Impact of diabetes, obesity, and dyslipidemia on the risk of hepatocellular carcinoma in patients with chronic liver diseases (original) (raw)

ABSTRACT

INTRODUCTION

Liver cancer is the sixth most common cancer worldwide, and hepatocellular carcinoma (HCC) accounts for 75–85% of all primary liver cancers [1]. HCC is the third leading cause of cancer-related death among solid cancers, leading to approximately 1 million deaths yearly [2]. Chronic liver diseases (CLDs), such as hepatitis B virus (HBV) infection and liver cirrhosis (LC), are well known risk factors for the development of HCC.

Recently, increasing evidence has suggested that metabolic factors, including diabetes mellitus (DM), obesity, dyslipidemia, and metabolic syndromes, are risk factors for HCC [3-6]. In populations with a low prevalence of viral hepatitis, the overall impact of metabolic factors on HCC was suggested to be greater than that of viral hepatitis [7]. The effects of metabolic factors on HCC have not been well established in individuals with CLDs. Observational studies have suggested the potential beneficial effects of some pharmaceutical agents for metabolic disorders (e.g., statins) in preventing HCC.

The purpose of this review is to summarize the associations between metabolic factors such as DM, dyslipidemia, and obesity and the risk of HCC in patients with CLDs. In this review, we focused on HBV, hepatitis C virus (HCV) infection, alcoholic liver disease (ALD), non-alcoholic fatty liver disease (NAFLD), and LC among CLDs. The associations in the general population were briefly summarized for comparison. A clearer understanding of the role of metabolic factors in the development of HCC will help in informed decision-making and patient management aimed at reducing the risk of HCC in patients with CLDs and comorbid metabolic disorders.

ROLE OF METABOLIC FACTORS IN CLDS: A SYNERGISTIC INTERACTION

Patients with CLDs, such as HBV/HCV, have a several dozen-fold higher risk of HCC development compared with individuals without CLDs, while metabolic disorders, such as diabetes and obesity, are associated with a 2–3-fold higher risk of HCC in the general population. If patients with CLD and comorbid metabolic disorders have a higher risk of HCC, then carcinogenic mechanisms by CLDs and metabolic disorders should have a strong synergistic interaction. Let us assume that HBV patients without diabetes have a 50-fold higher risk for HCC development compared to the general population without diabetes and that diabetic patients have a 2-fold higher risk compared with general population without diabetes. If HBV and diabetes are independent risk factors without synergistic interaction, compared to individuals without HBV and diabetes, and HBV patients without diabetes, HBV patients with diabetes will have approximately 51-fold and 1.02-fold higher risk, respectively, for HCC development (likewise, if HBV patients with diabetes have a 2-fold higher risk compared to those without diabetes, then they have a 100-fold higher risk compared with individuals without diabetes and HBV) [8,9]. Additionally, the stronger the impact of a certain CLD (such as LC) on HCC development is, the stronger synergistic interaction should exist in individuals with CLD for a metabolic factor to be identified as a risk factor in individuals with CLD. It is common that the associations of a factor are weaker in individuals with diseases than in the general population. Potential mechanisms should be focused on the synergistic interactions of mechanisms by CLDs and metabolic factors (rather than on the general mechanism of a factor per se). However, few detailed mechanisms have been proposed, and further research would be needed to explain potential synergic interactions.

DM

General population

In general population studies, DM has been associated with an increased risk of HCC. A Korean study using the National Health Insurance Service-Health Screening Cohort (NHIS-HEALS) reported that DM was associated with a higher risk of HCC (hazard ratio [HR], 1.82) [10]. Two meta-analyses showed that DM was associated with a risk ratio (RR) of 2.01 and 2.31 for HCC development [11,12]. However, liver diseases themselves increase the risk of DM development [13-15]. Thus, at least some part of the association between DM and HCC in the general population may be due to reverse causation.

Patients with viral hepatitis

DM has been associated with an increased risk of HCC in patients with CLD. The magnitude of association between DM and HCC somewhat differed for each liver disease (Table 1) [16-30]. A recent Korean study on 214,167 male patients with HBV reported that the presence of DM was significantly associated with a higher risk of HCC (HR, 1.23) [16]. In a recent systematic review and meta-analysis of 36 studies, DM was associated with an HR of 1.36 for HCC development in patients with HBV [17]. A systematic review of seven studies in 10,700 patients with HCV reported that DM increased the risk of HCC by approximately 2-fold [19]. The HCC risk associated with DM appears to be stronger in HCV than in HBV [19,29]. However, in a US study in 52,671 HCV-LC patients (7,605 HCC cases), DM was not associated with the risk of HCC [20].

Patients with LC

In ALD-LC patients, DM was associated with approximately 1.5-fold increased risk of HCC [20,23]. A recent systematic review reported that DM increases 2.65-fold the risk of HCC incidence in NAFLD patients [30]. In NAFLD patients with/without LC, HRs associated with DM for HCC were 1.24 (407 HCC cases), 1.93 (608 HCC cases), 1.30 (291 HCC cases), and 2.77 (253 HCC cases) in studies with >200 HCC cases, while the corresponding HRs were 4.18 (30 HCC cases), 2.90 (28 HCC cases), 4.72 (41 HCC cases), and 3.21(16 HCC cases) in studies with <50 HCC cases [20,23-28]. The magnitude of association between DM and the risk of HCC in NAFLD patients may be comparable to that in HCV patients or in general populations.

Patients with ALD and NAFLD

The HCC risk associated with DM seemed to be highest in NAFLD, followed by ALD and HBV. However, HCV and NAFLD are well known to increase the DM incidence by approximately 2-fold [14,15]. Comorbid DM is considered a marker of severity in NAFLD [31,32]. Therefore, the strongest associations do not necessarily mean strongest causal effect of DM on HCC development in NAFLD. Overall, diabetes may have causal effects on HCC in patients with CLDs. However, part of the effects of DM may reflect the underlying severity of the liver disease.

Anti-diabetic drugs in CLDs

Only a few studies on the effects of anti-diabetic drugs on the HCC risk in patients with CLDs have been conducted, and the results have been inconsistent (Table 2). In HBV patient, some recent cohort studies reported that metformin increased the HCC risk [33,34]. Whereas, another cohort study reported that metformin was not associated with HCC [35]. While in studies on patients with NAFLD, metformin did not have a significant association with HCC risk [24,36]. In contrast, thiazolidinediones reduced the risk of HCC in patients with HBV (HR, 0.46) [35]. Kramer et al. [37] reported that insulin and sulfonylureas alone had no effect on HCC risk, but insulin in combination with oral medication (metformin, sulfonylurea) had a higher risk of HCC with NAFLD. Overall, there is a lack of evidence, and it is unclear whether the use of anti-diabetic medication itself affects HCC development or whether different types of diabetic medications differently affect the development of HCC in patients with CLD.

OBESITY

In this review, the international standard of body mass index (BMI) classification, namely underweight (BMI <18.5 kg/m2), normal weight (18.5–24.9 kg/m2), overweight (25–29.9 kg/m2), and obesity (≥30 kg/m2), was used, instead of the classification by the Korean Society for the Study of Obesity that defines obesity as BMI ≥25 kg/m2.

General population

Obesity is considered a risk factor for HCC, especially in populations of European origin [38,39]. Meta-analyses in the general, mostly Western, populations showed that obesity increased the risk of HCC (or primary liver cancer) by approximately two times [40,41]. In Asian populations, despite their slim body shape, obesity-related HCC risks were substantially lower than Western populations and the summed relative risk was approximately 1.5 in a meta-analysis [10,40-44].

Patients with viral hepatitis

Most studies on patients with HBV have been conducted in Asia, especially Taiwan (Table 3) [45-49]. The Taiwanese studies did not show consistent association between obesity and HCC risk [45-48]. A recent study on Korean adults with chronic HBV showed that obesity, compared to a BMI of 18.5–22.9, was associated with 22% (men: HR, 1.22; 95% confidence interval [CI], 1.09–1.36]) and 46% (women: HR, 1.46; 95% CI, 1.24–1.71) higher risk of HCC [49]. This large-scale study using five categories of BMI identified relatively weak associations between obesity and HCC. Previous studies may not have found these subtle results consistently, probably due to the small number of participants, the crude categories of obesity generally used (such as obesity vs. non-obesity), and the inclusion of only men in some cases [45-48]. Obesity may be a risk factor for HCC in patients with HBV; however, the effect size seemed to be weaker than that in the general population.

Most studies reported that BMI or obesity were not associated with HCC in patients with HCV infections [21,22,50-54]. Two studies, one Japanese and one Taiwanese, which are commonly cited concerning the higher HCC risk associated with obesity, have a major limitation [55,56]. In the Japanese study, relative risks associated with obesity were calculated by comparing with underweight, not with normal weight [55]. In the Taiwanese study, the higher HCC risk associated with obesity was based on a very small number of HCC cases (five cases of obesity) [56]. Overall, the association between obesity and HCC has been inconsistent in patients with HCV. Considering the similarity of the impact of HBV and HCV on HCC, obesity might impact HCC in individuals with HCV infections similarly to that in individuals with HBV infection; that is, obesity may modestly increase HCC development in individuals with HCV.

Patients with LC

In individuals with LC, the associations have been complex and inconsistent and have shown the potential to differ according to LC etiology, which includes ALD, NAFLD, HBV, and HCV [20,23,24,57-63]. In two studies on NAFLD-LC patients who received care in the US Veterans Affairs healthcare system from the same authors, one showed a potentially higher HCC risk associated with high BMI, whereas the other did not [20,23]. The associations appeared to be clearer and stronger in alcoholic LC than in viral LC [20,57]. These results suggest a strong synergistic interaction between ALD (or alcohol consumption) and obesity, which was also observed in general populations [47].

Patients with NAFLD

The studies conducted among NAFLD patients of Western populations, including those without cirrhosis/fibrosis, found that BMI or obesity was not statistically significantly associated with HCC [25,59,64]. It is surprising that the impact of obesity on HCC development has not been clearly shown in NAFLD patients. However, these NAFLD studies, including those conducted mostly on non-cirrhotic patients, mainly used crude categories of BMI (obesity vs. non-obesity) and possibly lacked statistical power to detect the true association [25,59,64]. Additionally, probably because cirrhosis is a main mechanism of NAFLD-related HCC, the association between obesity and HCC might not be apparent in NAFLD-LC patients [20,23,24,59].

In summary, obesity may increase the risk of HCC in patients with CLD, but probably with a weaker effect size than that in the general population [21,52,53]. Liver damage caused by obesity and CLD itself may synergistically interact together and further facilitate the progression and/or development of HCC in patients with CLD. However, it should be noted that there is limited evidence even in NAFLD cases.

Central obesity

Compared to the lowest category, the highest waist circumference (WC) category had a 1.59 times higher risk of HCC development in a recent systematic review, mostly in the general population [65]. A recent Chinese study on chronic HBV patients reported that central obesity defined by waist-to-height ratio, not by WC and waist-hip ratio, was associated with an increased risk of HCC (HR, 1.63; 95% CI, 1.11–2.38) compared to non-central obesity. However, the risk of HCC with changes in the waist-to-height ratio was associated with an increased risk of HCC both with a decrease and an increase in the ratio [66]. In a small Korean study among 102 chronic HBV patients (seven HCC cases), central obesity, defined by WC, waist-hip ratio, and visceral fat area, was not associated with HCC risk [67]. Despite the potential, it is difficult to conclude whether central obesity is a better predictor and risk factor than general obesity in individuals with CLDs, due to the very small number of studies.

DYSLIPIDEMIA

General population

As each type of dyslipidemia may have different associations with HCC, we focused on studies reporting specific lipid profiles, such as total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and triglyceride (TG). Most studies revealed that low TC levels were strongly associated with an increased risk of HCC development [6,68-74]. In a Korean study that used the NHIS-HEALS, a nearly 50% reduced risk of HCC was observed for every 39 mg/dL (1 mmol/L) increase in TC (HR, 0.54; 95% CI, 0.51–0.58) [71]. Only a few studies have examined the associations of other lipids with HCC. Low levels of LDL-C and TG were generally associated with a higher risk of HCC, while HDL-C showed unclear inverse associations [69,70,74]. Overall, TC and potentially LDL-C and TG, seemed to have clear inverse associations with HCC in the general population, whereas the dose-response association of HDL-C with HCC was unclear.

Patients with CLDs: TC

In patients with CLDs, like in the general population, TC levels were inversely associated with the risk of HCC. Higher TC levels were generally associated with a lower risk of HCC in individuals with viral hepatitis (including HBV and HCV), LC, and NAFLD (Table 4) [36,70,75-77]. Compared to the studies on TC, a relatively small number of studies have reported associations between other lipid profiles and HCC.

While low TC levels may be a useful marker for HCC development in patients with CLD, the potential causal effects of low TC levels are uncertain. The presence of CLD leads to hepatic damage, which, in turn, is related to changes in lipid metabolism. Serum cholesterol levels in patients with HBV, HCV, and HCC were reported to be lower than those in the normal control group [78-80]. The decrement in cholesterol levels positively correlated with the severity of liver disease [78,79,81-85]. Statin affected lower levels of cholesterol levels were not associated with higher risk of HCC [71]. Low TC levels seemed to be more a strong predictor, but less a risk factor for HCC. Future studies are needed to clarify the role of lipid profiles, such as whether they have independent effects on HCC development.

Patients with CLDs: other lipid profile

The association between LDL-cholesterol and HCC was similar to that between TC and HCC [70,77]. A recent population-based Korean study on HBV patients showed that HCC was negatively associated with increased TG levels [86]. In a large-scale Korean study, TG showed inverse associations and low HDL-cholesterol (L-curve) levels were associated with a higher risk of HCC in individuals with chronic viral hepatitis or LC, whereas a smaller study (89 HCC cases) reported that high TG levels, but not low HDL-cholesterol levels, were associated with a higher risk of HCC in patients with HBV [70,77]. Dyslipidemia, defined by high TG and/or low HDL-cholesterol levels, was not associated with a higher risk of HCC in American patients with NAFLD [25]. Overall, similar to the results in the general population, TC, potentially LDL-cholesterol and TG, seemed to have inverse associations with HCC in individuals with CLDs, whereas the association of low HDL-cholesterol with HCC was unclear.

Statin therapy in CLDs

In several observational studies, statin therapy has generally been associated with a reduced risk of HCC development in patients with CLDs (Table 5). Statin use was associated with a materially lower risk of HCC in patients with HBV, HCV, and NAFLD [17,33,36,75,76,87-94]. A recent meta-analysis of studies on CLD reported that statin use was associated with an HR of 0.57 for HCC development [94].

Caution should be exercised when interpreting the beneficial effects of statins on HCC based on evidence from observational studies. The higher the cholesterol level, the higher the possibility of statin use. The natural low cholesterol level, not the attained level after statin use, is an important predictor and a potential risk factor for HCC [71]. Therefore, when the association between statin use and HCC risk is examined in observational studies, only patients newly initiated on statin treatment should be enrolled as statin users and their natural cholesterol levels before statin use should be measured. The preventive effect of statins disappeared in such studies on the general population [71,95]. Importantly, in the ad hoc analysis of randomized controlled trials for cardiovascular disease prevention, statins had no effect in preventing HCC (HR, 1.06 for statin use/more use compared with non-use/less use) [96,97]. Overall, whether statins have a preventive effect against HCC remains unclear. In the meantime, for the prevention of cardiovascular diseases, there is no need to apply different criteria for starting statins in patients with CLD and the general population.

SYNERGISTIC EFFECTS OF METABOLIC FACTORS

Several studies have examined whether metabolic factors have a synergistic effect on the development of HCC in patients with CLDs. In a study on patients with cirrhosis, the HR for obesity alone was 2.1, and the HR for DM alone was 1.4. The coexistence of obesity and DM (compared to BMI <25 kg/m2 and no diabetes) increased the risk of HCC by >6 times [60]. However, this study only adjusted for age, sex, and a cirrhosis etiology. In a Taiwanese study on patients with HBV, a single metabolic factor did not increase the risk of HCC; however, the coexistence of three or more metabolic factors increased the risk of HCC [98]. In a retrospective study on 271,906 patients with NAFLD in the USA, obesity, hypertension, diabetes, and dyslipidemia had no clear synergistic effects on the composite occurrence of HCC and LC [25]. In a recent study on Korean patients with HBV, a higher number of metabolic syndrome components was negatively associated with the risk of liver cancer [86]. Overall, the evidence of a synergistic interaction among metabolic factors is insufficient. Further studies are needed to elucidate this issue.

LIMITATION

First, only a limited number of studies are available on the association between lipids other than TC and anti-diabetic agents and HCC development, although the number of studies conducted on patients with CLDs is increasing. Second, statistical models seemed to adjust for potential mediators of the effects of metabolic factors in most studies that included information on the associations between metabolic factors and HCC. For example, it is inadequate to adjust for diabetes, blood pressure, and cholesterol when examining the association between obesity and cardiovascular disease/mortality because these factors are considered as mediators of the effects of obesity [99]. Although this is understandable, since most of the studies were not specifically examining metabolic factors. Nonetheless, these studies may not accurately capture the specific associations of metabolic factors. Third, it is difficult to compare potential regional differences (such as between Western and Eastern populations) in this association. For example, most studies on HBV have been conducted in East Asia, while NAFLD and non-viral LC have been mainly studied in the Western population. Finally, most of the previous studies reviewed were retrospective studies, which are more prone to bias than prospective studies. More prospective, large-scale studies are needed to better capture these associations.

CONCLUSION

Comorbid DM increases the risk of HCC in patients with CLD. However, at least some of the effects of DM on HCC reflect the underlying severity of CLDs (reverse causality), especially for HCV and NAFLD. Obesity may increase the risk of HCC in patients with CLD, although the magnitude of the effect is weaker than that in the general population. Although obesity increases the risk of NAFLD, there is limited evidence on whether obesity increases the HCC risk in patients with NAFLD. Low lipid levels, especially TC, and perhaps LDL-C and TG, rather than dyslipidemia, are associated with an increased risk of HCC. TC levels may be used as a marker of liver disease severity associated with the risk of HCC. Whether statins or anti-diabetic agents have a preventive effect against HCC in patients with CLD remains unclear.

FOOTNOTES

Authors’ contributions

Study concept and design: Baek Gyu Jun and Sang-Wook Yi; Data analysis and interpretation: Baek Gyu Jun, Hwang Sik Shin, and Sang-Wook Yi; Wrote the paper: Baek Gyu Jun, Sang-Wook Yi, and Hwang Sik Shin; All authors have read and approved the final version of the manuscript.

Conflicts of Interest

The authors have no conflicts to disclose.

Table 1.

Association between diabetes mellitus and hepatocellular carcinoma

Table 1.

Study Region Etiology Study design Patients/HCC cases Result (95% CI) _P_-value
Kim et al. [16] (2018) Korea HBV Population-based cohort study 214,167/11,241 HR, 1.23 (1.15–1.34) <0.01
Campbell et al. [17] (2021) HBV Meta-analysis 40 studies (536,456) HR, 1.36 (1.20–1.32) <0.01
Hsu et al. [18] (2018) Taiwan HBV Population-based cohort study 23,851/596 HR, 1.3 (1.1–1.6)
Dyal et al. [19] (2016) HCV Meta-analysis 9 studies HR, 1.73 (1.30–2.30); RR, 3.52 (1.29–9.24)
Yang et al. [29] (2022) HBV Meta-analysis 11 studies RR, 1.37 (1.24–1.51); OR, 1.99 (0.73–5.48)
HCV 15 studies RR, 1.76 (1.42–2.17); OR, 1.77 (1.18–2.64)
Chen et al. [30] (2022) NAFLD-LC Meta-analysis 8 studies HR, 4.55 (2.34–8.87)
NAFLD HR, 2.65 (2.02–3.49)
Ioannou et al. [20] (2018) USA HCV-LC Retrospective cohort study 44,007/7,605 HR, 1.02 0.6
ALD-LC 29,326/1,388 HR, 1.54 <0.01
NAFLD-LC 13,456/608 HR, 1.93 <0.01
Huang et al. [21] (2015) Taiwan HCV Population-based cohort study 2,187/82 HR, 1.91 (1.10–3.30) 0.02
Arase et al. [22] (2013) Japan HCV Retrospective cohort study 4,302/393 HR, 1.73 (1.30–2.30) <0.01
Ioannou et al. [23] (2019) USA ALD-LC Retrospective cohort study 16,175/871 HR, 1.46 <0.01
NAFLD-LC 7,068/407 HR, 1.24 0.10
Yang et al. [24] (2020) USA NAFLD-LC Hospital-based cohort study 354/30 HR, 4.18 (1.23–14.2) 0.02
Retrospective cohort study 6,630/291 HR, 1.30 (1.02–1.66) 0.03
Kanwal et al. [25] (2020) USA NAFLD Retrospective cohort study 271,906/253 HR, 2.77 (2.03–3.77)
Bertot et al. [26] (2018) Australia NAFLD Hospital-based cohort study 284/28 HR, 2.9 (1.2–7.3) 0.02
Vilar-Gomez et al. [27] (2018) Spain, Australia, Hong Kong, and Cuba NAFLD Cohort study 456/41 HR, 4.72 (2.13–10.45) <0.01
Kawamura et al. [28] (2012) Japan NAFLD Retrospective cohort study 6,508/16 HR, 3.21 (1.09–9.50) 0.04

Table 2.

Association between anti-diabetic drugs and hepatocellular carcinoma

Table 2.

Study Region Drug Etiology Study design Patients/HCC cases Result (95% CI) _P_-value
Chen et al. [33] (2015) Taiwan Metformin HBV Population-based cohort study 71,824/1,753 HR, 1.25 (1.06–1.47) <0.01
Hsu et al. [34] (2018) Taiwan Metformin HBV Population-based cohort study 27,820/802 HR, 2.20 (1.86–2.60) <0.01
Yip et al. [35] (2020) Hongkong Thiazolidinediones HBV Retrospective cohort study 28,999/2,307 HR, 0.46 (0.24–0.88) 0.19
Yang et al. [24] (2020) USA Metformin NAFLD Retrospective cohort study 354/30 HR, 1.93 (0.84–4.41) 0.12
Lee et al. [36] (2017) Taiwan Metformin NAFLD Population-based cohort study 18,080 HR, 1.29 (0.47–3.54) 0.62
Kramer et al. [37] (2022) USA Insulin NAFLD Retrospective cohort study 85,963/524 HR, 1.05 (0.88–1.27) 0.57
Sulfonylureas HR, 0.98 (0.84-.1.16) 0.84
Insulin-metformin HR, 1.53 (1.25-1.86) <0.01
Insulin-metformin-sulfonylureas HR, 1.71 (1.41-2.08) <0.01

Table 3.

Association between obesity and hepatocellular carcinoma

Table 3.

Study Region Etiology Study design Patients/HCC cases Result (95% CI) _P_-value
Yu et al. [45] (2008) Taiwan HBV Cohort study 2,903/134 HR, 1.48 (1.04–2.12); BMI 25–30†
HR, 1.96 (0.72–5.38); BMI >30†
Fu et al. [46] (2015) Taiwan HBV Cohort study 4,179/111 HR, 0.31 (0.04–2.24); obesity* 0.25
Loomba et al. [47] (2010) Taiwan HBV Cohort study 2,260/135 HR, 1.00 (0.93–1.06)[| ](#tfn5-cmh-2021-0383)
Chao et al. [48] (2011) Taiwan HBV Cohort study 1,142/124 HR, 1.58 (1.16–2.17); BMI ≥25* <0.01
Kim et al. [49] (2018) Korea HBV Population-based cohort study 370,322/14,609 HR, 1.04 (0.99–1.09); men; BMI 25.0–29.9‡
HR, 1.22 (1.09–1.36); men; BMI ≥30‡ <0.01
HR, 1.25 (1.15–1.36); women; BMI 25.0–29.9‡ <0.01
HR, 1.46 (1.24–1.71); women; BMI ≥30‡ <0.01
Huang et al. [21] (2015) Taiwan HCV Population-based cohort study 2,187/82 HR, 1.79 (0.43–7.44); obesity* 0.42
Arase et al. [22] (2013) Japan HCV Retrospective cohort study 4,302/393 uHR, 1.37 (1.12–1.66); BMI ≥22* <0.01
Veldt et al. [50] (2008) Europe, Canada HCV Prospective 541 HR, 0.94 (0.84–1.05); BMI[| ](#tfn5-cmh-2021-0383)
Arano et al. [51] (2011) Japan HCV Retrospective cohort study 325/122 HR, 1.01 (0.93–1.09); men[| ](#tfn5-cmh-2021-0383)
HR, 1.09 (0.99–1.19); women[| ](#tfn5-cmh-2021-0383) 0.06
Yang et al. [52] (2016) USA HCV Hospital-based cohort study 739/69 uHR, 1.0 (1.0–1.0); BMI[| ](#tfn5-cmh-2021-0383)
Lok et al. [53] (2009) USA HCV Cohort study 1,005/48 Lower BMI was associated with HCC (unadjusted) 0.04
McMahon et al. [54] (2017) USA HCV Population-based cohort study 1,080/40 uHR, 1.06 (0.5–2.1); BMI ≥30* 0.87
Ohki et al. [55] (2008) Japan HCV Retrospective cohort study 1,431/122 HR, 1.86 (1.09–3.16); BMI 25–30§ 0.02
HR, 3.10 (1.41–6.81); BMI >30§ 0.01
Hung et al. [56] (2011) Taiwan HCV Cohort study 1,470/87 HR, 1.31 (0.83–2.05); BMI ≥25* 0.24
Ioannou et al. [20] (2018) USA LC Retrospective study 116,404/10,042 HR, 1.0; BMI 24.5–28; HCV–LC†† 0.96
HR, 0.93; BMI 28–32; HCV–LC†† 0.04
HR, 0.84; BMI >32; HCV–LC <0.01
HR, 1.2; BMI 24.5–28; NAFLD–LC 0.48
HR, 1.52; BMI 28–32; NAFLD–LC 0.09
HR, 1.45; BMI >32; NAFLD–LC 0.10
HR, 1.59; BMI 24.5–28; ALD–LC†† <0.01
HR, 1.74; BMI 28–32; ALD–LC†† <0.01
HR, 1.88; BMI >32; ALD–LC†† <0.01
Ioannou et al. [23] (2019) USA LC Retrospective cohort study 23,243/1,278 HR, 1.32; BMI 25.2–29.3; ALD-LC‡‡ <0.01
HR, 1.39; BMI 29.3–33.8; ALD-LC ‡‡ <0.01
HR, 1.49; BMI 33.8–38.5; ALD-LC‡‡ <0.01
HR, 1.31; BMI >38.5; ALD-LC ‡‡ 0.06
HR, 0.91; BMI 25.2–29.3; NAFLD-LC ‡‡ 0.64
HR, 0.9; BMI 29.3–33.8; NAFLD-LC ‡‡ 0.60
HR, 0.8; BMI 33.8–38.5; NAFLD-LC‡‡ 0.28
HR, 0.77; BMI >38.5; NAFLD-LC ‡‡ 0.22
Yang et al. [24] (2020) USA LC Hospital-based cohort study 354/30 uHR, 0.99 (0.95–1.03); Mayo cohort; BMI[| ](#tfn5-cmh-2021-0383)
Cohort study 6,630/291 uHR, 1.00 (0.98–1.02); UNOS cohort; BMI[| ](#tfn5-cmh-2021-0383) 0.83
Nair et al. [57] (2002) USA LC Large-population database study 19,271/659 HR, 1.65 (1.22–2.22); BMI ≥30* <0.01
Pais et al. [58] (2015) France LC Retrospective study 110/29 OR, 8.24 (2.04–33.32); BMI >25* <0.01
Grimaudo et al. [59] (2020) Italy LC Prospective study 162/13 HR, 0.86 (0.21–3.51); BMI ≥30* 0.83
N’Kontchou et al. [60] (2006) France LC Retrospective cohort study 771/220 HR, 2.0 (1.4–2.7); BMI 25.0–29.9** <0.01
HR, 2.8 (2.0–4.0); BMI ≥30** <0.01
Archambeaud et al. [61] (2015) France LC Case-control study 905/282 OR, 1.56 (1.02–2.37); past obesity* 0.03
uOR, 1.04 (0.72–1.48); current obesity* 0.84
Brichler et al. [62] (2019) France LC Prospective cohort study 317/27 HR, 2.67 (1.04–6.84); BMI ≥30* 0.04
Ioannou et al. [63] (2007) USA LC Cohort study 2,126/173 HR, 2.8 (1.4–5.4); BMI 25.0–29.9**
HR, 2.5 (1.3–4.9); BMI ≥30
Kanwal et al. [25] (2020) USA NAFLD Retrospective cohort study 271,906/253 HR, 1.31 (0.98–1.74); BMI >30*
Ito et al. [64] (2020) Japan NAFLD Retrospective study 179/7 uHR, 1.07 (0.21–5.51); BMI ≥25* 0.94
Fan et al. [66] (2021) China HBV Cohort study 5,754/142 HR, 1.63 (1.11–2.38); waist-to-height ratio >0.50* 0.01
Lee et al. [67] (2016) Korea HBV Cohort study 102/7 uHR, 0.88 (0.17–4.53); WC ≥90 (men) or ≥80 (women)* 0.88
uHR, 1.94 (0.23–15.87); waist-to-hip ratio ≥0.9* 0.39
uHR, 1.69 (0.43–8.71), visceral fat area ≥100 cm2 * 0.50

Table 4.

Association between dyslipidemia and hepatocellular carcinoma

Table 4.

Study Region Etiology Study design Patients/HCC cases Result (95% CI) _P_-value
Lee et al. [36] (2017) Taiwan NAFLD Population-based cohort study 18,080/41 HR, 0.41 (0.15–1.11) 0.08
Cho et al. [70] (2021) Korea LC Large-population database study 6,399/561 HR, 0.78 (0.64–0.96); TC Q2*
HR, 0.65 (0.51–0.83); TC Q3
HR, 0.30 (0.21–0.44); TC Q4
HR, 0.78 (0.58–1.05); TG Q2
HR, 0.69 (0.44–1.07); TG Q3
HR, 0.47 (0.23–0.94); TG Q4
HR, 0.77 (0.63–0.94); LDL Q2
HR, 0.64 (0.51–0.81); LDL Q3
HR, 0.29 (0.21–0.40); LDL Q4
HR, 0.95 (0.75–1.20); HDL Q2
HR, 1.01 (0.80–1.28); HDL Q3
HR, 1.18 (0.94–1.49); HDL Q4
Viral hepatitis 86,368/806 HR, 0.69 (0.58–0.82); TC Q2*
HR, 0.50 (0.41–0.61); TC Q3
HR, 0.31 (0.24–0.39); TC Q4
HR, 0.78 (0.58–1.05); TG Q2
HR, 0.69 (0.44–1.07); TG Q3
HR, 0.47 (0.23–0.94); TG Q4
HR, 0.77 (0.63–0.94); LDL Q2
HR, 0.64 (0.51–0.81); LDL Q3
HR, 0.29 (0.21–0.40); LDL Q4
HR, 0.95 (0.75–1.20); HDL Q2
HR, 1.01 (0.80–1.28); HDL Q3
HR, 1.18 (0.94–1.49); HDL Q4
Goh et al. [75] (2020) Korea HBV Retrospective cohort study 7,713/702 HR, 0.77 (0.63–0.95); high TC 0.01
Wu et al. [76] (2014) Taiwan HBV Population-based cohort study 43,190/5,446 HR, 0.87 (0.61–1.23) 0.43
Tan et al. [77] (2019) Taiwan HBV Population-based cohort study 6,564/89 HR, 0.67 (0.39–1.16); high TG 0.15
HR, 0.99 (0.54–1.83); low HDL 0.98
Choe et al. [86] (2021) Korea HBV Large-population database study 1,504,880/29,412 HR, 0.46 (0.45-0.47); high TG
HR, 0.68 (0.66-0.70); low HDL
Kanwal et al. [25] (2020) USA NAFLD Retrospective cohort study 271,906/253 HR, 1.31 (0.84–2.04); high TG and/or low HDL-cholesterol

Table 5.

Association between statin use and hepatocellular carcinoma

Table 5.

Study Region Etiology Study design Patients/HCC cases Result (95% CI) _P_-value
Campbell et al. [17] (2020) HBV Meta-analysis 40 studies (536,456) HR, 0.36 (0.19–0.68) to HR, 0.81 (0.73–0.90)
Chen et al. [33] (2015) Taiwan HBV Population-based cohort study 71,824/1,735 HR, 0.34 (0.27–0.42) <0.001
Goh et al. [75] (2020) Korea HBV Retrospective cohort study 7,713/702 HR, 0.36 (0.19–0.68) 0.002
Wu et al. [76] (2014) Taiwan HBV Population-based cohort study 43,190/5,446 HR, 0.55 (0.47–0.63) <0.001
Tsan et al. [87] (2012) Taiwan HBV Population-based cohort study 33,413/1,021 HR, 0.47 (0.36–0.61) <0.001
Hsiang et al. [88] (2015) China HBV Cohort study 73,499/6,883 HR, 0.68 (0.48–0.97) 0.033
Tsan et al. [89] (2013) Taiwan HCV Population-based cohort study 260,864/27,883 HR, 0.53 (0.49–0.58) <0.001
Butt et al. [90] (2015) USA HCV Cohort study 7,248 HR, 0.52 (0.35–0.78) <0.001
Mohanty et al. [91] (2016) USA HCV Retrospective cohort study 2,747/173 HR, 0.42 (0.27–0.64) <0.001
Simon et al. [92] (2016) USA HCV Cohort study 9,135/239 HR, 0.85 (0.47–1.53) 0.580
Li et al. [93] (2020) HBV, HCV Meta-analysis 13 studies (519,707) RR, 0.54 (0.44–0.66) <0.001
Wong et al. [94] (2021) HBV, HCV, LC Meta-analysis 13 studies (1,742,260) HR, 0.57 (0.52–0.62) 0.060
Lee et al. [36] (2017) Taiwan NAFLD Population-based cohort study 18,080/41 HR, 0.29 (0.12–0.68) 0.005

Abbreviations

NAFLD

non-alcoholic fatty liver disease

NHIS-HEALS

National Health Insurance Service-Health Screening Cohort

REFERENCES

Citations

Citations to this article as recorded by Crossref logo