence of abdominal varices which, in turn, was used to plan the port placement during the operation. In addition, it was used to screen for the potential presence of concurrent hepatoma, which might change the operative plan and has become part of the practice.

The TissueLink electrocautery or argon beam coagulator was used in 90% of patients in this series. Whether or not this practice directly impacts upon blood loss or need for peri-operative transfusions is not clear. However, the use of cautery (normal or Tissuelink) near the liver hilum and, most importantly, the bile duct, should be avoided if possible in order to prevent potential thermal injury. Tissuelink cautery may penetrate the liver parenchyma and therefore should not be used in hilum. The majority of techniques reported in the literature in laparoscopic cholecystectomy employ some form of haemostatic device besides the traditional hook cautery in order to dissect the gallbladder of the gallbladder fossa. This has helped minimize bleeding, which is a major cause of complications after laparoscopic cholecystectomy in patients with cirrhosis patients. In the meta-analysis by Puggioni and Wong (2003), 4.3% of patients developed post-operative liver bleeding and 2.8% of patients developed post-operative intra-abdominal haematoma. In the present study, the use of topical sealants (collagen fleece, cyano-acrylate and fibrin sealants) was not required but the use of the TissueLink or argon beam was crucial and sufficient in achieving haemostasis at the gallbladder fossa after removal of the gallbladder.

In the present study, an additional 5th port was placed in the left upper quadrant in 10% of the patients to help elevate the left lateral segment, which was hypertrophied and heavy in these patients with cirrhosis. This manoeuver helps prevent excessive traction on the gallbladder and prevent potential bleeding (Table 5). Palanivelu et al. (2006) placed additional ports in 53.6% of their patients to improve exposure to the hilum by either retracting the duodenum or omentum downward or to lift the liver upward.²²

Table 5.

Recommended approach to laparoscopic cholecystectomy in patients with cirrhosis (based on literature and/or personal experience

1)Pre-operative

a)Medically optimize the patient: control ascites, correct coagulopathy with fresh frozen plasma (for INR > 1.5) and platelets (<50k/mm3).

b)Obtain pre-operative imaging to help identify abdominal wall varices or a recanalized umbilical vein and to rule-out hepatoma.

c)Consider cholecystostomy tube in patients with Child class C.

2)Intra-operative

a)If unsuspected cirrhosis is identified during a planned routine laparoscopic cholecystectomy, depending on the experience of the surgeon and the facilities; consider closing the fascia and skin and transfering the patient to a tertiary medical center with critical care, hepatology, and hepatobiliary and transplant surgery support.

b)Use the open Hasson approach via an infra-umbilical incision to avoid periumbilical wall varices.

c)Transilluminate with the laparoscope to identify vascular collaterals or measure placement based on preoperative imaging.

d)Subxiphoid port to the right of midline to avoid potential collaterals in the falciform ligament.

e)Consider a 5th port in the left upper quadrant to help lift the left lateral segment, which may be hypertrophied and heavy in cirrhotic patients.

f)Avoid excessive traction on the gallbladder itself during cephalic retraction of the gallbladder fundus and dissection of the triangle of Calot.

g)Use a TissueLink or argon beam to cauterize the gallbladder fossa and help prevent bleeding – represents an additional expense to case.

h)Consider laparoscopic needle biopsy of the liver to evaluate extent of cirrhosis and assist hepatologists in further medical management.

i)Close all fascia port sites, including all 5 mm port sites. Close all wounds with skin glue (e.g. Indermil ® Tissue Adhesive) or alternatively a running suture method.

In the present study, only two patients (3%) who underwent cholecystectomy were Child's class C cirrhotics. In fact, there has been very little reported regarding this subject. In the meta-analysis by Puggioni et al. (2003), only 1.7% of the patients were Child's class C cirrhotics.²⁰ Curro et al. (2005) reported a 50% mortality (due to severe liver failure and sepsis) and 75% morbidity rate, with double the length of hospital stay in only 4 Child's class C cirrhotic patients who undergo laparoscopic cholecystectomy compared with patients with Child's class A or B cirrhosis.¹⁸ This study was also limited by the number of Childs C cirrhotic patients reported; however, they recommended that cholecystectomy should only be entertained in Child's class C cirrhotic patients who develop acute cholecystitis that does not respond to conservative medical management. Under non-emergent situations, patients with Child's class C cirrhosis should be medically managed and downstaged before undergoing a cholecystectomy. Liver transplantation evaluation and listing before cholecystectomy in Child's B or C patients is recommended. Fortunately, none of the patients in the present study suffered acute decompensation necessitating organ transplantation. Alternatively, Schlenker et al. (2006) recommended endoscopic gallbladder stenting as a therapeutic option in high-risk cirrhotic patients.²⁴ Gallbladder stenting was performed in 23 patients, 38% of whom were diagnosed with acute cholecystitis. Twenty of the 23 patients (87%) remained asymptomatic until transplantion, death or conclusion of the study period.

Conclusion

Laparoscopic cholecystectomy in patients with cirrhosis is feasible with no mortality and low morbidity rates in experienced hands. The present study demonstrates that laparoscopic cholecystectomy is safe for Childs A however Childs B patients should be considered cautiously and Childs C should be approached only as a last attempt in life-threatening and failing medical management. To minimize the morbidity and mortality, the study proposes a number of recommendations for the pre-operative, intra-operative and post-operative approach to laparoscopic cholecystetomy in this patient population.