Abstract
Background and Aims
Metabolic (dysfunction)‐associated steatotic liver disease (MASLD) has been linked to an increased risk of extrahepatic cancers. Whether MASLD or metabolic (dysfunction)‐associated fatty liver disease (MAFLD) is associated with an increased risk of extrahepatic cancer incidence or death in Australia is unknown.
Methods
This was a study from the Crossroads 1 cohort, enrolling randomly selected adults from regional Australia between 2001 and 2003 to partake in a comprehensive evaluation of health. MASLD is defined as fatty liver index (FLI) ≥60, one metabolic risk factor and no alternate liver disease, while MAFLD was defined as FLI ≥60 with overweight/obesity, type 2 diabetes mellitus (T2DM) and/or metabolic dysfunction. Longitudinal outcomes linked through cancer registry, hospital admission and death registry datasets and defined according to the International Statistical Classification of Diseases and Related Health Problems, 10th revision codes.
Results
In total, 1324 and 1444 participants respectively are included in the MASLD and MAFLD analyses. Prevalence was 35.4% (MASLD) and 40.7% (MAFLD). Median follow‐up time was 19.7 years (interquartile range 19.1–20.1). Neither form of fatty liver disease (FLD) was associated with an increased risk of death related to extrahepatic cancer compared to those without FLD. Both forms of FLD were associated with an increased incidence of colorectal cancer (MASLD adjusted sub‐hazard ratio (sHR) 2.90, 95% CI 1.17–7.17; MAFLD: sHR 3.15, 95% CI 1.43–6.96) following adjustment for confounding factors. This was primarily influenced by a difference seen among females (sHR 3.89–4.32).
Conclusion
Neither MASLD nor MAFLD was associated with an increased risk of extrahepatic cancer‐related death, but they were associated with an increased risk of colorectal cancer incidence, particularly among females.
Keywords: MASLD, MAFLD, extrahepatic cancer, cancer mortality, cancer incidence
Introduction
Cancer is the second leading cause of death globally, with a complex pathophysiology related to both modifiable and non‐modifiable risk factors. 1 Behavioural risk factors that play a pivotal role in carcinogenesis include smoking, alcohol consumption, and dietary and lifestyle practices leading to obesity and insulin resistance. 1 , 2
Metabolic (dysfunction)‐associated steatotic liver disease (MASLD) and the related metabolic (dysfunction)‐associated fatty liver disease (MAFLD) are considered the hepatic manifestation of the metabolic syndrome (MetS). 3 Longitudinal outcome studies consistently demonstrate that deaths occur primarily due to extrahepatic cancer and cardiovascular disease (CVD) in those with fatty liver disease (FLD), markedly outweighing liver‐related mortality. 4 , 5 However, whether MASLD and/or MAFLD independently increases the risk of death from extrahepatic cancer remains unclear due to conflicting data. 6 , 7 , 8
Furthermore, the shift in nomenclature and accompanying diagnostic criteria may lead to differences in outcome(s) between FLD conditions. While MASLD is a diagnosis of exclusion, MAFLD is an inclusive one permitting the co‐existence of any additional cause of liver disease. 9 Notably, in higher‐income countries, a large proportion of the difference in prevalence between MASLD and MAFLD is accounted for by concurrent alcohol‐related liver disease in the latter. 10 This may accentuate the risk of extrahepatic cancer, given the known carcinogenic effect of ethanol.
To date, no study from Australia has evaluated the role of MASLD and/or MAFLD on incidence of or mortality from extrahepatic cancer. As such, we sought to determine whether either form of FLD was an independent risk factor for death from extrahepatic cancer and to establish the incidence of select solid organ extrahepatic cancers among those with FLD.
Methods
The Crossroads 1 study was a cross‐sectional survey‐based study of health and disease on community‐dwelling participants randomly selected from government household lists conducted between 2001 and 2003 in a major regional part of Victoria, Australia. It included a nested clinic sub‐study – the Crossroads Undiagnosed Disease Study (CUDS) – whose methodology was previously published in detail. 11 In summary, a randomly selected adult (aged ≥18 years) from households participating in Crossroads 1 was invited to partake in a clinic visit whereby comprehensive information on demographic details, anthropometry, detailed questionnaires on health, diet and lifestyle behavioural practices, and laboratory tests including full blood examination, liver and renal chemistries, and fasting lipid profile were collected. In addition, all participants without known diabetes undertook an oral glucose tolerance test. In total, n = 1454 enrolled into CUDS.
Definitions
FLD is defined by fatty liver index (FLI) ≥60. 12 MASLD was defined as per the original consensus criteria 13 : those with FLD in the absence of an alternative cause of liver disease including excessive alcohol consumption (≥210 g/week in men, ≥140 g/week in women) derived from a self‐reported lifestyle questionnaire and with at least one metabolic risk factor. MAFLD was defined as those with FLD in the presence of (1) overweight/obesity (body mass index (BMI) ≥23 kg/m2 in Asians or ≥25 kg/m2 in all other ethnicities), (2) type 2 diabetes mellitus (T2DM) and/or (3) ‘metabolic dysfunction’ as proposed in the original criteria, 9 except for high‐sensitivity C‐reactive protein or homeostatic model assessment for insulin resistance, as these were not measured.
T2DM was defined as per original CUDS coding and in accordance with the World Health Organisation and American Diabetes Association criteria. 11 Baseline prevalent cancer was determined through self‐report from the health questionnaire. Diet adequacy was defined as the consumption of sufficient serves of fruits and vegetables daily as per the Australian dietary guidelines (i.e., ≥4–5 serves of vegetables per day, ≥2 serves of fruit per day). 14 Those who answered yes to being vegetarian or vegan on diet questionnaire were considered as having no red meat intake, as there were no specific questions on amount or frequency of red meat consumption.
Outcomes
The primary aim of the study was to investigate the association between MASLD and MAFLD and extrahepatic cancer‐related mortality, defined as the primary cause of death. The secondary aims were to establish the specific cause of extrahepatic cancer‐related mortality and examine the incidence rates (non‐fatal and fatal) of specific extrahepatic cancers.
Data linkage
Longitudinal outcome data were acquired through linkage of CUDS with emergency presentation, hospital admission, and death and cancer registry datasets overseen by the Centre for Victorian Data Linkage, as previously described. 15 This is the government custodian for healthcare‐associated datasets within the state of Victoria, Australia, and collates notifications from all public and private hospitals in the state and cancer screening registries, allowing for comprehensive coverage of linked data within the state. 16 , 17 , 18 Underlying cause of death (UCOD) data were additionally obtained from the Australian Bureau of Statistics. Emergency presentation and death registry data were available from participant enrolment into CUDS, while hospital admission and cancer registry data were available from inception of those datasets (1 July 2007 and 1 January 2008 respectively). All long‐term outcomes were coded according to the International Statistical Classification of Diseases and Related Health Problems, 10th revision (ICD‐10) manual. Participants were considered alive until death or censored at final date for data linkage (31 October 2022).
Extrahepatic cancer‐related death was defined as UCOD C00‐C97 (except C22.0–C22.1) and D37‐D48. Deaths from hepatocellular carcinoma (C22.0) and intra‐hepatic cholangiocarcinoma (C22.1) were considered liver‐related deaths. Subcategories of extrahepatic cancer codes are presented in Table S1. 19 Incident cancer was considered as those occurring after enrolment into CUDS (i.e. those with self‐reported cancer at baseline were excluded when calculating incidence rates).
Statistical analysis
Continuous variables are presented as mean (± standard deviation) or median (interquartile range (IQR)) and between‐group differences investigated with independent student t‐test (or Welch's t‐test if equality of variance was not assumed on Levene's test) or Mann–Whitney U test, after normality assessment. Categorical variables are presented as frequency with percentage and Fisher's exact test or Pearson chi‐squared test applied for hypothesis testing, as appropriate. Incidence rates are presented per 100 000 person‐years (PY). Cox proportional hazards regression was used for univariate and multivariate prediction models and reported as hazard ratios (HRs) with 95% confidence intervals (CIs). Multivariate models were built a priori based on existing literature 1 , 2 , 20 and included age and sex (Model 1); ethnicity and birth location (dichotomised as Australian‐born or not) (Model 2, including Model 1 covariates); smoking (non‐smoker as reference), diet adequacy, red meat intake (yes/no) and excessive alcohol consumption (yes/no) (Model 3, including Model 2 covariates); and T2DM and aspirin use (Model 4, including Model 3 covariates). The relationship between FLD and incident cancer was also investigated using Cox regression models, using the fully adjusted Model 4, after excluding people with a previous diagnosis of the cancer type being investigated as the dependent outcome and considering non‐cancer death as a competing risk. For example, people with prevalent colorectal cancer at baseline were excluded from the incident colorectal cancer analysis. Sex‐specific malignancies (i.e. prostate, breast and gynaecological) were reported for each biological sex, with rates reported as per 100 000 male‐years or female‐years. Sub‐group analysis was conducted to investigate extrahepatic cancer mortality and incidence rates among sex, adjusted for Model 4 (except sex). Sensitivity analyses were undertaken comparing those with FLD to those with FLI <30 (i.e. control group definitively without FLD) and to calculate cancer incidence rates after excluding participants with prior diagnosis of any extrahepatic malignancy. Two‐tailed P‐value <0.05 is considered statistically significant. All statistical analyses were conducted using StataNow/SE version 18.5 for Windows (StataCorp, Texas, USA).
Ethics
The original Crossroads study was approved by the Goulburn Valley Health Human Research Ethics Committee (GCH‐3/99), while the current follow‐up study was approved by the Alfred Health Ethics Committee (project 310/22).
Results
After applying relevant exclusions, there were 1324 and 1444 participants remaining for MASLD and MAFLD analyses respectively (Fig. S1). The prevalence of MASLD was 35.4% (n = 469), and that of MAFLD was 40.7% (n = 588). Compared to those without FLD, people with either FLD were significantly older, more likely to be men, have MetS and each of its components, more likely to be smokers or ex‐smokers, and prescribed lipid‐lowering and diabetes medications (Table 1). Baseline prevalent extrahepatic cancer was essentially the same between those with and without FLD (Table 1).
Table 1.
Baseline demographic, clinical and laboratory covariates, including non‐invasive tests, according to fatty liver disease diagnosis
| Variable | No MASLD (n = 855) | MASLD (n = 469) | P‐value | No MAFLD (n = 856) | MAFLD (n = 588) | P‐value |
|---|---|---|---|---|---|---|
| Age, years | 50 (38.3–65.2) | 53.7 (42.9–65.8) | 0.001 | 50 (38.3–65.2) | 54.4 (44.3–66.5) | <0.001 |
| Male biologic sex | 294 (34.4) | 265 (56.5) | <0.001 | 295 (34.5) | 338 (57.5) | <0.001 |
| Ethnic background | 0.78 | 0.58 | ||||
| White | 833 (97.4) | 456/468 (97.4) | 834 (97.4) | 573/586 (97.8) | ||
| Asian | 10 (1.2) | 5/468 (1.1) | 10 (1.2) | 5/586 (0.9) | ||
| Indigenous | 4 (0.5) | 4/468 (0.9) | 4 (0.5) | 5/586 (0.9) | ||
| Other | 8 (0.9) | 3/468 (0.6) | 8 (0.9) | 3/586 (0.5) | ||
| Australian born | 752/854 (88.1) | 418 (89.1) | 0.56 | 753/855 (88.1) | 522 (88.8) | 0.68 |
| BMI, kg/m2 | 24.98 (± 3.0) | 32.37 (± 5.0) | <0.001 | 24.98 (± 3.0) | 32.15 (± 4.9) | <0.001 |
| BMI, kg/m2 | <0.001 | <0.001 | ||||
| < 25 | 439 (51.4) | 9 (1.9) | 440 (51.4) | 14 (2.4) | ||
| 25 to <30 | 378 (44.2) | 160 (34.1) | 378 (44.2) | 206 (35.0) | ||
| ≥ 30 | 38 (4.4) | 300 (64.0) | 38 (4.4) | 368 (62.6) | ||
| Waist circumference, cm | 86.21 (± 9.7) | 106.98 (± 10.2) | <0.001 | 86.22 (± 9.7) | 107.00 (± 10.1) | <0.001 |
| Hypertension | 398/854 (46.6) | 333 (71.0) | <0.001 | 398/855 (46.6) | 427 (72.6) | <0.001 |
| Dyslipidaemia | 467 (54.6) | 372 (79.3) | <0.001 | 468 (54.7) | 461 (78.4) | <0.001 |
| Type 2 diabetes mellitus | 30 (3.5) | 70 (14.9) | <0.001 | 30 (3.5) | 88 (15.0) | <0.001 |
| Metabolic syndrome | 110/854 (12.9) | 296 (63.1) | <0.001 | 110/855 (12.9) | 375 (63.8) | <0.001 |
| Excessive alcohol consumption | 148 (17.3) | 0 (0) | <0.001 | 147 (17.2) | 97 (16.5) | 0.74 |
| Viral hepatitis | 26 (3.0) | 0 (0) | <0.001 | 26 (3.0) | 25 (4.3) | 0.22 |
| Smoking status | <0.001 | <0.001 | ||||
| Non‐smoker | 456/852 (53.5) | 217/468 (46.4) | 456/853 (53.5) | 246/587 (41.9) | ||
| Ex‐smoker | 242/852 (28.4) | 180/468 (38.5) | 243/853 (28.5) | 240/587 (40.9) | ||
| Current | 154/852 (18.1) |
71/468 (15.2) |
154/853 (18.1) | 101/587 (17.2) | ||
| Adequate diet | 188/852 (22.1) |
114//469 (24.3) |
0.35 |
188/853 (22.0) |
143/587 (24.4) |
0.30 |
| No red meat intake |
26/847 (3.1) |
7/467 (1.5) |
0.08 |
26/848 (3.1) |
8/586 (1.4) |
0.037 |
| Aspirin |
84/847 (9.9) |
61/466 (13.1) |
0.08 |
84/848 (9.9) |
74/584 (12.7) |
0.10 |
| Lipid‐lowering medication | 44 (5.2) | 37 (7.9) | 0.046 | 44 (5.1) | 49 (8.3) | 0.015 |
| Diabetes medication |
22/843 (2.6) |
45/466 (9.7) |
<0.001 |
22/844 (2.6) |
52/584 (8.9) |
<0.001 |
| Baseline prevalent cancer | ||||||
| Digestive | 9/847 (1.1) |
4/467 (0.9) |
0.72 |
9/848 (1.1) |
5/585 (0.9) |
0.70 |
| Colorectal |
7/847 (0.8) |
3/467 (0.6) |
1.0 |
7/848 (0.8) |
4/585 (0.7) |
1.0 |
| Lung |
1/847 (0.1) |
0 (0) | ‐ |
1/848 (0.8) |
1/585 (0.2) |
1.0 |
| Prostate |
2/294 (0.7) |
7/265 (2.6) |
0.09 |
2/295 (0.7) |
8/338 (2.4) |
0.11 |
| Breast |
10/561 (1.8) |
4/204 (2.0) |
1.0 |
10/561 (1.8) |
4/250 (1.6) |
1.0 |
| Gynaecological |
10/561 (1.8) |
10/204 (4.9) |
0.017 |
10/561 (1.8) |
10/250 (4.0) |
0.06 |
| Urinary tract |
2/847 (0.2) |
3/467 (0.6) |
0.35 |
2/848 (0.2) |
5/585 (0.9) |
0.13 |
Note: Continuous variables presented as mean (± standard deviation) or median (interquartile range); categorical variables presented as frequency (%).
Abbreviations: BMI = body mass index; MASLD = metabolic (dysfunction)‐associated steatotic liver disease; MAFLD = metabolic (dysfunction)‐associated fatty liver disease.
Mortality
Over 24 112 and 26 111 PY follow‐up time for MASLD and MAFLD cohorts respectively (median follow‐up time 19.7 years (IQR 19.1–20.1)), there were a total of 298 deaths (MASLD n = 104 (22.2%), MAFLD n = 143 (24.3%), no FLD n = 155 (18.1%)), with 98 deaths attributed to extrahepatic cancer (MASLD n = 36 (7.7%), MAFLD n = 49 (8.3%), no FLD n = 49 (5.7%)). The most common specific causes of extrahepatic cancer‐related death in participants with either FLD were digestive tract and prostate cancer, along with lung cancer in those with MAFLD, while in those without FLD they were digestive tract and lung cancer (Table 2).
Table 2.
Specific causes of extrahepatic cancer‐related mortality according to fatty liver disease status
| Cancer type | No MASLD (n = 855) | MASLD (n = 469) | No MASLD (n = 855) | MASLD (n = 469) | Model 4 aHR | No MAFLD (n = 856) | MAFLD (n = 588) | No MAFLD (n = 856) | MAFLD (n = 588) | Model 4 aHR |
|---|---|---|---|---|---|---|---|---|---|---|
| Digestive tract | 10 (1.2) | 9 (1.9) |
63.8 (34.3–118.5) |
106.7 (55.5–205.1) |
1.27 (0.45–3.61) |
10 (1.2) |
13 (2.2) |
63.7 (34.3–118.4) |
124.9 (72.5–215.0) |
1.70 (0.69–4.19) |
| Colorectal |
4 (0.5) |
5 (1.0) |
25.5 (9.6–68.0) |
59.3 (24.7–142.5) |
2.80 (0.69–11.41) |
4 (0.5) |
7 (1.2) |
25.5 (9.6–67.9) |
67.2 (32.1–141.0) |
3.22 (0.88–11.81) |
| Lung |
12 (1.4) |
4 (0.9) |
76.5 (43.5–134.8) |
47.4 (17.8–126.4) |
0.43 (0.13–1.38) |
12 (1.4) |
10 (1.7) |
76.4 (43.4–134.6) |
96.0 (51.7–178.5) |
0.98 (0.41–2.37) |
| Prostate* |
1/294 (0.3) |
6/265 (2.3) |
18.8 (2.6–133.5) |
128.0 (57.5–284.9) |
6.07 (0.67–55.21) |
1/295 (0.3) |
8/338 (2.4) |
18.7 (2.6–132.9) |
136.4 (68.2–272.8) |
9.27 (1.09–78.83) |
| Breast* |
2/561 (0.4) |
2/204 (1.0) |
19.3 (4.8–77.2) |
53.4 (13.4–213.5) |
2.16 (0.27–17.30) |
2/561 (0.4) |
2/250 (0.8) |
19.3 (4.8–77.2) |
44.0 (11.0–175.9) |
2.11 (0.26–17.02) |
| Gynaecological* |
2/561 (0.4) |
2/204 (1.0) |
19.3 (4.8–77.2) |
53.4 (13.4–213.5) |
1.71 (0.21–14.17) |
2/561 (0.4) |
2/250 (0.8) |
19.3 (4.8–77.2) |
44.0 (11.0–175.9) |
1.61 (0.19–13.54) |
| Urinary tract |
4 (0.5) |
0 (0) |
25.5 (9.6–68.0) |
‐ | ‐ |
4 (0.5) |
1 (0.2) |
25.5 (9.6–67.9) |
9.6 (1.4–68.2) |
0.31 (0.03–2.86) |
| Other |
18 (2.1) |
13 (2.8) |
114.8 (72.3–182.2) |
154.2 (89.5–265.5) |
1.19 (0.54–2.63) |
18 (2.1) |
13 (2.2) |
114.7 (72.2–182.0) |
124.9 (72.5–215.0) |
0.95 (0.44–2.04) |
Note: Mortality rates presented as crude (n, %) and per 100 000 person‐years (*or male‐years/female‐years) (95% confidence interval). Model 4 presents adjusted hazard ratios (aHR) presented with 95% confidence interval for fatty liver disease adjusted for age, gender, ethnicity, birth location, smoking status, diet adequacy, red meat intake, excessive alcohol consumption, type 2 diabetes and aspirin use.
Abbreviations: MASLD = metabolic (dysfunction)‐associated steatotic liver disease; MAFLD = metabolic (dysfunction)‐associated fatty liver disease.
In Cox regression analysis for all‐cause extrahepatic cancer‐related death, MAFLD (HR 1.52, 95% CI 1.02–2.26) but not MASLD (HR 1.37, 95% CI 0.89–2.11) increased the risk of death on univariate analysis; however, MAFLD did not remain an independent risk factor for death on any multivariate model (Table 3). MAFLD participants had a significantly increased risk of death from prostate cancer compared to those without MAFLD; however, there was no significant difference between groups for any other cancer‐specific death (Table 2). There was no difference seen on sensitivity analysis comparing those with FLD to a control group consisting of only those with FLI <30 (Table S2), nor on subgroup analysis examining each sex (Table S3).
Table 3.
Univariate and multivariate Cox regression analysis exploring the relationship between fatty liver disease and extrahepatic cancer‐related death
| Model | MASLD | MAFLD |
|---|---|---|
| Univariate |
1.37 (0.89–2.11) |
1.52 (1.02–2.26) |
| Model 1 |
1.30 (0.84–2.02) |
1.42 (0.94–2.13) |
| Model 2 |
1.29 (0.83–2.01) |
1.42 (0.94–2.13) |
| Model 3 |
1.11 (0.70–1.75) |
1.34 (0.89–2.03) |
| Model 4 |
1.06 (0.66–1.69) |
1.29 (0.84–1.96) |
Note: Data presented as (adjusted) hazard ratios with 95% confidence intervals. Model 1 = adjusted for age and gender, Model 2 = Model 1 adjusted for ethnicity and birth location, Model 3 = Model 2 adjusted for smoking status, diet adequacy, red meat intake, and excessive alcohol consumption, Model 4 = Model 3 adjusted for type 2 diabetes and aspirin use.
Abbreviations: MASLD = metabolic (dysfunction)‐associated steatotic liver disease; MAFLD = metabolic (dysfunction)‐associated fatty liver disease.
Incidence
The most common incident cancers to occur in either form of FLD were digestive tract (primarily colorectal), breast and prostate cancer (Table 4). These were the most common incident extrahepatic cancers in those without FLD, along with lung cancer. After adjusting for relevant confounders using Model 4, those with either form of FLD had a higher risk for developing colorectal cancer versus those without FLD (MASLD: adjusted subhazard ratio (sHR) 2.90 (95% CI 1.17–7.17); MAFLD: 3.15 (95% CI 1.43–6.96)), while those with MAFLD had a higher risk of any digestive tract cancer (sHR 2.02 (95% CI 1.08–3.75)) (Table 4). Sensitivity analysis excluding participants with any prior extrahepatic cancer diagnosis did not alter the result (Table S4).
Table 4.
Incidence rates of extrahepatic cancer types according to fatty liver disease.
| Cancer type | No MASLD (n = 855) | MASLD (n = 469) | No MASLD (n = 855) | MASLD (n = 469) | Model 4 sHR | No MAFLD (n = 856) | MAFLD (n = 588) | No MAFLD (n = 856) | MAFLD (n = 588) | Model 4 sHR |
|---|---|---|---|---|---|---|---|---|---|---|
| Digestive tract | 20/846 (2.4) | 19/465 (4.1) |
128.8 (83.1–199.6) |
227.5 (145.1–356.6) |
1.65 (0.79–3.44) |
20/847 (2.4) |
30/583 (5.2) |
128.6 (83.0–199.3) |
290.9 (203.4–416.1) |
2.02 (1.08–3.75) |
| Colorectal |
10/848 (1.2) |
13/466 (2.8) |
64.3 (34.6–119.5) |
155.3 (90.2–267.4) |
2.90 (1.17–7.17) |
10/849 (1.2) |
20/584 (3.4) |
64.2 (34.5–119.3) |
193.6 (124.9–300.0) |
3.15 (1.43–6.96) |
| Lung |
19/854 (2.2) |
6/469 (1.3) |
121.3 (77.4–190.2) |
71.2 (32.0–158.4) |
0.57 (0.22–1.45) |
19/855 (2.2) |
12/587 (2.0) |
121.2 (77.3–190.0) |
115.3 (65.5–203.0) |
0.80 (0.39–1.63) |
| Prostate* |
14/292 (4.8) |
15/258 (5.8) |
265.2 (157.0–447.7) |
328.4 (198.0–544.7) |
1.32 (0.61–2.88) |
14/293 (4.8) |
21/330 (6.4) |
264.1 (156.4–446.0) |
366.9 (239.2–562.7) |
1.41 (0.72–2.73) |
| Breast* |
22/551 (4.0) |
9/200 (4.5) |
216.3 (142.4–328.5) |
245.0 (127.5–470.9) |
1.31 (0.56–3.10) |
22/551 (4.0) |
12/246 (4.9) |
216.3 (142.4–328.5) |
268.2 (152.3–472.2) |
1.17 (0.57–2.40) |
| Gynaecological* |
9/551 (1.6) |
3/194 (1.6) |
88.4 (46.0–169.9) |
83.9 (27.1–260.1) |
0.80 (0.18–3.55) |
9/551 (1.6) |
7/240 (2.9) |
88.4 (46.0–169.9) |
159.9 (76.2–335.4) |
1.74 (0.60–5.01) |
| Urinary tract |
7/853 (0.8) |
8/466 (1.7) |
44.7 (21.3–93.8) |
95.5 (47.7–190.9) |
1.45 (0.53–3.98) |
7/854 (0.8) |
10/583 (1.7) |
44.7 (21.3–93.7) |
96.9 (52.1–180.0) |
1.73 (0.66–4.53) |
Note: Incidence rates presented as crude (n/d, %) and per 100 000 person‐years (*or male‐years/female‐years) (95% confidence interval). Model 4 presents adjusted subhazard ratios (sHR) presented with 95% confidence interval for fatty liver disease adjusted for age, gender, ethnicity, birth location, smoking status, diet adequacy, red meat intake, excessive alcohol consumption, type 2 diabetes and aspirin use (gender not included in Model 4 for prostate, breast and gynaecological cancer as analysis conducted on men/women individually).
Abbreviations: MASLD = metabolic (dysfunction)‐associated steatotic liver disease; MAFLD = metabolic (dysfunction)‐associated fatty liver disease.
Women with either form of FLD had a higher risk of developing colorectal cancer after adjustment for confounding factors (MASLD: sHR 4.32 (95% CI 1.57–11.91); MAFLD: sHR 3.89 (95% CI 1.53–9.91)), while there were no differences seen among men (MASLD: sHR 1.22 (95% CI 0.25–5.82); MAFLD: sHR 2.29 (95% CI 0.51–10.35)). There were no other significant differences in incidence of extrahepatic cancer between those with and without FLD when stratifying by sex (data not shown). The differences in the risk factor profile between those with and without FLD was similar per sex; however, males with FLD were more likely to use aspirin than males without FLD, and females with FLD were more likely to have a diet adequate in fruits and vegetables but more likely to eat red meat than females without FLD (Supplementary Table 5).
Discussion
In this longitudinal study from a randomly selected community‐dwelling adult cohort in regional Australia, we have demonstrated that neither MASLD nor MAFLD independently increases the risk of death related to extrahepatic cancer. Incident extrahepatic cancers occur in similar frequency between those with and without FLD, with the notable exception being an increased incidence of colorectal cancer in those with either form of FLD. This difference was particularly seen among females.
While FLD – whether MASLD or MAFLD – is the pre‐eminent condition of the liver globally, with a rising prevalence related to the obesity epidemic, deaths related to liver disease pale in comparison to those attributed to cardiovascular disease and extrahepatic cancer. 4 , 5 Despite this, very few studies have specifically investigated the difference in mortality risk from extrahepatic cancer between those with and without FLD. One large, matched cohort study from Sweden including 10 568 biopsy‐proven adults with MASLD detected a 2.16‐fold increased risk of extrahepatic cancer mortality compared to matched controls, after controlling for relevant sociodemographic and clinical variables. 6 However, ~90% of the participants from this study were Nordic‐born, and thus there is poor generalisability to a multi‐ethnic population, such as Australia. What can be taken from this study is the large cohort (altogether >60 000 participants across cases and comparators) required to detect differences in this outcome. Another large (>300 000 subjects) Korean study demonstrated MASLD to be associated with a higher risk of cancer‐related mortality in women but not men, 7 while a smaller (~12 000 participants) cohort study from the US failed to detect any difference between those with and without MASLD. 8 Hence, it may be that our finding of no difference in extrahepatic cancer‐related mortality relates to a type II error due to insufficient study power. As such, whether FLD – either defined through MASLD or MAFLD criteria – independently increases the risk of extrahepatic cancer requires further investigation in other large cohorts from differing geographic regions or with multi‐ethnic cohorts.
Two large meta‐analyses explored the association between MASLD and incident extrahepatic cancer. 21 , 22 In both studies, the pooled analysis revealed MASLD was associated with an increased risk of several cancers of the digestive tract, thyroid, lung, breast, urinary tract and female genital tract, with HRs ranging from 1.30 to 2.60 for various cancer types. Once more, the applicability of these findings is limited by most included studies originating from South East Asian countries (9/10 studies in one meta‐analysis, 21 13/18 in the other 22 ) and the majority of the included studies were deemed to be at greater than low risk of bias. In our study, we were only able to detect a significant difference in adjusted rate of incident colorectal cancer, which may again relate to issues with the study power when assessing other cancer streams. The magnitude of effect was similar between MAFLD and MASLD. Whether MAFLD is more hazardous than MASLD on extrahepatic cancer risk due to the coexistence of liver disease, particularly excessive alcohol consumption and its associated behavioural risk factors, remains unknown. There is limited literature on the difference between MASLD and MAFLD on incident extrahepatic cancer. A single study from Korea demonstrated MAFLD only to carry the highest risk for colorectal cancer, with lower risk for those with concordant MASLD‐MAFLD and the lowest risk in those with MASLD only compared to non‐FLD controls, 23 while another study from China suggested those with concordant MASLD‐MAFLD had the highest risk of extrahepatic cancer, above those with MAFLD only. 24 Thus, further research is warranted to establish whether a true difference in extrahepatic cancer incidence exists between those with MASLD versus MAFLD. This ideally requires large cohorts sufficiently powered to detect such differences, if they are apparent at all, and to explore the underlying reasons for the difference. This would have public health implications in targeting high‐risk groups for surveillance programmes and possible lifestyle interventions to curb the risk, as well as lending weight to support the use of one term over another when describing FLD.
Our finding that females but not males had a higher incidence of colorectal cancer requires particular attention. We detected a three‐fold increased risk when comparing MASLD to no FLD and MAFLD to no FLD among females. Once again, this is an understudied area. In exploring differences in biological sex on clinical outcomes among those with MASLD, Zhou et al. demonstrated women to have a 44% higher risk than men for incident extrahepatic cancer upon meta‐analysis of four included studies. 25 Individual cancer streams were not investigated in this meta‐analysis but were in two of the included studies 26 , 27 and another study not included within the meta‐analysis, 28 with discordant findings to ours for incident colorectal cancer. In these studies, from Sweden, Korea and the US, the covariates in the multivariable analyses were vastly different to our models, with fewer behavioural and lifestyle risk factors included. Biological sex can influence outcome risk through many differing mechanisms, from genetic and hormonal differences to different patterns in lifestyle behaviours and environmental exposures. 1 Most studies concentrate on the ‘exposome’ between each biological sex as a mechanism for difference in colorectal cancer incidence, with men typically having a greater propensity for central obesity and a diet higher in red meat consumption and being more likely to consume alcohol and smoke, all risk factors for cancer incidence. 29 When investigating sex hormone influence on colorectal cancer in the non‐FLD population, oestrogen is thought to be a protective factor, with an attenuated influence of progesterone and testosterone on colorectal cancer risk. 30 In our cohort, females with FLD were more likely to eat red meat than females without FLD, while males with FLD were more likely to use aspirin than males without FLD. This may account for the difference in risk of incident colorectal cancer we detected between women and men, given the known links between red meat intake and colorectal cancer and the emerging chemoprophylactic role of aspirin in reducing colorectal cancer incidence. 31 Alternatively, there may be other unmeasured confounders (including differences in sex hormone levels) to explain the difference or the interaction between gender and measured risk factors may play a role in accentuating or attenuating risk. Finally, we cannot rule out a type 1 error in our finding, given this was a subgroup analysis. Further research is required to establish whether a true difference exists in men and women with FLD on colorectal cancer incidence and whether altered sex hormone profiles in those with FLD play a role in influencing this risk.
The strengths of this study include a randomly selected cohort from a community‐based cohort, the long follow‐up time to allow for detection of relevant outcomes and investigation of behavioural risk factors in multivariate models. While we have for the first time explored the influence of FLD – either MASLD or MAFLD – on extrahepatic cancer incidence and mortality, there are several limitations to the current study that require acknowledgement. Firstly, the current study is likely underpowered to detect significant differences, which likely accounts for the wide confidence intervals obtained. While we were able to adjust for diabetes, we were not able to adjust for obesity or waist circumference due to the risk of collinearity with the exposure variable of interest given these are both components of FLI, which was used to define FLD. Obesity itself is linked with several cancers, both hormone‐sensitive and hormone‐insensitive, and this itself may have contributed to the few differences detected. In addition, inherent in all linkage studies is the risk of miscoding or non‐coding. Twenty per cent of study participants were not linked, but the specific reason for this could not be established but ultimately relates to not utilising health care services within the state of Victoria, Australia, during the study period. Outcomes may also have been missed in the early follow‐up period when the cancer registry was not established. Finally, our study cohort was primarily from a single region of Australia, and whether this can be generalised to other geographic areas, particularly metropolitan populations, is uncertain.
Conclusions
Neither MASLD nor MAFLD appears to independently increase the risk of extrahepatic cancer‐related death, but they do confer an increased risk of incident colorectal cancer, particularly among women.
Supporting information
Figure S1. Study flow diagram.
Table S1. ICD‐10 codes utilised to define cancer types.
Table S2. Sensitivity analysis exploring relationship between FLD and extrahepatic cancer‐related death, using participants with FLI <30 as the control group.
Table S3. Subgroup analysis investigating relationship between FLD and extrahepatic cancer‐related death, according to gender.
Table S4. Sensitivity analysis investigating specific cause of incident extrahepatic cancer according to FLD status after excluding participants with prior diagnosis of extrahepatic cancer at baseline.
Table S5. Differences in risk and protective factors between those with and those without FLD among women and men.
Acknowledgements
The authors acknowledge the Australian Government Department of Health and Aged Care Rural Health Multidisciplinary Training Programme. We thank the Baker International Diabetes Institute for their support during Crossroads 1. We would also like to acknowledge the Victorian Department of Health as the source of Victorian Admitted Episodes Dataset (VAED), Victorian Emergency Minimum Dataset (VEMD), Victorian Integrated Non‐admitted Health (VINAH) dataset, Victorian Cancer Registry (VCR) and Victorian Cost Data Collection (VCDC) dataset for this study, and the Centre for Victorian Data Linkage (Victorian Department of Health) for provision of data linkage. The Australian Bureau of Statistics provided data for underlying cause of death. This research is supported by an Australian Government Research Training Program (RTP) scholarship.
David Simmons and Stuart K. Roberts are co‐senior authors for this manuscript.
Funding: None.
Conflict of interest: None.
Data availability statement
The data that support the findings of this study are available from CVDL. Restrictions apply to the availability of these data, which were used under license for this study. Data are available from the author(s) with the permission of CVDL.
References
- 1. Collaborators GBDCRF. The global burden of cancer attributable to risk factors, 2010‐19: a systematic analysis for the global burden of disease study 2019. Lancet 2022; 400: 563–591. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Whiteman DC, Webb PM, Green AC, Neale RE, Fritschi L, Bain CJ et al. Cancers in Australia in 2010 attributable to modifiable factors: summary and conclusions. Aust N Z J Public Health 2015; 39: 477–484. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Fu CE, Yong JN, Ng CH, Nah B, Chew NWS, Chin YH et al. The prognostic value of including non‐alcoholic fatty liver disease in the definition of metabolic syndrome. Aliment Pharmacol Ther 2023; 57: 979–987. [DOI] [PubMed] [Google Scholar]
- 4. Adams LA, Lymp JF, St Sauver J, Sanderson SO, Lindor KD, Feldstein A et al. The natural history of nonalcoholic fatty liver disease: a population‐based cohort study. Gastroenterology 2005; 129: 113–121. [DOI] [PubMed] [Google Scholar]
- 5. Lazo M, Hernaez R, Bonekamp S, Kamel IR, Brancati FL, Guallar E et al. Non‐alcoholic fatty liver disease and mortality among US adults: prospective cohort study. BMJ 2011; 343: d6891. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Simon TG, Roelstraete B, Khalili H, Hagstrom H, Ludvigsson JF. Mortality in biopsy‐confirmed nonalcoholic fatty liver disease: results from a nationwide cohort. Gut 2021; 70: 1375–1382. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Hwang YC, Ahn HY, Park SW, Park CY. Nonalcoholic fatty liver disease associates with increased overall mortality and death from cancer, cardiovascular disease, and liver disease in women but not men. Clin Gastroenterol Hepatol 2018; 16: 1131–1137. [DOI] [PubMed] [Google Scholar]
- 8. Alvarez CS, Graubard BI, Thistle JE, Petrick JL, McGlynn KA. Attributable fractions of nonalcoholic fatty liver disease for mortality in the United States: results from the third National Health and nutrition examination survey with 27 years of follow‐up. Hepatology 2020; 72: 430–440. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Eslam M, Newsome PN, Sarin SK, Anstee QM, Targher G, Romero‐Gomez M et al. A new definition for metabolic dysfunction‐associated fatty liver disease: An international expert consensus statement. J Hepatol 2020; 73: 202–209. [DOI] [PubMed] [Google Scholar]
- 10. Vaz K, Clayton‐Chubb D, Majeed A, Lubel J, Simmons D, Kemp W et al. Current understanding and future perspectives on the impact of changing NAFLD to MAFLD on global epidemiology and clinical outcomes. Hepatol Int 2023; 17: 1082–1097. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Simmons D, McKenzie A, Eaton S, Shaw J, Zimmet P. Prevalence of diabetes in rural Victoria. Diabetes Res Clin Pract 2005; 70: 287–290. [DOI] [PubMed] [Google Scholar]
- 12. Bedogni G, Bellentani S, Miglioli L, Masutti F, Passalacqua M, Castiglione A et al. The fatty liver index: a simple and accurate predictor of hepatic steatosis in the general population. BMC Gastroenterol 2006; 6: 33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Rinella ME, Lazarus JV, Ratziu V, Francque SM, Sanyal AJ, Kanwal F et al. A multi‐society Delphi consensus statement on new fatty liver disease nomenclature. J Hepatol 2023; 79: E93–E94. [DOI] [PubMed] [Google Scholar]
- 14. NHMRC . Australian dietary guidelines 2016. [cited 2022 12 December]. Available from URL: https://www.eatforhealth.gov.au.
- 15. Vaz K, Kemp W, Majeed A, Lubel J, Magliano DJ, Glenister KM et al. NAFLD and MAFLD independently increase the risk of major adverse cardiovascular events (MACE): a 20‐year longitudinal follow‐up study from regional Australia. Hepatol Int 2024; 18: 1135–1143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Canevari JT, Cheng AC, Wu L, Rowe SL, Wollersheim DE, West D et al. The relative effectiveness of three and four doses of COVID‐19 vaccine in Victoria, Australia: a data linkage study. Vaccine 2024; 42: 53–58. [DOI] [PubMed] [Google Scholar]
- 17. Clapperton A, Dwyer J, Millar C, Tolhurst P, Berecki‐Gisolf J. Sociodemographic characteristics associated with hospital contact in the year prior to suicide: a data linkage cohort study in Victoria, Australia. PLoS One 2021; 16: e0252682. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Mileshkin L, Dunn C, Cross H, Duffy M, Shaw M, Antippa P et al. The Victorian comprehensive cancer Centre lung cancer clinical audit: collecting the UK National Lung Cancer Audit data from hospitals in Australia. Intern Med J 2019; 49: 1001–1006. [DOI] [PubMed] [Google Scholar]
- 19. AIHW . Cancer in Australia 2021. Canberra: AIHW; 2021. [Google Scholar]
- 20. Thomas JA, Kendall BJ, El‐Serag HB, Thrift AP, Macdonald GA. Hepatocellular and extrahepatic cancer risk in people with non‐alcoholic fatty liver disease. Lancet Gastroenterol Hepatol 2024; 9: 159–169. [DOI] [PubMed] [Google Scholar]
- 21. Mantovani A, Petracca G, Beatrice G, Csermely A, Tilg H, Byrne CD et al. Non‐alcoholic fatty liver disease and increased risk of incident extrahepatic cancers: a meta‐analysis of observational cohort studies. Gut 2022; 71: 778–788. [DOI] [PubMed] [Google Scholar]
- 22. Zhou BG, Jiang X, She Q, Ding YB. Association of MASLD with the risk of extrahepatic cancers: a systematic review and meta‐analysis of 18 cohort studies. Eur J Clin Invest 2024; 54: e14276. [DOI] [PubMed] [Google Scholar]
- 23. Lee H, Lee HW, Kim SU, Chang KH. Metabolic dysfunction‐associated fatty liver disease increases colon cancer risk: a Nationwide cohort study. Clin Transl Gastroenterol 2022; 13: e00435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Yuan XW X, Wu S, Chen S, Wang Y, Wang J, Lu Y et al. Associations between metabolic dysfunction‐associated fatty liver disease and extrahepatic cancers: a cohort in China. HepatoBiliary Surg Nutr 2022; 12: 671–681. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Zhou H, Chen H, Lu H, Wu B, Zhang S, Gu Y et al. Sex differences in mortality and liver‐related events in non‐alcoholic fatty liver disease: a systematic review and meta‐analysis. Liver Int 2024; 44: 1600–1609. [DOI] [PubMed] [Google Scholar]
- 26. Bjorkstrom K, Widman L, Hagstrom H. Risk of hepatic and extrahepatic cancer in NAFLD: a population‐based cohort study. Liver Int 2022; 42: 820–828. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Kim GA, Lee HC, Choe J, Kim MJ, Lee MJ, Chang HS et al. Association between non‐alcoholic fatty liver disease and cancer incidence rate. J Hepatol 2017; 68: 140–146. [DOI] [PubMed] [Google Scholar]
- 28. Allen AM, Hicks SB, Mara KC, Larson JJ, Therneau TM. The risk of incident extrahepatic cancers is higher in non‐alcoholic fatty liver disease than obesity ‐ a longitudinal cohort study. J Hepatol 2019; 71: 1229–1236. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. White A, Ironmonger L, Steele RJC, Ormiston‐Smith N, Crawford C, Seims A. A review of sex‐related differences in colorectal cancer incidence, screening uptake, routes to diagnosis, cancer stage and survival in the UK. BMC Cancer 2018; 18: 906. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Abancens M, Bustos V, Harvey H, McBryan J, Harvey BJ. Sexual dimorphism in colon cancer. Front Oncol 2020; 10: 607909. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Murphy CC, Zaki TA. Changing epidemiology of colorectal cancer ‐ birth cohort effects and emerging risk factors. Nat Rev Gastroenterol Hepatol 2024; 21: 25–34. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1. Study flow diagram.
Table S1. ICD‐10 codes utilised to define cancer types.
Table S2. Sensitivity analysis exploring relationship between FLD and extrahepatic cancer‐related death, using participants with FLI <30 as the control group.
Table S3. Subgroup analysis investigating relationship between FLD and extrahepatic cancer‐related death, according to gender.
Table S4. Sensitivity analysis investigating specific cause of incident extrahepatic cancer according to FLD status after excluding participants with prior diagnosis of extrahepatic cancer at baseline.
Table S5. Differences in risk and protective factors between those with and those without FLD among women and men.
Data Availability Statement
The data that support the findings of this study are available from CVDL. Restrictions apply to the availability of these data, which were used under license for this study. Data are available from the author(s) with the permission of CVDL.