Risks of liver cirrhosis, hepatocellular carcinoma, hepatic-related complications, and mortality in patients with type 2 diabetes in Taiwan

Abstract

BACKGROUND

Hepatitis B and C and alcoholic liver disease are the principal causes of hepatic-related morbidity and mortality. However, evidence of the associations between diabetes without the above risk factors and hepatic-related study endpoints is not well understood. In addition, the effects of associated metabolic dysfunction and exercise on hepatic outcomes are still not clear.

AIM

To investigate the incidence and relative hazards of cirrhosis of the liver, hepatocellular carcinoma (HCC), hepatic-related complications and mortality in patients with type 2 diabetes (T2D) who were nonalcoholic and serologically negative for hepatitis B and C in Taiwan.

METHODS

A total of 33184 T2D patients and 648746 nondiabetic subjects selected from Taiwan’s adult preventive health care service were linked to various National Health Insurance databases, cancer registry, and death registry to identify cirrhosis of the liver, HCC, hepatic-related complications, and mortality. The Poisson assumption and Cox proportional hazard regression model were used to estimate the incidences and relative hazards of all hepatic-related study endpoints, respectively. We also compared the risk of hepatic outcomes stratified by age, sex, associated metabolic dysfunctions, and regular exercise between T2D patients and nondiabetic subjects.

RESULTS

Compared with nondiabetic subjects, T2D patients had a significantly greater incidence (6.32 vs 17.20 per 10000 person-years) and greater risk of cirrhosis of the liver [adjusted hazard ratio (aHR) 1.45; 95%CI: 1.30-1.62]. The aHRs for HCC, hepatic complications, and mortality were 1.81, 1.87, and 2.08, respectively. An older age, male sex, obesity, hypertension, and dyslipidemia further increased the risks of all hepatic-related study endpoints, and regular exercise decreased the risk, irrespective of diabetes status.

CONCLUSION

Patients with T2D are at increased risk of cirrhosis of the liver, HCC, hepatic-related complications, and mortality, and associated metabolic dysfunctions provide additional hazard. Coordinated interprofessional care for high-risk T2D patients and diabetes education, with an emphasis on the importance of physical activity, are crucial for minimizing hepatic outcomes.

Key Words: Type 2 diabetes mellitus; Liver cirrhosis; Hepatocellular carcinoma; Hepatic-related complications; Mortality; Metabolic dysfunction-associated steatotic liver disease; Metabolic dysfunction-associated steatohepatitis

Core Tip: Nonalcoholic, hepatitis B and C serologically negative patients with type 2 diabetes, one of the cardiometabolic risk factors for metabolic dysfunction-associated steatotic liver disease, exhibited independently increased risks of cirrhosis of the liver, hepatocellular carcinoma, hepatic-related complications, and mortality, irrespective of age or sex. Coexisting obesity, hypertension, and dyslipidemia worsened this risk, but regular exercise could ameliorate hepatic outcomes even after adjustment for confounders.

INTRODUCTION

The global burden of cirrhosis of the liver and hepatocellular carcinoma (HCC), major causes of hepatic-related complications and mortality, has been increasing worldwide. Hepatitis B and C and alcoholic liver disease are major etiologies of cirrhosis of the liver and HCC[1]. However, increasing evidence has shown the impacts of nonalcoholic fatty liver disease and nonalcoholic steatohepatitis (NASH), which are now replaced by new nomenclature metabolic dysfunction-associated steatotic liver disease (MASLD) and metabolic dysfunction-associated steatohepatitis (MASH)[2], on increasing the proportions of hepatic-related morbidity and mortality[1,3]. Type 2 diabetes (T2D), one of the five cardiometabolic risk factors for MASLD, can worsen these hepatic outcomes[4]. The measurement of liver stiffness by transient elastography in T2D patients in an outpatient setting showed that the prevalence of steatosis was 70% and that of fibrosis was 21%, although only a minority of patients had elevated liver function[5]. Increasing evidence indicates that the fibrosis stage at diagnosis is intimately related to the future risks of cirrhosis, HCC, and liver-related mortality[6,7], which necessitates health care providers to be more proactive in screening and detecting early-stage hepatic-related disorders in patients with T2D[8]. Obesity, hypertension and dyslipidemia, other cardiometabolic risk factors for MASLD, are prevalent in patients with T2D[9], but the literature regarding hepatic outcomes in T2D patients who are nonalcoholic and serologically negative for hepatitis B and C is scarce. Many of the previous reports determining the association between diabetes and hepatic-related study endpoints did not exclude alcoholism[10-15], viral hepatitis or other hepatic risk factors[10-12] in their protocol. One study that identified alcoholism and hepatitis in T2D patients reported that cirrhosis was uncommon in patients with steatotic liver disease[16]. In some studies, however, diabetes was reported to be a risk factor for nonviral, nonalcoholic liver cirrhosis[12,17], HCC[18], hepatic decompensation[18], and hepatic mortality[19]. However, many previous studies did not confirm viral hepatitis B and C statuses by serological markers[14,15,17-20], only examined hepatitis B surface antigen (HBsAg)[13] or restricted the study participants to hospitalized veteran patients[15,17].

The primary aim of our study was to ascertain the incidences and risks of cirrhosis of the liver, HCC, hepatic-related complications, and mortality in patients with T2D and nondiabetic study subjects who were serologically negative for hepatitis B and C from the nationwide Taiwan adult preventive health care database after the exclusion of International Classification of Diseases (ICD) diagnoses of risk factors for hepatic diseases, including alcohol-related disorders. Second, we determined the independent effects of age, sex, and associated metabolic dysfunction, such as obesity, hypertension, dyslipidemia, and regular exercise, on the risks of the abovementioned hepatic outcomes.

MATERIALS AND METHODS

Ascertainment of the study participants

For the early detection, prevention and treatment of chronic diseases, Taiwan’s Health Promotion Administration (HPA), Ministry of Health and Welfare (MHW), provides a nationwide preventive adult health care service for people aged 40-65 years once every three years[21]. This comprehensive medical check-up service includes questionnaires regarding baseline demographics; past history of chronic disease (diabetes, hypertension and dyslipidemia); personal history of the presence (≤ or > 150 minutes of exercise per week) or absence of regular exercise; medication history; a general physical examination, including measurements of height, weight, body mass index (BMI), and systolic and diastolic blood pressure (SBP and DBP, respectively); and laboratory examinations, such as fasting plasma glucose (FPG), aspartate transaminase (AST), alanine aminotransferase (ALT), total cholesterol (TC), triglyceride (TG), and high-density lipoprotein cholesterol (HDL-C) levels. In addition, this service provides a lifetime examination of serological markers for HBsAg and hepatitis C antibody (anti-HCV) in those aged 45-79 years. This study was approved by the Institutional Review Board of Fu Jen Catholic University (No. C110216), which waived the requirement for informed consent.

We selected 972766 people aged ≥ 40 and < 75 years who underwent medical check-ups administered by the HPA, MHW, with complete information on HBsAg and anti-HCV serology in this study between 2012 and 2019. We excluded 10792 people with missing sex information, 57236 people with missing BMI information, 2581 people with missing ALT or AST information, 204 people who died before the index date and 95071 people who were positive for HBsAg- or anti-HCV during the entire study period.

Information on the related diseases and prescribed medications of these study subjects was obtained from the National Health Insurance (NHI) ambulatory care or inpatient claims. Taiwan’s NHI program, administered by the NHI Administration, MHW, is a universal health program, and approximately 99.9% of Taiwanese people are covered by the NHI[22]. We identified patients with type 1 diabetes (T1D) from the NHI major illness database, as in Taiwan, the NHI Administration issues major illness cards to patients with T1D, and they are exempt from copayment. Information on patients with cancer diagnoses was obtained from a nationwide cancer registry database established by the MHW for comprehensive cancer surveillance. After linkage with the NHI major illness and cancer registry databases, we identified 174 patients with T1D and 27020 cancer patients, respectively, and these patients were also excluded from our study. Furthermore, we also excluded 47483 patients with a diagnosis of a risk factor for cirrhosis, hepatic complications, or HCC (Supplementary Table 1), which appeared ≥ 2 times in ambulatory claims or ≥ 1 time in inpatient claims in the NHI database.

Among the remaining 732205 people, if the study subject met the following criteria during the first visit of medical check-up, he or she was regarded as a T2D patient: FPG level ≥ 126 mg/dL or history of diabetes plus ≥ 2 ambulatory claims or ≥ 1 inpatient claim with an ICD diagnosis of diabetes (ICD-9: 250.x0 or 250.x2; ICD-10: E11) or at least ≥ 1 outpatient medical record of antidiabetic agent use within a year prior to the index date. We excluded 5275 people with FPG levels ≥ 126 mg/dL but no record of antidiabetic agent use or people with antidiabetic agent use only once during the year prior to the index date without a history or an ICD diagnosis of diabetes in either the ambulatory or inpatient claim database. Other people without T2D criteria were regarded as nondiabetic subjects. The final cohort consisted of 33184 patients with T2D and 648746 nondiabetic group. The index dates of the T2D patients and the control subjects were the first dates of medical check-ups between 2012 and 2019. The age of each study subject was evaluated by the difference between the date of birth and the index date. The cutoff point for BMI we used in this study was 30 kg/m², as this differentiation is most widely used to identify obesity in the literature[23]. Other metabolic dysfunctions we detected in this study included hypertension (BP ≥ 130/85 mmHg and/or the use of antihypertensive agents), dyslipidemia (TC level ≥ 200 mg/dL, TG level ≥ 150 mg/dL or HDL level ≤ 40 mg/dL for men or HDL level ≤ 50 mg/dL for women) and/or the use of antilipid medications.

Study outcomes

Using unique PINs, we linked both patients with T2D and those without diabetes to ambulatory and inpatient claims (2012-2020) to identify the first episode of cirrhosis of the liver and hepatic-related complications (Supplementary Table 2) as the endpoints of the study. Although there was lack of information regarding platelet data and abdominal ultrasonography in adult health care preventive service, we retrieved only those patients with diagnoses ≥ 2 within a year in the ambulatory claim or ≥ 1 in the inpatient claim to avoid the accidental inclusion of miscoded patients in the NHI claims database. In addition, we linked the study subjects to a cancer registry (2012-2020) and a national death registry (2012-2021) to identify HCC and hepatic-related mortality, respectively. Each outcome was analyzed separately. The follow-up period ranged from January 1, 2012, to December 31, 2021, an approximately 10-year period. If the study subject died from any cause, the date of censoring was considered the date of death. If a subject did not experience mortality, the date of censoring was the date of the end of this study (i.e., December 31, 2021).

Statistical analysis

Using the Poisson assumption, the incidence densities (IDs) of cirrhosis of the liver, HCC, hepatic-related complications, and mortality were determined with person-years (PYs) as the denominator. We also evaluated the relative hazards of the above study endpoints between T2D patients and nondiabetic subjects using a Cox proportional hazard regression model adjusted for age, sex, metabolic profile (BMI, SBP, DBP, TC, TG, and HDL-C levels), liver enzymes (ALT and AST), and the presence or absence of regular exercise. In addition, we also assessed the independent effects of age, sex, and associated metabolic dysfunctions, such as obesity, hypertension, dyslipidemia, and a history of regular exercise, on the risk of poor hepatic outcomes between nondiabetic subjects and patients with T2D by adjusting for the abovementioned covariates. Cumulative incidences of cirrhosis of the liver, HCC, and hepatic-related complications and survival plots of hepatic mortality related to the diabetes status and sex were generated using the Nelson-Aalen or Kaplan-Meier methods. We used the log-rank test to explore the statistical significance of differences among people with and without diabetes stratified by sex and various hepatic outcomes. All the statistical analyses were performed with SAS (version 9.4; SAS Institute, Cary, NC, United States). A P value < 0.05 was considered to indicate statistical significance.

RESULTS

The baseline general characteristics of the study subjects are presented in Table 1. Patients with T2D were older, and greater proportions (57.84%) of the patients were aged ≥ 50 years. A greater percentage of men with T2D (55.24%) was observed, whereas a greater proportion of nondiabetic subjects were women (60.41%). Patients with T2D had higher BMIs, SBP, DBP, TC levels, TG levels, ALT levels, and AST levels but lower HDL-C levels. The distributions of individuals who regularly exercised were comparable between patients with T2D and nondiabetic subjects. After a mean follow-up period of nearly 7 years, more patients with T2D suffered from all hepatic outcomes.

Table 1 General characteristics of the study subjects.

Variables1	Nondiabetic group		Type 2 diabetic group
	n	%	n	%
Demographics
Age (years)
< 50	433720	66.86	13991	42.16
≥ 50	215026	33.14	19193	57.84
mean ± SD	49.56	7.79	55.34	10.16
Sex
Women	391920	60.41	14853	44.76
Men	256826	39.59	18331	55.24
SBP (mmHg)
< 130	408227	62.93	14018	42.24
≥ 130	240519	37.07	19166	57.76
mean ± SD	125.00	18.65	133.50	18.67
DBP (mmHg)
< 85	465778	71.80	21673	65.31
≥ 85	182968	28.20	11511	34.69
mean ± SD	78.44	12.71	80.98	11.98
Medications
Anti-diabetic agents	0	0.00	33184	100.00
Antihypertensive agents	142453	21.96	21489	64.76
Antilipid agents	35626	5.49	19710	59.40
Laboratory results
FPG (mg/dL)
< 126	648746	100.00	8319	25.07
≥ 126	0	0	24865	74.93
mean ± SD	91.94	10.85	170.00	70.75
TC (mg/dL)
< 200	348098	53.66	22169	66.81
≥ 200	300648	46.34	11015	33.19
mean ± SD	199.00	37.63	187.00	45.14
HDL-C (mg/dL)
M: > 40; F > 50	471424	72.67	17571	52.95
M: ≤ 40; F ≤ 50	177322	27.33	15613	47.05
mean ± SD	56.04	16.69	47.73	15.91
TG (mg/dL)
< 150	470029	72.45	17411	52.47
≥ 150	178717	27.55	15773	47.53
mean ± SD	134.60	126.40	196.20	223.70
ALT (IU/L)
Normal	539660	83.19	24258	73.10
Above normal	109086	16.81	8.926	26.90
mean ± SD	25.86	31.40	31.76	34.01
AST (IU/L)
Normal	579562	89.34	27350	82.42
Above normal	69184	10.66	5834	17.58
mean ± SD	24.51	23.81	26.88	25.83
Metabolic dysfunctions
Obesity, kg/m2
BMI ≤ 30	600060	92.50	26247	79.10
BMI > 30	48686	7.5	6937	20.90
Hypertension
No	370062	57.04	12457	37.54
Yes	278684	42.96	20727	62.46
Dyslipidemia
No	207416	31.97	8227	24.79
Yes	441330	68.03	24957	75.21
Personal history
Regular exercise
No	359624	55.43	17272	52.05
Yes	289122	44.57	15912	47.95
Outcomes
Cirrhosis of liver	2697	0.42	385	1.16
Hepatic complications	1436	0.22	207	0.62
HCC	803	0.12	183	0.55
Hepatic mortality	517	0.08	134	0.40
Mean follow-up period (years) (± SD)	6.59	2.36	6.78	2.45
Total	648746	100.00	33184	100.00

¹The inconsistency between the total population and the population summed for individual variables was due to missing information.

ALT: Alanine transaminase; AST: Aspartate transaminase; BMI: Body mass index; DBP: Diastolic blood pressure; FPG: Fasting plasma glucose; HCC: Hepatocellular carcinoma; HDL-C: High-density lipoprotein cholesterol; SBP: Systolic blood pressure; TC: Total cholesterol; TG: Triglyceride; M: Male; F: Female.

Cumulative incidences of cirrhosis of the liver, HCC, and hepatic-related complications and survival plot of hepatic-related mortality

The cumulative incidence of cirrhosis of the liver was the highest in diabetic men, followed by diabetic women, after a mean follow-up of 7 years. Nondiabetic women had the lowest cumulative incidence of cirrhosis of the liver (P for log rank test < 0.0001; Figure 1A). The cumulative incidence of HCC was lower, and the highest incidence was also observed in diabetic men (Figure 1B). The difference in the incidence of HCC between diabetic women and nondiabetic men was small, and nondiabetic women had the lowest incidence (P for log rank test < 0.0001). The cumulative incidence rates of hepatic-related complications in diabetic men and women were similar (Figure 1C), whereas those in nondiabetic men and women were lower (P for log rank test < 0.0001). Diabetic men had the lowest hepatic-related survival rate, followed by diabetic women. Similarly, nondiabetic men and women had relatively higher hepatic-related survival rates (P for log rank test < 0.0001; Figure 1D).

Figure 1 Cumulative incidences and survival plot. A: Cumulative incidences of cirrhosis of the liver; B: Cumulative incidences of hepatocellular carcinoma; C: Cumulative incidences of hepatic-related complications; D: Survival plot of hepatic-related mortality.

Overall IDs and relative hazards of cirrhosis of the liver, HCC, hepatic-related complications, and mortality

The overall IDs of liver cirrhosis and hepatic-related complications in the nondiabetic group were 6.32 and 3.36/10000 PYs, respectively, but those in the T2D group were nearly three times higher, at approximately 17.20 and 9.22/10000 PYs, respectively. Similarly, the incidences of HCC and hepatic-related mortality in T2D patients were nearly four times higher (8.15 and 5.96/10000 PYs, respectively) than those in nondiabetic patients (1.88 and 1.21/10000 PYs, respectively) (Table 2).

Table 2 Overall incidence densities and relative hazards of cirrhosis of the liver, hepatocellular carcinoma, hepatic-related complications, and mortality in patients with type 2 diabetes and nondiabetic subjects.

Variables	Non-diabetic group			Patients with T2D			Model 1 crude HR (95%CI)	Model 2 adjusted HR (95%CI)2
	No. of patients	No. of events	Incidence densities (per 10000 person-years) (95%CI)1	No. of patients	No. of events	Incidence densities (per 10000 person-years) (95%CI)1
Cirrhosis of the liver	648746	2697	6.32 (6.08-6.56)	33184	385	17.20 (15.48-18.92)	2.71 (2.43-3.01)	1.45 (1.30-1.62)
Hepatocellular carcinoma	648746	803	1.88 (1.75-2.01)	33184	183	8.15 (6.97-9.33)	4.33 (3.69-5.09)	1.81 (1.52-2.15)
Hepatic complications	648746	1436	3.36 (3.19-3.54)	33184	207	9.22 (7.96-10.48)	2.75 (2.37-3.18)	1.87 (1.60-2.19)
Hepatic mortality	648746	517	1.21 (1.11-1.31)	33184	134	5.96 (4.95-6.97)	4.73 (3.91-5.72)	2.08 (1.70-2.54)

¹Based on the Poisson assumption.

²Based on the Cox proportional hazard regression model adjusted for age, body mass index, systolic blood pressure, diastolic blood pressure, total cholesterol, triglyceride, high-density lipoprotein cholesterol, alanine transaminase, and aspartate transaminase as continuous variables, with sex and regular exercise as categorical variables.

T2D: Type 2 diabetes; HR: Hazard ratio.

The relative hazard of cirrhosis of the liver was significantly higher in T2D patients, with a hazard ratio (HR) of 2.71 and a 95%CI of 2.43-3.01, than in nondiabetic patients. After adjustment for age, sex, various components of the metabolic profile (BMI, SBP, DBP, TC, TG, and HDL-C), and liver enzymes (ALT and AST) with regular exercise in the model, the adjusted HR (aHR) decreased but remained consequential (aHR: 1.45; 95%CI: 1.30-1.62). The aHR of T2D patients at risk of progressing to HCC was higher (aHR: 1.81; 95%CI: 1.52-2.15) after adjustment for covariates. Similarly, substantial risks of developing hepatic-related complications and mortality were observed in patients with T2D (aHR: 1.87 and 2.08, respectively) (Table 2).

Effects of age, sex, metabolic dysfunction, and exercise on the relative hazards of cirrhosis of the liver, HCC, hepatic-related complications, and mortality

Age and sex were significant effect modifiers for all hepatic outcomes. Compared with nondiabetic subjects aged < 50 years, nondiabetic subjects aged ≥ 50 years or patients with T2D aged either < 50 years or ≥ 50 years had considerable risks of progressing to hepatic-related study endpoints. Except hepatic-related complications, T2D patients aged < 50 years had lower risks of other hepatic outcomes than nondiabetic control subjects aged ≥ 50 years did. Additionally, T2D patients aged ≥ 50 years had the highest aHRs of all the study outcomes. Compared with nondiabetic women, men with and without diabetes had substantial relative hazards of hepatic outcomes, but diabetic men had the most elevated aHRs for all hepatic outcomes except for hepatic-related complications, for which both diabetic men and diabetic women had comparable risks (Table 3).

Table 3 Effects of age, sex, the metabolic profile, and exercise on the relative risks of cirrhosis of the liver, hepatocellular carcinoma, hepatic complications, and mortality in patients with type 2 diabetes and nondiabetic subjects.

Variables	Cirrhosis of the liver, HR (95%CI)1	HCC, HR (95%CI)1	Hepatic complications, HR (95%CI)1	Hepatic mortality, HR (95%CI)1
Age
Nondiabetic group
< 50 years	1.0	1.0	1.0	1.0
≥ 50 years	2.29 (2.12-2.48)	5.49 (4.67-6.45)	1.59 (1.43-1.78)	4.32 (3.57-5.24)
T2D patients
< 50 years	1.78 (1.46-2.17)	2.05 (1.30-3.23)	2.11 (1.64-2.71)	2.22 (1.37-3.59)
≥ 50 years	3.41 (2.98-3.91)	11.91 (9.61-14.77)	3.10 (2.56-3.74)	10.29 (7.99-13.26)
Sex
Nondiabetic group
Women	1.0	1.0	1.0	1.0
Men	2.11 (1.94-2.28)	2.67 (2.29-3.12)	1.53 (1.37-1.71)	2.56 (2.11-3.11)
T2D patients
Women	1.82 (1.53-2.18)	1.78 (1.30-2.43)	2.34 (1.87-2.94)	2.28 (1.61-3.23)
Men	2.70 (2.33-3.14)	4.86 (3.87-6.09)	2.45 (1.98-3.03)	5.10 (3.89-6.71)
Obesity (BMI ≤ 30 kg/m2vs > 30 kg/m2)
Nondiabetic group
No	1.0	1.0	1.0	1.0
Yes	1.30 (1.15-1.48)	1.51 (1.20-1.91)	0.89 (0.72-1.09)	1.55 (1.18-2.05)
T2D patients
No	1.43 (1.27-1.63)	1.87 (1.55-2.25)	1.65 (1.39-1.97)	2.04 (1.63-2.55)
Yes	1.98 (1.60-2.45)	2.47 (1.74-3.49)	2.25 (1.69-3.00)	3.25 (2.25-4.71)
Hypertension
Nondiabetic group
No	1.0	1.0	1.0	1.0
Yes	1.19 (1.09-1.29)	1.08 (0.94-1.25)	0.96 (0.59-1.54)	1.26 (1.05-1.52)
T2D patients
No	1.55 (1.28-1.87)	1.52 (1.11-2.07)	1.74 (1.48-2.04)	1.88 (1.30-2.73)
Yes	1.70 (1.47-1.96)	2.07 (1.66-2.57)	2.80 (1.65-4.77)	2.70 (2.08-3.51)
Dyslipidemia
Nondiabetic group
No	1.0	1.0	1.0	1.0
Yes	1.26 (1.17-1.37)	0.87 (0.75-1.00)	1.43 (1.29-1.60)	1.23 (1.03-1.47)
T2D patients
No	1.55 (1.29-1.85)	1.48 (1.13-1.93)	2.06 (1.59-2.67)	1.97 (1.42-2.74)
Yes	1.83 (1.59-2.11)	1.92 (1.56-2.37)	2.69 (2.22-3.26)	2.78 (2.16-3.57)
Regular exercise
Nondiabetic group
No	1.0	1.0	1.0	1.0
Yes	0.71 (0.66-0.77)	0.71 (0.62-0.82)	0.77 (0.69-0.85)	0.63 (0.53-0.75)
T2D patients
No	1.56 (1.35-1.80)	1.71 (1.35-2.17)	2.06 (1.69-2.51)	2.12 (1.63-2.75)
Yes	0.94 (0.79-1.11)	1.36 (1.08-1.71)	1.26 (0.99-1.59)	1.27 (0.95-1.71)

¹Based on the Cox proportional hazard regression model adjusted for age, body mass index, systolic blood pressure, diastolic blood pressure, high-density lipoprotein cholesterol, total cholesterol, triglyceride, alanine transaminase, and aspartate transaminase as continuous variables, with sex and regular exercise as categorical variables.

BMI: Body mass index; T2D: Type 2 diabetes; HR: Hazard ratio.

Nondiabetic control subjects with obesity (BMI > 30 kg/m²) had increased risks of liver cirrhosis, HCC or hepatic mortality but not hepatic complications. In contrast, in patients with T2D, BMIs ≤ 30 kg/m² or > 30 kg/m² were related to elevated risks of all hepatic outcomes. Accompanying hypertension and dyslipidemia in nondiabetic subjects increased the risks of cirrhosis of the liver and hepatic-related mortality. T2D patients with and without hypertension or dyslipidemia, on the other hand, had consequential risks of all hepatic endpoints, and patients with T2D comorbid with any metabolic dysfunctions had the highest aHRs of all hepatic-related study endpoints.

Interestingly, people with or without diabetes who exercised regularly had lower risks of cirrhosis of the liver, HCC, hepatic-related complications, and mortality. Compared with those without regular exercise, nondiabetic subjects who regularly exercised had a 30%-40% lower risk of developing all hepatic outcomes. In T2D patients who regularly exercised, the risks of all hepatic outcomes except HCC became neutral, but patients with T2D without exercise had a 56%-212% increased risk of hepatic-related endpoints compared with nondiabetic subjects without regular exercise.

DISCUSSION

After a mean follow-up period of 7 years, patients with T2D had increased risks of cirrhosis of the liver, HCC, hepatic-related complications, and mortality even after adjustment for various confounding factors. Compared with their nondiabetic counterparts, diabetic men and women had higher cumulative incidences of all study endpoints. Increasing age, male sex, and metabolic dysfunctions such as obesity, hypertension and dyslipidemia increase the risks of the above outcomes. Regular exercise, on the other hand, consistently reduced the risks of all hepatic-related study endpoints in study subjects both with and without diabetes.

To the best of our knowledge, only a few studies have analyzed the relationship between T2D and the risks of various hepatic outcomes. A direct comparison of the results of our study and those of previous studies is difficult because of the different study designs and dissimilar definitions of the study endpoints. Some of the previous studies identified viral hepatitis B and C by ICD diagnoses[14,15,17-20] rather than by biochemical confirmation. In one study, the authors examined HBsAg but not anti-HCV, and they did not exclude alcohol consumption or other hepatic risk factors[13]. Some studies recruited hospitalized patients with diabetes from the Department of Veterans Affairs, and the patients were older, with a male preponderance (98.4%)[15,17]. Furthermore, the liver cirrhosis identified in the above studies included unspecified chronic hepatitis[14,15,17]; hepatic-related complications such as hepatic encephalopathy and portal hypertension[13,17]; and biliary-[13-15,17], toxic liver-[20] or alcoholic liver-related cirrhosis[14,15]. Consequently, the aHRs of cirrhosis of the liver estimated in these studies[13,14,17,20] were higher than those in our study. We performed a separate analysis of the risks of hepatic-related complications in T2D patients, and the aHR estimated (1.87) in our study was still lower than that in some of the previous studies[14,17,18] but higher than that in a Canadian study[20] in which the authors recruited newly diagnosed patients with diabetes, which might have resulted in lower risk estimates. Similarly, because of the discrepancy in the criteria for the study participants, the risk of HCC in T2D patients in our study (aHR: 1.81) was lower than that estimated for veteran patients with diabetes (2.16)[17] and in the results of a meta-analysis[18] but higher than that reported in a Chinese study (1.52)[13]. Our risk estimate for hepatic-related mortality (2.08) was slightly lower than that of previous studies[10-12], in which alcoholic liver disease-related mortality[10-12], viral hepatitis-related mortality[11], toxic liver disease-related mortality[11], and biliary cirrhosis-related mortality[10,12] were included. Additionally, most of the above studies analyzed the relationship between diabetes and hepatic outcomes without adjusting for metabolic dysfunctions[10-13,17], which might have resulted in higher HR estimates for the corresponding study endpoints.

From a study of the Hong Kong electronic health care database, Zhang et al[24] reported that only 1.1% and 4.9% of T2D patients aged < 40 years and ≥ 40 to < 50 years, respectively, developed hepatic-related events, and the authors concluded that screening for hepatic-related events should start at the age of 50 years in clinical practice. In our study, however, even T2D patients aged < 50 years had an increased risk of all hepatic-related events compared with nondiabetic patients aged < 50 years. Age was a significant predictor of hepatic decompensation (P < 0.0001) or HCC (P = 0.029)[18], and an increase in age of 10 years increased the risks of chronic liver disease and HCC in veteran patients with diabetes[17]. Stratification by age (30-49, 50-64, and ≥ 65) in a Canadian study revealed a higher risk of severe liver disease in newly diagnosed patients with diabetes than in individuals without diabetes[20].

Sex was not a significant predictor of the development of either HCC or cirrhotic complications in the Hong Kong study (P values of 0.20 and 0.48, respectively)[24]. However, in our study, T2D patients of both sexes had significantly higher risks of all hepatic outcomes than nondiabetic patients, and diabetic men had the highest cumulative incidences and aHRs of all study endpoints. In an analysis of newly diagnosed patients with diabetes, both diabetic men and diabetic women had an excessive risk of severe liver disease compared with those without diabetes[20].

Few studies have analyzed the risks of metabolic dysfunctions and hepatic outcomes in patients with diabetes. A Canadian study revealed that patients with diabetes with and without obesity had a higher risk of serious liver disease than did those with neither obesity nor diabetes[20], but BMI did not predict hepatic-related events in T2D patients in a Hong Kong study[24]. In our study, nondiabetic study subjects with obesity had increased risks of liver cirrhosis, HCC, and hepatic-related mortality but not hepatic-related complications, probably due to the high prevalence of sarcopenia and frailty in patients with decompensated cirrhosis[25,26]. However, T2D patients with and without obesity had significantly higher risks of all hepatic outcomes, similar to the results of a Canadian study[20], and the risk estimates were higher than those of nondiabetic patients with obesity. A previous study revealed that newly diagnosed patients with diabetes with and without hypertension or dyslipidemia had an increased risk of severe liver disease[20]. Similarly, our T2D patients with and without hypertension or dyslipidemia presented higher risks of all study outcomes than nondiabetic subjects without hypertension or dyslipidemia did, and the risk estimates were the highest for T2D patients with hypertension or dyslipidemia.

Regular exercise has been shown to ameliorate MASLD and MASH[27] and subsequently reduce the risks of hepatobiliary cancers[28] and liver-related mortality[29] in some studies, but studies on the effect of regular exercise on the risk of liver outcomes in patients with diabetes are scarce. In our study, both nondiabetic subjects and T2D patients who regularly exercised, irrespective of the duration of physical activity, had reduced risks of all hepatic outcomes compared with nondiabetic subjects without regular exercise. Although T2D patients who regularly exercised still had a modestly increased risk of HCC, the estimated aHR was lower than that of T2D patients who did not exercise regularly.

The exact mechanisms underlying the associations between diabetes and increased risks of poor hepatic outcomes have not been elucidated. Even in T2D patients without MASLD, a 2-fold increase in the fat percentage and a nearly 5-fold increase in the number fat-containing cells were observed, irrespective of age and sex[30]. Hepatic fat accumulation leads to mitochondrial injury, resulting in free radical and peroxisome release[31]. Subsequent insulin resistance, lipotoxicity, oxidative stress, cytokine injury, and hepatic iron and/or lipid peroxidation[32,33] promote and accelerate inflammasome activation and apoptotic cell death. Successive inflammation and regeneration lead to the progression of liver fibrosis and cirrhosis[34,35]. Some researchers have hypothesized that alterations in the gut microbiota and lipopolysaccharide or short-chain fatty acids produced by bacteria or derived from bile acid metabolism could impact liver inflammation and fibrosis progression[33,36]. The high prevalence of other metabolic derangements, such as obesity, hypertension, and dyslipidemia, in T2D patients may worsen and accelerate the course of the pathogenetic process[23], resulting in an increased risk of adverse hepatic outcomes. Exercise enhances lipophagy, promotes peroxisome proliferator-activated receptor-γ expression, attenuates hepatocyte apoptosis, increases insulin sensitivity, and decreases downstream inflammatory processes[37-39], which might have resulted in an inverse relationship between regular exercise and any of the hepatic endpoints in our study subjects with and without diabetes.

The significant risks of various hepatic-related outcomes in T2D patients who are nonalcoholic and serologically negative for hepatitis B and C revealed in this study indicate that not only regular screening of micro- and macrovascular complications but also an awareness of MASLD and MASH in patients with diabetes are important in daily clinical practice. In addition to the optimal control of diabetes, hypertension and dyslipidemia, facilitating positive health behaviors for weight management and encouraging engagement in physical exercise are also essential to reduce the hepatic endpoints in T2D patients.

Our study had several methodological strengths. First, our study participants were recruited from a nationwide adult preventive health care service, and the outcomes were obtained from a highly representative NHI database, which is population-based with little likelihood of recall bias. In addition, both the adult preventive health care service and the NHI databases contain longitudinal records for a large sample of patients from different geographical areas with little possibility of nonresponse or loss to follow-up. Second, we identified the viral hepatitis B and C status by detecting serological markers and excluded potential risk factors for hepatic disease, including alcoholic-related disorders and alcoholic liver disease, so that we could obtain relatively accurate estimates of the incidence and risk of metabolic dysfunction-associated hepatic-related endpoints in our study. Third, the large number of study subjects enabled us to perform separate analyses with stratification for several metabolic dysfunctions and regular exercise without compromising the sample size. Fourth, we identified diagnoses of liver cirrhosis, hepatic-related complications from ambulatory and inpatient claims, T1D from the major illness registry, and HCC and hepatic mortality from the national cancer and death registries, which might have improved the precision of the outcome diagnoses. Fifth, we determined several metabolic dysfunctions, such as BMI, blood pressure, laboratory profiles, and antidiabetic, antihypertensive, and antilipid medication use, from our national databases.

Despite these strengths, our study has several limitations. First, the exclusive reliance on NHI database analyses might have excluded individuals who did not seek medical care, which could have potentially caused selection bias and misclassification bias and reduced the representativeness of our results. However, we strictly limited the definition of T2D to patients with FPG data or a history of diabetes plus ICD diagnoses and/or antidiabetic agent use, which might have reduced the possibility of disease misclassification. However, nondiabetic subjects might have had undiagnosed or newly emerging diabetes cases, which could have undermined the validity of comparisons between diabetic and nondiabetic groups. Second, our reliance on ambulatory and inpatient claims for diagnoses of liver cirrhosis and hepatic-related complications exclusively would have led to the exclusion of some patients who did not receive ambulatory or inpatient care. This approach resulted in an underestimation of the incidence rate, but it would have little influence on the estimation of the HR of our hepatic-related study endpoints. Third, as this study was entirely based on Taiwanese national databases, the generalizability and applicability of our results to individuals of other ethnicities and regions needs further evaluation because there might be different metabolic and genetic diseases in other societies. Fourth, although our study highlights the benefits of regular exercise, we could only collect data on durations of exercise ≤ or > 150 minutes and lacked detailed information on the modality, frequency, and intensity of physical activity, making a full understanding the impacts of exercise on hepatic-related morbidities and mortality difficult to achieve. Fifth, our study was retrospective, and we could not fully control for all other confounding variables. Therefore, our results should be interpreted with caution, and future prospective studies with more comprehensive data from diverse populations are necessary to validate our conclusions.

CONCLUSION

Compared with nondiabetic subjects, T2D patients had increased risks of cirrhosis of the liver, HCC, hepatic-related complications, and mortality, irrespective of age or sex. Coexisting obesity, hypertension, and dyslipidemia worsened these risks, but regular exercise could ameliorate these outcomes. Diabetologists should be aware of the excessive risks of harmful hepatic outcomes in T2D patients, and a comprehensive medical evaluation with a meticulous assessment of the risks, improvement of associated metabolic dysfunctions, and enhancement of diabetes self-management education emphasizing the importance of regular physical activity are mandatory to reduce the hazards of hepatic-related diseases and mortality in T2D patients.

ACKNOWLEDGEMENTS

The authors are grateful to the Fu-Jen University Foundation for their helpful comments and recommendations. The authors also thank San-Lin You, PhD, Department of Public Health, College of Medicine, Fu Jen Catholic University, for helpful discussions during the development of the study. Dr. You received no financial support for his participation.