Colorectal Surgery in Cirrhotics

Author

Year

Study design

Diagnosis

Patient number

Overall morbidity rate (%)

Overall mortality rate (%)

Anastomotic leak rate (%)

Metcalf [1]

1984

Retrospective, single-institution

Mixed

19^a

Wind [3]

1994

Retrospective, single-institution

Mixed

Gervaz [4]

2003

Retrospective, single-institution

Colorectal cancer

Martinez [5]

2004

Retrospective, single-institution

Mixed

Meunier [6]

2008

Retrospective, single-institution

Mixed

Nguyen [7]

2009

Administrative database

Mixed

4042

43–55^b

N/A

Ghaferi [8]

2010

Administrative database

Mixed

1565

N/A

Montomoli [9]

2013

Nationwide Cohort

Colorectal cancer

158

N/A

Sabbagh [10]

2016

Case-Matched, single-institution

Colorectal cancer

^aReported as “intrabdominal sepsis”.

^bMorbidity rates reported separately for patients with cirrhosis and cirrhosis associated with portal hypertensionGhaferi used the definition of “chronic liver disease” instead of cirrhosis (see text).

Earlier assessments of the relationship between CTP scores and mortality after a variety of abdominal operations indicated a proportional relationship between increased CTP score and mortality rates, which was approximately 10% for CTP A class, 31–31% for CTP B class, and 76–82% for CTP C class [14, 15]. However, subsequent assessments of CTP when specifically tested in colorectal surgery have been less consistent. In at least two single-institutional reports, CTP score was not statistically associated with mortality, nor the absolute mortality rates were highest among CTP C patients [4, 6], which might depend on patient selection and the greater impact of other relevant factors on postoperative outcomes. For example, in a series on 41 such patients, CTP was not significantly associated with either postoperative morbidity or mortality. the only independent factor significantly associated with mortality was the onset of postoperative infection, while preoperative ascites was the only factor independently associated with postoperative morbidity [6]. Although widely used to evaluate the severity of cirrhosis, the CTP score also carries significant inherent limitations. Two of the variables are very subjective, namely encephalopathy and ascites, and can also be affected by the use of diuretics and lactulose. In addition, all variables in CTP are weighted equally in the construction of the ultimate score. The MELD score should ideally be the alternative predictive model addressing such limitations. It is a mathematical formula based on the values of serum creatinine, bilirubin, and International Normalized Ratio (INR). The MELD score was initially developed as a statistical model predicting survival in patients undergoing TIPS [16], but was later prospectively validated as a predictive model for patients awaiting orthotopic liver transplantation (OLT) [12], and has subsequently become widely used as a predictive model of surgical risk associated with nontransplant surgery [17].

The MELD score has been shown to predict 1-year and 5-year mortality in a large cohort of cirrhotic patients having a variety of underlying liver disease etiologies and severity [18] and has also been specifically tested on patients undergoing colorectal surgery included in large administrative databases. In a study on over 10,000 patients included in the National Surgical Quality Improvement Program (NSQIP) database, almost 70% of the patients undergoing colorectal surgery had a MELD score of less than 10, which was associated with a perioperative mortality of 2.9%. However, an increased crude MELD score in the whole study population was associated with a substantially increased perioperative mortality. For example, a MELD score between 10 and 14 was associated with a perioperative mortality of almost 10%, while a MELD score between 15 and 19 was associated with a mortality rate exceeding 20%. The authors constructed multivariate regression models on the predicted probability of death, including a number of other relevant variables, stratified by the MELD score. When applying such models to elective procedures, the risk of death increased by 0.5% with each point increase in the MELD score up to 20, after which the risk of death increased by 1% for each additional score point increase. Even more direct was the association between mortality and MELD score following emergent procedures, in which the predicted morality increased by 1% with each MELD score point increase, up to a score of 15, after which each point score increase resulted in a 2% increase in the probability of death. As the authors suggested, while perhaps less sophisticated than other established models to predict perioperative mortality, the MELD score can be easily calculated and practically used to discuss with patients the risk of surgery [19]. It is notable that this study was not limited to patients with known liver disease.

When specifically looking at the adverse outcomes in patients with chronic liver disease, Ghaferi et al. examined 30,927 NSQIP patients undergoing colorectal resections, 1,565 of whom had chronic liver disease. The definition of chronic liver disease used was based on clinical findings indicative of liver dysfunction, namely ascites, esophageal varices and serum total bilirubin greater than 2 mg/dL. The MELD score was used for further risk stratification in the patient cohort thus defined. Patients with chronic liver disease had almost a 6.5-fold increased risk of mortality following colorectal operations. In particular, the mortality among patients without chronic liver disease was 3.2% versus 21.5% among those with chronic liver disease. The same group was also associated with a significantly increased risk of major complications by a relative risk factor of 2.72, based on the absolute percentages of 41.9% among patients with chronic liver disease vs. 15.4% among patients without liver disease. Mortality following major complications (failure to rescue rate) was also markedly increased among patients with chronic liver disease (relative risk 2.27, 34.2%, vs. 15.1%). Stratification of chronic liver disease patients by MELD score demonstrated a significantly higher rate of complications, failure to rescue and mortality in patients with chronic liver disease having a MELD score of 15 or greater compared with patients having a MELD score of less than 15 [8]. This threshold is close to the MELD score cutoff of 14, which has been proposed as a more accurate equivalent of the CTP C class as a predictor of very high risk status after intrabdominal surgery [20].

There is no conclusive evidence indicating the superiority of one predictive score over the other among cirrhotic patients undergoing surgery or other interventional procedures [21]. At this time either score should therefore be considered acceptable for risk stratification and to facilitate the preoperative discussion of patient prognosis.

Preoperative Transjugular Intrahepatic Portosystemic Shunt Placement

TIPS is a minimally invasive procedure, which accomplishes a diversion of the portal system blood flow into the hepatic veins. TIPS has been proposed as a method to correct portal hypertension and allow extrahepatic abdominal surgery in patients initially deemed inoperable due to their liver disease. In their seminal paper, Azoulay et al. reported on seven patients undergoing preoperative TIPS, also referred to as “neoadjuvant TIPS”, followed by a waiting period ranging from 1 to 5 months before abdominal surgery, during which the mean venous pressure gradient successfully decreased from 18 to 9 mmHg. Three out of the seven patients underwent colorectal surgery procedures (left hemicolectomy in two, Hartmann’s reversal in one), complicated by one death secondary to liver failure, 36 days after colonic resection [22]. More recently, Menahem et al. reported successful placement of TIPS in eight patients, 1–9 weeks prior to colorectal surgery (three with right colectomy, two with left colectomy, and three with procotectomy), which also resulted in a significant reduction of the hepatic portal venous pressure. There were two postoperative deaths, which occurred in the two patients with the highest pre-TIPS MELD scores [23]. A number of reports, generally including only few selected patients, have confirmed the feasibility of preoperative TIPS before colorectal surgery (Table 15.2). However, not all studies concur in supporting the preoperative use of TIPS to optimize patients undergoing extrahepatic surgery. In particular, Vinet et al. analyzed 18 patients, 10 of whom underwent colectomy, who received preoperative TIPS placement followed by elective extrahepatic surgery after a median interval of 72 ± 21 days. Patients undergoing preoperative TIPS placement were case-matched with a control group of 17 cirrhotic patients who underwent surgery without TIPS placement based on age, etiology of cirrhosis, indications for surgery, type of surgery and coagulation parameters. While the CTP score was significantly higher in the TIPS group (7.7 versus 6.2), no significant differences were observed with respect to operative blood loss, postoperative complications, duration of hospitalization, 30-day and 1-year mortality. A Cox proportional hazard model failed to demonstrate that either CTP score or TIPS placement had any significant association with survival. While the present study cautioning against the routine use of preoperative TIPS includes a relatively large number of patients, it is important to point out that despite the case-matched design, the CTP score was significantly decreased among patients who did not receive preoperative TIPS. In addition, the portal hepatic pressure gradient was not measured in all patients so that it is possible that portal hypertension was less severe in the control group. With a relatively limited number of events, even a multivariate analysis could fail to account for important differences in the severity of baseline cirrhosis between the two groups, favoring the control group [24]. The use of preoperative TIPS has also been reported among patient undergoing surgery for ulcerative colitis (UC). In a retrospective study on 50 patients with primary sclerosing cholangitis (PSC) requiring surgery for synchronous UC, 13 patients receiving preoperative TIPS were compared with the remaining 37. Duration of UC at the time of surgery, preoperative medical management, or indications for colectomy were similar between the two groups. Not surprisingly, the study group was associated with a significantly more severe liver disease and in particular increased MELD scores, longer aPTT, lower hemoglobin and platelet levels, and decreased mean albumin serum levels. Postoperative mortality only occurred in one patient treated with preoperative TIPS and overall morbidity was statistically comparable, albeit quite common (100% vs. 88% in the TIPS vs. control group, respectively). Patients undergoing preoperative TIPS suffered significantly increased rates of wound infection, wound dehiscence, blood transfusion, and experienced longer hospital stay (8 vs. 5 days, p = 0.041) and increased readmission rates (57% vs. 19%, p = 0.032). A total of 95% of the patients in the control group ultimately achieved restoration of the intestinal continuity compared with only 43% after preoperative TIPS. This study once more underscores the selection bias for preoperative TIPS toward patients with more severe disease, so that it is difficult to conclusively attribute any increase in postoperative complications mainly to the baseline disease severity or whether preoperative TIPS placement may have played an additional role [27].

Table 15.2

Studies on preoperative placement of TIPS prior to colorectal surgery

Author	Year	Study design	Patient number	Operations performed	TIPS-surgery interval	Perioperative morbidity n (%)	Perioperative mortality, n (%)
Azoulay [22]	2001	Retrospective	3	Left Hemicolectomy (2) Hartmann’s Reversal (1)	1–5 months^a	1(33)	1(33)
Vinet [24]	2006	Case-matched	18	Colectomy	72 ± 21 days^b(mean)	13^b(72)	2^b(11)
Kim [25]	2009	Retrospective	3	Subtotal Colectomy (2) Total Colectomy (1)	33,12 and 46 days	2 (67)	0
Menahem [23]	2014	Retrospective	8	Right hemicolectomy (2) Left hemicolectomy (3) Proctectomy (3)	1–9 weeks	6(75)	2(25)
Kochar [26]	2014	Retrospective Case Control	9	Colectomy for UC with PSC	2.3 months (median)	8 (89)	1 (11)

^aInterval based on eight procedures as a whole, including the three colorectal operations

^bData based on colectomy and other procedures, namely antrectomy (n = 5), pancreatectomy (n = 1), small bowel resection (n = 1), and nephrectomy (n = 1)

Surgery for Inflammatory Bowel Disease

In an analysis from the Nationwide Inpatient Sample from 1988 to 2006 to determine the frequency of chronic liver disease among patients with IBD and their in-hospital outcomes, the age-adjusted rate of chronic liver disease among hospitalized patients with IBD increased from 4.35 per 100,000 persons during the period between 1988 and 2001 to 7.45 per 100,000 persons in the period between 2004 and 2006. Not surprisingly, the presence of cirrhosis was a significant factor associated with mortality, even among patients receiving medical treatment without undergoing surgery. In particular, there was more than a twofold increase in mortality in patients with concomitant IBD and liver disease compared to other patients with IBD. Not surprisingly, the most common etiology of chronic liver disease was PSC, with over 50% of the cases among patients with UC and 26.5% among patients with Crohn’s disease [28]. PSC is a progressive disease associated with the development of cholestasis, which predisposes to an increased risk of cholangiocarcinoma and ultimately leads to end-stage liver disease. Patients with UC have a 2.5–10% risk of synchronous PSC [29–31], while up to 90% of PSC patients have underlying UC [32–36]. The association between Crohn’s disease and PSC is less well studied, but at least one study reported that the prevalence of PSC among patients with Crohn’s disease was 3.4% [37], while the prevalence of Crohn’s disease among PSC patients has been estimated at 16% [38]. With respect to implications for the colon and rectum, PSC is also associated with an increased risk of colorectal dysplasia, progression to colorectal cancer, and pouchitis among patients undergoing ileal pouch-anal anastomosis (IPAA) for UC. In a study comparing 65 patients with PSC and IBD undergoing restorative proctocolectomy vs. 260 IBD patients without associated PSC undergoing the same operation, Gorgun and colleagues identified an increased incidence of cancer and dysplasia in the resected specimen among patients with concurrent PSC, an increased risk of postoperative pelvic sepsis and a higher long-term mortality (35% versus 4%, p < 0.001) after a mean follow-up of 68 months, which in the majority of cases was related to liver disease. The 5-year survival in the synchronous PSC group was also significantly decreased, while no significant differences were noted in functional and quality of life results between the two groups. These findings indicate that despite the inherently increased risks, patients with UC and synchronous PSC do benefit from IPAA [39].

Only gold members can continue reading. Log In or Register to continue