C L I N I C A L S T U D I E S
Cryptogenic pyogenic liver abscess as a sign of colorectal cancer:
a population-based 5-year follow-up study
Hsueh-Chou Lai1,2and Herng-Ching Lin2
1 Department of Internal Medicine, Division of Gastroenterology, China Medical University Hospital, Taichung, Taiwan 2 School of Health Care Administration, Taipei Medical University, Taipei, Taiwan
Keywords
colorectal cancer – cryptogenic pyogenic liver abscess – diabetes
Correspondence
Herng-Ching Lin, School of Health Care Administration, Taipei Medical University, 250 Wu-Hsing St., Taipei 110, Taiwan Tel: 1886 2 2736 1661, ext. 3613 Fax: 1886 2 2378 9788 e-mail: [email protected] Received 18 January 2010 Accepted 23 July 2010 DOI:10.1111/j.1478-3231.2010.02327.x Abstract
Background/aims: No large-scale population-based study has ever been con-ducted to examine the relationship between cryptogenic pyogenic liver abscesses (PLA) and the subsequent risk of colorectal cancer. This study aimed to estimate the risk for colorectal cancer following a diagnosis of cryptogenic PLA over a 5-year period. Methods: The study group comprised 274 patients who visited an outpatient care centre or were hospitalized with a diagnosis of cryptogenic PLA between 2001 and 2003. The comparison group included 1370 randomly selected subjects. Cox proportional hazard regressions were performed to compare the 5-year colorectal cancer-free survival rates for these two groups. Results: Of the total sample, 40 patients from the study group (2.43%) had colorectal cancer during the 5-year follow-up period: 15 (5.45% of those with cryptogenic PLA) and 25 from the comparison group (1.82% of the comparison group). After adjusting for patients’ age, sex, monthly income, level of urbanization and geographical location, the hazard of colorectal cancer during the 5-year period was 3.36 times greater for patients with cryptogenic PLA than for the comparison group [95% confidence interval (CI) = 1.72–6.56, Po 0.001]. The adjusted hazard of colorectal cancer during the 5-year follow-up period was 5.54 times higher for cryptogetic PLA patients with diabetes (95% CI = 2.11–14.56, Po 0.001) than the comparison group and 2.64 times higher among PLA patients without diabetes (95% CI = 1.19–5.85, Po 0.05). Conclusions: We conclude that cryptogenic PLA is an alarm that may signal colorectal cancer, especially among female patients with diabetes.
The aetiology and pathogenesis of liver abscesses have changed drastically in recent years (1, 2). In the begin-ning, pyogenic liver abscess (PLA) was most often associated with hepatobiliary tract diseases or intra-abdominal infections, including cholecystitis, suppura-tive cholangitis, suppurasuppura-tive pylephlebitis, appendicitis, diverticulitis and peritonitis (1–5). Some PLAs were idiopathic – namely, cryptogenic abscesses. In many reports, cryptogenic PLA has been regarded as the first manifestation of silent colorectal cancers and colonic tubulovillous adenomas (5–13). If a patient has a pre-vious colon lesion (e.g. advanced colon polyp or colon cancer), then bacteria may move from the mucosal defect into the portal system and result in cryptogenic PLA.
In addition, multiple liver abscesses caused by anaero-bic bacteria such as peptostreptococcus or streptococcus spp., clostridium septicum or bacteroides fragilis are the presenting forms of unsuspected colorectal cancer with-out liver metastases (13–19). Furthermore, there have been a growing number of reports about the relationship
between diabetes mellitus and liver abscesses (20, 21). Recent evidence indicates that metabolic abnormalities such as impaired glucose tolerance and diabetes mellitus are also associated with increased incidence of colonic adenoma and cancer (22, 23). More recently, occult colonic cancer and tubulovillous adenomas have been reported in K. pneumoniae liver abscesses (7, 12, 24, 25). However, to our knowledge, no large-scale population-based study has ever been conducted to examine the relationship between cryptogenic PLA and the subse-quent risk of colorectal cancer.
Using a nationwide population-based dataset in Taiwan, this study aims to estimate the risk of colorectal cancer following a diagnosis of cryptogenic PLA over a 5-year follow-up period. The annual incidence of PLA has increased steadily in Taiwan from 11.15/100 000 individuals in 1996 to 17.59/100 000 individuals in 2004 (24). According to the 2006 cancer report released by the Taiwan Department of Health, the incidence of colorectal cancer was 32.17/100 000 for females and 42.71/100 000
for males. Colorectal cancer has become one of the five most common cancers, and is also the third most common cause of cancer-related death in Taiwan (26). The high relative incidences of both PLA and colorectal cancer in Taiwan provide a unique opportunity to examine the association between these two conditions while adjusting for clinical and sociodemographical factors. We hypothesize that the risk of colorectal cancer is increased following a diagnosis of cryptogenic PLA.
Materials and methods Database
In this study, we used the ‘Longitudinal Health Insurance Database 2001’ (LHID2001), a nationwide population-based dataset released by the Taiwan National Health Research Institute (NHRI) in 2006. Taiwan launched a single-payer National Health Insurance Program in 1995 to provide affordable healthcare for all the island’s residents. As of 2007, 22.60 million of Taiwan’s 22.96 million people were enrolled in this programme. The LHID2001 includes all original medical claims data as well as a registry of 1 000 000 beneficiaries, randomly sampled from all enrollees included in the National Health Insurance (NHI) programme. The Taiwan NHRI reports that no statistically significant differences were found in the distributions of age, gender or healthcare costs between the sample group of the LHID and all enrollees. Therefore, the LHID2001 provides an excep-tional opportunity to examine the risk of colorectal cancer following PLA.
As the dataset used for this study includes de-identi-fied secondary data released to the public for research purposes, this study was exempt from a full review by the Institutional Review Board.
Study sample
For the study group, we selected patients who visited outpatient care centres or were hospitalized with a principal diagnosis of liver abscess (ICD-9-CM code 572.0) between 1 January 2001 and 31 December 2003 (n = 325). We excluded patients who were below 18 years of age in order to limit our sample to the adult popula-tion (n = 16). In addipopula-tion, we excluded patients who had any type of cancer (ICD-9-CM codes 140–239) (n = 19) or PLA diagnosed before 2001 (n = 11). We also excluded patients with a history of cholangitis, common bile duct stones, pancreatitis or inflammatory bowel disease before 2001 (n = 5). Finally, a total of 274 patients with crypto-genic PLA were included in the study group. We assigned the first ambulatory care visit or hospitalization in which they received a principal diagnosis of cryptogenic PLA as the index ambulatory care visit.
The comparison group for this study was extracted from the remaining beneficiaries of the LHID2001. We excluded patients who had PLA during the period from 2001 to 2008. Using the SAS program, we randomly
selected 1370 enrollees from the registry of beneficiaries (five for every patient with PLA) matched with the study group in terms of age (o 45, 45–54, 55–64, 65–74 and 474 years), sex and the year of index ambulatory care visit. We assigned their first ambulatory care visit in the index year as the index ambulatory care visit. We excluded patients who had any type of cancer before their index ambulatory care visit. Finally, 1644 patients were included in our study. Each patient was individually tracked for five years following their index ambulatory care visit to identify all those who developed colorectal cancer (ICD-9-CM codes 153 or 154) during the follow-up period.
Statistical analysis
We used the SAS statistical package (SAS System for
Windows, Version 8.2) to perform all analyses in this study. Pearson’s w2 tests were carried out to examine differences in the distributions of socio-demographic characteristics [age, sex, monthly income, level of urba-nization and the geographical location of the community in which the patient resided (Northern, Central, Eastern and Southern Taiwan)] between the study and the comparison groups. Monthly income was categorized into four groups: 0, NT$1-NT$15 840, NT$15 841-NT$25 000 and Z841-NT$25 001 (US$1.00 = NT$ 33.00 in 2003). For this study, urbanization levels in Taiwan were classified into five strata, with level 1 referring to the ‘most urbanized’ and level 5 referring to the ‘least urbanized’ communities, in accordance with prior stu-dies in Taiwan (27).
The 5-year colorectal cancer-free survival rate was computed using the Kaplan–Meier method, and the difference in the risk of colorectal cancer between the two groups was estimated by a log-rank test. In addition, stratified Cox proportional hazard regressions (stratified on age group, sex and year of index hospitalization) were carried out to calculate the hazard of colorectal cancer for these two groups, after adjusting for monthly income, level of urbanization and the geographical location of the sampled patients. The sampled patients were censored if individuals died from non-colorectal cancer causes dur-ing the 5-year follow-up period (246 patients had already died, comprising 71 from the study group and 175 from the comparison group). We have tested our data and found that the data meet the proportionality assumption (survival curves for two strata (patients with cryptogenic PLA and patients in the comparison group) have hazard functions that are proportional over time). The differ-ences were considered significant if a two-sided P value was less than or equal to 0.05.
Results
Of the total sample of 1644 patients, the mean age was 53.6 years, with a standard deviation of 17.1 years. Table 1 shows the distribution of the sociodemographic
characteristics and comorbidities between patients with and without cryptogenic PLA. The majority of the patients were male (55.5%), and almost a half (48.9%) were below 54 years of age. After matching patients by age and gender, a significant difference was observed between these two groups in the level of urbanization of the community where the patient resided (P = 0.028). In addition, patients with cryptogenic PLA were more likely to have diabetes (Po 0.001) and renal disease (P = 0.009) than patients without cryptogenic PLA.
Table 2 presents the distribution of colorectal cancer during the 5-year follow-up period for patients with and without cryptogenic PLA. Of the total sample, 40 patients from the study group (2.43%) developed color-ectal cancer during the 5-year period, 15 (5.45% of the patients with cryptogenic PLA), and 25 from the com-parison group (1.82% of patients without cryptogenic PLA). The log-rank test revealed a significant difference in the 5-year colorectal cancer-free survival rates between patients with and without cryptogenic PLA (Po 0.001). Figure 1 displays the results of the Kaplan–Meier survival analysis.
Table 2 also shows the crude and adjusted hazard ratios (HRs) of colorectal cancer within the 5-year period following the index ambulatory care visits. The crude HR for colorectal cancer for patients with cryptogenic PLA was 3.12 [95% confidence interval (CI) = 1.62–5.99, Po 0.001] as compared with patients without crypto-genic PLA. Similarly, stratified Cox proportional hazard regressions (stratified on age group, sex and year of index hospitalization) showed that after adjusting for patients’ monthly income, the level of urbanization and the geographical location of the patient’s community, the HR for colorectal cancer for patients with cryptogenic PLA was 3.36 times as high as the comparison group in the 5-year follow-up period (95% CI = 1.72–6.56, Po 0.001). In addition, among cryptogenic PLA pa-tients with and without diabetes, the adjusted hazard of colorectal cancer during the 5-year follow-up period was 5.54 (95% CI = 2.11–14.56, Po 0.001) and 2.64 (95% CI = 1.19–5.85, Po 0.05) times higher, respectively, than that in the comparison group.
The crude and adjusted HRs for colorectal cancer stratified by patient sex and age are presented in Table 3. Among male and female patients, the adjusted hazard of colorectal cancer during the 5-year follow-up period was 2.82 (95% CI = 1.14–6.99, P = 0.024) and 4.08 times greater (95% CI = 1.42–11.73, P = 0.009), respectively, for those with cryptogenic PLA than for those in the comparison group. Furthermore, we analysed the risk of colorectal cancer for patients with cryptogenic PLA by age group (o 65 vs. Z65). We found that the adjusted HR for colorectal cancer during the 5-year period was 3.09 (95% CI = 1.31–7.33, P = 0.019) and 3.70 times higher (95% CI = 1.21–11.27, P = 0.001) for cryptogenic PLA patients Z65 years old ando 65 years old, respec-tively, than that for the same age categories in the comparison group.
Table 1. Patients with cryptogenic pyogenic liver abscess and com-parison group in relation to sociodemographical characteristics and comorbid conditions in Taiwan, 2001–2003 (n = 1644)
Variables Patients with cryptogenic pyogenic liver abscess (n = 274) Comparison group (n = 1370) P value Number % Number % Gender 1.000 Male 152 55.5 760 55.5 Female 122 44.5 610 44.5 Age 1.000 o 45 74 27.0 370 27.0 45–54 60 21.9 300 21.9 55–64 54 19.7 270 19.7 65–74 54 19.7 270 19.7 474 32 11.7 160 11.7 Monthly income 0.652 0 147 53.7 743 54.2 NT$1–15 840 23 8.4 132 9.6 NT$15 841–25 000 79 28.8 352 25.7 ZNT$25 001 25 9.1 143 10.4 Urbanization level 0.028 1 45 16.4 245 17.9 2 40 14.6 251 18.3 3 21 7.7 137 10.0 4 32 11.7 93 6.8 5 136 49.6 644 47.0 Geographical region 0.239 Northern 178 65.0 929 67.8 Central 49 17.9 186 13.6 Southern 45 16.4 235 17.2 Eastern 2 0.7 20 1.5 Diabetes o 0.001 Yes 57 20.8 117 8.5 No 217 79.2 1253 91.5 Hyperlipidaemia 0.948 Yes 8 2.9 41 2.9 No 266 97.1 1329 97.1 Hypertension 0.464 Yes 51 18.6 230 16.8 No 223 81.4 1140 83.2
Coronary heart disease 0.162
Yes 20 7.3 71 5.2 No 254 92.7 1299 94.8 Heart failure 0.203 Yes 5 1.8 13 1.0 No 269 98.2 1357 99.0 Cerebrovascular disease 1.000 Yes 9 3.3 45 3.3 No 265 96.7 1325 96.7 Renal disease 0.009 Yes 10 3.7 19 1.4 No 96.4 264 1351 98.6 Diabetes with ophthalmic manifestations 0.654 Yes 1 0.4 3 0.2 No 273 99.6 1367 99.8
Table 4 shows the distribution of concurrent micro-organism diagnoses among patients with cryptogenic PLA according to colorectal cancer. We also found that there was no significant relationship between whether patients had bacteraemia or septicaemia and the risk of colorectal cancer during the 5-year follow-up period among patients with cryptogenic PLA (3.3 vs. 6.0%; P = 0.393 respectively).
Discussion
To our knowledge, this is the first attempt to investigate the risk of colorectal cancer following cryptogenic PLA using a nationwide population-based dataset. We found that 5.45% of patients with cryptogetnic PLA, and 1.82%
of patients without, developed colorectal cancer during the 5-year follow-up period. The likelihood of develop-ing colorectal cancer was 3.36 times greater among cryptogenic PLA patients than for patients without the condition, during the 5-year follow-up period. Further-more, cryptogenic PLA patients with diabetes (adjusted HR, 5.54; 95% CI = 2.11–14.56, Po 0.001) were also at a higher risk of developing colorectal cancer than such patients without diabetes, relative to the comparison group (adjusted HR, 2.64:95% CI = 1.19–5.85, Po 0.001). These findings support our hypothesis that the risk of colorectal cancer is increased following a diagnosis of cryptogenic PLA.
Recent reports assert that cryptogenic PLA might be a presentation of occult or silent colorectal cancer or adenoma without any gastrointestinal symptoms on admission (5, 7, 8, 14). Studies by Cohen et al. (6), Nosher et al. (28) and Yeh et al. (29) revealed that 5% (1/20 patients), 6.7% (1/15 patients) and 11.5% (6/52 patients) of patients with PLA had colorectal cancer respectively. Recently, Hiraoka et al. reported that three of 41 patients with PLA had the condition because of colon cancer (12). Colorectal cancer or large adenomas can induce loss of integrity of the normal mucosal barrier (11, 18). The mechanism that explains how colorectal cancer or adenoma can be involved in the aetiology of cryptogenic PLA is probably that mucosal defects offer a route for bacterial invasion into the portal system, which consequently arrives at the liver (11, 18, 30). In addition, in a report from the United Kingdom, one out of four patients with a definitive diagnosis of cryptogenic PLA developed colorectal cancer during a 2-year follow-up period (10). This suggests that colorectal cancer could coexist with or pre-exist cryptogenic PLA. Our study thus confirms prior findings that cryptogenic PLA is a warning sign of colorectal cancer.
In previous reports, multiple hepatic abscesses caused by anaerobic bacteria are the presenting signs of unsus-pected colon cancer without liver metastases (13, 14). Several bacteria, such as Bacteroides, anaerobic Strepto-coccus and K. pneumoniae cause PLA with an underlying
Table 2. Crude and adjusted hazard ratio for colorectal cancer during the 5-year follow-up period for patients in Taiwan with cryptogenic pyogenic liver abscess and patients in the comparison group (n = 1644)
Presence of Colorectal Cancer Total sample Comparison Cryptogenic pyogenic liver abscess (all) Cryptogenic pyogenic liver abscess, (with comorbid diabetes)
Cryptogenic pyogenic liver abscess, (without comorbid diabetes)
No. % No. % No. % No. % No. %
5-year follow-up period
Yes 40 2.43 25 1.82 15 5.45 6 10.53 9 4.15
No 1604 97.57 1345 98.18 259 94.53 51 89.47 208 95.85
Crude HR (95% CI) – 1.00 3.12 (1.62–5.99) 6.33 (2.49–16.11) 2.33 (1.07–5.06)
Adjustedw HR (95% CI) – 1.00 3.36 (1.72–6.56) 5.54 (2.11–14.56) 2.64 (1.19–5.85)
wStratified Cox proportional hazard regression was used to stratify by age group and year of index hospitalization and to adjust for monthly income and urbanization level. Po 0.05. Po 0.001. 1.00 Comparison P < 0.001 Cryptogenic pyogenic liver abscess 0.99 0.98 0.97 Sur viv al rate 0.96 0.95 0.00 0 500 1000
Days after index ambulatory care visit
1500 2000
Fig. 1. Colorectal cancer-free survival rates for patients with cryptogenic pyogenic liver abscess and the comparison group in Taiwan, 2001–2003.
diagnosis of colon cancer (9, 14, 15, 31, 32). In our earlier study, we described two cases, one with cryptogenic PLA and K. pneumoniae infection and another with diabetes and K. pneumoniae infection, which were found to be the presenting manifestation of colonic tubulovillous adeno-ma (7). These bacteria are colonizers in the huadeno-man gastrointestinal tract. The defect in the mucosal barrier caused by colorectal cancer or large adenomas predispose to microbial seeding to the portal venous system and into the liver, resulting in abscess formation (11, 18).
Patients with diabetes are known to show defects in neutrophil chemotaxis and phagocytosis, which are thought to be the most important predisposing factors for K. pneumoniae-pathogenic liver abscess (33). Our previous survey also showed that diabetes mellitus was a major risk factor for PLA in Taiwanese (21). Similarly, Thomsen et al. reported that 11.2% of PLA cases were accompanied by diabetes in the United States (20). In the present study, we found that about 22% of patients with PLA had diabetes. It is proven that diabetes is a risk factor for Gram-negative bacteraemia, including episodes caused by abdominal foci of infection (34). Potential biological mechanisms may include tissue hyperglycae-mia and preference for some microorganisms, such as Escherichia coli and K. pneumoniae (24, 35, 36).
Many risk factors have been identified for colorectal cancer including ageing, a high-fat diet, obesity, familial polyposis syndrome and inflammatory bowel disease. Recent evidence indicates that metabolic abnormalities such as impaired glucose tolerance and diabetes mellitus are also associated with increased incidence of colonic adenoma and cancer (22, 23). One study by Seow et al. (37) reported that diabetes was significantly associated with colorectal cancer risk in both men [relative risk (RR) = 1.5, 95% CI = 1.2–2.1] and women (RR = 1.4, 95% CI = 1.0–1.9) in Singapore Chinese. Similarly, a study by Larsson et al also showed a significant risk of colorectal cancer (RR 1.49, 95% CI = 1.14–1.96) in Swedish men (38). Metabolic syndrome was found to be
Table 4. The distribution of concurrent microorganism diagnoses according to colorectal cancer among patients with cryptogenic pyogenic liver abscess (n = 274)
Concurrent microorganism diagnosis
(ICD-9-CM code) n Colorectal cancer Yes, n (%) Streptococcus (038.0, 041.0X) 6 0 Staphylococcus (038.1X, 041.1X) 1 0 Pneumococcus (038.2, 041.2) 1 0 Escherichia (038.42, 041.4) 3 0 Klebsiella (041.3) 17 0 Proteus (041.6) 1 0 Gram-negative bacteria (038.40, 038.49, 041.85) 5 1 (20.0) Other/unspecified bacteria (038.8, 038.9, 041.81) 25 1 (4.0) None 215 13 (6.0) Table 3. Crude and Adjuste d haza rd ra tios for color ectal cance r during the five-year follow-up perio d for patien ts with cryptog enic pyogenic liver abscess an d patien ts in the compari son gr oup, strat ified by patient age and gender Pr esenc e o f col or ectal ca ncer Pati ent sex w Pati ent age z (yea rs) Male Female o 65 Z 65 Com parison num ber (% ) Cryp togenic pyoge nic liver abs cess nu mber (% ) Comp arison numbe r (%) Cryp tog enic pyo genic liver abs cess numbe r (%) Com parison num ber (% ) Cryp togenic pyoge nic liver abs cess nu mber (%) Comp arison numbe r (%) Cryp tog enic pyo genic liver abs cess numbe r (%) 5-yea r fol low-up peri od Y e s 1 6 (2.14) 8 (5.26) 9 (1.4 5) 7 (5.74) 8 (0.85) 6 (3.19) 17 (3 .93) 9 (10.47 ) No 732 (97.86) 149 (94.74) 613 (98. 55) 115 (94.26) 929 (99.15 ) 182 (96.81) 416 (9 6.07) 77 (89.53 ) Cru de HR (95% CI) 1.00 2.54 (1.07–6. 05) 1.00 4. 15 (1.51– 11.36) 1.00 3.83 (1.31– 11.17) 1.00 2. 86 (1.23– 6.65) Ad justed HR (95% CI) 1.00 2.82 (1.14–6. 99) 1.00 4. 08 (1.42– 11.73) 1.00 3.70 (1.21– 11.27) 1.00 3. 09 (1.31– 7.33) wHazar d ratio was calculated by Stratified Cox pr oportional hazar d regression which was stratified on age group and year of index hospitalization and a d justed for monthly income, urbanization level and geographical regio n. zHazard ratio was calculated by Stratified Cox pr oportional hazard regr ession which was stratified on sex and year of index hospitalization a nd adjuste d for monthly income, urbanization level and geographical regio n. P o 0.05. P o 0.001.
associated with an increased risk of colorectal adenoma (odds ratio, 1.51, 95% CI = 1.18–1.93) in a study by Kim (39). Those studies support our finding that PLA patients with diabetes had a higher risk of developing colorectal cancer during a 5-year follow-up than patients without diabetes (adjusted HR. 5.54 vs. 2.64).
The strengths of this study include its large nationwide population-based sample, the availability of data on patients with a wide range of demographic characteristics and complete histories of medical service utilization for the sampled patients. Although we tried to overcome the inherent limitations of using administrative data, this study suffers from two limitations that should be ad-dressed. The first is the lack of detailed information on body mass index data, smoking, alcohol use, a physically active or a sedentary lifestyle, high-fat diet and a family history of colorectal cancer, all of which have been shown to be associated with the risk of colorectal cancer. In addition, microbiological data that might be related to the risk for colorectal cancer were also unavailable in our dataset.
Secondly, there may be a surveillance bias in that patients with a cryptogenic PLA diagnosis are likely to have more frequent check-ups, and thus to have their colorectal cancer detected by a physician. Indeed, we found that 19.3% of the patients with cryptogenic PLA and 8.2% of the patients without had received endo-scopic procedures or faecal testing analysis during the 5-year follow-up period respectively. Therefore, it is possible that receiving medical attention for cryptogenic PLA may contribute to the elevated odds of being diagnosed with colorectal cancer.
Conclusions
Despite the above limitations, we have found that during a 5-year follow-up period, the risk for colorectal cancer was 5.54 times higher than the comparison group among cryptogenic PLA patients and 2.64 times higher for PLA patients without diabetes. Furthermore, the odds of colorectal cancer were 4.08 times higher for females with cryptogenic PLA than the comparison group and 2.82 times higher among males with cryptogenic PLA. Our findings revealed that cryptogenic PLA is an alarm suggesting colorectal cancer, especially for those with diabetes and females. We suggest that the patients with cryptogenic PLA, an underlying cause of gastrointestinal tract, should be surveilled aggressively; we therefore recommend that diabetes patients with cryptogenic PLA undergo a survey of the colon to detect the presence of neoplasms.
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, Taiwan and managed by the National Health
Research Institutes. The interpretations and conclusions contained herein do not represent those of the Bureau of National Health Insurance, Department of Health, or the National Health Research Institutes.
None of the authors have any financial conflicts to disclose.
No funding source.
References
1. Sherman JD, Robbin SL. Changing trends in casuistics of hepatic abscess. Am J Med 1960; 28: 943–50.
2. Ledeman ER, Crum NF. Pyogenic liver abscess with a focus on Klebsiella pneumoniae as a primary pathogen: an emer-ging disease with unique clinical characteristics. Am J Gatroenterol 2005; 100: 322–31.
3. Miedema BW, Dineen P. The diagnosis and treatment of pyogenic liver abscess. Ann Surg 1984; 200: 328–35. 4. Stain SC, Yellin AE, Donovan AJ, et al. Pyogenic liver
abscess: modern treatment. Arch Surg 1991; 126: 991–6. 5. Leiba A, Apter S, Avni I, et al. Pyogenic liver abscess-an
unusual presentation of colonic villous adenoma. Harfuah (Israel) 2003; 142: 336–7.
6. Cohen JL, Martin FM, Rossi RL, et al. Liver abscess: the need for complete gastrointestinal evaluation. Arch Surg 1989; 124: 561–4.
7. Lai HC, Chan CY, Peng CY, et al. Pyogenic liver abscess associated with large colonic tubulovillous adenoma. World J Gastroenterol 2006; 72: 880–4.
8. Lee KJ, Kum J, Ghosh P. Nonmetastatic Cancer of the colon associated with pyogenic liver abscess. Am J Gastroenterol 2008; 103: 798–9.
9. Lim WC, Lim CC. Silent colorectal carcinoma and pyo-genic liver abscess. J Gastroenterol Hepatol 2004; 19: 945–6.
10. Mohsen AH, Green ST, Read RC, et al. Liver abscess in adults: ten years experience in a UK centre. QJM 2002; 95: 797–802.
11. Zakout R, Santos JM, Ferreira C, et al. Colonoscopy for ‘cryptogenic’ pyogenic liver abscess? Colorectal Disease 2010; 12: 71–3.
12. Hiraoka A, Yamashita Y, Uesugi K, et al. Three cases of liver abscesses complicated with colon cancer without liver metastasis: importance of screening for digestive disease. Intern Med 2007; 46: 2013–7.
13. Alvarez J, Baldonedo R, Bear I, et al. Anaerobic liver abscesses as initial presentation of silent colonic cancer. HPB (Oxgord) 2004; 6: 41–2.
14. Lonardo A, Grisendi A, Pulvirenti M, et al. Right colon adenocarcinoma presenting as Bacteroides fragilis liver abscess. J Clin Gastroenterol 1992; 14: 335–8.
15. Teitz S, Guidetti-Sharon A, Manor H, et al. Pyogenic liver abscess: warning indicator of silent colonic cancer. Dis Colon Rectum 1995; 38: 1220–3.
16. Klein RS, Recco RA, Catalano MT, et al. Association of Streptococcus bovis with carcinoma of the colon. N Engl J Med 1977; 297: 800–2.
17. Trump DL, Fahnestock R, Cloutier CT, et al. Anaerobic liver abscess and intrahepatic metastases: a case report and review of the literature. Cancer 1978; 41: 682–6.
18. Simmang CL, Senatore P, Lowry A, et al. Practice para-meters for detection of colorectal neoplasms. The Stan-dards Committee, The American Society of Colon and Rectal Surgeons. Dis Colon Rectum 1999; 42: 1123–9. 19. Harnik IG. Pyogenic liver abscess presenting after
malig-nant polypectomy. Dig Dis Sci 2007; 52: 3524–5.
20. Thomsen RW, Jepsen P, Sorensen HT. Diabetes mellitus and pyogenic liver abscess: risk and prognosis. Clin infect Dis 2007; 44: 1194–210.
21. Chen W, Chen CH, Liu KL, et al. Clinical outcome and prognostic factors or patients with pyogenic liver abscess requiring intensive care. Crit Care Med 2008; 36: 1184–8. 22. Nilsen TI, Vatten LJ. Prospective study of colorectal cancer
risk and physical activity, diabetes, blood glucose and BMI: exploring the hyperinsulinaemia hypothesis. Br J Cancer 2001; 84: 417–22.
23. Giovannucci E, Michaud D. The role of obesity and related metabolic disturbances in cancers of the colon, prostate, and pancreas. Gastroenterology 2007; 132: 2208–25. 24. Tsai FC, Huang YT, Chang LY, et al. Pyogenic liver abscess
as endemic disease, Taiwan. Emerging Infect Dis 2008; 14: 1592–600.
25. Hsu WH, Yu FJ, Chuang CH, Chen CF. Occult colon cancer in a patient with diabetes and recurrent Klebsiella pneumo-niae liver abscess. Kaohsiung J Med Sci 2009; 25: 98–103. 26. Taiwan Cancer Registry. Available at http://crs.cph.ntu.
edu.tw/main.php?Page=A5 (accesssed March 2010). 27. Lin HC, Chao PZ, Lee HC. Sudden sensorineural hearing
loss increases the risk of stroke: a 5-year follow-up study. Stroke 2008; 39: 2744–8.
28. Nosher JL, Giudici M, Needell GS, et al. Elective one stage abdominal operations after percutaneous catheter drainage of pyogenic liver abscess. Am Surg 1993; 59: 658–63.
29. Yeh TS, Jan YY, Jeng LB, et al. Pyogenic liver abscesses in patients with malignant disease. Arch Surg 1998; 242–45. 30. The liver in infections. In: Sherlock S, Doley J, eds. Diseases
of the Liver and Biliary System. Oxford: Blackwell Science, 1997; 497–500.
31. Tzur T, Liberman S, Felzenstein I, et al. Liver abscesses caused by Streptococcus milleri: an uncommon present-ing sign of silent colonic cancer. Irs Med Assoc J 2003; 5: 206–7.
32. Matsushita M, Hajiro K, Okazaki K, et al. Endophthalmitis with brain, lung, and liver abscesses associated with an occult colon cancer. Am J Gastroenterol 2000; 95: 3664–5. 33. Delamaire M, Maugendre D, Moreno M, et al. Impaired
leucocyte functions in diabetic patients. Diabet Med 1997; 14: 29–34.
34. Thomsen RW, Hungborg HH, Lervang HH, et al. Diabetes mellitus as a risk and prognostic factor for community-acquired bacteremia due to enterobacteria: a 10-year, population-based study among adults. Clin Infect Dis 2005; 40: 628–31.
35. Rayfield EJ, Ault MJ, Keusch GT, et al. Infection and diabetes: the case for glucose control. Am J Med 1982; 72: 439–50.
36. Hansen DS, Gottschau A, Kolmos HJ. Epidemiology of Klebsiella bacteremia: a case control study using Escher-icha coli bacteremia as control. J Hosp Infect 1998; 38: 119–32.
37. Seow A, Yuan JM, Koh WP, et al. Diabetes mellitus and risk of colorectal cancer in the Singapore Chinese Health study. J Natl Cancer Inst 2006; 98: 135–8.
38. Larsson SC, Giovannuci E, Work A. Diabetes and colorectal cancer incidence in the cohort of swedish men. Diabetes Care 2005; 28: 1805–7.
39. Kim JH, Lim JY, Kim YH, et al. Is metabolic syndrome a risk factor for colorectal adenoma. Cancer Epidemiol Bio-markers Prev 2007; 16: 1543–6.
