Page 33 - Read Online

Benhammou et al. Hepatoma Res 2020;6:35 I http://dx.doi.org/10.20517/2394-5079.2020.16 Page 5 of 15

Beyond germ-like mutations that may predispose to disease, which has largely been the focus of genetic
studies in NAFLD, heritable epigenetic changes also have an important role in NAFLD-associated
HCC which include but are not limited to DNA methylation, chromosomal looping interactions, RNA
modifications and the emerging role of non-coding RNAs [54-57] . Additionally, several studies have applied
circulating tumor DNA (cfDNA) methods to further study HCC [58,59] , which has not only shed some light
on the biology of HCC but is also promising for use as a biomarker in the future. The identification of these
epigenetic modifications points to further understanding of gene regulation changes, which are influenced
[54]
by their environment .

NON-GENETIC ENVIRONMENTAL FACTORS FOR NAFLD-HCC: CLINICAL,
PHARMACOLOGICAL AND LIFESTYLE FACTORS
Diabetes
[51]
NAFLD is a complex trait with common and rare variants that are influenced by the environment. Clinical
studies have demonstrated that the features of MetS affect NAFLD-associated HCC development. T2D has
been known to affect the risk of HCC as early as when NAFLD and NASH were starting to become more
[60]
recognized. The seminal VA study by El-Serag et al. demonstrated in a cohort of 173,643 Veterans followed
for 10 years that T2D significantly increased the risk of HCC. Studies that followed not only corroborated
these findings [61,62] but also demonstrated an increase in the risk of HCC when more features of MetS were
[63]
present . Using a prospectively collected cohort in the Nurses’ Health Study and Health Professionals’
[63]
Health Study, Simon et al. found that the adjusted hazard ratio (HR) for HCC in patients with diabetes
was 5.8 (95%CI: 3.49-9.64) and 5.49 (95%CI: 3.16-9.51) compared to non-diabetic patients in women
and men after adjusting for baseline characteristics, respectively. Interestingly, the risk of HCC was also
dependent on the duration of diabetes, solidifying the effects of T2D and insulin resistance on hepatocellular
carcinogenesis. Compared to patients without diabetes, diabetic patients had an adjusted HR of 7.52 (95%CI:
3.88-14.6) if they have had the disease for 10 years or more.

Obesity
Obesity has been associated with an increased risk of developing many cancers and this association is
strongest for HCC [64,65] . By convention, BMI has been used to measure obesity in epidemiological studies.
Although readily available in the clinical setting, BMI does not inform adipose distribution, specifically
visceral versus peripheral, which have different implications on metabolic health. Early studies in patients
with cirrhosis demonstrated that those with visceral adiposity were at higher risk of death compared to
[66]
those with peripheral adipose tissue. Ioannou et al. elegantly demonstrated these associations using the
National Health and Nutritional Examination Survey where patients were categorized based on central or
peripheral adipose distribution. Among patients with central adipose distribution, cirrhosis-related death
2
and hospitalizations were more common in the obese group (BMI ≥ 30 kg/m ) (adjusted HR = 2.2, 95%CI:
2
1.1-4.6) compared to normal-weight individuals (BMI < 25 kg/m ), which was not observed in patients with
increased peripheral adipose distribution. In NAFLD and NAFLD-associated HCC, central obesity, a key
feature of MetS, is also more physiologically informative of metabolic health [67,68] . This is also relevant in the
setting of studying NAFLD in groups of patients that may not have similar body compositions such as the
[69]
Asian population, which tends to have a higher percentage of body fat compared to White patients .
Hypertension and dyslipidemia
The evidence for hypertension, which is included in many definitions of MetS, is inconsistent and has been
shown to be a risk factor in some studies but not others [70,71] . Many studies also use collective features of MetS
to assess the attendant risk. Thus, the true effects of hypertension in isolation without other features of MetS
[70]
are unclear. Similarly, the data on dyslipidemia is conflicting. For instance, Welzel et al. demonstrated
that patients with a diagnosis of dyslipidemia based on ICDs demonstrated an adjusted odds ratio (OR)
[63]
of 1.35 (95%CI: 1.26-1.45) for NAFLD-HCC development, similar to others . Other studies have shown