Japanese Journal of Clinical Oncology Advance Access published November 6, 2014 Japanese Journal of Clinical Oncology, 2014, 1–9 doi: 10.1093/jjco/hyu178 Original Article
Original Article
Effect of liver cirrhosis on metastasis in colorectal cancer patients: a nationwide population-based cohort study Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
Wen-Yen Chiou1,2,3, Chun-Ming Chang2,4, Kuo-Chih Tseng2,5, Shih-Kai Hung1,2, Hon-Yi Lin1,2, Yi-Chun Chen2,6, Yu-Chieh Su2,7, Chih-Wei Tseng2,5, Shiang-Jiun Tsai1, Moon-Sing Lee1,2,*, and Chung-Yi Li3,* 1
Department of Radiation Oncology, Buddhist Dalin Tzu Chi Hospital, Chiayi, 2School of Medicine, Tzu Chi University, Hualien, 3Department of Public Health, College of Medicine, National Cheng Kung University, Tainan, 4Department of Surgery, Buddhist Dalin Tzu Chi Hospital, Chiayi, 5Division of Gastroenterology, Department of Internal Medicine, Buddhist Dalin Tzu Chi Hospital, Chiayi, 6Division of Nephrology, Department of Internal Medicine, Buddhist Dalin Tzu Chi Hospital, Chiayi, and 7Division of Hematology-Oncology, Department of Internal Medicine, Buddhist Dalin Tzu Chi Hospital, Chiayi, Taiwan *For reprints and all correspondence: Moon-Sing Lee, Department of Radiation Oncology, Buddhist Dalin Tzu Chi Hospital, No. 2, Ming Sheng Road, Dalin, Chiayi, Taiwan. E-mail: [email protected]. Chung-Yi Li, Department of Public Health, College of Medicine, National Cheng Kung University, No.1, University Road, Tainan City 701, Taiwan. E-mail: armstrong_washington@ hotmail.com Received 8 August 2014; Accepted 14 October 2014
Abstract Objective: The aim of this study is to evaluate the liver metastasis risk among colorectal cancer patients with liver cirrhosis. Methods: This was a nationwide population-based cohort study of 2973 newly diagnosed colorectal cancer patients with liver cirrhosis and 11 892 age–sex matched controls enrolled in Taiwan between 2000 and 2010. The cumulative risk by Kaplan–Meier method, hazard ratio by the multivariate Cox proportional model and the incidence density were evaluated. Results: The median time interval from the colorectal cancer diagnosis to the liver metastasis event was 7.42 months for liver cirrhosis group and 7.67 months for non-liver cirrhosis group. The incidence density of liver metastasis was higher in the liver cirrhosis group (61.92/1000 person-years) than in the non-liver cirrhosis group (47.48/1000 person-years), with a significantly adjusted hazard ratio of 1.15 (95% CI = 1.04–1.28, P = 0.007). The 10-year cumulative risk of liver metastasis for the liver cirrhosis and the non-liver cirrhosis group was 27.1 and 23.6%, respectively (P = 0.006). For early cancer stage with locoregional disease patients receiving surgery alone without adjuvant anti-cancer treatments, patients with liver cirrhosis (10-year cumulative risk 23.9 vs. 15.7%, P < 0.001) or cirrhotic symptoms (10-year cumulative risk 25.6 vs. 16.6%, P = 0.009) both still had higher liver metastasis risk compared with their counterparts. For etiologies of liver cirrhosis, the 10-year cumulative risk for hepatitis B virus and hepatitis C virus, hepatitis B virus, hepatitis C virus, other causes and nonliver cirrhosis were 29.5, 28.9, 27.5, 26.7 and 23.4%, respectively, (P = 0.03). Conclusions: Our study found that liver metastasis risk was underestimated and even higher in colorectal cancer patients with liver cirrhosis.
© The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
1
2
Effect of liver cirrhosis on metastasis in colorectal cancer patients
Key words: liver cirrhosis, colorectal cancer, liver metastasis
Introduction
Patients and methods Ethics statement This study was reviewed and approved by the National Health Insurance Research Database (NHIRD) research committee and by the institutional review board (IRB no. B10104024) of Buddhist Dalin Tzu Chi Hospital. Informed consent was waived because the patients’ personal identification numbers were de-identified.
Source of data Patients were drawn from the NHIRD, released for research purposes by the National Health Research Institutes, Taipei, Taiwan (16). This dataset provided all original claims data from the Taiwan National Health Insurance program, lunched since 1995, a mandatory-enrollment, single-payment system that financed healthcare for all Taiwanese citizens
Study design and patients This was a population-based cohort study, with the study period spanning from 1 January 2000 through 31 December 2010. We first identified patients with a diagnosis of CRC (ICD-9-CM codes 153.0 to 153.9 and 154.0 to 154.2) from inpatient and outpatient claims during the study period. These CRC patients must also have been issued catastrophic illness certificates that further ensure the accuracy of cancer diagnosis (21). To ensure the inclusion of only newly diagnosed CRC between 2000 and 2010, patients who had a diagnosis of CRC between 1995 and 1999 were excluded. We further excluded patients with claims for any other malignancy (ICD-9-CM codes 140–208, excluding CRC) from 1995 to 2010, such as hepatocellular carcinoma. A total of 91 441 patients with newly diagnosed CRC were identified. Among the 91 441 CRC patients, 2973 patients who had ICD-9-CM codes 571.2, alcoholic cirrhosis of liver, 571.5, cirrhosis of liver without mention of alcohol and 571.6, biliary cirrhosis, made in at least one inpatient or two outpatient medical records to make sure not a miscoding, before the initial diagnosis of CRC during the study period were defined to have a diagnosis of underlying liver cirrhosis (LC) in this study (22). Because LC was prevalent in Taiwan, the government paid close attention to this health issue and has established specific guidelines for diagnosis and treatment of LC (23). For each CRC patient with underlying LC, four CRC patients without LC were selected using frequency matching on age and gender, leading to 11 892 CRC patients without LC (Fig. 1). Each study participant was followed until the occurrence of the following outcomes whichever comes first: (i) a diagnosis of LM (ICD-9-CM code 197.7) was made; (ii) the participant withdrew from the NHI system mainly due to death; and (iii) 31 December 2010. LC severity and residual liver function information were evaluated by cirrhotic symptoms including ascites, jaundice, hepatic encephalopathy or albumin, frozen plasma (FP), flesh frozen plasma (FFP) transfusion.
Measurements of end-points and covariates The primary dependent variable of interest was liver metastasis, including both synchronous and metachronous, which was identified
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
Liver cirrhosis (LC) is a serious disease that results in permanent damage to the liver. There are a variety of diseases and conditions that can cause cirrhosis of the liver. Alcoholic liver disease and hepatitis C are the most common causes in developed countries, whereas hepatitis B is the prevailing cause in most parts of Asia and sub-Saharan Africa (1). The major clinical consequences of cirrhosis are impaired liver function, an increased intrahepatic resistance ( portal hypertension) and the development of hepatocellular carcinoma (1). However, there are limited data in the literature on the incidence, prevalence and management of non-hepatic cancers (NHC) in cirrhotic patients (2,3). Previous studies suggested that LC might play a role in NHC metastasis, but the studies yield inconsistent results mainly due to limited sample sizes (4–6). Colorectal cancer is a worldwide health problem, particularly in developed countries, comprising 9.4% of all cancer cases with ∼1 million diagnoses annually worldwide (7). It is also one of the leading causes of cancer related deaths in America, Europe and Asia (8). In Taiwan, like in many other nations, colorectal cancer (CRC) is also a major public health problem. According to the report of the Bureau of Health Promotion, Taiwan, the CRC has become the most common malignancy in Taiwan since 2006 and the crude incidence rate was 60.62 per 100 000 people in 2010 (9). CRC cells travel to the liver through the hepatic portal circulation, and liver is the most common anatomical site for hematogenous metastasis (10). Unlike other malignancy, the CRC patients with limited metastatic disease confined to the liver who are able to undergo resection may still have 25–37%, even 58%, chance of long-term survival (11,12). The National Comprehensive Cancer Network (NCCN) clinical practice guidelines in Oncology also suggested that hepatic resection is the treatment of choice for resectable liver metastasis (LM) from CRC (13,14). It was found that patients with an untreated single LM had a median survival of 19 months, with no patients surviving 5 years, while patients with a resected single LM had a median survival of 36 months with 25% of patients surviving 5 years (15). By identifying risk factors of LM, it could be expected to find more resectable LM and achieve better outcomes. The aim of this study was to estimate the effect of LC on risk of LM in CRC, which was one of the cancers likely to cause LM.
(>99% of the 24 million inhabitants of Taiwan) through a registry of board certified specialists and contracted medical facilities (∼97% of all Taiwan medical facilities) (17). The NHIRD is one of the largest and most comprehensive databases in the world and has been used extensively in various studies (18). For the precision of the claims data, the Bureau of National Health Insurance (BNHI) performs expert reviews on a random sample of every 50–100 ambulatory and every 20 inpatient claims in each hospital and clinic quarterly and interviews patients in order to verify the accuracy of diagnosis (19). False reports of diagnosis would yield a severe penalty from the BNHI (20). Therefore, information obtained from the NHIRD is considered to be complete and accurate (21). The NHIRD provided information on encrypted patient identification numbers, sex, birthday, dates of admission and discharge, medical institutions providing the services, the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes of up to five diagnoses and procedures, respectively, outcome at hospital discharge (recovered, died or transferred out), order codes and the fees charged for medical services.
Jpn J Clin Oncol 2014
3
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
Figure 1. Study design flowchart. NHI, National Health Insurance; CRC, colorectal cancer; ICD-9, International Classification of Diseases, Ninth Revision; LC, liver cirrhosis.
from one inpatient or two outpatient claims. As mentioned above, we also excluded patients with claims for any other malignancy from 1995 to 2010 to rule out false LM events from hepatocellular carcinoma or other NHCs. The covariates considered included gender, age at CRC diagnosis, hepatitis, comorbidities, clinical examinations, cancer treatment modalities, geographic region of living, urbanization level of residence and socioeconomic status. Hepatitis types included hepatitis B (ICD-9-CM codes 070.22, 070.23, 070.32, 070.33, V02.61) and hepatitis C (070.44, 070.54, V02.62). Comorbidities included diabetes mellitus (ICD-9-CM code 250), hypertension (ICD-9-CM codes 401–405), hyperlipidemia (ICD-9-CM codes 272.0–272.4)
and chronic kidney disease (ICD-9-CM code 585). These comorbidities were adjusted for their possible role in liver microenvironment (24–26). Geographic regions, urbanization level of residence and socioeconomic status information were adjusted for possible uneven medical resources to prevent information bias. There were four geographic regions (northern, central, southern, eastern and other) in Taiwan. The urbanization level of residence was classified as seven levels based on five indices: population density, percentages of residences with college or higher education, percentages of residents >65 years old, percentages of residents who were agriculture workers and the number of physicians per 100 000 people (27). Urbanization level of residence was further recorded as urban
4
Effect of liver cirrhosis on metastasis in colorectal cancer patients
Statistical analysis The statistical software packages SAS (version 9.2; SAS Institute, Inc., Cary, NC, USA) were used for data analysis. To assess the independent effects of LC on the risk of LM, we used Cox proportional hazard regression models, adjusting for the aforementioned potential confounders. The incidence density was calculated with person-years as the denominator under the Poisson assumption. Denominator of the incidence density was the sum of individual’s person-years at-risk, namely from date of first ever CRC diagnosis to date of LM or censored. The Kaplan–Meier method was used to estimate the cumulative risk of LM. A two-sided P value < 0.05 was defined as statistically significant.
Results Subject characteristics The distributions of demographic characteristics and comorbidities for the CRC patients with and without LC are shown in Table 1. Over 66% CRC patients with LC were male. The mean (±SD) age of patients with LC was 67.59 (±11.64). Compared with the non-LC patients, patients with LC had a significantly higher prevalence of comorbidities, which may have effect on liver microenvironment, including hepatitis, diabetes mellitus, hypertension, hyperlipidemia and chronic kidney disease. Most patients, both LC and non-LC groups, received surgery alone for CRC without adjuvant therapy. Compared with those without LC, lower percentage of patients with LC had cancer follow-up examinations, such as CEA, abdominal sonography and computed tomography examination, except CA199. Patients with LC were more likely to reside in the rural areas. A higher percentage of non-LC patients had high socioeconomic status and lived in northern Taiwan where more medical resources were available.
Risk of LM in CRC subjects with LC The median time intervals from the CRC diagnosis to the LM event were 7.42 months for the LC group and 7.67 months for the
Table 1. Characteristics of colorectal cancer patients with and without liver cirrhosis Variables
With liver cirrhosis N = 2973 (%)
Characteristics Gender Female 33.54 Male 66.46 Age 0–44 2.76 45–54 11.10 55–64 21.49 65–74 32.53 75+ 32.12 Mean age (years) ± SD 67.59 ± 11.64 Comorbidities Hepatitis B 5.99 Hepatitis C 19.17 Diabetes mellitus 44.70 Hypertension 65.52 Hyperlipidemia 37.77 Chronic kidney disease 11.47 Cancer treatment modalities Surgery alone 53.18 Surgery with adjuvant 16.38 chemotherapy 5.45 Surgery with adjuvant CCRT Surgery with adjuvant 3.50 radiotherapy CCRT 1.48 Chemotherapy alone 2.52 Radiotherapy alone 1.35 Supportive care 16.15 Clinical cancer follow-up examinations CEA 64.51 CA199 16.82 Abdominal sonography 10.39 Computed tomography 43.39 Geographic region Northern 39.02 Central 26.85 Southern 31.11 Eastern 3.03 Urbanization level Urban 23.98 Suburban 41.20 Rural 34.81 Socioeconomic status High 22.37 Middle 52.37 Low 25.26
Without liver cirrhosis N = 11 892a (%)
33.54 66.46 2.76 11.10 21.49 32.53 32.12 67.59 ± 11.98 0.56 1.7 30.76 57.94 33.77 5.46 50.63 21.33 6.54 3.67 1.46 1.42 1.07 13.87 67.13 14.93 44.48 64.83 45.66 23.44 28.83 2.07 27.61 44.11 28.28 28.05 45.86 26.08
a
Frequency match on age and gender with 4 : 1 ratio. CCRT, concurrent chemoradiotherapy; CEA, carcinoembryonic antigen; CA199, carbohydrate antigen 19-9.
non-LC group. The overall LM rates were 17.4 and 16.9% for the LC and non-LC group, respectively. Table 2 shows the incidence densities and hazard ratios (HRs) of LM in both groups. The incidence density of LM was higher in the LC group (61.92/1000 person-years) than in the non-LC group (47.48/1000 person-years), with a significantly increased adjusted HR of 1.15 (95% CI = 1.04–1.28, P = 0.007). The
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
(Urbanization level 1), suburban (Urbanization levels 2–3) and rural (Urbanization levels 4–7) as per the National Statistics of Regional Standard Classification (28). This study also used enrollee category, which can be used as a proxy measure of individual socioeconomic status (29). All patients were categorized as high socioeconomic status (civil servants, full-time or regular paid personnel with a government affiliation, employees of privately owned institutions), middle (selfemployed individuals, other employees and members of the farmers’ or fishermen’s associations) and low socioeconomic status (veterans, members of low-income families and substitute service draftees) (30). Main and adjuvant cancer treatment modalities received by CRC patients depended on disease severity according to the NCCN clinical practice guidelines. This study adjusted cancer treatment modality types to reduce the potential for confounding by disease severity (Tables 1 and 2). Adjusted treatment modality types included surgery alone, surgery combining with concurrent chemoradiotherapy (CCRT), chemotherapy or radiotherapy or definite CCRT, chemotherapy alone, radiotherapy alone or supportive care alone. Clinical cancer follow-up examinations were adjusted for avoiding potential disease surveillance/detection bias that patients with LC may be more likely receive frequent clinical examinations, such as abdominal sonography or computed tomography, than their non-LC counterparts, leading to an elevated risk of LM. Other clinical cancer examinations, such as tumor marker CEA and CA199 were also adjusted.
Jpn J Clin Oncol 2014
5
Table 2. Incidence density and hazard ratio of liver metastasis according to liver cirrhosis status and cirrhotic symptoms
Control group Liver cirrhosis group Ascites Jaundice Hepatic encephalopathy Hypoalbuminb PT prolongc
N
Person-years
Liver metastasis events (%)
ID ( per 1000 person-years)
Adjusted HRa
(95% CI)
P value
11 892 2973 336 112 23 176 399
42 393 8334 625 252 44 317 804
2013 (16.9) 516 (17.4) 54 (16.1) 21 (18.8) 4 (17.4) 26 (14.8) 52 (13.0)
47.48 61.92 86.38 83.44 91.91 81.94 64.66
1.00 1.15 1.18 1.38 1.17 1.08 1.07
Reference (1.04–1.28) (0.88–1.58) (0.98–1.95) (0.52–2.61) (0.76–1.52) (0.85–1.34)
– 0.007** 0.28 0.06 0.70 0.67 0.56
Figure 2. Comparison of Kaplan–Meier failure estimates of liver metastasis between the LC and non-LC patients groups. (10-year cumulative risks were 27.1 vs. 23.6%, respectively, P = 0.006). Dotted line = colorectal cancer (CRC) patients without liver cirrhosis (non-LC patients), continuous line = CRC patients with liver cirrhosis (LC patients).
10-year cumulative risk of LM for the LC and the non-LC group was 27.1 and 23.6%, respectively (P = 0.006, Fig. 2). Table 2 also shows HRs of LM according cirrhotic symptoms. Patients with various LC symptoms, including ascites, jaundice, hepatic encephalopathy or ever receiving albumin, frozen plasma (FP), flesh frozen plasma (FFP) transfusion all had higher incidence density of LM (64.66–91.91/1000 person-years) and increased adjusted HR of 1.07–1.38. Significantly higher HRs of LM were associated with all other treatment modality, with a magnitude of HR ranging from 1.59 (surgery and chemotherapy, radiotherapy only) to 5.44 (chemotherapy alone), except surgery only (adjusted HR = 0.78, 95% CI = 0.68– 0.89). For patients who received surgery alone without adjuvant treatment, patients with LC had higher 10-year cumulative risk of LM (23.9 vs. 15.7%, P < 0.001; Fig. 3A). We also compared patients with and without liver cirrhotic symptoms. We found that in the
surgical alone group, CRC patients with cirrhotic symptoms, such as ascites, jaundice, hepatic encephalopathy or albumin, frozen plasma, flesh frozen plasma transfusion, had higher 10-year cumulative risk of LM (25.6 vs. 16.6%, P = 0.009; Fig. 3B). For etiology of LC, the 10-year cumulative risk for HBV and HCV, HBV, HCV, other causes and non-LC were 29.5, 28.9, 27.5, 26.7 and 23.4%, respectively, (P = 0.03; Fig. 4). Patients with both HBV and HCV infections showed the highest risk of LM.
Discussion In this study, we found that non-LC patients had higher socioeconomic status and lived in northern Taiwan where more medical resources were available. Besides, patients with LC were also more likely to reside in rural area where the medical resources were less than in urban area. Patients with LC seemed to be a medical vulnerable group. For the cancer treatment, we did find that lower percentage of LC group
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
ID, incidence density; HR, hazard ratio; CI, confidence interval. a Based on Cox proportional hazard regression with adjustment for gender, age, hepatitis types, diabetes mellitus, hypertension, hyperlipidemia, chronic kidney disease, clinical cancer follow-up examinations, cancer treatment modalities, geographic region, urbanization of residence and socioeconomic status. b Patients who had received albumin or frozen plasma (FP) transfusion. c Patients who had received flesh frozen plasma (FFP) transfusion. **P < 0.01.
6
Effect of liver cirrhosis on metastasis in colorectal cancer patients
ever received abdominal sonography or computed tomography than non-LC group, which was opposite to our thought that patients with LC should receive more examinations. We also found patients with LC had a significantly higher prevalence of comorbidities, including hepatitis, diabetes mellitus, hypertension, hyperlipidemia and chronic kidney disease. These comorbidities could further worse the health inequality.
The risk of LM in CRC patients with LC and cirrhotic symptoms in a hepatitis endemic nation Taiwan is a nation characterized by high prevalence rates of infections by hepatitis B virus (HBV) and hepatitis C virus (HCV). It was
estimated that some 17.3 and 4.4% of the general population were with seroprevalence of HBV and HCV, respectively (31). Our data showed that 5.79% of CRC patients in Taiwan were also with clinical LC, which was much higher than the figure reported in an earlier US study in 1976–80 (0.15%) (32). Liver is both the most frequent synchronous and metachronous metastatic site of CRC. A French study ever reported that the proportion of CRC patients with synchronous LM and the 5-year cumulative metachronous LM rate in CRC patients without LC mentioned were both 14.5% (33). However, very few studies ever explored the relationship between LC and LM in patients with CRC (4–6) and suggested that LC is rare in CRC patients with LM and, vice versa, LM is rare in CRC patients with LC. A Japanese study showed that no LM was found in surgical CRC patients who
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
Figure 3. (A) Comparison of Kaplan–Meier failure estimates of liver metastasis between surgical alone colorectal cancer (CRC) patients with and without liver cirrhosis (LC) (10-year cumulative risk were 23.9 vs. 15.7%, respectively, P < 0.001). Dotted line = surgical alone CRC patients without LC, continuous line = surgical alone CRC patients with LC. (B) Comparison of Kaplan–Meier failure estimates of liver metastasis between surgical alone CRC patients with any LC symptoms, which included ascites, jaundice, hepatic encephalopathy or albumin, frozen plasma, flesh frozen plasma transfusion and surgical alone CRC patients without any symptom of above. (10-year cumulative risk were 25.6 vs. 16.6%, respectively, P = 0.009). Dotted line = surgical alone CRC patients without any LC symptom, continuous line = surgical alone CRC patients with LC symptom.
Jpn J Clin Oncol 2014
7
suffered from LC (5). In that study, LC was not found among the 40 patients with metastases but was present in 46 of the 210 nonmetastatic patients in their surgery department during past 5 years. Another Italian study which included 747 patients undergoing surgery for CRC in the surgery department, noted that synchronous LM were present in 32% of normal livers, but in 15% of fatty livers and 4.7% of patients with LC (4). However, for patients with clinical obvious cirrhotic symptoms, no studies yet ever studied the LM rate in this kind of patients. This cohort study showed the 10-year cumulative LM rate, including synchronous and metachronous, was 23.6% in normal liver, while 27.1% in cirrhotic liver. In further stratified analysis, surgery alone CRC patients with LC symptoms had higher 10-year cumulative LM risk than without symptoms (25.6 vs. 16.6%, P = 0.009).
Possible reason for low LM rate in other surgical studies: selection bias Several possible reasons may be responsible for the low LM rate observed in previous studies. First, both Italian and Japanese studies included only patients who received surgery in the surgery department. CRC patients with severe LC may not be referred to surgical department for hepatic metastasectomy because of poor reserved liver function (34). Our nationwide population-based study would not have such selection bias problem. Secondary, the Italian study had more early cancer stage in LC patients while more Stage IV and nodal involvement in non-LC patients (P < 0.02 and P < 0.001), which would lead to less LM noted in cirrhotic liver just simply because a greater proportion of LC patients had earlier cancer stage disease. In this study, the cancer treatment modalities received by patients have been adjusted (Table 2).
Possible underlying mechanisms LC is an ongoing and chronic disease that can result in grave or lethal complications of other vital organs and body systems, such as digestive, circulatory, central nervous and immune systems. Our result implied a positive link between LC and LM. The etiology may include local microenvironment change or immune reaction change.
Mechanical factors may play a role such as that the mesenteric circulation and fenestrated capillaries of the liver facilitate hepatic metastasis (35). Several studies have found a significantly increased intestinal permeability of small intestinal, gastroduodenal and whole intestine in patients with LC when compared with healthy subjects (36,37). There had been several indications for the presence of intestinal epithelial barrier dysfunction in patients with LC, especially in those with cirrhosis complications (38). LC is also an advanced stage of liver fibrosis accompanied by vascular remodeling and distortion, which could lead to shunting of the portal and arterial blood supply directly into the hepatic outflow (central veins). Vascular distortion and blood speed deceleration may further facilitate cell seeding (1, 38). Further investigations for the mechanism with which cirrhotic microenvironment change affects metastasis would help better understand the cancer metastasis process.
Study strength Our study had several methodological strengths. First, the study subjects were collected from a nationwide population-based database. Our sample included all CRC patients receiving all kinds of cancer treatment modalities and eliminated the selection bias that severe LC patients may not receive hepatic surgery by clinical consideration. Second, nationwide insurance coverage minimized the potential for other selection bias, such as lost follow-up. Third, insurance claim records had information of medical treatments so that residual liver function and cirrhotic symptoms could be evaluated. There was no study yet ever investigating the relationship of cirrhotic symptoms and LM before this study.
Study limitations Despite the above strengths, limitations were noted in our study. First, as other previous studies lacking cancer stage, such as the Japanese study etc., this study still lack such information. However, main and adjuvant cancer treatment modalities patients received were adjusted in this study. Furthermore, subgroup analysis showed that for patients receiving surgery alone without adjuvant anti-cancer treatments, who were mostly early cancer stage with only locoregional disease, patients
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
Figure 4. Comparison of Kaplan–Meier failure estimates of liver metastasis between colorectal cancer (CRC) patients with liver cirrhosis (LC) of various etiologies. (10-year cumulative risk for HBV and HCV, HBV, HCV, other causes and non-LC were 29.5, 28.9, 27.5, 26.7, and 23.4%, respectively, P = 0.03). Very-short-dash line = HBV and HCV related LC, long-dash line = HBV related LC, dotted line = HCV related LC, continuous line = other causes LC, dash line = without LC.
8
Effect of liver cirrhosis on metastasis in colorectal cancer patients
Conclusion Our study found that CRC patients with LC was a possible medical vulnerable group. LC prevalence is underestimated among CRC patients with LM in Taiwan and vice versa, LM risk is severely underestimated and even higher in CRC patients with LC. CRC patients with LC should be paid more attention about LM in clinical practice.
Acknowledgements This study is based in part on data from the National Health Insurance Research Database provided by the Bureau of National Health Insurance, Department of Health, and managed by National Health Research Institutes (registry number 99029). The interpretation and conclusions contained herein are not those of the Bureau of National Health Insurance, Department of Health, or National Health Research Institutes.
Funding This study was supported by research grants of Buddhist Dalin Tzu Chi Hospital (DTCRD-103(2)-I-21). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Conflict of interest statement The authors have declared that no competing interests exist.
References 1. Schuppan D, Afdhal NH. Liver cirrhosis. Lancet 2008;371:838–51. 2. Cabibbo G, Palmeri L, Palmeri S, Craxi A. Should cirrhosis change our attitude towards treating non-hepatic cancer? Liver Int 2012;32:21–7. 3. Sorensen HT, Friis S, Olsen JH, et al. Risk of liver and other types of cancer in patients with cirrhosis: a nationwide cohort study in Denmark. Hepatology 1998;28:921–5. 4. Iascone C, Ruperto M, Barillari P. [Occurrence of synchronous colorectal cancer metastasis in the cirrhotic or fatty liver]. Minerva Chir 2005;60:185–90. 5. Uetsuji S, Yamamura M, Yamamichi K, Okuda Y, Takada H, Hioki K. Absence of colorectal cancer metastasis to the cirrhotic liver. Am J Surg 1992;164:176–7. 6. Qian HG, Zhang J, Leng JH, et al. [Association of hepatitis B virus infection and cirrhosis with liver metastasis in colorectal cancer]. Zhonghua Wei Chang Wai Ke Za Zhi 2010;13:202–4.
7. World Health Organisation. International Agency for Research on Cancer. GLOBOCAN. Cancer incidence and mortality worldwide in 2008. http:// globocan.iarc.fr/Pages/fact_sheets_cancer.aspx (27 October 2014, date last accessed). 8. Edwards BK, Ward E, Kohler BA, et al. Annual report to the nation on the status of cancer, 1975–2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates. Cancer 2010;116:544–73. 9. Report of Taiwan Cancer Registry. 2010. http://www.hpa.gov.tw/BHPNet/ Web/Stat/StatisticsShow.aspx?No=201305060001 (27 October 2014, date last accessed). 10. Fong Y, Cohen AM, Fortner JG, et al. Liver resection for colorectal metastases. J Clin Oncol 1997;15:938–46. 11. Choti MA, Sitzmann JV, Tiburi MF, et al. Trends in long-term survival following liver resection for hepatic colorectal metastases. Ann Surg 2002;235:759–66. 12. Gallagher DJ, Kemeny N. Metastatic colorectal cancer: from improved survival to potential cure. Oncology 2010;78:237–48. 13. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) Rectal Cancer Version 3. 2013. http:// www.nccn.org/professionals/physician_gls/f_guidelines.asp (27 October 2014, date last accessed). 14. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) Colon Cancer. Version 2. 2013. http:// www.nccn.org/professionals/physician_gls/f_guidelines.asp (27 October 2014, date last accessed). 15. Wanebo HJ, Semoglou C, Attiyeh F, Stearns MJ Jr. Surgical management of patients with primary operable colorectal cancer and synchronous liver metastases. Am J Surg 1978;135:81–5. 16. Bureau of National Health Insurance: National Health Insurance Annual Statistical Report. Taipei, Taiwan, 1995–2004. 17. NHI. NHI profile. http://www.nhi.gov.tw/english/webdata.asp?menu= 11&menu_id=290&webdata_id=1884. 2008 (7 July 2008, date last accessed). 18. Lu JF, Hsiao WC. Does universal health insurance make health care unaffordable? Lessons from Taiwan. Health Aff 2003;22:77–88. 19. Tseng CH. Mortality and causes of death in a national sample of diabetic patients in Taiwan. Diabetes Care 2004;27:1605–9. 20. Bureau of National Health Insurance Website (2000) Methods for estimating false claims. http://www.nhi.gov.tw/ information/bulletin_file/421– 0890036465-19.doc (1 October 2009, date last accessed). 21. Chen HF, Chen P, Li CY. Risk of malignant neoplasms of liver and biliary tract in diabetic patients with different age and sex stratifications. Hepatology 2010;52:155–63. 22. Wu PH, Lin YT, Kuo CN, Chang WC, Chang WP. No increased risk of herpes zoster found in cirrhotic patients: a nationwide population-based study in Taiwan. PLoS One 2014;9:e93443. 23. Center for Health Policy Research and Development, National Health Research Institutes: Evidence Based Practice Guidelines 2009: Treatment Guidelines for Liver Cirrhosis. http://ebpg.nhri.org.tw/check/9.pdf (27 October 2014, date last accessed). 24. Sharabi Y, Eldad A. Nonalcoholic fatty liver disease is associated with hyperlipidemia and obesity. Am J Med 2000;109:171. 25. Fabrizi F, Dixit V, Martin P, Jadoul M, Messa P. Meta-analysis: the impact of nutritional status on the immune response to hepatitis B virus vaccine in chronic kidney disease. Dig Dis Sci 2012;57:1366–72. 26. Iascone C, Ruperto M, Barillari P. [Colorectal carcinoma metastasis in livers infected with hepatitis B or C virus]. Minerva Chir 2005;60:77–81. 27. Liu CY, Hung YT, Chuang YL, et al. Incorporating development stratification of Taiwan townships into sampling design of large scale health interview survey. J Health Manag 2006;4:1–22. 28. Chen HF, Chen P, Li CY. Risk of malignant neoplasm of the pancreas in relation to diabetes: a population-based study in Taiwan. Diabetes Care 2011;34:1177–9. 29. Lee CC, Chien SH, Hung SK, Yang WZ, Su YC. Effect of individual and neighborhood socioeconomic status on oral cancer survival. Oral Oncol 2012;48:253–61.
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
with LC or cirrhotic symptoms still had higher LM risk. Second, in this study, only age and sex were matched without other characteristics such as comorbidities matched. Although this study design could allow these other characteristics to be analyzed, and discussed after regression adjustment to realize the difference of these two groups with and without LC, this was still a limitation for conclusion. Third, the diagnosis of LM was ICD-9-CM code 197.7, which was made by clinical physicians according to histologically defined or radiological criteria sometimes. However, LM misjudgment should have been minimized and could be ignored under the cancer treatment policy in Taiwan that cancer status dispute or treatment strategy changes would be confirmed later by multidisciplinary experts meeting to decide the subsequent treatment choices. Moreover, LM was confirmed by catastrophic illness certificates that further ensure the accuracy of cancer diagnosis. Other malignancy (ICD-9-CM codes 140–208, excluding CRC), such as hepatocellular carcinoma was also excluded to prevent the double malignancy condition.
Jpn J Clin Oncol 2014 30. Chen CY, Liu CY, Su WC, Huang SL, Lin KM. Factors associated with the diagnosis of neurodevelopmental disorders: a population-based longitudinal study. Pediatrics 2007;119:e435–43. 31. Chen CH, Yang PM, Huang GT, Lee HS, Sung JL, Sheu JC. Estimation of seroprevalence of hepatitis B virus and hepatitis C virus in Taiwan from a large-scale survey of free hepatitis screening participants. J Formos Med Assoc 2007;106:148–55. 32. Everhart J, editor. Digestive Diseases in the United States: Epidemiology and Impact. US Department of Health and Human Services, Public Health Service, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases. Washington, DC: US Government Printing Office, 1994; NIH publication no. 94–1447. 33. Manfredi S, Lepage C, Hatem C, Coatmeur O, Faivre J, Bouvier AM. Epidemiology and management of liver metastases from colorectal cancer. Ann Surg 2006;244:254–9.
9
34. Hanje AJ, Patel T. Preoperative evaluation of patients with liver disease. Nat Clin Pract Gastroenterol Hepatol 2007;4:266–76. 35. Lalor PF, Lai WK, Curbishley SM, Shetty S, Adams DH. Human hepatic sinusoidal endothelial cells can be distinguished by expression of phenotypic markers related to their specialised functions in vivo. World J Gastroenterol 2006;12:5429–39. 36. Benjamin J, Singla V, Arora I, Sood S, Joshi YK. Intestinal permeability and complications in liver cirrhosis: a prospective cohort study. Hepatol Res 2013;43:200–7. 37. Zuckerman MJ, Menzies IS, Ho H, et al. Assessment of intestinal permeability and absorption in cirrhotic patients with ascites using combined sugar probes. Dig Dis Sci 2004;49:621–6. 38. Pijls KE, Jonkers DM, Elamin EE, Masclee AA, Koek GH. Intestinal epithelial barrier function in liver cirrhosis: an extensive review of the literature. Liver Int 2013;33:1457–69.
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
Original Article
Effect of liver cirrhosis on metastasis in colorectal cancer patients: a nationwide population-based cohort study Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
Wen-Yen Chiou1,2,3, Chun-Ming Chang2,4, Kuo-Chih Tseng2,5, Shih-Kai Hung1,2, Hon-Yi Lin1,2, Yi-Chun Chen2,6, Yu-Chieh Su2,7, Chih-Wei Tseng2,5, Shiang-Jiun Tsai1, Moon-Sing Lee1,2,*, and Chung-Yi Li3,* 1
Department of Radiation Oncology, Buddhist Dalin Tzu Chi Hospital, Chiayi, 2School of Medicine, Tzu Chi University, Hualien, 3Department of Public Health, College of Medicine, National Cheng Kung University, Tainan, 4Department of Surgery, Buddhist Dalin Tzu Chi Hospital, Chiayi, 5Division of Gastroenterology, Department of Internal Medicine, Buddhist Dalin Tzu Chi Hospital, Chiayi, 6Division of Nephrology, Department of Internal Medicine, Buddhist Dalin Tzu Chi Hospital, Chiayi, and 7Division of Hematology-Oncology, Department of Internal Medicine, Buddhist Dalin Tzu Chi Hospital, Chiayi, Taiwan *For reprints and all correspondence: Moon-Sing Lee, Department of Radiation Oncology, Buddhist Dalin Tzu Chi Hospital, No. 2, Ming Sheng Road, Dalin, Chiayi, Taiwan. E-mail: [email protected]. Chung-Yi Li, Department of Public Health, College of Medicine, National Cheng Kung University, No.1, University Road, Tainan City 701, Taiwan. E-mail: armstrong_washington@ hotmail.com Received 8 August 2014; Accepted 14 October 2014
Abstract Objective: The aim of this study is to evaluate the liver metastasis risk among colorectal cancer patients with liver cirrhosis. Methods: This was a nationwide population-based cohort study of 2973 newly diagnosed colorectal cancer patients with liver cirrhosis and 11 892 age–sex matched controls enrolled in Taiwan between 2000 and 2010. The cumulative risk by Kaplan–Meier method, hazard ratio by the multivariate Cox proportional model and the incidence density were evaluated. Results: The median time interval from the colorectal cancer diagnosis to the liver metastasis event was 7.42 months for liver cirrhosis group and 7.67 months for non-liver cirrhosis group. The incidence density of liver metastasis was higher in the liver cirrhosis group (61.92/1000 person-years) than in the non-liver cirrhosis group (47.48/1000 person-years), with a significantly adjusted hazard ratio of 1.15 (95% CI = 1.04–1.28, P = 0.007). The 10-year cumulative risk of liver metastasis for the liver cirrhosis and the non-liver cirrhosis group was 27.1 and 23.6%, respectively (P = 0.006). For early cancer stage with locoregional disease patients receiving surgery alone without adjuvant anti-cancer treatments, patients with liver cirrhosis (10-year cumulative risk 23.9 vs. 15.7%, P < 0.001) or cirrhotic symptoms (10-year cumulative risk 25.6 vs. 16.6%, P = 0.009) both still had higher liver metastasis risk compared with their counterparts. For etiologies of liver cirrhosis, the 10-year cumulative risk for hepatitis B virus and hepatitis C virus, hepatitis B virus, hepatitis C virus, other causes and nonliver cirrhosis were 29.5, 28.9, 27.5, 26.7 and 23.4%, respectively, (P = 0.03). Conclusions: Our study found that liver metastasis risk was underestimated and even higher in colorectal cancer patients with liver cirrhosis.
© The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
1
2
Effect of liver cirrhosis on metastasis in colorectal cancer patients
Key words: liver cirrhosis, colorectal cancer, liver metastasis
Introduction
Patients and methods Ethics statement This study was reviewed and approved by the National Health Insurance Research Database (NHIRD) research committee and by the institutional review board (IRB no. B10104024) of Buddhist Dalin Tzu Chi Hospital. Informed consent was waived because the patients’ personal identification numbers were de-identified.
Source of data Patients were drawn from the NHIRD, released for research purposes by the National Health Research Institutes, Taipei, Taiwan (16). This dataset provided all original claims data from the Taiwan National Health Insurance program, lunched since 1995, a mandatory-enrollment, single-payment system that financed healthcare for all Taiwanese citizens
Study design and patients This was a population-based cohort study, with the study period spanning from 1 January 2000 through 31 December 2010. We first identified patients with a diagnosis of CRC (ICD-9-CM codes 153.0 to 153.9 and 154.0 to 154.2) from inpatient and outpatient claims during the study period. These CRC patients must also have been issued catastrophic illness certificates that further ensure the accuracy of cancer diagnosis (21). To ensure the inclusion of only newly diagnosed CRC between 2000 and 2010, patients who had a diagnosis of CRC between 1995 and 1999 were excluded. We further excluded patients with claims for any other malignancy (ICD-9-CM codes 140–208, excluding CRC) from 1995 to 2010, such as hepatocellular carcinoma. A total of 91 441 patients with newly diagnosed CRC were identified. Among the 91 441 CRC patients, 2973 patients who had ICD-9-CM codes 571.2, alcoholic cirrhosis of liver, 571.5, cirrhosis of liver without mention of alcohol and 571.6, biliary cirrhosis, made in at least one inpatient or two outpatient medical records to make sure not a miscoding, before the initial diagnosis of CRC during the study period were defined to have a diagnosis of underlying liver cirrhosis (LC) in this study (22). Because LC was prevalent in Taiwan, the government paid close attention to this health issue and has established specific guidelines for diagnosis and treatment of LC (23). For each CRC patient with underlying LC, four CRC patients without LC were selected using frequency matching on age and gender, leading to 11 892 CRC patients without LC (Fig. 1). Each study participant was followed until the occurrence of the following outcomes whichever comes first: (i) a diagnosis of LM (ICD-9-CM code 197.7) was made; (ii) the participant withdrew from the NHI system mainly due to death; and (iii) 31 December 2010. LC severity and residual liver function information were evaluated by cirrhotic symptoms including ascites, jaundice, hepatic encephalopathy or albumin, frozen plasma (FP), flesh frozen plasma (FFP) transfusion.
Measurements of end-points and covariates The primary dependent variable of interest was liver metastasis, including both synchronous and metachronous, which was identified
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
Liver cirrhosis (LC) is a serious disease that results in permanent damage to the liver. There are a variety of diseases and conditions that can cause cirrhosis of the liver. Alcoholic liver disease and hepatitis C are the most common causes in developed countries, whereas hepatitis B is the prevailing cause in most parts of Asia and sub-Saharan Africa (1). The major clinical consequences of cirrhosis are impaired liver function, an increased intrahepatic resistance ( portal hypertension) and the development of hepatocellular carcinoma (1). However, there are limited data in the literature on the incidence, prevalence and management of non-hepatic cancers (NHC) in cirrhotic patients (2,3). Previous studies suggested that LC might play a role in NHC metastasis, but the studies yield inconsistent results mainly due to limited sample sizes (4–6). Colorectal cancer is a worldwide health problem, particularly in developed countries, comprising 9.4% of all cancer cases with ∼1 million diagnoses annually worldwide (7). It is also one of the leading causes of cancer related deaths in America, Europe and Asia (8). In Taiwan, like in many other nations, colorectal cancer (CRC) is also a major public health problem. According to the report of the Bureau of Health Promotion, Taiwan, the CRC has become the most common malignancy in Taiwan since 2006 and the crude incidence rate was 60.62 per 100 000 people in 2010 (9). CRC cells travel to the liver through the hepatic portal circulation, and liver is the most common anatomical site for hematogenous metastasis (10). Unlike other malignancy, the CRC patients with limited metastatic disease confined to the liver who are able to undergo resection may still have 25–37%, even 58%, chance of long-term survival (11,12). The National Comprehensive Cancer Network (NCCN) clinical practice guidelines in Oncology also suggested that hepatic resection is the treatment of choice for resectable liver metastasis (LM) from CRC (13,14). It was found that patients with an untreated single LM had a median survival of 19 months, with no patients surviving 5 years, while patients with a resected single LM had a median survival of 36 months with 25% of patients surviving 5 years (15). By identifying risk factors of LM, it could be expected to find more resectable LM and achieve better outcomes. The aim of this study was to estimate the effect of LC on risk of LM in CRC, which was one of the cancers likely to cause LM.
(>99% of the 24 million inhabitants of Taiwan) through a registry of board certified specialists and contracted medical facilities (∼97% of all Taiwan medical facilities) (17). The NHIRD is one of the largest and most comprehensive databases in the world and has been used extensively in various studies (18). For the precision of the claims data, the Bureau of National Health Insurance (BNHI) performs expert reviews on a random sample of every 50–100 ambulatory and every 20 inpatient claims in each hospital and clinic quarterly and interviews patients in order to verify the accuracy of diagnosis (19). False reports of diagnosis would yield a severe penalty from the BNHI (20). Therefore, information obtained from the NHIRD is considered to be complete and accurate (21). The NHIRD provided information on encrypted patient identification numbers, sex, birthday, dates of admission and discharge, medical institutions providing the services, the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes of up to five diagnoses and procedures, respectively, outcome at hospital discharge (recovered, died or transferred out), order codes and the fees charged for medical services.
Jpn J Clin Oncol 2014
3
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
Figure 1. Study design flowchart. NHI, National Health Insurance; CRC, colorectal cancer; ICD-9, International Classification of Diseases, Ninth Revision; LC, liver cirrhosis.
from one inpatient or two outpatient claims. As mentioned above, we also excluded patients with claims for any other malignancy from 1995 to 2010 to rule out false LM events from hepatocellular carcinoma or other NHCs. The covariates considered included gender, age at CRC diagnosis, hepatitis, comorbidities, clinical examinations, cancer treatment modalities, geographic region of living, urbanization level of residence and socioeconomic status. Hepatitis types included hepatitis B (ICD-9-CM codes 070.22, 070.23, 070.32, 070.33, V02.61) and hepatitis C (070.44, 070.54, V02.62). Comorbidities included diabetes mellitus (ICD-9-CM code 250), hypertension (ICD-9-CM codes 401–405), hyperlipidemia (ICD-9-CM codes 272.0–272.4)
and chronic kidney disease (ICD-9-CM code 585). These comorbidities were adjusted for their possible role in liver microenvironment (24–26). Geographic regions, urbanization level of residence and socioeconomic status information were adjusted for possible uneven medical resources to prevent information bias. There were four geographic regions (northern, central, southern, eastern and other) in Taiwan. The urbanization level of residence was classified as seven levels based on five indices: population density, percentages of residences with college or higher education, percentages of residents >65 years old, percentages of residents who were agriculture workers and the number of physicians per 100 000 people (27). Urbanization level of residence was further recorded as urban
4
Effect of liver cirrhosis on metastasis in colorectal cancer patients
Statistical analysis The statistical software packages SAS (version 9.2; SAS Institute, Inc., Cary, NC, USA) were used for data analysis. To assess the independent effects of LC on the risk of LM, we used Cox proportional hazard regression models, adjusting for the aforementioned potential confounders. The incidence density was calculated with person-years as the denominator under the Poisson assumption. Denominator of the incidence density was the sum of individual’s person-years at-risk, namely from date of first ever CRC diagnosis to date of LM or censored. The Kaplan–Meier method was used to estimate the cumulative risk of LM. A two-sided P value < 0.05 was defined as statistically significant.
Results Subject characteristics The distributions of demographic characteristics and comorbidities for the CRC patients with and without LC are shown in Table 1. Over 66% CRC patients with LC were male. The mean (±SD) age of patients with LC was 67.59 (±11.64). Compared with the non-LC patients, patients with LC had a significantly higher prevalence of comorbidities, which may have effect on liver microenvironment, including hepatitis, diabetes mellitus, hypertension, hyperlipidemia and chronic kidney disease. Most patients, both LC and non-LC groups, received surgery alone for CRC without adjuvant therapy. Compared with those without LC, lower percentage of patients with LC had cancer follow-up examinations, such as CEA, abdominal sonography and computed tomography examination, except CA199. Patients with LC were more likely to reside in the rural areas. A higher percentage of non-LC patients had high socioeconomic status and lived in northern Taiwan where more medical resources were available.
Risk of LM in CRC subjects with LC The median time intervals from the CRC diagnosis to the LM event were 7.42 months for the LC group and 7.67 months for the
Table 1. Characteristics of colorectal cancer patients with and without liver cirrhosis Variables
With liver cirrhosis N = 2973 (%)
Characteristics Gender Female 33.54 Male 66.46 Age 0–44 2.76 45–54 11.10 55–64 21.49 65–74 32.53 75+ 32.12 Mean age (years) ± SD 67.59 ± 11.64 Comorbidities Hepatitis B 5.99 Hepatitis C 19.17 Diabetes mellitus 44.70 Hypertension 65.52 Hyperlipidemia 37.77 Chronic kidney disease 11.47 Cancer treatment modalities Surgery alone 53.18 Surgery with adjuvant 16.38 chemotherapy 5.45 Surgery with adjuvant CCRT Surgery with adjuvant 3.50 radiotherapy CCRT 1.48 Chemotherapy alone 2.52 Radiotherapy alone 1.35 Supportive care 16.15 Clinical cancer follow-up examinations CEA 64.51 CA199 16.82 Abdominal sonography 10.39 Computed tomography 43.39 Geographic region Northern 39.02 Central 26.85 Southern 31.11 Eastern 3.03 Urbanization level Urban 23.98 Suburban 41.20 Rural 34.81 Socioeconomic status High 22.37 Middle 52.37 Low 25.26
Without liver cirrhosis N = 11 892a (%)
33.54 66.46 2.76 11.10 21.49 32.53 32.12 67.59 ± 11.98 0.56 1.7 30.76 57.94 33.77 5.46 50.63 21.33 6.54 3.67 1.46 1.42 1.07 13.87 67.13 14.93 44.48 64.83 45.66 23.44 28.83 2.07 27.61 44.11 28.28 28.05 45.86 26.08
a
Frequency match on age and gender with 4 : 1 ratio. CCRT, concurrent chemoradiotherapy; CEA, carcinoembryonic antigen; CA199, carbohydrate antigen 19-9.
non-LC group. The overall LM rates were 17.4 and 16.9% for the LC and non-LC group, respectively. Table 2 shows the incidence densities and hazard ratios (HRs) of LM in both groups. The incidence density of LM was higher in the LC group (61.92/1000 person-years) than in the non-LC group (47.48/1000 person-years), with a significantly increased adjusted HR of 1.15 (95% CI = 1.04–1.28, P = 0.007). The
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
(Urbanization level 1), suburban (Urbanization levels 2–3) and rural (Urbanization levels 4–7) as per the National Statistics of Regional Standard Classification (28). This study also used enrollee category, which can be used as a proxy measure of individual socioeconomic status (29). All patients were categorized as high socioeconomic status (civil servants, full-time or regular paid personnel with a government affiliation, employees of privately owned institutions), middle (selfemployed individuals, other employees and members of the farmers’ or fishermen’s associations) and low socioeconomic status (veterans, members of low-income families and substitute service draftees) (30). Main and adjuvant cancer treatment modalities received by CRC patients depended on disease severity according to the NCCN clinical practice guidelines. This study adjusted cancer treatment modality types to reduce the potential for confounding by disease severity (Tables 1 and 2). Adjusted treatment modality types included surgery alone, surgery combining with concurrent chemoradiotherapy (CCRT), chemotherapy or radiotherapy or definite CCRT, chemotherapy alone, radiotherapy alone or supportive care alone. Clinical cancer follow-up examinations were adjusted for avoiding potential disease surveillance/detection bias that patients with LC may be more likely receive frequent clinical examinations, such as abdominal sonography or computed tomography, than their non-LC counterparts, leading to an elevated risk of LM. Other clinical cancer examinations, such as tumor marker CEA and CA199 were also adjusted.
Jpn J Clin Oncol 2014
5
Table 2. Incidence density and hazard ratio of liver metastasis according to liver cirrhosis status and cirrhotic symptoms
Control group Liver cirrhosis group Ascites Jaundice Hepatic encephalopathy Hypoalbuminb PT prolongc
N
Person-years
Liver metastasis events (%)
ID ( per 1000 person-years)
Adjusted HRa
(95% CI)
P value
11 892 2973 336 112 23 176 399
42 393 8334 625 252 44 317 804
2013 (16.9) 516 (17.4) 54 (16.1) 21 (18.8) 4 (17.4) 26 (14.8) 52 (13.0)
47.48 61.92 86.38 83.44 91.91 81.94 64.66
1.00 1.15 1.18 1.38 1.17 1.08 1.07
Reference (1.04–1.28) (0.88–1.58) (0.98–1.95) (0.52–2.61) (0.76–1.52) (0.85–1.34)
– 0.007** 0.28 0.06 0.70 0.67 0.56
Figure 2. Comparison of Kaplan–Meier failure estimates of liver metastasis between the LC and non-LC patients groups. (10-year cumulative risks were 27.1 vs. 23.6%, respectively, P = 0.006). Dotted line = colorectal cancer (CRC) patients without liver cirrhosis (non-LC patients), continuous line = CRC patients with liver cirrhosis (LC patients).
10-year cumulative risk of LM for the LC and the non-LC group was 27.1 and 23.6%, respectively (P = 0.006, Fig. 2). Table 2 also shows HRs of LM according cirrhotic symptoms. Patients with various LC symptoms, including ascites, jaundice, hepatic encephalopathy or ever receiving albumin, frozen plasma (FP), flesh frozen plasma (FFP) transfusion all had higher incidence density of LM (64.66–91.91/1000 person-years) and increased adjusted HR of 1.07–1.38. Significantly higher HRs of LM were associated with all other treatment modality, with a magnitude of HR ranging from 1.59 (surgery and chemotherapy, radiotherapy only) to 5.44 (chemotherapy alone), except surgery only (adjusted HR = 0.78, 95% CI = 0.68– 0.89). For patients who received surgery alone without adjuvant treatment, patients with LC had higher 10-year cumulative risk of LM (23.9 vs. 15.7%, P < 0.001; Fig. 3A). We also compared patients with and without liver cirrhotic symptoms. We found that in the
surgical alone group, CRC patients with cirrhotic symptoms, such as ascites, jaundice, hepatic encephalopathy or albumin, frozen plasma, flesh frozen plasma transfusion, had higher 10-year cumulative risk of LM (25.6 vs. 16.6%, P = 0.009; Fig. 3B). For etiology of LC, the 10-year cumulative risk for HBV and HCV, HBV, HCV, other causes and non-LC were 29.5, 28.9, 27.5, 26.7 and 23.4%, respectively, (P = 0.03; Fig. 4). Patients with both HBV and HCV infections showed the highest risk of LM.
Discussion In this study, we found that non-LC patients had higher socioeconomic status and lived in northern Taiwan where more medical resources were available. Besides, patients with LC were also more likely to reside in rural area where the medical resources were less than in urban area. Patients with LC seemed to be a medical vulnerable group. For the cancer treatment, we did find that lower percentage of LC group
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
ID, incidence density; HR, hazard ratio; CI, confidence interval. a Based on Cox proportional hazard regression with adjustment for gender, age, hepatitis types, diabetes mellitus, hypertension, hyperlipidemia, chronic kidney disease, clinical cancer follow-up examinations, cancer treatment modalities, geographic region, urbanization of residence and socioeconomic status. b Patients who had received albumin or frozen plasma (FP) transfusion. c Patients who had received flesh frozen plasma (FFP) transfusion. **P < 0.01.
6
Effect of liver cirrhosis on metastasis in colorectal cancer patients
ever received abdominal sonography or computed tomography than non-LC group, which was opposite to our thought that patients with LC should receive more examinations. We also found patients with LC had a significantly higher prevalence of comorbidities, including hepatitis, diabetes mellitus, hypertension, hyperlipidemia and chronic kidney disease. These comorbidities could further worse the health inequality.
The risk of LM in CRC patients with LC and cirrhotic symptoms in a hepatitis endemic nation Taiwan is a nation characterized by high prevalence rates of infections by hepatitis B virus (HBV) and hepatitis C virus (HCV). It was
estimated that some 17.3 and 4.4% of the general population were with seroprevalence of HBV and HCV, respectively (31). Our data showed that 5.79% of CRC patients in Taiwan were also with clinical LC, which was much higher than the figure reported in an earlier US study in 1976–80 (0.15%) (32). Liver is both the most frequent synchronous and metachronous metastatic site of CRC. A French study ever reported that the proportion of CRC patients with synchronous LM and the 5-year cumulative metachronous LM rate in CRC patients without LC mentioned were both 14.5% (33). However, very few studies ever explored the relationship between LC and LM in patients with CRC (4–6) and suggested that LC is rare in CRC patients with LM and, vice versa, LM is rare in CRC patients with LC. A Japanese study showed that no LM was found in surgical CRC patients who
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
Figure 3. (A) Comparison of Kaplan–Meier failure estimates of liver metastasis between surgical alone colorectal cancer (CRC) patients with and without liver cirrhosis (LC) (10-year cumulative risk were 23.9 vs. 15.7%, respectively, P < 0.001). Dotted line = surgical alone CRC patients without LC, continuous line = surgical alone CRC patients with LC. (B) Comparison of Kaplan–Meier failure estimates of liver metastasis between surgical alone CRC patients with any LC symptoms, which included ascites, jaundice, hepatic encephalopathy or albumin, frozen plasma, flesh frozen plasma transfusion and surgical alone CRC patients without any symptom of above. (10-year cumulative risk were 25.6 vs. 16.6%, respectively, P = 0.009). Dotted line = surgical alone CRC patients without any LC symptom, continuous line = surgical alone CRC patients with LC symptom.
Jpn J Clin Oncol 2014
7
suffered from LC (5). In that study, LC was not found among the 40 patients with metastases but was present in 46 of the 210 nonmetastatic patients in their surgery department during past 5 years. Another Italian study which included 747 patients undergoing surgery for CRC in the surgery department, noted that synchronous LM were present in 32% of normal livers, but in 15% of fatty livers and 4.7% of patients with LC (4). However, for patients with clinical obvious cirrhotic symptoms, no studies yet ever studied the LM rate in this kind of patients. This cohort study showed the 10-year cumulative LM rate, including synchronous and metachronous, was 23.6% in normal liver, while 27.1% in cirrhotic liver. In further stratified analysis, surgery alone CRC patients with LC symptoms had higher 10-year cumulative LM risk than without symptoms (25.6 vs. 16.6%, P = 0.009).
Possible reason for low LM rate in other surgical studies: selection bias Several possible reasons may be responsible for the low LM rate observed in previous studies. First, both Italian and Japanese studies included only patients who received surgery in the surgery department. CRC patients with severe LC may not be referred to surgical department for hepatic metastasectomy because of poor reserved liver function (34). Our nationwide population-based study would not have such selection bias problem. Secondary, the Italian study had more early cancer stage in LC patients while more Stage IV and nodal involvement in non-LC patients (P < 0.02 and P < 0.001), which would lead to less LM noted in cirrhotic liver just simply because a greater proportion of LC patients had earlier cancer stage disease. In this study, the cancer treatment modalities received by patients have been adjusted (Table 2).
Possible underlying mechanisms LC is an ongoing and chronic disease that can result in grave or lethal complications of other vital organs and body systems, such as digestive, circulatory, central nervous and immune systems. Our result implied a positive link between LC and LM. The etiology may include local microenvironment change or immune reaction change.
Mechanical factors may play a role such as that the mesenteric circulation and fenestrated capillaries of the liver facilitate hepatic metastasis (35). Several studies have found a significantly increased intestinal permeability of small intestinal, gastroduodenal and whole intestine in patients with LC when compared with healthy subjects (36,37). There had been several indications for the presence of intestinal epithelial barrier dysfunction in patients with LC, especially in those with cirrhosis complications (38). LC is also an advanced stage of liver fibrosis accompanied by vascular remodeling and distortion, which could lead to shunting of the portal and arterial blood supply directly into the hepatic outflow (central veins). Vascular distortion and blood speed deceleration may further facilitate cell seeding (1, 38). Further investigations for the mechanism with which cirrhotic microenvironment change affects metastasis would help better understand the cancer metastasis process.
Study strength Our study had several methodological strengths. First, the study subjects were collected from a nationwide population-based database. Our sample included all CRC patients receiving all kinds of cancer treatment modalities and eliminated the selection bias that severe LC patients may not receive hepatic surgery by clinical consideration. Second, nationwide insurance coverage minimized the potential for other selection bias, such as lost follow-up. Third, insurance claim records had information of medical treatments so that residual liver function and cirrhotic symptoms could be evaluated. There was no study yet ever investigating the relationship of cirrhotic symptoms and LM before this study.
Study limitations Despite the above strengths, limitations were noted in our study. First, as other previous studies lacking cancer stage, such as the Japanese study etc., this study still lack such information. However, main and adjuvant cancer treatment modalities patients received were adjusted in this study. Furthermore, subgroup analysis showed that for patients receiving surgery alone without adjuvant anti-cancer treatments, who were mostly early cancer stage with only locoregional disease, patients
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
Figure 4. Comparison of Kaplan–Meier failure estimates of liver metastasis between colorectal cancer (CRC) patients with liver cirrhosis (LC) of various etiologies. (10-year cumulative risk for HBV and HCV, HBV, HCV, other causes and non-LC were 29.5, 28.9, 27.5, 26.7, and 23.4%, respectively, P = 0.03). Very-short-dash line = HBV and HCV related LC, long-dash line = HBV related LC, dotted line = HCV related LC, continuous line = other causes LC, dash line = without LC.
8
Effect of liver cirrhosis on metastasis in colorectal cancer patients
Conclusion Our study found that CRC patients with LC was a possible medical vulnerable group. LC prevalence is underestimated among CRC patients with LM in Taiwan and vice versa, LM risk is severely underestimated and even higher in CRC patients with LC. CRC patients with LC should be paid more attention about LM in clinical practice.
Acknowledgements This study is based in part on data from the National Health Insurance Research Database provided by the Bureau of National Health Insurance, Department of Health, and managed by National Health Research Institutes (registry number 99029). The interpretation and conclusions contained herein are not those of the Bureau of National Health Insurance, Department of Health, or National Health Research Institutes.
Funding This study was supported by research grants of Buddhist Dalin Tzu Chi Hospital (DTCRD-103(2)-I-21). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Conflict of interest statement The authors have declared that no competing interests exist.
References 1. Schuppan D, Afdhal NH. Liver cirrhosis. Lancet 2008;371:838–51. 2. Cabibbo G, Palmeri L, Palmeri S, Craxi A. Should cirrhosis change our attitude towards treating non-hepatic cancer? Liver Int 2012;32:21–7. 3. Sorensen HT, Friis S, Olsen JH, et al. Risk of liver and other types of cancer in patients with cirrhosis: a nationwide cohort study in Denmark. Hepatology 1998;28:921–5. 4. Iascone C, Ruperto M, Barillari P. [Occurrence of synchronous colorectal cancer metastasis in the cirrhotic or fatty liver]. Minerva Chir 2005;60:185–90. 5. Uetsuji S, Yamamura M, Yamamichi K, Okuda Y, Takada H, Hioki K. Absence of colorectal cancer metastasis to the cirrhotic liver. Am J Surg 1992;164:176–7. 6. Qian HG, Zhang J, Leng JH, et al. [Association of hepatitis B virus infection and cirrhosis with liver metastasis in colorectal cancer]. Zhonghua Wei Chang Wai Ke Za Zhi 2010;13:202–4.
7. World Health Organisation. International Agency for Research on Cancer. GLOBOCAN. Cancer incidence and mortality worldwide in 2008. http:// globocan.iarc.fr/Pages/fact_sheets_cancer.aspx (27 October 2014, date last accessed). 8. Edwards BK, Ward E, Kohler BA, et al. Annual report to the nation on the status of cancer, 1975–2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates. Cancer 2010;116:544–73. 9. Report of Taiwan Cancer Registry. 2010. http://www.hpa.gov.tw/BHPNet/ Web/Stat/StatisticsShow.aspx?No=201305060001 (27 October 2014, date last accessed). 10. Fong Y, Cohen AM, Fortner JG, et al. Liver resection for colorectal metastases. J Clin Oncol 1997;15:938–46. 11. Choti MA, Sitzmann JV, Tiburi MF, et al. Trends in long-term survival following liver resection for hepatic colorectal metastases. Ann Surg 2002;235:759–66. 12. Gallagher DJ, Kemeny N. Metastatic colorectal cancer: from improved survival to potential cure. Oncology 2010;78:237–48. 13. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) Rectal Cancer Version 3. 2013. http:// www.nccn.org/professionals/physician_gls/f_guidelines.asp (27 October 2014, date last accessed). 14. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) Colon Cancer. Version 2. 2013. http:// www.nccn.org/professionals/physician_gls/f_guidelines.asp (27 October 2014, date last accessed). 15. Wanebo HJ, Semoglou C, Attiyeh F, Stearns MJ Jr. Surgical management of patients with primary operable colorectal cancer and synchronous liver metastases. Am J Surg 1978;135:81–5. 16. Bureau of National Health Insurance: National Health Insurance Annual Statistical Report. Taipei, Taiwan, 1995–2004. 17. NHI. NHI profile. http://www.nhi.gov.tw/english/webdata.asp?menu= 11&menu_id=290&webdata_id=1884. 2008 (7 July 2008, date last accessed). 18. Lu JF, Hsiao WC. Does universal health insurance make health care unaffordable? Lessons from Taiwan. Health Aff 2003;22:77–88. 19. Tseng CH. Mortality and causes of death in a national sample of diabetic patients in Taiwan. Diabetes Care 2004;27:1605–9. 20. Bureau of National Health Insurance Website (2000) Methods for estimating false claims. http://www.nhi.gov.tw/ information/bulletin_file/421– 0890036465-19.doc (1 October 2009, date last accessed). 21. Chen HF, Chen P, Li CY. Risk of malignant neoplasms of liver and biliary tract in diabetic patients with different age and sex stratifications. Hepatology 2010;52:155–63. 22. Wu PH, Lin YT, Kuo CN, Chang WC, Chang WP. No increased risk of herpes zoster found in cirrhotic patients: a nationwide population-based study in Taiwan. PLoS One 2014;9:e93443. 23. Center for Health Policy Research and Development, National Health Research Institutes: Evidence Based Practice Guidelines 2009: Treatment Guidelines for Liver Cirrhosis. http://ebpg.nhri.org.tw/check/9.pdf (27 October 2014, date last accessed). 24. Sharabi Y, Eldad A. Nonalcoholic fatty liver disease is associated with hyperlipidemia and obesity. Am J Med 2000;109:171. 25. Fabrizi F, Dixit V, Martin P, Jadoul M, Messa P. Meta-analysis: the impact of nutritional status on the immune response to hepatitis B virus vaccine in chronic kidney disease. Dig Dis Sci 2012;57:1366–72. 26. Iascone C, Ruperto M, Barillari P. [Colorectal carcinoma metastasis in livers infected with hepatitis B or C virus]. Minerva Chir 2005;60:77–81. 27. Liu CY, Hung YT, Chuang YL, et al. Incorporating development stratification of Taiwan townships into sampling design of large scale health interview survey. J Health Manag 2006;4:1–22. 28. Chen HF, Chen P, Li CY. Risk of malignant neoplasm of the pancreas in relation to diabetes: a population-based study in Taiwan. Diabetes Care 2011;34:1177–9. 29. Lee CC, Chien SH, Hung SK, Yang WZ, Su YC. Effect of individual and neighborhood socioeconomic status on oral cancer survival. Oral Oncol 2012;48:253–61.
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
with LC or cirrhotic symptoms still had higher LM risk. Second, in this study, only age and sex were matched without other characteristics such as comorbidities matched. Although this study design could allow these other characteristics to be analyzed, and discussed after regression adjustment to realize the difference of these two groups with and without LC, this was still a limitation for conclusion. Third, the diagnosis of LM was ICD-9-CM code 197.7, which was made by clinical physicians according to histologically defined or radiological criteria sometimes. However, LM misjudgment should have been minimized and could be ignored under the cancer treatment policy in Taiwan that cancer status dispute or treatment strategy changes would be confirmed later by multidisciplinary experts meeting to decide the subsequent treatment choices. Moreover, LM was confirmed by catastrophic illness certificates that further ensure the accuracy of cancer diagnosis. Other malignancy (ICD-9-CM codes 140–208, excluding CRC), such as hepatocellular carcinoma was also excluded to prevent the double malignancy condition.
Jpn J Clin Oncol 2014 30. Chen CY, Liu CY, Su WC, Huang SL, Lin KM. Factors associated with the diagnosis of neurodevelopmental disorders: a population-based longitudinal study. Pediatrics 2007;119:e435–43. 31. Chen CH, Yang PM, Huang GT, Lee HS, Sung JL, Sheu JC. Estimation of seroprevalence of hepatitis B virus and hepatitis C virus in Taiwan from a large-scale survey of free hepatitis screening participants. J Formos Med Assoc 2007;106:148–55. 32. Everhart J, editor. Digestive Diseases in the United States: Epidemiology and Impact. US Department of Health and Human Services, Public Health Service, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases. Washington, DC: US Government Printing Office, 1994; NIH publication no. 94–1447. 33. Manfredi S, Lepage C, Hatem C, Coatmeur O, Faivre J, Bouvier AM. Epidemiology and management of liver metastases from colorectal cancer. Ann Surg 2006;244:254–9.
9
34. Hanje AJ, Patel T. Preoperative evaluation of patients with liver disease. Nat Clin Pract Gastroenterol Hepatol 2007;4:266–76. 35. Lalor PF, Lai WK, Curbishley SM, Shetty S, Adams DH. Human hepatic sinusoidal endothelial cells can be distinguished by expression of phenotypic markers related to their specialised functions in vivo. World J Gastroenterol 2006;12:5429–39. 36. Benjamin J, Singla V, Arora I, Sood S, Joshi YK. Intestinal permeability and complications in liver cirrhosis: a prospective cohort study. Hepatol Res 2013;43:200–7. 37. Zuckerman MJ, Menzies IS, Ho H, et al. Assessment of intestinal permeability and absorption in cirrhotic patients with ascites using combined sugar probes. Dig Dis Sci 2004;49:621–6. 38. Pijls KE, Jonkers DM, Elamin EE, Masclee AA, Koek GH. Intestinal epithelial barrier function in liver cirrhosis: an extensive review of the literature. Liver Int 2013;33:1457–69.
Downloaded from http://jjco.oxfordjournals.org/ at Ondokuz Mayis University on November 7, 2014
