Differential risk of 23 site-specific incident cancers and cancer-related mortality among patients with metabolic dysfunction-associated fatty liver disease: a population-based cohort study with 9.7 million Korean subjects

Abstract

Introduction: Although an association between metabolic dysfunction-associated fatty liver disease (MAFLD) and cardiovascular disease or overall mortality has been reported, it is unclear whether there is an association between MAFLD and cancer incidence or mortality. We aimed to investigate the differential risk of all- and site-specific cancer incidence and mortality according to MAFLD subgroups categorized by additional etiologies of liver disease.

Methods: Using the Korean National Health Insurance Service database, we stratified the participants into three groups: (1) single-etiology MAFLD (S-MAFLD) or MAFLD of pure metabolic origin; (2) mixed-etiology MAFLD (M-MAFLD) or MAFLD with additional etiological factor(s) (i.e., concomitant liver diseases and/or heavy alcohol consumption); and (3) non-MAFLD. Hepatic steatosis and fibrosis were defined using the fatty liver index and the BARD score, respectively. Cox proportional hazards regression was performed to estimate the risk of cancer events.

Results: Among the 9,718,182 participants, the prevalence of S-MAFLD and M-MAFLD was 29.2% and 6.7%, respectively. During the median 8.3 years of follow-up, 510,330 (5.3%) individuals were newly diagnosed with cancer, and 122,774 (1.3%) cancer-related deaths occurred among the entire cohort. Compared with the non-MAFLD group, the risk of all-cancer incidence and mortality was slightly higher among patients in the S-MAFLD group (incidence, adjusted hazard ratio [aHR] = 1.03; 95% confidence interval [CI]: 1.02-1.04; mortality, aHR = 1.06; 95% CI: 1.04-1.08) and highest among patients with M-MAFLD group (incidence, aHR = 1.31; 95% CI: 1.29-1.32; mortality, aHR = 1.45; 95% CI: 1.42-1.48, respectively). The M-MAFLD with fibrosis group (BARD score ≥ 2) showed the highest relative risk of all-cancer incidence (aHR = 1.38, 95% CI = 1.36-1.39), followed by the M-MAFLD without fibrosis group (aHR = 1.09, 95% CI = 1.06-1.11). Similar trends were observed for cancer-related mortality.

Conclusions: MAFLD classification, by applying additional etiologies other than pure metabolic origin, can be used to identify a subgroup of patients with poor cancer-related outcomes.

Keywords: cancer; etiology; metabolic; mortality; steatosis.

FIGURE 1

Flow chart of the selection of the study population.

FIGURE 2

Association between MAFLD subtypes and site‐specific cancer incidence. Incidence rates were presented per 100,000 person‐years. The hazard ratio and its 95% confidence interval were derived from the Cox regression model. Adjusted for age, sex, income, smoking, exercise, estimated glomerular filtration rate, Charlson comorbidity index score, waist circumference, glucose, total cholesterol and systolic blood pressure. Abbreviations: MAFLD, metabolic‐associated fatty liver disease; S‐MAFLD, single etiology MAFLD; M‐MAFLD, mixed etiology MAFLD; HR, hazard ratio; CI, confidence interval. *P for trend was estimated across MAFLD categories (modeled continuously), compared to non‐MAFLD; for details, see Methods.

FIGURE 3

Association between MAFLD subtypes and site‐specific cancer mortality. Mortality rates were presented per 100,000 person‐years. The hazard ratio and its 95% confidence interval were derived from the Cox regression model. Adjusted for age, sex, income, smoking, exercise, estimated glomerular filtration rate, Charlson comorbidity index score, waist circumference, glucose, total cholesterol and systolic blood pressure. Abbreviations: MAFLD, metabolic‐associated fatty liver disease; S‐MAFLD, single etiology MAFLD; M‐MAFLD, mixed etiology MAFLD; HR, hazard ratio; CI, confidence interval. *P for trend was estimated across MAFLD categories (modeled continuously), compared to non‐MAFLD; for details, see Methods.