- Email: [email protected]
A Greater Proportion of Liver Transplant Candidates Have Colorectal Neoplasia Than in the Healthy Screening Population
Clinical Gastroenterology and Hepatology 2015;13:956–962
A Greater Proportion of Liver Transplant Candidates Have Colorectal Neoplasia Than in the Healthy Screening Population Philip Jeschek,*,‡ Arnulf Ferlitsch,*,‡ Petra Salzl,*,‡ Georg Heinze,§ Georg Györi,k Karoline Reinhart,*,‡ Elisabeth Waldmann,*,‡ Martha Britto-Arias,*,‡ Michael Trauner,* and Monika Ferlitsch*,‡ *Department of Internal Medicine III, Division of Gastroenterology and Hepatology, §Section for Clinical Biometrics, Center for Medical Statistics, Informatics and Intelligent Systems, kDepartment of Surgery, Division of Transplantation, Medical University of Vienna, ‡Quality Assurance Working Group of the Austrian Society of Gastroenterology and Hepatology, Vienna, Austria BACKGROUND & AIMS:
Various types of liver disease are associated with an increased prevalence of colorectal adenomas. We investigated whether cirrhosis is a risk factor for colorectal neoplasia by analyzing colonoscopy findings from 2 cohorts of patients awaiting liver transplantation.
METHODS:
We performed a retrospective analysis to compare findings from colorectal cancer screenings of 567 adult patients with cirrhosis placed on the waitlist for liver transplantation with those from controls (matched for age, sex, body mass index, smoking, and diabetes). Rates of adenoma and advanced adenoma detection were adjusted owing to differences in rates of polypectomies performed in the 2 cohorts.
RESULTS:
Adenomas were detected in a significantly higher percentage of patients with cirrhosis (29.3%) than in controls (21.5%) (P [ .0057; relative risk [RR], 1.36; 95% confidence interval [CI], 1.09– 1.69); and patients with cirrhosis had a higher rate of advanced adenoma detection than controls (13.9% vs 7.7%; P [ .0015; relative risk, 1.82; 95% CI, 1.25–2.64). A greater percentage of patients with alcoholic cirrhosis had neoplasias (34.3%) than controls (25.3%; P [ .0350; RR, 1.36), and rates of advanced adenoma detection were 16.7% vs 10.2% (P [ .0409; RR, 1.63). Adenomas were detected in 27.8% of patients with viral cirrhosis vs 15.9% of controls (P [ .0061; RR, 1.74), with rates of advanced adenoma detection of 13.6% vs 5.0% (P [ .0041; RR, 2.73). Similar proportions of patients with cirrhosis of other etiologies and controls were found to have colorectal neoplasias.
CONCLUSIONS:
Based on a retrospective analysis of colonoscopy findings from patients awaiting liver transplantation, those with alcoholic or viral cirrhosis are at higher risk of developing colorectal neoplasia and should be considered for earlier colonoscopy examination.
Keywords: Colon Cancer; Screening Colonoscopy; Early Detection.
n patients with advanced cirrhosis, colonoscopy is performed as part of the evaluation to assess eligibility for orthotopic liver transplantation (OLT) to rule out malignancies that would exclude patients from receiving a graft. Although there are indications that immunosuppression may shorten the interval of progression from premalignant lesions to colorectal cancer (CRC),1–3 there is evidence that CRC risk after OLT is increased only in patients with concomitant primary sclerosing cholangitis because of the strong association with inflammatory bowel disease.4 Because of shared risk factors such as obesity, diabetes, and metabolic syndrome, patients with nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH) are at higher risk of colorectal
I
neoplasia, both adenomas and carcinomas, which remained after adjustment for risk factors.5,6 Although Wong et al5 found a higher detection rate of advanced neoplasia (45%) in the right-sided colon of NAFLD patients of both sexes, another study6 found comparatively more adenomas in the rectum of male, and in the proximal colon of female, NAFLD patients. Abbreviations used in this paper: AADR, advanced adenoma detection rate; ADR, adenoma detection rate; ALCI, alcoholic cirrhosis; BMI, body mass index; CI, confidence interval; CRC, colorectal cancer; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; OLT, orthotopic liver transplantation; PDR, polyp detection rate; RR, relative risk. © 2015 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2014.08.018
May 2015
Patients who are at high risk of developing CRC should undergo earlier screening colonoscopy with closer monitoring at the age of 40, and in some cases even earlier.7 Cirrhosis itself has not been considered an independent risk factor for colorectal neoplasia to date and therefore is not included in CRC screening guidelines. Only a few studies have examined the prevalence of such lesions in OLT candidates, but were restricted either to relatively small cohorts, sigmoidoscopy only, or lacked a proper control group.8–11 Therefore, the aims of our study were to evaluate cirrhosis and its different etiologies as risk factors for colorectal neoplasia in liver transplant candidates and to determine if those patients with end-stage liver disease would benefit from earlier referral to colonoscopy.
Methods Study Design and Patients Our study was designed as a retrospective cohort observational study. Two cohorts were examined. The first cohort included cirrhotic patients who were listed for liver transplantation from 2000 to early 2012, in whom colonoscopy was performed as part of the evaluation for liver transplantation eligibility before placement on the waitlist of organ recipients. Data from 567 patients in the cohort of cirrhotic patients listed for OLT were provided by the Division of Transplantation of the Department of Surgery and the Division of Gastroenterology and Hepatology of the Department of Internal Medicine III at the Medical University of Vienna (Vienna, Austria). The second cohort was composed of general screening population subjects who underwent a screening colonoscopy within the Certificate of Quality Assurance for Colorectal Cancer Screening program.12 A total of 80,793 patients were screened within that program, of whom 19,904 also attended a health check reporting data on body mass index (BMI), diabetes, and smoking. A total of 567 of those patients were matched for confounding variables (namely age, sex, BMI, smoking, and diabetes) with the other cohort by randomly selecting 1 match for each cirrhotic patient out of the pool of matched patients of the screening population using SAS (2011; SAS Institute, Inc, Cary, NC). The majority of patients were at least 50 years old, and ascribed to current recommendations to perform screening colonoscopy starting at the age of 50. Advanced adenomas were defined as adenomas containing villous or tubulovillous features, as tubular adenomas of at least 1 cm in diameter, or as high-grade dysplasia. Because of liver disease–associated coagulopathy not all polyps could be resected, however, we assumed that the proportion of adenomas in polyps that did not undergo a biopsy were comparable with the ones that did undergo a biopsy. In this study, we therefore calculated and used an adjusted adenoma detection rate
Liver Transplant and Colorectal Neoplasia
957
and an adjusted advanced adenoma detection rate as described in the Statistical Analysis section.
Statistical Analysis Both cohorts were described by sex, age, BMI, smoking, and diabetes; transplant candidates also were described by etiology and stage of cirrhosis. Categoric variables were described by absolute and relative frequency, and continuous variables were described by medians and 25th and 75th percentiles. Primary outcome measures were the adenoma detection rate (ADR) and the advanced adenoma detection rate (AADR), which were adjusted owing to a variable rate of polypectomies performed in the 2 cohorts: the reason being an increased bleeding tendency in cirrhotic patients as a result of severe coagulopathy. We assumed the same rate of adenomas among resected and nonresected polyps. Thus, we computed a polypectomyadjusted ADR as follows: adjusted ADR ¼ ADR ðresected polyps þ nonresected polypsÞ resected polyps Those adjusted rates were compared between the 2 cohorts and further stratified into age groups for complementary age-specific analyses by the calculation of an adjusted relative risk (RR) with a 95% confidence interval and by the use of a chi-square test. Within the cohort of cirrhotic patients, the polyp detection rate (PDR), ADR, and AADR were compared between the different etiologies of cirrhosis (alcoholic, viral, and other causes). Secondary outcome measures were the number (1, 2–4, >4) and localization of polyps (proximal colon, sigmoid/rectum, and distal and proximal colon), as well as the cecal intubation, sedation, and complication rates, although for the latter we only had data for the cohort of transplant candidates. Variables mentioned earlier were compared between the 2 cohorts by the calculation of absolute and relative frequencies and by the use of chi-square tests. The significance level was considered 5%. All statistical analyses were conducted as 2-sided tests using the SPSS 19.0 (2010, IBM Corp, Armonk, NY) and Microsoft Excel 2007 (2007, Microsoft Corp, Redmond, WA) software packages. The study protocol was approved by the ethics commission of the Medical University of Vienna in February 2012 (EK063/2012).
Results Patient Characteristics As shown in Table 1, a total of 567 patients in each cohort, adjusted for age, sex, BMI, smoking, and diabetes,
958
Jeschek et al
Clinical Gastroenterology and Hepatology Vol. 13, No. 5
Table 1. Patient Characteristics Patient characteristics
Cirrhotic patients
Total 567 Male sex 410 (72.3%) Median age at screening, y 54 Median BMI 25.5 Smokers 248 (43.7%) Diabetic subjects 129 (22.8%) Cirrhosis stage Median MELD score 15.6 Child A 61 (10.8%) Child B 213 (37.7%) Child C 291 (51.5%) Indication for OLT/etiology of cirrhosis Alcoholic cirrhosis 277 (48.9%) Viral cirrhosis 206 (36.3%) Hepatocellular carcinoma 145 (25.6%) Biliary duct diseases 44 (7.8%) Other etiologies 64 (11.3%)
Healthy controls 567 410 (72.3%) 54 25.7 221 (39.0%) 97 (17.1%)
MELD, model for end-stage liver disease.
were included. Their median age was 54 years and there were 157 women (27.7%) and 410 men (72.3%) in both cohorts, with a median BMI of 25.5 in cirrhotic patients and 25.7 in screening patients. Among cirrhotic patients, 43.7% were smokers and 22.8% had diabetes compared with 39.0% and 17.1%, respectively, in healthy controls. Cirrhotic patients presented with a median model for end-stage liver disease score of 15.6 by the time of colonoscopy, and Child–Pugh score was stage A in 10.8%, stage B in 37.7%, and stage C in 51.5%. Of the 567 cirrhotic patients with one or multiple indications for transplantation, cirrhosis as a result of alcohol abuse was diagnosed in 277 (48.9%), viral hepatitis was diagnosed in 206 (36.3%) (among whom 165 [80.1%] had hepatitis C virus, 29 [14.1%] had hepatitis B virus, and 12 [5.8%] had both types of viruses), hepatocellular carcinoma and cirrhosis were diagnosed in 145 (25.6%), and biliary duct diseases were diagnosed in 44 (7.8%) patients. In 64 (11.3%) patients there were other causes for cirrhosis.
polyps, compared with 18.7% (P ¼ .0017) of controls, had polyps proximal to the sigmoid. In cirrhotic patients only 40.2% of lesions were found in the sigmoid colon and/or rectum compared with 57.4% in the screening cohort (P ¼ .0002). A total of 28.3% of cirrhotic patients vs 23.9% of controls had polyps both in the distal and proximal parts of the colon (P ¼ .29).
Number of Polyps Among cirrhotic patients with endoscopically detected colorectal polyps 16.2% had more than 4 polyps, 42.6% had 2 to 4 polyps, and 41.1% had only 1 polyp. Healthy screening patients with colonic polyps had more than 4 polyps in 6.7%, 2 to 4 polyps in 42.6%, and only a single polyp in 50.7%.
Histologic Findings The ADR was significantly higher in cirrhotic patients compared with the screening cohort (29.3%, CI, 25.0–33.9; vs 21.5%; CI, 18.2–25.2; P ¼ .0057) as shown in Figure 1 and Table 2. For the AADR the difference was even higher (13.9%; CI, 10.8–17.6; vs 7.7%; CI, 5.6–10.2; P ¼ .0015). Therefore, the RR for adenomas was 1.36 (CI, 1.09–1.69) and 1.82 (CI, 1.25–2.64) for advanced adenomas. Transplant candidates aged 40 to 49, 50 to 59, and 60 to 69 years all had a higher ADR and AADR than screening subjects, but the difference did not reach statistical significance because the number of patients in each subgroup was too low. Cirrhotic patients aged 40 to 49 years, for instance, had an ADR of 21.8% vs 13.2% (P ¼ .0913), and an AADR of 12.0% vs 7.7% (P ¼ .29) compared with controls of the same age group.
Macroscopic Findings Among 567 liver transplant candidates endoscopic polyps were found in 46.7% (n ¼ 265; confidence interval [95% CI], 42.6%–50.9%) compared with 36.9% (n ¼ 209; CI, 32.9%–41.0%) in the screening cohort (P ¼ .0007; RR, 1.27; 95% CI, 1.10–1.46).
Localization of Macroscopic Lesions Isolated endoscopic lesions proximal to the sigmoid were significantly more frequent in cirrhotic patients. A total of 31.4% of cirrhotic patients with endoscopic
Figure 1. PDRs, adjusted ADRs, and adjusted AADRs in cirrhotic patients compared with healthy controls matched for age, sex, body mass index, smoking, and diabetes. Data are shown as means.
May 2015
Liver Transplant and Colorectal Neoplasia
959
Table 2. PDRs, ADRs, and AADRs in Cirrhotic Patients and Controls Overall Cirrhotic patients
Healthy controls
Diagnosis
Rate (%)
CI (%)
Rate (%)
CI (%)
P value
RR
Endoscopic polyps Adenomas, adjusted rates Advanced adenomas, adjusted rates
46.74 29.27 13.93
42.57–50.94 24.98–33.86 10.77–17.59
36.86 21.53 7.66
32.88–40.98 18.16–25.21 5.58–10.22
.0007 .0057 .0015
1.27 1.36 1.82
Sex-Specific Rates for Colorectal Neoplasia In both sexes colorectal neoplasia detection rates were higher in cirrhotic patients than in controls (Table 3). Among females, cirrhotic women had a higher PDR (42.7% vs 28.0%; P ¼ .0066) than female controls. The ADR was 27.9% in cirrhotic vs 16.7% in healthy women (RR, 1.67; P ¼ .0257) and 11.5% vs 6.0%, respectively, for the AADR (P ¼ .11; RR, 1.91). In cirrhotic men the PDR was 48.3% compared with 40.2% in healthy male screening subjects (P ¼ .0203). The ADR was 29.8% vs 23.4%, respectively (P ¼ .0572; RR, 1.27), and only the AADR was significantly higher in cirrhotic men (14.9% vs 8.3%, respectively; P ¼ .0061; RR, 1.79).
Association Between Etiologies of Cirrhosis and Colorectal Neoplasia Among all etiologies of cirrhosis, in patients with alcoholic cirrhosis (ALCI), colorectal neoplasia rates were highest (Figure 2). ALCI patients had a higher prevalence of adenomas (34.3% vs 25.3%; P ¼ .0350; RR, 1.36) and advanced adenomas (16.7% vs 10.2%, P ¼ .0409; RR, 1.63) than matched controls from the screening cohort. Among men, male ALCI patients had a higher rate of adenomas (35.1% vs 25.3%; P ¼ .0408; RR, 1.39) and advanced adenomas (18.5% vs 10.0%; P ¼ .0190; RR, 1.85) than matched screening subjects. Also in female ALCI patients we found a nonsignificantly higher rate of adenomas than in matched controls (31.4% vs 25.7%; P ¼ .54; RR, 1.22), whereas detection rates for advanced
adenomas were comparable between those 2 quite small subgroups (9.7% vs 11.0%; P ¼ .90; RR, 0.88). Patients with cirrhosis caused by viral hepatitis C and hepatitis B had more adenomas (27.8% vs 15.9%; P ¼ .0061; RR, 1.74) and advanced adenomas (13.6% vs 5.0%; P ¼ .0041; RR, 2.73) than matched screening subjects (Figure 3). Among men, male patients with viral cirrhosis had a higher rate of adenomas (27.7% vs 17.7%; P ¼ .0440; RR, 1.56) and nonsignificantly more advanced adenomas (12.7% vs 6.3%; P ¼ .0655; RR, 2.00) than matched screening subjects. In addition, in female patients with viral cirrhosis we found a significantly higher rate of adenomas (27.9% vs 9.3%; P ¼ .0314; RR, 3.00) and advanced adenomas (16.7% vs 0.0%; P ¼ .0180) than in matched screening subjects. Cirrhotic patients with entities other than alcoholic and viral cirrhosis had no significant variations in colorectal neoplasia rates compared with matched controls. Adenoma rates in this subgroup were 20.1% vs 19.0% (P ¼ .84; RR, 1.06), and advanced adenoma rates were 8.5% vs 5.6% (P ¼ .44; RR, 1.46) in matched screening subjects.
Cecal Intubation, Sedation, and Complication Rate In screening patients the cecum was reached in 96.6% of colonoscopies compared with 93.6% in cirrhotic patients (P ¼ .0181). Regarding the sedation rate, 84.3% of screening patients were sedated with propofol and/or midazolam during the colonoscopy procedure vs 65.3% of transplant candidates (P < .0001). Postinterventional
Table 3. PDRs, ADRs, and AADRs in Men and Women Cirrhotic patients Diagnosis by sex Men Endoscopic polyps Adenomas, adjusted rates Advanced adenomas, adjusted rates Women Endoscopic polyps Adenomas, adjusted rates Advanced adenomas, adjusted rates
Healthy controls
Rate (%)
CI (%)
Rate (%)
CI (%)
P value
RR
48.29 29.77 14.88
43.36–53.25 24.67–35.27 11.07–19.41
40.24 23.39 8.30
35.46–45.17 19.32–27.87 5.78–11.46
.0203 .0572 .0061
1.20 1.27 1.79
42.68 27.90 11.49
34.83–50.81 20.16–36.76 6.43–18.53
28.03 16.68 6.01
21.16–35.74 11.10–23.64 2.78–11.09
.0066 .0257 .1063
1.52 1.67 1.91
960
Jeschek et al
Figure 2. PDRs, adjusted ADRs, and adjusted AADRs in patients with alcoholic cirrhosis compared with healthy controls matched for age, sex, body mass index, smoking, and diabetes. Data are shown as means.
hemorrhage in 1.4% of cirrhotic patients was the only complication that occurred during colonoscopy and was controlled endoscopically in all cases.
Discussion This study investigated colonoscopy outcomes in a large cohort of patients with cirrhosis under evaluation for liver transplantation, and it compared the prevalence rates of colorectal neoplasia with age- and sex-matched controls that were adjusted for BMI, smoking, and diabetes. We evaluated the colonoscopy findings of 567 patients listed for liver transplantation and found a significantly higher rate of colorectal adenomas and
Figure 3. PDRs, adjusted ADRs, and adjusted AADRs in patients with viral cirrhosis compared with healthy controls matched for age, sex, body mass index, smoking, and diabetes. Data are shown as means.
Clinical Gastroenterology and Hepatology Vol. 13, No. 5
advanced adenomas than in healthy controls. On closer examination, alcoholic cirrhosis constituted the strongest predictor of finding colorectal neoplasia in this study. These findings are consistent with the conclusion published by the International Agency for Research on Cancer, which stated that alcohol consumption is related to CRC.13 A meta-analysis of 27 cohort and 34 casecontrol studies estimated relative risks for alcohol consumption leading to CRC, which were 1.21 for moderate and 1.52 for heavy drinking (RR for adenomas among patients with ALCI in our study was 1.36 and was 1.63 for advanced adenoma) and concluded that there is strong evidence for an association between drinking more than 1 alcoholic drink per day and an increased colorectal cancer risk.14 Under those circumstances it is uncertain whether ALCI was partly causative by itself, or just a significant predictor of the presence of colorectal neoplasia indicating a history of alcohol consumption. The only study comparing ALCI patients and noncirrhotic alcoholic subjects with regards to colorectal neoplasia rates was a French cross-sectional prevalence study in 1992. Among 100 ALCI patients, colorectal adenoma was found in 58%, whereas among 100 alcoholic patients without cirrhosis colorectal adenoma was detected by colonoscopy in only 26% (P < .001). However, in this study patients were not sex- and age-matched and the mean duration of reported alcohol abuse in ALCI patients was 5 years longer, which limits those findings.15 Furthermore, our study also found higher rates of colorectal neoplasia in cirrhotic patients with hepatitis C virus and hepatitis B virus infection as the underlying etiology of cirrhosis. Some viral infections such as human immunodeficiency virus and subtypes of polyomaviruses including John Cunningham (JC) virus have been connected with the facilitation of colorectal carcinogenesis, however, data on the impact of viral infections are controversial.16,17 Interestingly, it has been shown that chronic hepatic viral infection might be protective for the occurrence of liver metastases in patients with CRC.18 However, there is no evidence supporting an association between hepatitis C virus/ hepatitis B virus infection and colorectal neoplasia, and the high prevalence of colorectal neoplasia in patients with viral cirrhosis in our study might be owing to the chronic inflammatory state in those patients. After age stratification we found higher rates for colorectal neoplasia in all age groups between 40 and 69 years in cirrhotic patients, although those results did not reach statistical significance. A recent study performed by Weismüller et al8 prospectively evaluated 243 OLT candidates who underwent colonoscopy and found neoplastic polyps in 21% of patients, but did not compare those findings with age- and sex-matched controls. Nevertheless, they compared their findings with 2 general population–based studies and found a higher rate of neoplastic polyps in cirrhotic patients aged 40 to 49 (14.7% vs 8.7%–9.5%) and 50 to 59 years (23.4% vs 14.9%) than in healthy patients of similar ages, whereas
May 2015
rates of advanced neoplasia were comparable. The higher rates of colorectal neoplasia in our study might be explained by significant variations in the patient cohorts because our study included more males (72.3% vs 63%) and more patients with alcoholic cirrhosis (48.6% vs 29.2%). Because of the nature of our study we did not involve patients with nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH) without cirrhosis, however, it is worth mentioning that Wong et al5 found a significantly higher prevalence of colorectal neoplasia in such patients. In patients with NAFLD they found a higher rate of adenomas (34.7% vs 21.5%) and advanced adenomas (18.6% vs 5.5%) than in the healthy control cohort, comparable with the colorectal neoplasia rates we found among ALCI patients. The higher prevalence of neoplasia in NASH but not in simple steatosis remained after adjusting for demographic and metabolic factors, however, because the study was retrospective we were only able to adjust for sex, age, BMI, smoking, and diabetes, but not for alcohol consumption in our study. It has been confirmed in large general population–based studies that men are more likely to have colorectal neoplasia than women.12,19 In our study we still found higher rates of colorectal neoplasia in cirrhotic men than in women (29.8% vs 27.9% for adenomas, 14.9% vs 11.5% for advanced adenomas), however, the difference between cirrhotic men and women was not as drastic as between men and women in routinely screened general population subjects (24.9% vs 14.8% for adenomas, 8.0% vs 4.7% for advanced adenomas).12 These findings indicate that an increased risk for colorectal neoplasia in cirrhotic patients, and especially in patients with ALCI and viral cirrhosis, is more obvious in women than in men. In consideration of full colonoscopy still being controversial as the screening method of choice, we evaluated the localization of endoscopic polyps and found that a large proportion of cirrhotic patients had lesions proximal to the sigmoid colon or in proximal as well as distal parts of the colon (60%) compared with 43% of controls. In NAFLD patients, Wong et al5 found a higher prevalence of advanced neoplasia in the rightsided colon whereas Stadlmayr et al6 only found more adenomas in the proximal colon of female, but not of male, patients. In our cirrhotic cohort, only 1.4% had a postinterventional hemorrhage that could always be controlled endoscopically. A study by Jeon et al20 confirmed our finding that polypectomy was comparably safe in patients with cirrhosis, probably also because procoagulatory and anticoagulatory factors are rebalanced in most cirrhotic patients, which corresponds to low postinterventional bleeding rates.21,22 We therefore propose performing polypectomy in patients with balanced hemostatic markers, because a post-transplantation condition increases the risk for colorectal cancer (odds ratio, 1.4),
Liver Transplant and Colorectal Neoplasia
961
which would increase even more in patients with underlying alcoholic liver disease (odds ratio, 2.3).23 Limitations included the retrospective nature of this study, which did not allow matching for all potentially confounding factors such as alcohol consumption; however, patients were adjusted for sex, age, BMI, smoking, and diabetes. A study with potential transplant candidates usually does not allow conclusions about colorectal cancer rates because patients diagnosed with colorectal cancer are not evaluated further for liver transplantation. Furthermore, our data set only contained data on the number and localization of endoscopic polyps overall, including hyperplastic and neoplastic polyps, so we could not differentiate between adenomas and hyperplastic polyps regarding number and localization, and only described those characteristics for endoscopic polyps overall. Moreover, in some patients polypectomy or biopsy was not performed owing to severe coagulopathy, and in another small portion of patients polyps could not be retrieved with the endoscope. Nevertheless, our data on adenomas and advanced adenomas were adjusted for those missing findings. Although bowel preparation and withdrawal time both are relevant factors that ensure the quality of colonoscopies, data on those factors could not be provided in our study. In conclusion, patients with alcoholic and viral cirrhosis had a higher prevalence of colorectal neoplasia than matched controls. Although colonoscopy is not mandatory in every pre–liver transplant evaluation system, we clearly would recommend performing it in patients 40 years or older, and especially in those with alcoholic and viral cirrhosis. Because of the high prevalence of colorectal neoplasia in those patients it seems reasonable to start colorectal cancer screening earlier than recommendations in general population–based guidelines suggest, regardless of whether patients are potential transplant recipients or not.
References 1. Nicolaas JS, de Jonge V, Steyerberg EW, et al. Risk of colorectal carcinoma in post-liver transplant patients: a systematic review and meta-analysis. Am J Transplant 2010;10:868–876. 2. Buell JF, Papaconstantinou HT, Skalow B, et al. De novo colorectal cancer: five-year survival is markedly lower in transplant recipients compared with the general population. Transplant Proc 2005;37:960–961. 3. Selgrad M, Koornstra JJ, Fini L, et al. JC virus infection in colorectal neoplasia that develops after liver transplantation. Clin Cancer Res 2008;14:6717–6721. 4. Singh S, Varayil JE, Loftus EV Jr, et al. Incidence of colorectal cancer after liver transplantation for primary sclerosing cholangitis: a systematic review and meta-analysis. Liver Transpl 2013;19:1361–1369. 5. Wong VW-S, Wong GL-H, Tsang SW-C, et al. High prevalence of colorectal neoplasm in patients with non-alcoholic steatohepatitis. Gut 2011;60:829–836. 6. Stadlmayr A, Aigner E, Steger B, et al. Nonalcoholic fatty liver disease: an independent risk factor for colorectal neoplasia. J Intern Med 2011;270:41–49.
962
Jeschek et al
7. Qaseem A, Denberg TD, Hopkins RH, et al. Screening for colorectal cancer: a guidance statement from the American College of Physicians. Ann Intern Med 2012;156:378–386. 8. Weismüller TJ, Bleich F, Negm AA, et al. Screening colonoscopy in liver transplant candidates: risks and findings. Clin Transplant 2013;27:E161–E168. 9. Zaman A, Hapke R, Flora K, et al. Prevalence of upper and lower gastrointestinal tract findings in liver transplant candidates undergoing screening endoscopic evaluation. Am J Gastroenterol 1999;94:895–899. 10. Díaz-Sánchez A, Núñez-Martínez O, González-Asanza C, et al. [Results of colonoscopic examination in the evaluation of cirrhotic candidates for liver transplantation]. Gastroenterol Hepatol 2009;32:327–333. 11. Selingo JA, Herrine SK, Weinberg DS, et al. Role of screening colonoscopy in elective liver transplantation evaluation. Transplant Proc 1997;29:2506–2508. 12. Ferlitsch M. Sex-specific prevalence of adenomas, advanced adenomas, and colorectal cancer in individuals undergoing screening colonoscopy. JAMA 2011;306:1352. 13. World Cancer Research Fund/American Institute for Cancer Research. Food, nutrition, physical activity, and the prevention of cancer: a global perspective. Washington, DC: AICR, 2007. 14. Fedirko V, Tramacere I, Bagnardi V, et al. Alcohol drinking and colorectal cancer risk: an overall and dose-response metaanalysis of published studies. Ann Oncol 2011;22:1958–1972. 15. Naveau S, Chaput JC, Bedossa P, et al. Cirrhosis as an independent risk factor for colonic adenomas. Gut 1992;33: 535–540. 16. Lin OS. Acquired risk factors for colorectal cancer. Methods Mol Biol 2009;472:361–372.
Clinical Gastroenterology and Hepatology Vol. 13, No. 5 17. Giuliani L, Ronci C, Bonifacio D, et al. Detection of oncogenic DNA viruses in colorectal cancer. Anticancer Res 2008;28:1405–1410. 18. Zeng Y, He X-W, He X-S, et al. Chronic hepatic viral infection could be a protective factor for colorectal cancer liver metastases: analysis in a single institute. Hepatogastroenterology 2013;60:37–41. 19. Regula J, Rupinski M, Kraszewska E, et al. Colonoscopy in colorectal-cancer screening for detection of advanced neoplasia. N Engl J Med 2006;355:1863–1872. 20. Jeon JW, Shin HP, Lee JI, et al. The risk of postpolypectomy bleeding during colonoscopy in patients with early liver cirrhosis. Surg Endosc 2012;26:3258–3263. 21. Lisman T, Leebeek FWG. Hemostatic alterations in liver disease: a review on pathophysiology, clinical consequences, and treatment. Dig Surg 2007;24:250–258. 22. Ferlitsch M, Reiberger T, Hoke M, et al. Von Willebrand factor as new noninvasive predictor of portal hypertension, decompensation and mortality in patients with liver cirrhosis. Hepatology 2012;56:1439–1447. 23. Schrem H, Kurok M, Kaltenborn A, et al. Incidence and long-term risk of de novo malignancies after liver transplantation with implications for prevention and detection: cancer After LT: incidence, prevention, detection. Liver Transplant 2013;19:1252–1261.
Reprint requests Address requests for reprints to: Monika Ferlitsch, MD, Division of Gastroenterology and Hepatology, Medical University of Vienna, Waehringer Guertel 18-20, A-1090 Vienna, Austria. e-mail: [email protected]; fax: (43) 1-40400-47350. Conflicts of interest The authors disclose no conflicts.
A Greater Proportion of Liver Transplant Candidates Have Colorectal Neoplasia Than in the Healthy Screening Population Philip Jeschek,*,‡ Arnulf Ferlitsch,*,‡ Petra Salzl,*,‡ Georg Heinze,§ Georg Györi,k Karoline Reinhart,*,‡ Elisabeth Waldmann,*,‡ Martha Britto-Arias,*,‡ Michael Trauner,* and Monika Ferlitsch*,‡ *Department of Internal Medicine III, Division of Gastroenterology and Hepatology, §Section for Clinical Biometrics, Center for Medical Statistics, Informatics and Intelligent Systems, kDepartment of Surgery, Division of Transplantation, Medical University of Vienna, ‡Quality Assurance Working Group of the Austrian Society of Gastroenterology and Hepatology, Vienna, Austria BACKGROUND & AIMS:
Various types of liver disease are associated with an increased prevalence of colorectal adenomas. We investigated whether cirrhosis is a risk factor for colorectal neoplasia by analyzing colonoscopy findings from 2 cohorts of patients awaiting liver transplantation.
METHODS:
We performed a retrospective analysis to compare findings from colorectal cancer screenings of 567 adult patients with cirrhosis placed on the waitlist for liver transplantation with those from controls (matched for age, sex, body mass index, smoking, and diabetes). Rates of adenoma and advanced adenoma detection were adjusted owing to differences in rates of polypectomies performed in the 2 cohorts.
RESULTS:
Adenomas were detected in a significantly higher percentage of patients with cirrhosis (29.3%) than in controls (21.5%) (P [ .0057; relative risk [RR], 1.36; 95% confidence interval [CI], 1.09– 1.69); and patients with cirrhosis had a higher rate of advanced adenoma detection than controls (13.9% vs 7.7%; P [ .0015; relative risk, 1.82; 95% CI, 1.25–2.64). A greater percentage of patients with alcoholic cirrhosis had neoplasias (34.3%) than controls (25.3%; P [ .0350; RR, 1.36), and rates of advanced adenoma detection were 16.7% vs 10.2% (P [ .0409; RR, 1.63). Adenomas were detected in 27.8% of patients with viral cirrhosis vs 15.9% of controls (P [ .0061; RR, 1.74), with rates of advanced adenoma detection of 13.6% vs 5.0% (P [ .0041; RR, 2.73). Similar proportions of patients with cirrhosis of other etiologies and controls were found to have colorectal neoplasias.
CONCLUSIONS:
Based on a retrospective analysis of colonoscopy findings from patients awaiting liver transplantation, those with alcoholic or viral cirrhosis are at higher risk of developing colorectal neoplasia and should be considered for earlier colonoscopy examination.
Keywords: Colon Cancer; Screening Colonoscopy; Early Detection.
n patients with advanced cirrhosis, colonoscopy is performed as part of the evaluation to assess eligibility for orthotopic liver transplantation (OLT) to rule out malignancies that would exclude patients from receiving a graft. Although there are indications that immunosuppression may shorten the interval of progression from premalignant lesions to colorectal cancer (CRC),1–3 there is evidence that CRC risk after OLT is increased only in patients with concomitant primary sclerosing cholangitis because of the strong association with inflammatory bowel disease.4 Because of shared risk factors such as obesity, diabetes, and metabolic syndrome, patients with nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH) are at higher risk of colorectal
I
neoplasia, both adenomas and carcinomas, which remained after adjustment for risk factors.5,6 Although Wong et al5 found a higher detection rate of advanced neoplasia (45%) in the right-sided colon of NAFLD patients of both sexes, another study6 found comparatively more adenomas in the rectum of male, and in the proximal colon of female, NAFLD patients. Abbreviations used in this paper: AADR, advanced adenoma detection rate; ADR, adenoma detection rate; ALCI, alcoholic cirrhosis; BMI, body mass index; CI, confidence interval; CRC, colorectal cancer; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; OLT, orthotopic liver transplantation; PDR, polyp detection rate; RR, relative risk. © 2015 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2014.08.018
May 2015
Patients who are at high risk of developing CRC should undergo earlier screening colonoscopy with closer monitoring at the age of 40, and in some cases even earlier.7 Cirrhosis itself has not been considered an independent risk factor for colorectal neoplasia to date and therefore is not included in CRC screening guidelines. Only a few studies have examined the prevalence of such lesions in OLT candidates, but were restricted either to relatively small cohorts, sigmoidoscopy only, or lacked a proper control group.8–11 Therefore, the aims of our study were to evaluate cirrhosis and its different etiologies as risk factors for colorectal neoplasia in liver transplant candidates and to determine if those patients with end-stage liver disease would benefit from earlier referral to colonoscopy.
Methods Study Design and Patients Our study was designed as a retrospective cohort observational study. Two cohorts were examined. The first cohort included cirrhotic patients who were listed for liver transplantation from 2000 to early 2012, in whom colonoscopy was performed as part of the evaluation for liver transplantation eligibility before placement on the waitlist of organ recipients. Data from 567 patients in the cohort of cirrhotic patients listed for OLT were provided by the Division of Transplantation of the Department of Surgery and the Division of Gastroenterology and Hepatology of the Department of Internal Medicine III at the Medical University of Vienna (Vienna, Austria). The second cohort was composed of general screening population subjects who underwent a screening colonoscopy within the Certificate of Quality Assurance for Colorectal Cancer Screening program.12 A total of 80,793 patients were screened within that program, of whom 19,904 also attended a health check reporting data on body mass index (BMI), diabetes, and smoking. A total of 567 of those patients were matched for confounding variables (namely age, sex, BMI, smoking, and diabetes) with the other cohort by randomly selecting 1 match for each cirrhotic patient out of the pool of matched patients of the screening population using SAS (2011; SAS Institute, Inc, Cary, NC). The majority of patients were at least 50 years old, and ascribed to current recommendations to perform screening colonoscopy starting at the age of 50. Advanced adenomas were defined as adenomas containing villous or tubulovillous features, as tubular adenomas of at least 1 cm in diameter, or as high-grade dysplasia. Because of liver disease–associated coagulopathy not all polyps could be resected, however, we assumed that the proportion of adenomas in polyps that did not undergo a biopsy were comparable with the ones that did undergo a biopsy. In this study, we therefore calculated and used an adjusted adenoma detection rate
Liver Transplant and Colorectal Neoplasia
957
and an adjusted advanced adenoma detection rate as described in the Statistical Analysis section.
Statistical Analysis Both cohorts were described by sex, age, BMI, smoking, and diabetes; transplant candidates also were described by etiology and stage of cirrhosis. Categoric variables were described by absolute and relative frequency, and continuous variables were described by medians and 25th and 75th percentiles. Primary outcome measures were the adenoma detection rate (ADR) and the advanced adenoma detection rate (AADR), which were adjusted owing to a variable rate of polypectomies performed in the 2 cohorts: the reason being an increased bleeding tendency in cirrhotic patients as a result of severe coagulopathy. We assumed the same rate of adenomas among resected and nonresected polyps. Thus, we computed a polypectomyadjusted ADR as follows: adjusted ADR ¼ ADR ðresected polyps þ nonresected polypsÞ resected polyps Those adjusted rates were compared between the 2 cohorts and further stratified into age groups for complementary age-specific analyses by the calculation of an adjusted relative risk (RR) with a 95% confidence interval and by the use of a chi-square test. Within the cohort of cirrhotic patients, the polyp detection rate (PDR), ADR, and AADR were compared between the different etiologies of cirrhosis (alcoholic, viral, and other causes). Secondary outcome measures were the number (1, 2–4, >4) and localization of polyps (proximal colon, sigmoid/rectum, and distal and proximal colon), as well as the cecal intubation, sedation, and complication rates, although for the latter we only had data for the cohort of transplant candidates. Variables mentioned earlier were compared between the 2 cohorts by the calculation of absolute and relative frequencies and by the use of chi-square tests. The significance level was considered 5%. All statistical analyses were conducted as 2-sided tests using the SPSS 19.0 (2010, IBM Corp, Armonk, NY) and Microsoft Excel 2007 (2007, Microsoft Corp, Redmond, WA) software packages. The study protocol was approved by the ethics commission of the Medical University of Vienna in February 2012 (EK063/2012).
Results Patient Characteristics As shown in Table 1, a total of 567 patients in each cohort, adjusted for age, sex, BMI, smoking, and diabetes,
958
Jeschek et al
Clinical Gastroenterology and Hepatology Vol. 13, No. 5
Table 1. Patient Characteristics Patient characteristics
Cirrhotic patients
Total 567 Male sex 410 (72.3%) Median age at screening, y 54 Median BMI 25.5 Smokers 248 (43.7%) Diabetic subjects 129 (22.8%) Cirrhosis stage Median MELD score 15.6 Child A 61 (10.8%) Child B 213 (37.7%) Child C 291 (51.5%) Indication for OLT/etiology of cirrhosis Alcoholic cirrhosis 277 (48.9%) Viral cirrhosis 206 (36.3%) Hepatocellular carcinoma 145 (25.6%) Biliary duct diseases 44 (7.8%) Other etiologies 64 (11.3%)
Healthy controls 567 410 (72.3%) 54 25.7 221 (39.0%) 97 (17.1%)
MELD, model for end-stage liver disease.
were included. Their median age was 54 years and there were 157 women (27.7%) and 410 men (72.3%) in both cohorts, with a median BMI of 25.5 in cirrhotic patients and 25.7 in screening patients. Among cirrhotic patients, 43.7% were smokers and 22.8% had diabetes compared with 39.0% and 17.1%, respectively, in healthy controls. Cirrhotic patients presented with a median model for end-stage liver disease score of 15.6 by the time of colonoscopy, and Child–Pugh score was stage A in 10.8%, stage B in 37.7%, and stage C in 51.5%. Of the 567 cirrhotic patients with one or multiple indications for transplantation, cirrhosis as a result of alcohol abuse was diagnosed in 277 (48.9%), viral hepatitis was diagnosed in 206 (36.3%) (among whom 165 [80.1%] had hepatitis C virus, 29 [14.1%] had hepatitis B virus, and 12 [5.8%] had both types of viruses), hepatocellular carcinoma and cirrhosis were diagnosed in 145 (25.6%), and biliary duct diseases were diagnosed in 44 (7.8%) patients. In 64 (11.3%) patients there were other causes for cirrhosis.
polyps, compared with 18.7% (P ¼ .0017) of controls, had polyps proximal to the sigmoid. In cirrhotic patients only 40.2% of lesions were found in the sigmoid colon and/or rectum compared with 57.4% in the screening cohort (P ¼ .0002). A total of 28.3% of cirrhotic patients vs 23.9% of controls had polyps both in the distal and proximal parts of the colon (P ¼ .29).
Number of Polyps Among cirrhotic patients with endoscopically detected colorectal polyps 16.2% had more than 4 polyps, 42.6% had 2 to 4 polyps, and 41.1% had only 1 polyp. Healthy screening patients with colonic polyps had more than 4 polyps in 6.7%, 2 to 4 polyps in 42.6%, and only a single polyp in 50.7%.
Histologic Findings The ADR was significantly higher in cirrhotic patients compared with the screening cohort (29.3%, CI, 25.0–33.9; vs 21.5%; CI, 18.2–25.2; P ¼ .0057) as shown in Figure 1 and Table 2. For the AADR the difference was even higher (13.9%; CI, 10.8–17.6; vs 7.7%; CI, 5.6–10.2; P ¼ .0015). Therefore, the RR for adenomas was 1.36 (CI, 1.09–1.69) and 1.82 (CI, 1.25–2.64) for advanced adenomas. Transplant candidates aged 40 to 49, 50 to 59, and 60 to 69 years all had a higher ADR and AADR than screening subjects, but the difference did not reach statistical significance because the number of patients in each subgroup was too low. Cirrhotic patients aged 40 to 49 years, for instance, had an ADR of 21.8% vs 13.2% (P ¼ .0913), and an AADR of 12.0% vs 7.7% (P ¼ .29) compared with controls of the same age group.
Macroscopic Findings Among 567 liver transplant candidates endoscopic polyps were found in 46.7% (n ¼ 265; confidence interval [95% CI], 42.6%–50.9%) compared with 36.9% (n ¼ 209; CI, 32.9%–41.0%) in the screening cohort (P ¼ .0007; RR, 1.27; 95% CI, 1.10–1.46).
Localization of Macroscopic Lesions Isolated endoscopic lesions proximal to the sigmoid were significantly more frequent in cirrhotic patients. A total of 31.4% of cirrhotic patients with endoscopic
Figure 1. PDRs, adjusted ADRs, and adjusted AADRs in cirrhotic patients compared with healthy controls matched for age, sex, body mass index, smoking, and diabetes. Data are shown as means.
May 2015
Liver Transplant and Colorectal Neoplasia
959
Table 2. PDRs, ADRs, and AADRs in Cirrhotic Patients and Controls Overall Cirrhotic patients
Healthy controls
Diagnosis
Rate (%)
CI (%)
Rate (%)
CI (%)
P value
RR
Endoscopic polyps Adenomas, adjusted rates Advanced adenomas, adjusted rates
46.74 29.27 13.93
42.57–50.94 24.98–33.86 10.77–17.59
36.86 21.53 7.66
32.88–40.98 18.16–25.21 5.58–10.22
.0007 .0057 .0015
1.27 1.36 1.82
Sex-Specific Rates for Colorectal Neoplasia In both sexes colorectal neoplasia detection rates were higher in cirrhotic patients than in controls (Table 3). Among females, cirrhotic women had a higher PDR (42.7% vs 28.0%; P ¼ .0066) than female controls. The ADR was 27.9% in cirrhotic vs 16.7% in healthy women (RR, 1.67; P ¼ .0257) and 11.5% vs 6.0%, respectively, for the AADR (P ¼ .11; RR, 1.91). In cirrhotic men the PDR was 48.3% compared with 40.2% in healthy male screening subjects (P ¼ .0203). The ADR was 29.8% vs 23.4%, respectively (P ¼ .0572; RR, 1.27), and only the AADR was significantly higher in cirrhotic men (14.9% vs 8.3%, respectively; P ¼ .0061; RR, 1.79).
Association Between Etiologies of Cirrhosis and Colorectal Neoplasia Among all etiologies of cirrhosis, in patients with alcoholic cirrhosis (ALCI), colorectal neoplasia rates were highest (Figure 2). ALCI patients had a higher prevalence of adenomas (34.3% vs 25.3%; P ¼ .0350; RR, 1.36) and advanced adenomas (16.7% vs 10.2%, P ¼ .0409; RR, 1.63) than matched controls from the screening cohort. Among men, male ALCI patients had a higher rate of adenomas (35.1% vs 25.3%; P ¼ .0408; RR, 1.39) and advanced adenomas (18.5% vs 10.0%; P ¼ .0190; RR, 1.85) than matched screening subjects. Also in female ALCI patients we found a nonsignificantly higher rate of adenomas than in matched controls (31.4% vs 25.7%; P ¼ .54; RR, 1.22), whereas detection rates for advanced
adenomas were comparable between those 2 quite small subgroups (9.7% vs 11.0%; P ¼ .90; RR, 0.88). Patients with cirrhosis caused by viral hepatitis C and hepatitis B had more adenomas (27.8% vs 15.9%; P ¼ .0061; RR, 1.74) and advanced adenomas (13.6% vs 5.0%; P ¼ .0041; RR, 2.73) than matched screening subjects (Figure 3). Among men, male patients with viral cirrhosis had a higher rate of adenomas (27.7% vs 17.7%; P ¼ .0440; RR, 1.56) and nonsignificantly more advanced adenomas (12.7% vs 6.3%; P ¼ .0655; RR, 2.00) than matched screening subjects. In addition, in female patients with viral cirrhosis we found a significantly higher rate of adenomas (27.9% vs 9.3%; P ¼ .0314; RR, 3.00) and advanced adenomas (16.7% vs 0.0%; P ¼ .0180) than in matched screening subjects. Cirrhotic patients with entities other than alcoholic and viral cirrhosis had no significant variations in colorectal neoplasia rates compared with matched controls. Adenoma rates in this subgroup were 20.1% vs 19.0% (P ¼ .84; RR, 1.06), and advanced adenoma rates were 8.5% vs 5.6% (P ¼ .44; RR, 1.46) in matched screening subjects.
Cecal Intubation, Sedation, and Complication Rate In screening patients the cecum was reached in 96.6% of colonoscopies compared with 93.6% in cirrhotic patients (P ¼ .0181). Regarding the sedation rate, 84.3% of screening patients were sedated with propofol and/or midazolam during the colonoscopy procedure vs 65.3% of transplant candidates (P < .0001). Postinterventional
Table 3. PDRs, ADRs, and AADRs in Men and Women Cirrhotic patients Diagnosis by sex Men Endoscopic polyps Adenomas, adjusted rates Advanced adenomas, adjusted rates Women Endoscopic polyps Adenomas, adjusted rates Advanced adenomas, adjusted rates
Healthy controls
Rate (%)
CI (%)
Rate (%)
CI (%)
P value
RR
48.29 29.77 14.88
43.36–53.25 24.67–35.27 11.07–19.41
40.24 23.39 8.30
35.46–45.17 19.32–27.87 5.78–11.46
.0203 .0572 .0061
1.20 1.27 1.79
42.68 27.90 11.49
34.83–50.81 20.16–36.76 6.43–18.53
28.03 16.68 6.01
21.16–35.74 11.10–23.64 2.78–11.09
.0066 .0257 .1063
1.52 1.67 1.91
960
Jeschek et al
Figure 2. PDRs, adjusted ADRs, and adjusted AADRs in patients with alcoholic cirrhosis compared with healthy controls matched for age, sex, body mass index, smoking, and diabetes. Data are shown as means.
hemorrhage in 1.4% of cirrhotic patients was the only complication that occurred during colonoscopy and was controlled endoscopically in all cases.
Discussion This study investigated colonoscopy outcomes in a large cohort of patients with cirrhosis under evaluation for liver transplantation, and it compared the prevalence rates of colorectal neoplasia with age- and sex-matched controls that were adjusted for BMI, smoking, and diabetes. We evaluated the colonoscopy findings of 567 patients listed for liver transplantation and found a significantly higher rate of colorectal adenomas and
Figure 3. PDRs, adjusted ADRs, and adjusted AADRs in patients with viral cirrhosis compared with healthy controls matched for age, sex, body mass index, smoking, and diabetes. Data are shown as means.
Clinical Gastroenterology and Hepatology Vol. 13, No. 5
advanced adenomas than in healthy controls. On closer examination, alcoholic cirrhosis constituted the strongest predictor of finding colorectal neoplasia in this study. These findings are consistent with the conclusion published by the International Agency for Research on Cancer, which stated that alcohol consumption is related to CRC.13 A meta-analysis of 27 cohort and 34 casecontrol studies estimated relative risks for alcohol consumption leading to CRC, which were 1.21 for moderate and 1.52 for heavy drinking (RR for adenomas among patients with ALCI in our study was 1.36 and was 1.63 for advanced adenoma) and concluded that there is strong evidence for an association between drinking more than 1 alcoholic drink per day and an increased colorectal cancer risk.14 Under those circumstances it is uncertain whether ALCI was partly causative by itself, or just a significant predictor of the presence of colorectal neoplasia indicating a history of alcohol consumption. The only study comparing ALCI patients and noncirrhotic alcoholic subjects with regards to colorectal neoplasia rates was a French cross-sectional prevalence study in 1992. Among 100 ALCI patients, colorectal adenoma was found in 58%, whereas among 100 alcoholic patients without cirrhosis colorectal adenoma was detected by colonoscopy in only 26% (P < .001). However, in this study patients were not sex- and age-matched and the mean duration of reported alcohol abuse in ALCI patients was 5 years longer, which limits those findings.15 Furthermore, our study also found higher rates of colorectal neoplasia in cirrhotic patients with hepatitis C virus and hepatitis B virus infection as the underlying etiology of cirrhosis. Some viral infections such as human immunodeficiency virus and subtypes of polyomaviruses including John Cunningham (JC) virus have been connected with the facilitation of colorectal carcinogenesis, however, data on the impact of viral infections are controversial.16,17 Interestingly, it has been shown that chronic hepatic viral infection might be protective for the occurrence of liver metastases in patients with CRC.18 However, there is no evidence supporting an association between hepatitis C virus/ hepatitis B virus infection and colorectal neoplasia, and the high prevalence of colorectal neoplasia in patients with viral cirrhosis in our study might be owing to the chronic inflammatory state in those patients. After age stratification we found higher rates for colorectal neoplasia in all age groups between 40 and 69 years in cirrhotic patients, although those results did not reach statistical significance. A recent study performed by Weismüller et al8 prospectively evaluated 243 OLT candidates who underwent colonoscopy and found neoplastic polyps in 21% of patients, but did not compare those findings with age- and sex-matched controls. Nevertheless, they compared their findings with 2 general population–based studies and found a higher rate of neoplastic polyps in cirrhotic patients aged 40 to 49 (14.7% vs 8.7%–9.5%) and 50 to 59 years (23.4% vs 14.9%) than in healthy patients of similar ages, whereas
May 2015
rates of advanced neoplasia were comparable. The higher rates of colorectal neoplasia in our study might be explained by significant variations in the patient cohorts because our study included more males (72.3% vs 63%) and more patients with alcoholic cirrhosis (48.6% vs 29.2%). Because of the nature of our study we did not involve patients with nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH) without cirrhosis, however, it is worth mentioning that Wong et al5 found a significantly higher prevalence of colorectal neoplasia in such patients. In patients with NAFLD they found a higher rate of adenomas (34.7% vs 21.5%) and advanced adenomas (18.6% vs 5.5%) than in the healthy control cohort, comparable with the colorectal neoplasia rates we found among ALCI patients. The higher prevalence of neoplasia in NASH but not in simple steatosis remained after adjusting for demographic and metabolic factors, however, because the study was retrospective we were only able to adjust for sex, age, BMI, smoking, and diabetes, but not for alcohol consumption in our study. It has been confirmed in large general population–based studies that men are more likely to have colorectal neoplasia than women.12,19 In our study we still found higher rates of colorectal neoplasia in cirrhotic men than in women (29.8% vs 27.9% for adenomas, 14.9% vs 11.5% for advanced adenomas), however, the difference between cirrhotic men and women was not as drastic as between men and women in routinely screened general population subjects (24.9% vs 14.8% for adenomas, 8.0% vs 4.7% for advanced adenomas).12 These findings indicate that an increased risk for colorectal neoplasia in cirrhotic patients, and especially in patients with ALCI and viral cirrhosis, is more obvious in women than in men. In consideration of full colonoscopy still being controversial as the screening method of choice, we evaluated the localization of endoscopic polyps and found that a large proportion of cirrhotic patients had lesions proximal to the sigmoid colon or in proximal as well as distal parts of the colon (60%) compared with 43% of controls. In NAFLD patients, Wong et al5 found a higher prevalence of advanced neoplasia in the rightsided colon whereas Stadlmayr et al6 only found more adenomas in the proximal colon of female, but not of male, patients. In our cirrhotic cohort, only 1.4% had a postinterventional hemorrhage that could always be controlled endoscopically. A study by Jeon et al20 confirmed our finding that polypectomy was comparably safe in patients with cirrhosis, probably also because procoagulatory and anticoagulatory factors are rebalanced in most cirrhotic patients, which corresponds to low postinterventional bleeding rates.21,22 We therefore propose performing polypectomy in patients with balanced hemostatic markers, because a post-transplantation condition increases the risk for colorectal cancer (odds ratio, 1.4),
Liver Transplant and Colorectal Neoplasia
961
which would increase even more in patients with underlying alcoholic liver disease (odds ratio, 2.3).23 Limitations included the retrospective nature of this study, which did not allow matching for all potentially confounding factors such as alcohol consumption; however, patients were adjusted for sex, age, BMI, smoking, and diabetes. A study with potential transplant candidates usually does not allow conclusions about colorectal cancer rates because patients diagnosed with colorectal cancer are not evaluated further for liver transplantation. Furthermore, our data set only contained data on the number and localization of endoscopic polyps overall, including hyperplastic and neoplastic polyps, so we could not differentiate between adenomas and hyperplastic polyps regarding number and localization, and only described those characteristics for endoscopic polyps overall. Moreover, in some patients polypectomy or biopsy was not performed owing to severe coagulopathy, and in another small portion of patients polyps could not be retrieved with the endoscope. Nevertheless, our data on adenomas and advanced adenomas were adjusted for those missing findings. Although bowel preparation and withdrawal time both are relevant factors that ensure the quality of colonoscopies, data on those factors could not be provided in our study. In conclusion, patients with alcoholic and viral cirrhosis had a higher prevalence of colorectal neoplasia than matched controls. Although colonoscopy is not mandatory in every pre–liver transplant evaluation system, we clearly would recommend performing it in patients 40 years or older, and especially in those with alcoholic and viral cirrhosis. Because of the high prevalence of colorectal neoplasia in those patients it seems reasonable to start colorectal cancer screening earlier than recommendations in general population–based guidelines suggest, regardless of whether patients are potential transplant recipients or not.
References 1. Nicolaas JS, de Jonge V, Steyerberg EW, et al. Risk of colorectal carcinoma in post-liver transplant patients: a systematic review and meta-analysis. Am J Transplant 2010;10:868–876. 2. Buell JF, Papaconstantinou HT, Skalow B, et al. De novo colorectal cancer: five-year survival is markedly lower in transplant recipients compared with the general population. Transplant Proc 2005;37:960–961. 3. Selgrad M, Koornstra JJ, Fini L, et al. JC virus infection in colorectal neoplasia that develops after liver transplantation. Clin Cancer Res 2008;14:6717–6721. 4. Singh S, Varayil JE, Loftus EV Jr, et al. Incidence of colorectal cancer after liver transplantation for primary sclerosing cholangitis: a systematic review and meta-analysis. Liver Transpl 2013;19:1361–1369. 5. Wong VW-S, Wong GL-H, Tsang SW-C, et al. High prevalence of colorectal neoplasm in patients with non-alcoholic steatohepatitis. Gut 2011;60:829–836. 6. Stadlmayr A, Aigner E, Steger B, et al. Nonalcoholic fatty liver disease: an independent risk factor for colorectal neoplasia. J Intern Med 2011;270:41–49.
962
Jeschek et al
7. Qaseem A, Denberg TD, Hopkins RH, et al. Screening for colorectal cancer: a guidance statement from the American College of Physicians. Ann Intern Med 2012;156:378–386. 8. Weismüller TJ, Bleich F, Negm AA, et al. Screening colonoscopy in liver transplant candidates: risks and findings. Clin Transplant 2013;27:E161–E168. 9. Zaman A, Hapke R, Flora K, et al. Prevalence of upper and lower gastrointestinal tract findings in liver transplant candidates undergoing screening endoscopic evaluation. Am J Gastroenterol 1999;94:895–899. 10. Díaz-Sánchez A, Núñez-Martínez O, González-Asanza C, et al. [Results of colonoscopic examination in the evaluation of cirrhotic candidates for liver transplantation]. Gastroenterol Hepatol 2009;32:327–333. 11. Selingo JA, Herrine SK, Weinberg DS, et al. Role of screening colonoscopy in elective liver transplantation evaluation. Transplant Proc 1997;29:2506–2508. 12. Ferlitsch M. Sex-specific prevalence of adenomas, advanced adenomas, and colorectal cancer in individuals undergoing screening colonoscopy. JAMA 2011;306:1352. 13. World Cancer Research Fund/American Institute for Cancer Research. Food, nutrition, physical activity, and the prevention of cancer: a global perspective. Washington, DC: AICR, 2007. 14. Fedirko V, Tramacere I, Bagnardi V, et al. Alcohol drinking and colorectal cancer risk: an overall and dose-response metaanalysis of published studies. Ann Oncol 2011;22:1958–1972. 15. Naveau S, Chaput JC, Bedossa P, et al. Cirrhosis as an independent risk factor for colonic adenomas. Gut 1992;33: 535–540. 16. Lin OS. Acquired risk factors for colorectal cancer. Methods Mol Biol 2009;472:361–372.
Clinical Gastroenterology and Hepatology Vol. 13, No. 5 17. Giuliani L, Ronci C, Bonifacio D, et al. Detection of oncogenic DNA viruses in colorectal cancer. Anticancer Res 2008;28:1405–1410. 18. Zeng Y, He X-W, He X-S, et al. Chronic hepatic viral infection could be a protective factor for colorectal cancer liver metastases: analysis in a single institute. Hepatogastroenterology 2013;60:37–41. 19. Regula J, Rupinski M, Kraszewska E, et al. Colonoscopy in colorectal-cancer screening for detection of advanced neoplasia. N Engl J Med 2006;355:1863–1872. 20. Jeon JW, Shin HP, Lee JI, et al. The risk of postpolypectomy bleeding during colonoscopy in patients with early liver cirrhosis. Surg Endosc 2012;26:3258–3263. 21. Lisman T, Leebeek FWG. Hemostatic alterations in liver disease: a review on pathophysiology, clinical consequences, and treatment. Dig Surg 2007;24:250–258. 22. Ferlitsch M, Reiberger T, Hoke M, et al. Von Willebrand factor as new noninvasive predictor of portal hypertension, decompensation and mortality in patients with liver cirrhosis. Hepatology 2012;56:1439–1447. 23. Schrem H, Kurok M, Kaltenborn A, et al. Incidence and long-term risk of de novo malignancies after liver transplantation with implications for prevention and detection: cancer After LT: incidence, prevention, detection. Liver Transplant 2013;19:1252–1261.
Reprint requests Address requests for reprints to: Monika Ferlitsch, MD, Division of Gastroenterology and Hepatology, Medical University of Vienna, Waehringer Guertel 18-20, A-1090 Vienna, Austria. e-mail: [email protected]; fax: (43) 1-40400-47350. Conflicts of interest The authors disclose no conflicts.