| |
Can Dietary Fish Intake Prevent Liver Cancer? Editorial
|
| |
| |
Download the PDF here
Gastroenterology June 2012
Hepatocellular carcinoma (HCC) is a common cancer worldwide, with poor 5-year survival. An estimated 748,300 new cases and 695,900 cancer deaths occur per year, ranking it fifth among cancers for incidence and third among cancers for mortality.1 There is considerable geographic variation in the incidence of HCC. The largest concentration of HCC cases in the world is in Asia, followed by Africa, Europe, and North and South America.2 Chronic hepatitis B (HBV) infection is the most important risk factor for HCC worldwide, especially in Asia. In Asian and African countries, >80% of patients with HCC have underlying chronic HBV infection.3 The 1 exception in Asia is Japan, where the prevalence of HCC has been related to chronic hepatitis C (HCV) infection.4 In Western countries, however, chronic HCV infection has been determined to be present in about 60% of patients with HCC, and is the main etiologic agent leading to HCC,5, 6 with obesity and insulin resistance also thought to be important risk factors. The incidence of HCC is expected to continue to rise in Western countries as the cohort of patients infected with HCV ages, concurrent with the increasing prevalence of obesity and diabetes.
Within the context of chronic HCV and HBV infection, the presence of cirrhosis is the most important risk factor in the development of HCC.7 Nonmodifiable risk factors include older age, male gender, and family history of HCC (Figure 1). There are several modifiable risk factors in HCC, of which the most important are alcohol and tobacco. There is evidence of a dose-dependent effect of alcohol and tobacco and HCC,8, 9 including a synergistic effect with viral hepatitis.10
However, identifying additional modifiable risk factors, including diet, is important. A number of hypotheses link different aspects of diet with HCC, including coffee and tea, fructose, iron, red and white meats, types of fat, selenium, vitamin D, and vitamin E. Of these exposures, the most data is available for coffee, where associations have been observed for HCC,11 progression from fibrosis to cirrhosis and clinical outcomes,12 and, in the context of HCV, response to peginterferon and ribavirin therapy.13 However, associations with other dietary components remain unclear.
Studies of diet are challenging for several methodologic reasons. First, diet is complex and interrelated with other aspects of lifestyle. Individuals who eat a lot of fish, for example, likely also have many other behaviors that are associated with cancer. Also, assessing diet is difficult. Most studies use food frequency questionnaires, requiring study participants to answer questions about their typical diet over the past year. As might be expected, such instruments assess diet with error.14 Cross-sectional, case-control studies suffer from an additional potential methodologic limitation, because cases may overestimate behaviors they consider harmful and underestimate those they consider helpful. A stronger methodologic design is the prospective cohort, in which participants' diet and other behaviors are assessed before disease diagnosis. Although diet and other behaviors are still measured with error, such assessments are thought to be less biased, because participants complete their questionnaires without knowledge of their future disease diagnosis. However, even in prospective cohorts, challenges remain. Most such studies have been conducted in populations with low rates of liver cancer, limiting statistical power. Studies also tend to lack information on important HCC risk factors, including HBV and HCV infection status, and underlying liver disease.
Within this context, the manuscript by Sawada et al15 detailing inverse associations between fish and n-3 fatty acid consumption with HCC risk in 90,296 participants of the Japan Public Health Center-based Prospective Study, published in this issue of Gastroenterology, is of particular interest. Fish is a rich source of n-3 fatty acids and micronutrients including selenium and vitamin D. A large literature supports a protective effect for fish and n-3 fatty acid intake on cardiovascular disease,16, 17 and suggested mechanisms, including lower plasma triglycerides and reduced inflammation, could also apply to the liver. Relative to previous studies, the current analysis has a number of important advantages, including that it was conducted in a population that eats a lot of fish. The authors assessed important HCC risk factors, including alcohol, body mass index, diabetes, tobacco smoking, and, in a subset of the cohort, HBV, HCV, and the liver enzyme alanine aminotransferase.
In this cohort, the authors observe an inverse association between fish and incident HCC, with participants in the highest quintile of fish intake having 0.64 times the risk (95% confidence interval, 0.41-1.02) of participants in the lowest quintile; similar results were observed in those in the highest quintile of eating n-3 fatty acid rich fish (0.64; 95% confidence interval, 0.42-0.96), with evidence of a dose-response (P for trend across categories = .04). Associations were also observed for specific fatty acids, including eicosapentaenoic acid, docosapentaenoic acid, and docosahexaenoic acid. Although previous studies have not specifically investigated n-3 fatty acids and HCC, the current results are consistent with 2 previous cohort studies.18, 19, 20 One recent US cohort observed an inverse association with incident HCC, but was conducted in a population with limited fish consumption and lacked information on HBV and HCV status.18, 19 A second cohort from Japan also observed an inverse association, but presented only univariate, unadjusted, analyses.20 A number of laboratory studies have also linked n-3 fatty acids with liver health21; for example, a 2007 study observed a prevention effect of n-3 fatty acid intake on acute hepatitis in a transgenic mouse model,22 whereas a 2004 study observed decreased hepatic triglycerides in rats fed fish oil relative to control animals.23
Although the current results are tantalizing, care must be taken before fish and n-3 fatty acids should be recommended as HCC preventive agents. Existing results from 3 prospective cohorts are consistent and each suggests an inverse association. However, the current studies have limitations and further replication is needed. As observational studies, despite careful adjustment for liver cancer risk factors, associations with fish and n-3 fatty acid intake may simply reflect other lifestyle or environmental exposures. Also, in each study, diet (and other important risk factors such as alcohol, tobacco, and coffee) was assessed by questionnaire at a single time point and thus may not reflect use over the entire lifetime. Preexisting liver disease is a concern, because it typically occurs well before liver cancer diagnosis and may affect dietary intake. The authors adjust for self-reported liver disease and in a subset of their cohort, ALT levels, which correlated with chronic liver disease, are not a sensitive marker. These analyses are important given that patients with chronic liver disease, especially those with cirrhosis, are at high risk for developing HCC. Because ALT levels and self-report are imperfect proxies for underlying liver disease, the observed attenuation of the authors' results after excluding those with self-reported liver disease is of concern. In addition, the measurement of chronic HCV infection in only 19% of the entire cohort is problematic, given the high prevalence of this infection among patients with HCC in Japan, although among this subset of participants, results were similar to those overall after adjustment for HCV infection.
Diet was also assessed with error; for example, correlations for fish and n-3 fatty acid intake within a subset of participants completing a food frequency questionnaire and recording their diet over a 14- or 28-day period only ranged from 0.21 to 0.45. However, because error on the questionnaire is unlikely to be correlated with future disease risk, such errors would likely attenuate any observed risk estimates. The current study also examined associations for fish and n-3 fatty acids in Japan, where the major risk factor is chronic HCV infection.4 Yet, most mechanistic and laboratory studies of n-3 fatty acids in liver disease have occurred within the context of nonalcoholic fatty liver disease.24 Future studies are needed, therefore, to evaluate and confirm inverse associations between fish and n-3 fatty acids and liver cancer in populations with other spectrums of liver cancer risk factors. Studies with detailed information on preexisting, underlying liver disease are particularly needed. In addition, even if fish were truly associated with protection from HCC, it is not clear from the current analysis whether the protection stems from n-3 fatty acids or from another component, such as vitamin D or selenium. Additional studies, therefore, are needed to define which aspects of dietary fish intake may be inversely associated with liver cancer. Further mechanistic and laboratory studies are also needed to support these epidemiologic findings.
In conclusion, this study shows an inverse relationship between fish intake and incident HCC. Given the multiple risk factors identified in HCC, it highlights the difficulty in performing prospective studies properly adjusting for these risk factors. Although fish and n-3 fatty acid intake cannot be recommended for HCC prevention currently, these provocative results merit future study.
-------------------------------
Consumption of n-3 Fatty Acids and Fish Reduces Risk of Hepatocellular Carcinoma
"In conclusion, our large prospective study indicated that high consumption of n-3 PUFA-rich fish and n-3 PUFAs was associated with a reduced risk of HCC, even among a high-risk population. Given that the prognosis for HCC is extremely poor, our results would, if confirmed, have important implications for public health. Greater consumption of n-3 PUFA-rich fish and n-3 PUFAs may modify the development of HCC among HBV- and/or HCV-infected subjects.....Fish is a rich source of n-3 polyunsaturated fatty acids (PUFAs), such as eicosapentaenoic acid (EPA), docosapentaenoic acid (DPA), and docosahexaenoic acid (DHA), and several studies have documented a protective effect of dietary n-3 PUFA on the development of several cancers....Fish consumption might reflect other lifestyle factors. In particular, subjects with higher fish consumption tended to drink less alcohol and coffee, and tended to have a past history of diabetes. Although we also assessed the effect of fish consumption according to alcohol, coffee drinking status, or history of diabetes, an inverse association between fish and n-3 PUFA-rich fish and HCC risk was shown in both regular (1-2 times/wk) and nonregular (<1 time/wk) alcohol drinkers, in both daily and nondaily coffee drinkers, and in both those with and without a history of diabetes (data not shown). Interaction between n-3 PUFA-rich fish and alcohol, coffee drinking status, or history of diabetes was not detected (Pinteraction =.25, .57, and .58, respectively)."
Background & Aims
Fish is a rich source of n-3 polyunsaturated fatty acids (PUFAs), such as eicosapentaenoic acid (EPA), docosapentaenoic acid (DPA), and docosahexaenoic acid (DHA). Although consumption of fish and n-3 PUFA has been reported to protect against the development of some types of cancer, little is known about its association with hepatocellular carcinoma (HCC).
Methods
We investigated the association between fish and n-3 PUFA consumption and HCC incidence (n = 398) in a population-based prospective cohort study of 90,296 Japanese subjects (aged, 45-74 y). Hazard ratios and 95% confidence intervals (CIs) for the highest vs the lowest quintile were estimated from multivariable adjusted Cox proportional hazards regression models. We also conducted subanalyses of subjects with known hepatitis B virus (HBV) or hepatitis C virus (HCV) status, and of subjects who were anti-HCV and/or hepatitis B surface antigen positive. All tests of statistical significance were 2-sided.
Results
Among all subjects, consumption of n-3 PUFA-rich fish and individual n-3 PUFAs was associated inversely with HCC, in a dose-dependent manner. Hazard ratios for the highest vs lowest quintiles were 0.64 (95% CI, 0.42-0.96) for n-3 PUFA-rich fish, 0.56 (95% CI, 0.36-0.85) for EPA, 0.64 (95% CI, 0.41-0.98) for DPA, and 0.56 (95% CI, 0.35-0.87) for DHA. These inverse associations were similar irrespective of HCV or HBV status.
Conclusions
Consumption of n-3 PUFA-rich fish or n-3 PUFAs, particularly EPA, DPA, and DHA, appears to protect against the development of HCC, even among subjects with HBV and/or HCV infection.
The most important risk factor in the development of hepatocellular carcinoma (HCC) in human beings is chronic infection with hepatitis B virus (HBV) or hepatitis C virus (HCV).1 The markedly poor prognosis of HCC, with a 5-year survival rate in Japan of less than 20%,2 emphasizes the need for effective preventive measures, particularly in hepatitis virus carriers. Although dietary factors also might be risk factors, the role of diet in the etiology of HCC remains unclear, except with regard to alcohol consumption and aflatoxin contamination.3
A recent prospective study showed an inverse association between white meat, including fish, and liver cancer.4 Inverse associations with the consumption of white meat or fish were observed in some studies,5, 6, 7, 8 but were not confirmed in others.9, 10, 11 Moreover, except for 2 case-control studies,5, 7 most previous epidemiologic studies of white meat or fish and HCC did not consider HCV or HBV infection status.
Fish is a rich source of n-3 polyunsaturated fatty acids (PUFAs), such as eicosapentaenoic acid (EPA), docosapentaenoic acid (DPA), and docosahexaenoic acid (DHA), and several studies have documented a protective effect of dietary n-3 PUFA on the development of several cancers.12, 13 However, less is known about the influence of n-3 PUFA on HCC.
Here, we investigated the presence of an association between fish and n-3 PUFA consumption and HCC in a large-scale, population-based, cohort study in Japan, with consideration for HCV and HBV infection status.
Discussion
Here, we investigated the relationship between fish and n-3 PUFA consumption and the risk of HCC in a population-based prospective study in Japan. Results showed a decreased risk in those with a higher consumption of n-3 PUFA-rich fish and n-3 PUFAs, particularly EPA, DPA, and DHA. Of particular note was the inverse association even when analysis was confined to subjects who were also either or both HCV and HBV positive.
A recent prospective study in the United States also reported that the consumption of white meat, including fish, was significantly inversely associated with the risk of HCC (HR for the highest vs lowest quintile of 0.52, Ptrend < .001), but this study lacked information about HBV and HCV.4 In a previous study of the association between fish intake and HCC, results from a prospective study in Japan showed a significantly decreased risk of HCC mortality in the second category (3-4 times/wk), albeit in univariate analysis.6 In a case-control study in China, liver cancer mortality was associated with a curvilinear reduction of fish intake.8 Another case-control study in China also showed that the frequent intake of fresh fish (³3 times/week) decreased risk of HCC, with an odds ratio after adjustment for confounding factors, including HBV, of 0.32.5 In contrast, several case-control studies showed no association between HCC and fermented fish in Thailand10 or fish in Japan11 or Italy.9 Further, although adjusted by HBV and HCV, fish intake showed no association with HCC in a case-control study in Italy.7 This inconsistency may be owing to errors in exposure measurement and limited variation in fish. Given that Japanese consume large quantities of fish, the inverse association between fish and HCC in our study might have been clarified by the comprehensive questionnaire and wide range of consumption.
Although fish are the principal source of n-3 PUFAs, we are unaware of any study of the association between n-3 PUFA intake and HCC. In the present study, we also observed that consumption of n-3 PUFAs, particularly EPA, DPA, and DHA, was associated inversely with HCC. In clinical trials, dietary supplementation with n-3 PUFAs for 1-3 months was associated with a decreased release of interleukin-1ß and interleukin-6.22, 23, 24, 25 Given that HCC is an inflammation-related cancer that has a background of chronic inflammation, triggered by exposure to hepatitis virus infection or toxic compounds, such as ethanol,26, 27 the anti-inflammatory properties of n-3 PUFAs might decrease the risk of HCC. Of note, we showed that the risk of HCC was decreased with greater consumption of fish and n-3 PUFAs in subjects who were either or both anti-HCV or HBsAg positive. The intake of n-3 PUFA-rich fish might reduce the risk of HCC through the anti-inflammatory effects of n-3 PUFAs on chronic hepatitis.
Another possibility is that fish and n-3 PUFAs also might be associated with HCC through an improvement in insulin sensitivity. Given that recent epidemiologic data have suggested that diabetes and obesity are associated with an increased risk of HCC,28, 29, 30, 31 insulin resistance is now recognized as an independent risk factor for the development of HCC.29 Animal experiments indicate that the intake of n-3 fatty acids from fish oils has a beneficial effect on insulin sensitivity in rats,32 but not in human beings.33, 34, 35 High concentration of n-3 PUFAs in human skeletal muscle cells have been associated with improved insulin sensitivity.36 n-3 PUFAs from fish therefore might improve insulin resistance. In addition, a clinical study has shown the induction of plasma adiponectin in response to a daily intake of EPA and DHA.37 Thus, the induction of adiponectin also might contribute to the beneficial effect of n-3 PUFA on systemic insulin sensitivity. However, there was no difference in association between fish and n-3 PUFAs and HCC in participants with and without self-reported diabetes.
In contrast, ALA, which is another component of n-3 PUFAs, was weakly or not associated with HCC, although ALA might be converted to EPA and DHA. Other than fish, the other source of n-3 PUFA in this study population was vegetable oil, in which ALA is the only n-3 PUFA (EPA, DPA, and DHA are not included in vegetable oil). On adjustment for vegetable oil, results were not changed substantially. Therefore, EPA, DPA, or DHA among n-3 PUFA from fish might play particularly important roles as factors that lower the risk of HCC.
The strengths of the present study were its prospective design and negligible proportion of loss to follow-up evaluation (0.4%). Information on fish consumption was collected before the subsequent diagnosis of HCC, thereby diminishing the probability of the recall bias that is inherent to case-control studies. Another strength was that virus infection status was available at baseline, allowing us to clarify the association between n-3 PUFAs and HCC in a high-risk population, albeit the sample size was small. Further, dietary information was ascertained using a validated FFQ and the validity of fish and n-3 PUFAs intake was moderate.
Several limitations also warrant mention. First, because we estimated the consumption of fish and associated nutrients from self-reports and at one time point only, some measurement error in the assessment of consumption is inevitable. If present, however, this probably was nondifferential and likely would have lead to the underestimation of results. Second, we had no information on the clinical severity of hepatitis or on the treatment of subjects with hepatitis virus infection before or during the study period. If infected subjects had received treatment, the occurrence of HCC might have been decreased. However, this might have lead to the underestimation of HCC occurrence, which also would have biased the results toward the null. Finally, our study subjects were a middle-aged population, and caution accordingly is required in generalizing the present results to the young and elderly.
In conclusion, our large prospective study indicated that high consumption of n-3 PUFA-rich fish and n-3 PUFAs was associated with a reduced risk of HCC, even among a high-risk population. Given that the prognosis for HCC is extremely poor, our results would, if confirmed, have important implications for public health. Greater consumption of n-3 PUFA-rich fish and n-3 PUFAs may modify the development of HCC among HBV- and/or HCV-infected subjects.
Results
During an average follow-up period of 11.2 years, a total of 398 HCC cases were identified in total subjects. Baseline characteristics of subjects according to total fish consumption are shown in Table 1. Subjects with higher fish consumption tended to be older, smoke less, and drink less alcohol and coffee. Body mass index and soybean intake was not substantially different according to consumption. Intake of vegetables, iron, and fatty acid increased as fish intake increased. The proportion of subjects positive for anti-HCV, HBsAg, or both among quintiles of fish consumption was similar. The pattern of characteristics was similar according to intake of n-3 PUFA-rich fish (data not shown).
Spearman correlation coefficients for the associations between total fish, n-3 PUFA-rich fish, n-3 PUFA, EPA, DPA, and DHA were analyzed. There were strong correlations between fish and n-3 PUFA (r = 0.73), EPA (r = 0.85), DPA (r = 0.83), and DHA (r = 0.87) and between n-3 PUFA-rich fish and n-3 PUFA (r = 0.73), EPA (r = 0.86), DPA (r = 0.87), and DHA (r = 0.84).
Table 2 presents hazard ratios in relation to fish and n-3 PUFA consumption for HCC cases. Total fish consumption had a weak inverse association with the risk of HCC, with a multivariable HR for the highest vs lowest quintile of 0.64 (95% CI, 0.41-1.02; Ptrend = .07). n-3 PUFA-rich fish consumption was dose-dependently associated with a decreased risk of HCC, with a multivariable HR for the highest vs lowest quintile of 0.64 (95% CI, 0.42-0.96; Ptrend = .04). In addition, inverse associations were seen between EPA, DPA, DHA, and HCC, with multivariable HRs for the highest vs lowest quintile of 0.56 (95% CI, 0.36-0.85; Ptrend = .01) for EPA, 0.64 (95% CI, 0.41-0.98; Ptrend = .05) for DPA, and 0.56 (95% CI, 0.35-0.87; Ptrend = .03) for DHA. n-3 PUFA and ALA did not show statistically significant inverse associations with HCC, with respective multivariable HRs for the highest vs lowest quintile of 0.63 (95% CI, 0.36-1.10) and 0.78 (95% CI, 0.48-1.28). No substantial change in results was seen on additional analyses for HCC stratified by sex, smoking status, or body mass index (data not shown). Furthermore, our analyses did not change when restricted to cases that occurred after the first 3 years of follow-up evaluation (122 cases excluded) and when cases identified by death certificate only were excluded (42 cases excluded) (data not shown). Moreover, when subjects with self-reported pre-existing liver diseases were excluded (133 cases excluded), the results were attenuated but not substantially changed. The prevalence of fish oil supplement use was 0.06%; no change was seen when these users were excluded.
Fish consumption might reflect other lifestyle factors. In particular, subjects with higher fish consumption tended to drink less alcohol and coffee, and tended to have a past history of diabetes. Although we also assessed the effect of fish consumption according to alcohol, coffee drinking status, or history of diabetes, an inverse association between fish and n-3 PUFA-rich fish and HCC risk was shown in both regular (1-2 times/wk) and nonregular (<1 time/wk) alcohol drinkers, in both daily and nondaily coffee drinkers, and in both those with and without a history of diabetes (data not shown). Interaction between n-3 PUFA-rich fish and alcohol, coffee drinking status, or history of diabetes was not detected (Pinteraction =.25, .57, and .58, respectively).
To adjust for HCV and HBsAg status, we also analyzed the association between fish and n-3 PUFAs and HCC risk among subjects who had information on HCV and HBV infection status (Table 3). Although statistical significance was diminished because of a small sample size, similar results were seen, with multivariable HRs for the highest vs lowest tertile of 0.54 (95% CI, 0.23-1.24) for fish, 0.73 (95% CI, 0.35-1.53) for n-3 PUFA-rich fish, 0.51 (95% CI, 0.20-1.32) for n-3 PUFA, 0.70 (95% CI, 0.29-1.71) for ALA, 0.62 (95% CI, 0.28-1.39) for EPA, 0.80 (95% CI, 0.34-1.85) for DPA, and 0.63 (95% CI, 0.27-1.49) for DHA.
To clarify the association between fish and n-3 PUFAs and HCC risk among HBV- and/or HCV-infected subjects, we restricted analysis to subjects who were either or both anti-HCV or HBsAg positive (n = 1303) and anti-HCV positive (n = 911) (Table 4). Total fish consumption was not statistically significantly associated with the risk of HCC, with a multivariable HR for the highest vs lowest quintile of 0.52 (95% CI, 0.20-1.32; Ptrend = .31), and the inverse association between total fish and HCC was strengthened when subjects were limited to those who were anti-HCV positive, with a multivariable HR for the highest vs lowest quintile of 0.30 (95% CI, 0.11-0.82; Ptrend = .03). Higher n-3 PUFA-rich fish and n-3 PUFA consumption appeared to decrease the risk of HCC, but without statistical significance. Multivariable HRs for the highest vs lowest tertile among subjects who were either or both anti-HCV or HBsAg positive was 0.60 (95% CI, 0.25-1.40) for n-3 PUFA-rich fish and 0.41 (95% CI, 0.14-1.19) for n-3 PUFA, whereas the HR among subjects who were anti-HCV positive was 0.42 (95% CI, 0.16-1.12) for n-3 PUFA-rich fish and 0.44 (95% CI, 0.13-1.42) for n-3 PUFA. ALA, EPA, DPA, and DHA consumption also tended to be associated with a decreased risk of HCC among subjects who were either or both anti-HCV or HBsAg positive, albeit without statistical significance (highest vs lowest: multivariable HR, 0.69; 95% CI, 0.26-1.86; HR, 0.55; 95% CI, 0.22-1.39; HR, 0.55; 95% CI, 0.21-1.42, and HR, 0.59; 95% CI, 0.22-1.57, respectively). When subjects were restricted to those who were anti-HCV positive, a dose-dependent inverse association was seen, with multivariable HRs for the highest vs lowest tertile of 0.33 (95% CI, 0.12-0.92; Ptrend = .03) for EPA, 0.30 (95% CI, 0.10-0.88; Ptrend = .02) for DHA, and 0.37 (95% CI, 0.13-1.05; Ptrend = .06) for DPA. ALA showed no association with HCC.
|
|
| |
| |
|
|
|
