 |
|
|
| |
NAFLD and cancer: More cause for concern?
|
| |
| |
"The study population comprised 25,947 subjects who met the inclusion criteria (Fig. 1). The prevalence of NAFLD was 33.6% (n = 8,721).
....Compared to subjects without NAFLD, those with NAFLD were older (mean age 50 vs. 47 years, p <0.001).....more likely to be male (71.1% vs. 45.1%, p <0.001), smokers (58.4% vs. 38.5%, p <0.001), diabetic (16.2% vs. 4.4%, p <0.001), and hypertensive (32.3%vs. 16.9%, p <0.001); and more likely to have higher levels of fasting glucose (102.8 vs. 93.1 mg/dl, p <0.001), total cholesterol (200.7 vs.187.0 mg/dl, p <0.001), serum ALT (30.7 vs. 17.8 U/L, p <0.001), and GGT (36.5 vs. 23.3 U/L, p <0.001).......the cancer incidence rate of the NAFLD group was higher than that of the non-NAFLD group (782.9 vs. 592.8 per 100,000 person-years; hazard ratio [HR] 1.32; 95% confidence interval [CI] 1.17-1.49; p <0.001)...When demographic and metabolic factors were adjusted for, NAFLD showed a strong association with three cancers: hepatocellular carcinoma ([HCC]; HR 16.73; 95% CI 2.09-133.85; p = 0.008),colorectal cancer in males (HR 2.01; 95% CI 1.10-3.68; p = 0.02), and breast cancer in females (HR 1.92; 95% CI 1.15-3.20;p = 0.01).....
A high NAFLD fibrosis score (NFS) and a high fibrosis-4 (FIB-4) score were associated with the development of all cancers and HCC.....On univariate analysis, the subjects with NAFLD showed higher association with the development of all cancers than those without NAFLD (hazard ratio [HR] 1.32; 95% CI 1.17-1.49; p <0.001; Table 4). After adjusting for age and sex, NAFLD had a significant association with three specific cancers: HCC (HR 15.86; 95% CI 2.07-121.33; p = 0.008), colorectal cancer in males (HR 2.13; 95% CI 1.22-3.74; p = 0.008), and breast cancer in females (HR 1.90; 95% CI 1.20-3.01; p = 0.006; Table 4). The results of multivariable analysis were consistent.
EDITORIAL - Kim et al. did find that, among patients with NAFLD, those with the higher fibrosis scores had a higher incidence of HCC, but not a significantly increased incidence of colorectal or breast cancer. Similarly, moderate to severe steatosis detected on ultrasound was directly related to the development of HCC only. While confirming the association and the pathogenic mechanism of NAFLD with HCC, the lack of a dose-effect relationship between markers of disease severity and rates of breast and colorectal cancer casts additional doubt on the causal effect of NAFLD on these cancer sites."
--------------------
NAFLD and cancer: More cause for concern? Editorial
Jan 18 Jnl of Hepatology - Peter Jepsen1,2,, Federica Turati3, Carlo La Vecchia3
1Department of Hepatology and Gastroenterology, Aarhus University Hospital, Aarhus, Denmark; 2Department of Clinical Epidemiology,
Aarhus University Hospital, Aarhus, Denmark; 3Department of Clinical Sciences and Community Health, University of Milan, 20133 Milan, Italy
Non-alcoholic fatty liver disease (NAFLD) is likely to become the most prevalent liver disease in many countries,[1], [2] yet clinicians are struggling to determine exactly why they should care about NAFLD. Is it merely a bystander - a manifestation of the metabolic syndrome resulting in cardiovascular disease - or is it a liver disease in its own right?3
In their cohort study from the Republic of Korea, Kim et al. present convincing data that patients with NAFLD, without cirrhosis, have a substantially (over 15-fold) higher incidence of hepatocellular carcinoma (HCC); thatmen with NAFLD have a 2-fold higher incidence of colorectal cancer; and that women with NAFLD have a 1.9-fold higher incidence of breast cancer.4 These estimates were based on 14 incident cases of HCC, 76 incident cases of colorectal cancer, and 91 incident cases of breast cancer.
The study was based on a cohort of 25,947 subjects who had a health check-up in a tertiary hospital in Korea between September 2004 and December 2005, and were followed-up to the end of 2015. Therefore, it includes a comparable reference population, since diagnostic ascertainment was similar in the entire cohort, unlike earlier studies which used general population controls.5 Moreover, participants with or without NAFLD were likely comparable with respect to several characteristics, such as job situation, socioeconomic status, and environmental exposures. All participants went through the same examinations for liver conditions. As many as 33.6% of the participants had NAFLD, and these NAFLD patients also had a higher body mass index, fasting glucose, LDL cholesterol, and triglyceride, as well as a higher prevalence of hypertension and diabetes, highlighting the fact that NAFLD is a manifestation of the metabolic syndrome.
These syndromic characteristics call into question whether NAFLD caused cancer development, or if cancer development was a result of the metabolic syndrome. When Kim et al. controlled for confounding by age, gender, smoking status, diabetes, hypertension, GGT, LDL cholesterol, HDL cholesterol, and triglycerides, the hazard ratio for HCC for patients with vs. without NAFLD fell from 25 to 17. Surprisingly, the same effect reduction was seen when they controlled for gender and age alone. In any case, an over 15-fold excess rate of HCC (and a lower confidence limit >2) indicates that NAFLD has a direct, causal effect on liver cancer, likely and largely due to its inflammatory status in the liver tissue. This conclusion is supported by the fact that the effects of diabetes, obesity, and hypertension - with relative risks in the order of two to four[6], [7], [8], [9] - are too weak to fully explain the strong association between NAFLD and HCC. In short, there are systemic mechanisms related to the metabolic syndrome, which partly account for the huge increase in liver cancer risk among patients with NAFLD, but the key mechanism remains the inflammatory status induced by NAFLD on the liver. Thus, the relation between NAFLD and HCC can be considered causal, although given the wide confidence interval (CI) from 2 to over 100 - it remains poorly quantified.
Inference on the twofold excess rates of male colorectal and female breast cancer is more difficult. Colorectal cancer is related to diabetes, abdominal obesity (more common in men than in women), and hence to the metabolic syndrome.[10], [11], [12] Consequently, after confounder adjustment the hazard ratio for colorectal cancer for both sexes combined decreased from 2.04 to 1.45. This leaves open the issue that confounding from components of the metabolic syndrome could explain the association between NAFLD and colorectal cancer. In addition, the absence of association in women (the hazard ratio was 0.63 in females, based on 23 cases vs. 53 in males), and hence the lack of significant association between NAFLD and colorectal cancer in both sexes combined (HR 1.45, 95% CI 0.88-2.38) leaves any inference on causality open to discussion.
In contrast, adjustment for components of the metabolic syndrome had little effect on the hazard ratio estimates for breast cancer. This is expected, since diabetes has a limited, or no, effect on breast cancer development, and the key component of the metabolic syndrome associated with breast cancer is being overweight in post-menopause.[13], [14] In stratified analyses, the association of NAFLD with breast cancer was restricted to non-obese women (BMI <25 kg/m2). Kim et al. argue that, in those women, NAFLD has the same carcinogenic effect as obesity. That might be true, and can be tested. Presently, it remains uncertain whether NAFLD is a distinct cause of breast cancer.
Kim et al. used ultrasound to diagnose NAFLD. This adds validity to their report and is consistent with European Association for the Study of the Liver (EASL) clinical practice guidelines,15 which further recommend that patients with abnormal liver enzymes and/or medium or high risk of fibrosis, according to the NAFLD fibrosis score or fibrosis-4 scores, should be referred to a specialist for a full workup. Kim et al. did find that, among patients with NAFLD, those with the higher fibrosis scores had a higher incidence of HCC, but not a significantly increased incidence of colorectal or breast cancer.
Similarly, moderate to severe steatosis detected on ultrasound was directly related to the development of HCC only. While confirming the association and the pathogenic mechanism of NAFLD with HCC, the lack of a dose-effect relationship between markers of disease severity and rates of breast and colorectal cancer casts additional doubt on the causal effect of NAFLD on these cancer sites. It would have been a helpful addition to the paper if Kim et al. had compared the three groups defined by the EASL guidelines: i) abnormal liver enzymes, or steatosis and medium/high fibrosis score; ii) steatosis, normal liver enzymes, and low fibrosis score; and iii) no steatosis and normal liver enzymes. Such an analysis could have validated the diagnostic flowchart presented in the EASL guidelines.15
Should we screen patients with NAFLD for HCC? Well, we do not have to understand exactly what causes HCC in patients with NAFLD, as long as they are at increased risk, which they certainly are. Still the answer is undefined, given the current body of evidence, and this is consistent with current EASL and American Association for the Study of Liver Diseases guidelines.[15], [16] Randomized controlled trials are necessary to determine whether screening reduces mortality in patients with NAFLD.
What is the risk of HCC in patients with NAFLD? The analysis gives the incidence rate of HCC among subjects with NAFLD (23 per 100,000 person-years), and among those without NAFLD (0.9 per 100,000 person-years). The very large confidence interval around the corresponding incidence rate ratio reflects the low number of HCC cases in the cohort (14 in the NAFLD and no NAFLD groups combined). In addition, it is impossible to estimate the risk of HCC without knowing the rate of death without HCC, the competing outcome in Kim et al.'s analyses.17 It is possible, likely in fact, that Kim et al.'s patients with NAFLD have a higher rate of death without HCC than their non-NAFLD counterparts because of their higher prevalence of smoking, diabetes, high LDL cholesterol, etc. If so, the higher rates of colorectal cancer and breast cancer in NAFLD patients might not result in higher risks of those cancers.18 For HCC, the difference in rates is so great that the absolute risk of HCC is likely be greater, too. However, it would have strengthened Kim et al.'s study considerably to have data on mortality without cancer.
Thus, it is now clear that patients with NAFLD have an increased risk of HCC, but their relative and absolute risk remains imprecisely quantified. It is unclear whether NAFLD increases the risk of colorectal cancer among men, and even less clear whether it increases the risk of breast cancer among women. Having more detailed patient data, such as data from liver biopsies, would have been helpful for narrowing down the subset of NAFLD patients who are at the greatest risk of developing HCC, and for clarifying the associations between NAFLD and colorectal and breast cancer. It would also have been helpful to supplement the baseline characteristics with data demonstrating how those characteristics changed during the follow-up. Importantly, despite the strong association between NAFLD and HCC development, it is unclear whether we should offer HCC surveillance to patients with NAFLD - a randomized trial should be conducted first.
Still, despite several open issues, the Korean cohort study by Kim et al. (1) provides relevant and valid information on the association between NAFLD and cancer risk, and evidence of its strong association with HCC. This may well become more relevant as a major cause of HCC in the future, considering our increasing control of HBV and HCV infections (i.e. the historic major causes of HCC) on a global population level.[19], [20]
-------------------------------------
Association between non-alcoholic fatty liver disease and cancer incidence rate
Highlights
⋅NAFLD was associated with HCC development.
⋅NAFLD was associated with colorectal cancer development in males.
⋅NAFLD was associated with breast cancer development in females.
⋅High NFS and high FIB-4 score were associated with developing all cancers and HCC.
Background & Aims
Little is known about the association between non-alcoholic fatty liver disease (NAFLD) and cancer development. This study investigated the cancer incidence rates in NAFLD and analysed the association between NAFLD and cancer development.
Methods
This historical cohort study included subjects who were followed up for >1 year after having a heath checkup at a tertiary hospital in Korea from September 1, 2004 to December 31, 2005. NAFLD was diagnosed by ultrasonographic detection of hepatic steatosis in the absence of other known liver disease, including alcoholic or viral hepatitis. Cox proportional hazards regression model was conducted to assess the association between NAFLD and cancer development.
Results
Of 25,947 subjects, 8,721 (33.6%) had NAFLD. During the total follow-up of 164,671 person-years (median 7.5 years), the cancer incidence rate of the NAFLD group was higher than that of the non-NAFLD group (782.9 vs. 592.8 per 100,000 person-years; hazard ratio [HR] 1.32; 95% confidence interval [CI] 1.17-1.49; p <0.001). When demographic and metabolic factors were adjusted for, NAFLD showed a strong association with three cancers: hepatocellular carcinoma ([HCC]; HR 16.73; 95% CI 2.09-133.85; p = 0.008), colorectal cancer in males (HR 2.01; 95% CI 1.10-3.68; p = 0.02), and breast cancer in females (HR 1.92; 95% CI 1.15-3.20; p = 0.01). A high NAFLD fibrosis score (NFS) and a high fibrosis-4 (FIB-4) score were associated with the development of all cancers and HCC.
Conclusion
NAFLD was associated with the development of HCC, colorectal cancer in males, and breast cancer in females. A high NFS and a high FIB-4 score showed a strong association with the development of all cancers and HCC.
Lay summary
Non-alcoholic fatty liver disease (NAFLD) is associated with developing hepatocellular carcinoma (HCC). There have been limited data on the association between NAFLD and extrahepatic cancers. This study demonstrated that patients with NAFLD showed a higher association with the development of HCC, colorectal cancer in males, and breast cancer in females. A high NAFLD fibrosis score and a high fibrosis-4 score showed a strong association with the development of all cancers and HCC.
Introduction
Non-alcoholic fatty liver disease (NAFLD) is one of the most common chronic liver diseases globally, with an estimated prevalence of 25.2%.1 NAFLD may progress to non-alcoholic steatohepatitis (NASH), cirrhosis, and hepatocellular carcinoma (HCC).[1], [2], [3]Moreover, NAFLD is strongly associated with insulin resistance, the metabolic syndrome, diabetes, and cardiovascular disease, indicating that NAFLD is a multisystem disease with extrahepatic complications.[4], [5], [6]
Several studies have shown that the second-most frequent cause of death after cardiovascular disease among patients with NAFLD is malignancy.[2], [7] Recent data have shown that NAFLD is an attributable cause of HCC and indicated that the number of cases of NAFLD-related HCC in the United States increased 9% each year from 2004 to 2009.8 Moreover, a cross-sectional study has demonstrated that patients with NAFLD, particularly those with NASH, are more likely to develop advanced colorectal neoplasms than healthy controls.9 Although previous studies have confirmed the association between NAFLD and the development of colorectal cancers, the association between them over a long period of follow-up has not been demonstrated.[9], [10] In addition, little attention has been given to an association between NAFLD and other extrahepatic cancers.
The aim of this study was to identify the incidence rates of various cancers in patients with NAFLD and the association between NAFLD and cancer development. Furthermore, the study aimed to investigate whether the severity of NAFLD based on noninvasive fibrosis scores is related to cancer development.
Results
Characteristics of the study population
The study population comprised 25,947 subjects who met the inclusion criteria (Fig. 1). The prevalence of NAFLD was 33.6% (n = 8,721). The baseline characteristics of the study subjects are presented (Table 1). Compared to subjects without NAFLD, those with NAFLD were older (mean age 50 vs. 47 years, p <0.001); more likely to be male (71.1% vs. 45.1%, p <0.001), smokers (58.4% vs.38.5%, p <0.001), diabetic (16.2% vs. 4.4%, p <0.001), and hypertensive (32.3% vs. 16.9%, p <0.001); and more likely to have higher levels of fasting glucose (102.8 vs. 93.1 mg/dl, p <0.001), total cholesterol (200.7 vs. 187.0 mg/dl, p <0.001), serum ALT (30.7 vs.17.8 U/L, p <0.001), and GGT (36.5 vs. 23.3 U/L, p <0.001).
Incidence rates of cancer
The median follow-up duration was 7.5 years (interquartile range [IQR] 3.2-9.3 years), contributing 164,671 person-years of follow-up. During the follow-up period, 440 subjects (5.0%) with NAFLD and 643 subjects (3.7%) without NAFLD developed malignancies. The overall cancer incidence rate was significantly higher in the NAFLD group (782.9 [95% confidence interval (CI) 711.5-859.7] per 100,000 person-years) than in the non-NAFLD group (592.8 [95% CI 547.8-640.4] per 100,000 person-years; IRR 1.32; 95% CI 1.17-1.49; p <0.001; Table 2).
Details of the cancer incidence rates are provided (Table 2). Subjects with NAFLD had significantly higher cancer incidence rates in three specific cancers than those without NAFLD: HCC (23.1 vs. 0.9 per 100,000 person-years; IRR 25.09; 95% CI 3.28-191.83;p = 0.002), colorectal cancer (69.4 vs. 34.1 per 100,000 person-years; IRR 2.04; 95% CI 1.30-3.19; p = 0.002), and breast cancer in female subjects (181.6 vs. 102.5 per 100,000 person-years; IRR 1.77; 95% CI 1.15-2.74; p = 0.01). There was no significant difference in the incidence rates of cancers of the thyroid, oesophagus, stomach, pancreas, biliary tract, lung, kidney, bladder, or uterus; non-Hodgkin lymphoma; leukemia; or other rare tumors.
When stratified by gender, male subjects with NAFLD had significantly higher incidence rates of HCC (30.3 vs. 2.0 per 100,000 person-years; IRR 14.80; 95% CI 1.93-113.85; p = 0.01; Table 3) and colorectal cancer (85.7 vs. 38.8 per 100,000 person-years; IRR 2.21; 95% CI 1.26-3.87; p = 0.006), whereas female subjects with NAFLD had a significantly higher incidence rate of breast cancer (181.6 vs.102.5 per 100,000 person-years; IRR 1.77; 95% CI 1.15-2.74; p = 0.01). Among female subjects, the incidence rate of colorectal cancer did not differ between the NAFLD and control group.
Association between NAFLD and cancer development
On univariate analysis, the subjects with NAFLD showed higher association with the development of all cancers than those without NAFLD (hazard ratio [HR] 1.32; 95% CI 1.17-1.49; p <0.001; Table 4). After adjusting for age and sex, NAFLD had a significant association with three specific cancers: HCC (HR 15.86; 95% CI 2.07-121.33; p = 0.008), colorectal cancer in males (HR 2.13; 95% CI 1.22-3.74; p = 0.008), and breast cancer in females (HR 1.90; 95% CI 1.20-3.01; p = 0.006; Table 4). The results of multivariable analysis were consistent. After adjusting for demographic and metabolic factors, NAFLD had a strong association with HCC (HR 16.73; 95% CI 2.09-133.85; p = 0.008; Table 4), colorectal cancer in males (HR 2.01; 95% CI 1.10-3.68; p = 0.02), and breast cancer in females (HR 1.92; 95% CI 1.15-3.20; p = 0.01). To further investigate the impact of obesity on the development of breast cancer, the subjects were stratified as obese (≥25 kg/m2) or non-obese according to their BMI. In multivariable analyses, there were no significant associations between NAFLD and breast cancer among the obese subjects (HR 1.06; 95% CI 0.45-2.49; p = 0.90), whereas the non-obese subjects showed a substantially greater association between them (HR 2.40; 95% CI 1.27-4.54; p = 0.007).
|
| |
|
|
|
|
|
