Abstract
Background
Patients with chronic hepatitis B (CHB) have a higher risk of developing hepatocellular carcinoma (HCC), causing poor prognosis. Type 2 diabetes mellitus (T2DM), a common comorbidity of CHB, further elevates HCC risk. Insulin and insulin secretagogues are commonly used antidiabetic medications (ADMs) that lower blood glucose by elevating circulating insulin levels. This study aimed to explore the association of insulin and/or insulin secretagogues with the risk of HCC and all-cause mortality among patients with CHB and T2DM.
Patients and Methods
This study included 696 patients diagnosed with CHB and T2DM between January 2016 and December 2019, who were divided into four groups. Group 1 received insulin and/or insulin secretagogues; Group 2 received no ADMs; Group 3 received other ADMs except for insulin and insulin secretagogues; Group 4 received insulin and/or insulin secretagogues in combination with other ADMs. Cox proportional hazards regression models were constructed to assess the risk of developing HCC and death. Subgroup analysis and sensitivity analysis were conducted to assess the robustness of the findings.
Results
Throughout the follow-up period (median 77.5 months for HCC, 80 months for all-cause mortality), Group 1 had a significantly higher incidence of both HCC and all-cause mortality compared with other groups (both P < 0.001). After adjusting for potential covariates, compared with Group 2, Group 1 was associated with a higher risk of developing HCC (HR = 4.42, 95% CI: 1.75–11.17, P = 0.002) and all-cause mortality (HR = 2.46, 95% CI: 1.35–4.50, P = 0.003). Subgroup and sensitivity analyses confirmed robustness.
Conclusion
This study found that the use of insulin and/or insulin secretagogues was associated with a higher incidence of HCC and all-cause mortality among patients with CHB and T2DM. However, when combining insulin and/or insulin secretagogues with other types of ADMs, no increased risk was observed.
Keywords: antidiabetic medications, hepatitis B Virus, diabetes complications, liver cirrhosis, liver cancer
Introduction
Hepatitis B virus (HBV) infection is a significant global health concern, with approximately 30% of individuals worldwide showing evidence of current or past infection.1 HBV infection has always been widely acknowledged as a major risk factor for hepatocellular carcinoma (HCC).2 HCC is the third leading cause of cancer-related mortality.3 Type 2 diabetes mellitus (T2DM) is a metabolic disorder characterized by insulin resistance and hyperglycemia.4 Studies have shown that T2DM is linked to an increased risk of HCC among patients with chronic hepatitis B (CHB).5,6
It is very important to control blood glucose to meet standards for preventing complications and comorbidities related to diabetes,7,8 and various antidiabetic medications (ADMs) are frequently used in clinical practice. Existing studies have indicated that different ADMs may influence the risk of HCC. Meta-analyses have demonstrated that metformin use is associated with a 50% reduction in HCC incidence, whereas sulfonylurea or insulin use increases HCC incidence by 62% and 161%, respectively.9 A study from Taiwan with a mean follow-up of 5.84 years reported that, among patients with T2DM and compensated cirrhosis, insulin use was associated with a higher risk of mortality and liver-related complications compared to non-users.10 The National Health Insurance Service-National Sample Cohort of South Korea investigated the impact of different sulfonylureas on HCC incidence. Among 47,738 newly diagnosed diabetic patients aged ≥40 years, 241 incident HCC cases were identified. Patients treated with a sulfonylurea had a 1.7-fold increased risk of HCC compared to those who never used sulfonylureas.11 These studies suggest that the use of insulin and sulfonylureas in patients with T2DM is associated with an increased risk of HCC. As commonly prescribed glucose-lowering agents, particularly in patients with underlying liver disease, their potential adverse effects on the liver and overall health warrant attention. However, other researches have shown no significant association between these drugs and HCC. Two prospective U.S. cohort studies, including 120,826 women and 50,284 men, found that T2DM was associated with an increased risk of HCC, but neither insulin use nor oral hypoglycemics use was associated with HCC risk.12
To date, no study has specifically focused on the association between insulin and/or insulin secretagogues and HCC risk in patients with both CHB and T2DM. Moreover, official associations or organizations do not currently issue specific pharmacological guidance for diabetes management in CHB patients. Therefore, we conducted this study to investigate the effects of insulin and insulin secretagogues on HCC risk and all-cause mortality in patients with CHB and T2DM. Both insulin and insulin secretagogues elevate circulating insulin levels, and through the activation of the insulin-like growth factor signaling pathway, insulin may accelerate hepatocarcinogenesis.10,13 Therefore, considering both the biological rationale and the limited sample size, we grouped insulin and insulin secretagogues into a single exposure category.
Materials and Methods
Study Design and Setting
The present study was a single-center retrospective cohort study. The study participants were identified from The Third People’s Hospital of Changzhou, a tertiary referral hospital, from January 2016 to December 2019. Patients were divided into four groups: Group 1 received insulin and/or insulin secretagogues; Group 2 received no ADMs, and controlled blood sugar through dietary and exercise interventions; Group 3 received other ADMs except for insulin and insulin secretagogues; Group 4 received insulin and/or insulin secretagogues in combination with other ADMs.
Data Collection
For each patient, the following variables were collected: demographic data (age, sex), clinical data (diagnosis of cirrhosis, CHB, and T2DM), laboratory results (liver function, triglycerides [TG], Glycated Hemoglobin A1c [HbA1c], renal function, complete blood count and coagulation function), and treatment details (types of ADMs and antiviral drugs). Data were captured through automated extraction from electronic medical records (EMR) and manual chart review where necessary. The follow-up period was defined from the admission date of each patient with CHB and T2DM to the date of HCC diagnosis and death. The primary outcomes were HCC incidence and all-cause mortality.
Study Definitions
Chronic Hepatitis B
CHB is a chronic inflammatory liver disease caused by persistent infection with the HBV, characterized by the presence of HBsAg for more than 6 months.1
Type 2 Diabetes Mellitus
T2DM is a chronic metabolic disorder characterized by insulin resistance and relative insulin deficiency. Diagnostic criteria for T2DM in our study are as follows: Fasting Plasma Glucose (FPG) ≥ 7.0 mmol/L or 2-h PG ≥ 11.1 mmol/L during an oral glucose tolerance test (OGTT), HbA1c ≥ 6.5%, or in an individual with classic symptoms of hyperglycemia or hyperglycemic crisis, a random plasma glucose ≥ 11.1 mmol/L.14
Hepatocellular Carcinoma
In this study, HCC was diagnosed per contemporary AASLD guidance.15 For at-risk patients, arterial phase hyperenhancement (APHE) and washout on the portal venous or delayed phases of contrast-enhanced multiphase CT or MRI are considered radiological hallmarks of HCC. For liver nodules in patients without cirrhosis or without HBV infection, a pathological diagnosis of HCC should be obtained.
Participants
Inclusion Criteria
Age ≥18 years
Patients diagnosed with CHB
Patients diagnosed with T2DM
Exclusion Criteria
Diagnosis of HCC or other cancers within 6 months (n = 28)
Died within 6 months (n = 26)
Missing data (n = 33)
Coinfection with hepatitis C virus or human immunodeficiency virus (n = 0)
Diagnosis of autoimmune liver disease, drug-induced liver injury, alcoholic liver injury, and genetic metabolic liver diseases (n = 0)
Statistical Analyses
Software and Sample Size
All statistical analyses were conducted using SPSS (Version 27) and R (Version 4.5.2). Due to the retrospective study design, the sample size was derived from all available data to maximize statistical power.
Descriptive Statistics
Continuous variables were tested for normality with the Shapiro–Wilk test. Data conforming to normality (p > 0.05) are presented as mean ± standard deviation and analyzed with parametric tests; otherwise, data are presented as median (interquartile range) and analyzed with non-parametric tests. Continuous variables were compared using the Kruskal–Wallis test. For multiple comparisons, Dunn’s test with Benjamini-Hochberg correction was applied for group comparison. Categorical variables were expressed as frequency (percentage) and compared using the chi-square test.
Survival Analysis
The Kaplan-Meier curve and Log rank test were used to initially assess the associations between treatment strategies and the risk of HCC and all-cause mortality. Subsequently, Cox proportional hazards regression models (CPH) were constructed to scrutinize the associations. The potential violation of the proportional hazards assumption was examined using the Schoenfeld residual test. The variance inflation factor (VIF) was calculated to assess multicollinearity in model 3. Model 1 was a crude model without any adjustment. Model 2 was adjusted for sex and age. Model 3 was a fully adjusted model that further accounted for clinical parameters, including HbA1c, cirrhosis, antiviral medication use, and HBV-DNA.
Additional Analysis
Subgroup analyses were performed by stratifying participants according to sex, age, and cirrhosis status to explore potential variations in the observed association across these subgroups. A series of sensitivity analyses was conducted to assess the robustness of the findings. Participants with follow-up less than 9 and 12 months were excluded to reduce potential reverse causality. TG and body mass index (BMI) were further adjusted to examine their impacts on the observed associations. Furthermore, inverse probability of treatment weighting (IPTW) was used to create a pseudo-population balanced on predefined covariates. The absolute standardized mean difference (SMD) in unweighted and IPTW-weighted fashions for each covariate was visualized, where an absolute SMD ≤ 0.1 was defined as an adequate balance. The main analysis was re-conducted after IPTW weighting. Lastly, the E-value was computed to account for the effect of unmeasured confounders on the observed associations.
Results
Baseline Characteristics of the Study Participants
After exclusion, 696 participants were included in the study (Figure 1). In Group 1, patients using insulin alone accounted for 77%. In Group 3, 58% of patients were treated with metformin (alone or in combination), while 31.2% received DPP4 inhibitors (DPP-4i) (alone or in combination). The characteristics of the included participants are displayed in Table 1. The baseline median age of our cohort was 55.0 years. Most were male (71.4%). The median BMI was 23.9 kg/m2. At baseline, the median HbA1c was 7.6%. 44.4% of patients had cirrhosis, and 67.5% of the included participants used entecavir. Compared to participants in other groups, those in Group 1 had a significantly lower BMI (P < 0.001) and lower albumin levels (P < 0.001). Compared to participants in Group 3 and Group 4, individuals in Group 1 had lower platelet levels (P < 0.001 for Group 2, P = 0.005 for Group 3) and a higher prevalence of cirrhosis (P < 0.001 for Group 2, P = 0.028 for Group 3). Additionally, compared to participants in Group 2, those in Group 1 had higher HbA1c levels (P < 0.001). Moreover, compared to participants in Group 3, individuals in Group 1 were older (P = 0.005), had lower alanine aminotransferase (ALT) levels (P = 0.013), lower TG levels (P < 0.001), and a higher prevalence of antiviral agent usage (P = 0.011). Compared with participants in Group 4, individuals in Group 1 had higher total bilirubin (TBIL) levels (P < 0.001). During a median follow-up of 77.5 months (IQR: 63–90), 75 cases of HCC occurred, with 55, 5, 10, and 5 cases in groups 1, 2, 3, and 4, respectively. The incidence of HCC was significantly higher in Group 1 (19.0%) compared with others (P<0.001). Additionally, 112 participants died, including 79, 13, 7, and 13 in groups 1, 2, 3, and 4, respectively. The incidence of death was also significantly higher in Group 1 (27.3%) compared with others (P < 0.001).
Figure 1.
Participants inclusion process.
Abbreviation: HCC, hepatocellular carcinoma; ADMs, antidiabetic medications.
Table 1.
Baseline Characteristics of the Included Participants (n=696)
| Baseline Characteristics | Total (n=696) | G1 (n=289) | G2 (n=125) | G3 (n=154) | G4 (n=128) | P-valuea |
|---|---|---|---|---|---|---|
| Age(years), M (IQR) | 55.0(47.0,63.0) | 57.0(49.0, 64.0)c | 54.0(46.0, 62.0) | 53.5(45.0, 61.0)c | 55.0(47.0, 63.0) | 0.013 |
| Male, n (%) | 497(71.4) | 215(74.4) | 85(68.0) | 113(73.4) | 84(65.6) | 0.225 |
| Hypertension, n (%) | 323(46.4) | 121(41.9) | 61(48.8) | 73(47.4) | 68(53.1) | 0.167 |
| Cirrhosis, n (%) | 309(44.4) | 157(54.3)c,d | 53(42.4) | 48(31.2)c | 51(39.8)d | <0.001 |
| BMI, kg/m2, M (IQR) | 23.9(22.0,26.0) | 23.3(21.2,25.3)b,c,d | 24.5(22.7,26.4)b | 24.2(22.4,26.7)c | 24.5(22.5,25.9)d | <0.001 |
| Hepatitis e antigen positivity, n (%) | 113(16.2) | 48(16.6) | 22(17.6) | 26(16.9) | 17(13.3) | 0.784 |
| ALT (IU/L), M(IQR) | 55.0(29.0,142.8) | 48.0(26.0,121.5)c | 60.0(28.0,154.9) | 75.5(36.0,177.3)c | 58.5(27.0,127.9) | 0.023 |
| TBIL (umol/L), M (IQR) | 19.1(13.9,28.8) | 20.6(15.0,32.8)d | 19.2(14.3,30.8) | 18.0(13.8,27.9) | 16.7(12.3,23.7)d | <0.001 |
| Albumin (g/dL), M (IQR) | 39.5(34.6,43.9) | 37.4(32.5,42.1)b,c,d | 40.0(36.8,44.1)b | 41.7(37.0,45.1)c | 41.0(37.0,44.8)d | <0.001 |
| Creatinine, mg/dL, M (IQR) | 76.2(67.3,87.5) | 75.8(65.5,88.0) | 77.0(69.5,88.4) | 78.2(70.0,87.5) | 74.2(65.8,86.5) | 0.194 |
| HbA1c (%), M (IQR) | 7.6(6.8,9.1) | 8.0(7.0,9.8)b | 7.3(6.5,8.2)b | 7.6(6.9,8.9) | 7.6(6.6,9.1) | <0.001 |
| Platelet (109/L), M (IQR) | 120.5(75.3,176.0) | 106.0(64.0,159.5)c,d | 126.0(80.5,176.5) | 148.0(104.5,191.5)c | 131.0(84.8,180.8)d | <0.001 |
| TG (mmol/L), M(IQR) | 1.1(0.8,1.6) | 1.0(0.7,1.5)c | 1.2(0.9,1.6) | 1.3(0.9,1.9)c | 1.2(0.8,1.6) | <0.001 |
| AFP (ng/mL), M (IQR) | 4.8(2.8,11.4) | 5.3(2.8,13.6) | 4.7(2.9,12.7) | 4.1(2.6,8.4) | 5.0(2.6,16.8) | 0.265 |
| log10HBV DNA, M(IQR) | 3.3(1.7,5.3) | 3.2(1.7,5.4) | 3.0(1.7,5.2) | 3.3(1.7,5.4) | 3.7(1.7,5.1) | 0.950 |
| Antiviral drugs, n (%) | 0.004 | |||||
| Entecavir | 470(67.5) | 205(70.9)c | 79(63.2) | 95(61.7)c | 91(71.1) | |
| Tenofovir | 98(14.1) | 48(16.6)c | 16(12.8) | 17(11.0)c | 17(13.3) | |
| Untreated | 128(18.4) | 36(12.5)c | 30(24.0) | 42(27.3)c | 20(15.6) |
Notes: aContinuous data are presented as median (interquartile range) and compared using Kruskal–Wallis and Dunn’s test. Categorical variables were expressed as frequency (percentage) and compared using the chi-square test. Fisher’s test with Bonferroni correction was applied to perform group-wise comparisons. b1 vs. 2, P <0.050; c1 vs. 3, P < 0.050; d 1 vs. 4, P<0.050.
Abbreviations: BMI, body mass index; TG, triglycerides; ALT, alanine aminotransferase; TBIL, total bilirubin; AFP, alpha-fetoprotein.
Associations Between Antidiabetic Drugs and HCC Risk and Death
The cumulative survival curves demonstrated differences between patients treated with different antidiabetic drugs in the risk of HCC and death (Figure 2). Group 1 had a significantly higher risk of both HCC and all-cause mortality compared to the other groups (P value for Log rank test < 0.001).
Figure 2.
Comparisons of HCC (A) and all-cause mortality (B) between participants in Group 1 and other groups.
Abbreviation: HCC, hepatocellular carcinoma.
Multivariable Cox proportional hazards regression models were constructed between different groups, with results being summarized in Tables 2 and 3. In the crude models (Model 1) without any adjustment, compared with no treatment, insulin and insulin secretagogues were associated with an increased risk of HCC (HR = 5.54, 95% CI: 2.22–13.85, P < 0.001) and death (HR = 2.87, 95% CI: 1.60–5.16, P < 0.001). After adjusting for all potential covariates (Model 3), insulin and insulin secretagogues remained associated with increased risk of HCC (HR = 4.42, 95% CI: 1.75–11.17, P = 0.002) and death (HR = 2.46, 95% CI: 1.35–4.50, P = 0.003). In addition, we found that the combination of insulin and/or insulin secretagogues with other antidiabetic drugs does not increase the risk of HCC and death.
Table 2.
Associations Between Antidiabetic Medications and HCC Among the Included Participants (n = 696)
| Exposures | Model1a,b | Model2a,b | Model3a,b,c | |||
|---|---|---|---|---|---|---|
| HR (95% CI) P value | HR (95% CI) P value | HR (95% CI) P value | ||||
| Group2 | 1.00 (Reference) | — | 1.00 (Reference) | — | 1.00 (Reference) | — |
| Group1 | 5.54 (2.22,13.85) | <0.001 | 4.99 (2.00,12.49) | <0.001 | 4.42 (1.75,11.17) | 0.002 |
| Group3 | 1.62 (0.55,4.74) | 0.378 | 1.67 (0.57,4.88) | 0.352 | 1.70 (0.58,5.00) | 0.333 |
| Group4 | 1.00 (0.29,3.44) | 0.995 | 0.96 (0.28,3.31) | 0.944 | 0.91 (0.26,3.13) | 0.874 |
Notes: aModel 1 was the crude Cox proportional hazards regression model without any adjustment. Model 2 was adjusted for age and sex. On the basis of model 2, model 3 was further adjusted for HbA1c, cirrhosis, HBV DNA and antiviral drugs. bNo evidence of proportional hazards assumption violation was observed for the variables and global models (all P values for Schoenfeld residual test > 0.050). cNo evidence of collinearity was observed (maximum variance inflation factor = 3.28).
Abbreviations: HCC, hepatocellular carcinoma; HR, hazard ratio; CI, confidence interval.
Table 3.
Associations Between Antidiabetic Medications and Death Risk Among the Included Participants
| Exposures | Model1a,b | Model2a,b | Model3a,b,c | |||
|---|---|---|---|---|---|---|
| HR (95% CI) P value | HR (95% CI) P value | HR (95% CI) P value | ||||
| Group2 | 1.00 (Reference) | — | 1.00 (Reference) | — | 1.00 (Reference) | — |
| Group1 | 2.87 (1.60,5.16) | <0.001 | 2.59 (1.44,4.67) | 0.001 | 2.46 (1.35,4.50) | 0.003 |
| Group3 | 0.43 (0.17,1.08) | 0.072 | 0.46 (0.18,1.14) | 0.094 | 0.47 (0.19,1.17) | 0.103 |
| Group4 | 1.00 (0.46,2.15) | 0.989 | 0.94 (0.44,2.03) | 0.881 | 0.93 (0.43,2.00) | 0.843 |
Notes: aModel 1 was the crude Cox proportional hazards regression model without any adjustment. Model 2 was adjusted for age and sex. On the basis of model 2, model 3 was further adjusted for HbA1c, cirrhosis, HBV DNA and antiviral drugs. bNo evidence of proportional hazards assumption violation was observed for the variables and global models (all P values for Schoenfeld residual test > 0.050). cNo evidence of collinearity was observed (maximum variance inflation factor = 2.20).
Abbreviations: HCC, hepatocellular carcinoma; HR, hazard ratio; CI, confidence interval.
Subgroup analyses comparing Group 1 versus Group 2 demonstrated differential associations with HCC and all-cause mortality (Figure 3). In the overall comparison, Group 1 showed significantly elevated risks for both HCC (HR = 4.36, 95% CI: 1.71–11.13, P = 0.002) and all-cause mortality (HR = 2.51, 95% CI: 1.37–4.61, P = 0.003) relative to Group 2. A significant interaction by sex was observed for HCC risk (P for interaction = 0.022) and all-cause mortality (P for interaction = 0.012). The association was particularly pronounced in male patients, where Group 1 carried substantially higher risks of HCC (HR = 10.06, 95% CI: 2.40–42.13, P = 0.002) and mortality (HR = 4.77, 95% CI: 1.89–12.04, P = 0.001) compared to Group 2, whereas the result was not statistically significant in female patients. Subgroup analyses revealed no statistically significant interaction effects by either age or cirrhosis status (P for interaction > 0.05 for both outcomes). For age and cirrhosis, the results were broadly consistent with the previous findings.
Figure 3.
Summarized results of subgroup analyses and sensitivity analysis.
Abbreviations: HCC, hepatocellular carcinoma; BMI, body mass index; TG, triglycerides; IPTW, inverse probability of treatment weighting.
Several sensitivity analyses were conducted to further analyze and examine the robustness of the main results (Figure 3). First, participants with follow-up times < 9 months and < 12 months were excluded to reduce inverse causality. The results were generally similar to those in the primary analysis. Second, after further adjustment for TG and BMI, the associations remained statistically significant. Last, after IPTW, acceptable balances were achieved for all covariates, as indicated by the reduction in absolute SMD to within 0.1 (Figure 4). In IPTW-adjusted Cox proportional hazards regression models, insulin and insulin secretagogues remained associated with a higher risk of HCC (HR = 4.52, 95% CI: 1.18–17.30, P = 0.028) and death (HR = 2.90, 95% CI: 1.39–6.06, P = 0.005). The E-value for associations of insulin and insulin secretagogues with HCC and all-cause mortality risk were 8.30 and 4.36, respectively. The above analyses proved the robustness of the findings in the current study.
Figure 4.
Covariate balance after IPTW for the analysis of HCC risk (A) and all-cause mortality (B).
Abbreviations: HCC, hepatocellular carcinoma; IPTW, inverse probability of treatment weighting; AMU, antiviral medication use.
Discussion
This study found that, during a median follow-up of 77.5 and 80 months, patients receiving insulin and/or insulin secretagogues had a significantly higher incidence of HCC and all-cause mortality than other groups. Cumulative hazard curves showed a significantly higher risk of HCC and all-cause mortality in the group compared with the other groups. After adjusting for all potential covariates in multivariable Cox proportional hazards models, the use of insulin and/or insulin secretagogues remained associated with increased risks of HCC and mortality. Moreover, the combination of insulin and/or insulin secretagogues with other AMDs did not increase the risk of HCC or mortality.
Multiple meta-analyses have consistently demonstrated a significant association between the use of insulin and/or insulin secretagogues and an elevated risk of HCC in patients with T2DM.16,17 A study from Taiwan involving 2,903 male patients with HBV infection found that elevated insulin levels were an independent risk factor for HCC among HBV carriers.18 Research by Li et al demonstrated that treatment of T2DM with insulin ± sulfonylurea significantly increased the risk of hepatocarcinogenesis.6 This study is consistent with previous research findings, suggesting that insulin and/or insulin secretagogues may increase the risk of HCC and all-cause mortality. Insulin provides exogenous hormones directly, while secretagogues stimulate β-cells to release endogenous insulin; both directly or indirectly elevate systemic insulin levels. Hyperinsulinaemia may directly affect HCC development and progression by promoting the metabolism, proliferation, and survival of cancer cells.19 Insulin induces the up-regulation of PKM2.20 PKM2 enhances mitochondrial permeability, generating more ATP for cell survival under conditions of nutritional depletion.21 Moreover, the sumoylation of PKM2 induces its plasma membrane targeting and subsequent ectosomal excretion, which promotes HCC by inducing macrophage differentiation and remodeling the tumor microenvironment.22
In addition, we observed that when insulin and insulin secretagogues were used in combination with other non-secretagogues (metformin, Thiazolidinediones, sodium-glucose cotransporter-2 inhibitors (SGLT-2is), GLP-1 receptor agonists (GLP-1RAs), DPP-4i and α-glucosidase inhibitors), the risks of HCC and all-cause mortality did not reach statistical significance. Metformin and Thiazolidinediones have been shown to reduce the risk of HCC in patients with T2DM.23,24 Nishina and Kawaguchi et al proposed that DPP-4i may lower HCC risk by activating lymphocyte chemotaxis and downregulating the pentose phosphate pathway.25,26 Particularly, multiple studies have reported that metformin is associated with a lower incidence of HCC. In this study, 63% of patients in Group 4 were treated with metformin. Research demonstrates that AMPK activation is negatively associated with HCC occurrence in vivo.27 Metformin binds electrostatically to the mitochondrial protein VDAC1, leading to increased cytosolic calcium levels and thereby contributing to autophagy induction.28 Metformin also inhibits HCC progression by upregulating FOXO3, which activates NLRP3 transcription, thereby inducing both apoptosis and pyroptosis.29 Furthermore, a large follow-up study by Giorda et al involving 137,158 patients with T2DM suggested that SGLT-2is, DPP-4is, and GLP-1RAs were strongly associated with a lower frequency of HCC.30 A retrospective cohort study that included 1,890,020 patients with T2DM found that GLP-1RAs were associated with a reduced risk of incident HCC, with hazard ratios of 0.20 [95% CI: 0.14–0.31], 0.39 [95% CI: 0.21–0.69], and 0.63 [95% CI: 0.26–1.50] compared with insulin, sulfonylureas, and metformin, respectively.31 A mouse model showed the improvement in glucose homeostasis induced by the GLP-1RAs could contribute to its suppressive effects on hepatocarcinogenesis.32 A retrospective cohort study conducted in Hong Kong involving 62,699 T2DM patients showed that, SGLT-2is use was associated with a lower risk of HCC compared with DPP-4is use.33 SGLT-2is selectively inhibit glucose uptake by SGLT-2is-expressing human liver cancer cells, reducing intracellular adenosine triphosphate (ATP) levels.34 Investigating the associations between different classes of ADMs and HCC may influence drug selection for T2DM patients. Due to the limited sample size, future large-scale studies are needed to evaluate the impact of combining insulin and/or insulin secretagogues with other antidiabetic drugs on HCC risk.
In the subgroup analyses, no significant interaction was observed for age or cirrhosis, with results consistent with the primary analysis. A significant interaction was found for sex. Among male patients, insulin and/or insulin secretagogues remained significantly associated with an increased risk of both HCC and all-cause mortality. The wide 95% CI (HCC: HR = 10,06, 95% CI: 2.40–42.13, P = 0.002; all-cause mortality: HR = 4.77, 95% CI: 1.89–12.04, P = 0.001) in the male subgroup may be attributable to the limited sample size. In contrast, among female patients, this association did not reach statistical significance. This may be attributed to the protective effect of estrogen.35 Estradiol suppresses hepatic fibrosis by inhibiting the generation of reactive oxygen species, thereby attenuating hepatocyte apoptosis and hepatic stellate cell activation. In addition, the small sample size of female patients (28.6%) may have contributed to the lack of statistical significance. The observed sex difference may also be related to the higher rates of smoking and alcohol use in male patients, both of which are established risk factors for HCC.
Our effect estimate is larger than estimates from larger nationwide cohorts and meta-analyses. Several biases may have contributed to the overestimation of the observed effect. First, there were substantial baseline differences between Group 1 and Group 2. Compared with Group 2, Group 1 had lower BMI, lower albumin levels, and higher HbA1c, which may be attributable to more advanced disease severity. Although we adjusted for cirrhosis status, HbA1c and BMI using multivariable Cox regression and IPTW, residual confounding by unmeasured indicators of liver function reserve (eg., Child-Pugh score) cannot be fully excluded. Second, as Group 2 is healthier at baseline,the choice of Group 2 as reference may inflate the apparent effect size for Group 1. Third, selection bias may arise from our single-center, hospitalized cohort, which may not represent the broader CHB+T2DM population. Last, surveillance bias cannot be excluded. Compared to other groups, Group 1 had a markedly higher prevalence of cirrhosis (54.3%) and may undergo more frequent liver imaging, leading to earlier or more frequent HCC detection.
We also acknowledge several limitations. First, this study was conducted at a single center, which may limit the generalizability of our findings to other populations. Second, due to the limitations of the electronic medical record system, we were unable to definitively determine the exact start time of antidiabetic treatment or the insulin dosage for all patients. Consequently, it is difficult to interpret the temporal relationship between treatment exposure and outcomes, and a dose–response analysis could not be performed. Third, exogenous insulin and agents that simulate endogenous insulin release were combined in the present study. It is important to note that it might introduce potential mechanism heterogeneity. Fourth, sarcopenia frequently occur in patients with T2DM. However, due to the retrospective nature of our study and the lack of relevant data, we were unable to account for these factors. Future prospective studies incorporating comprehensive assessments of metabolism and muscle mass are needed to further explore these associations. Moreover, HBV reactivation has been shown to be a critical risk factor influencing the long-term prognosis of hepatitis B-related HCC patients.36–38 However, our study only utilized baseline data and did not account for dynamic changes in HBV status or reactivation events. Last, Due to the limited sample size and the small number of events, we did not perform a subgroup analysis stratified by antiviral treatment status in patients with CHB. To further explore the relationship between antidiabetic medications and the risk of HCC, future studies should pay attention to the following aspects. First, prospective multicenter studies incorporating detailed treatment data, such as medication duration, and refined classification of ADMs use. Second, future studies should focus on investigating the impact of novel ADMs—such as GLP-1RAs and SGLT-2is—on the incidence of HCC and all-cause mortality. Third, considering that differences in baseline liver disease severity may lead to variations in outcomes, stratified analysis should be performed.
Conclusion
In conclusion, our study suggests that the use of insulin and/or insulin secretagogues was associated with approximately 4.4-fold higher HCC incidence and 2.5-fold higher mortality among in hospitalized Chinese patients with CHB and T2DM. Due to the inherent limitations of our single-center retrospective design, causality cannot be established. Prospective multicenter studies with larger samples are needed to confirm these findings.
Funding Statement
This research was supported by the Nanjing Medical University Changzhou Medical Center Key Project (CMCM202311), Changzhou Science and Technology Bureau Basic Research Project (CJ20210088), and Wujieping Medical Foundation Clinical Research Special Fund (320.6750.2022-06-29).
Abbreviations
ADMs, antidiabetic medications; ALT, alanine aminotransferase; APHE, arterial phase hyperenhancement; BMI, body mass index; CHB, chronic hepatitis B; DPP-4i, DPP-4 inhibitors; EMR, electronic medical records; GLP-1RAs, GLP-1 receptor agonists; HbA1c, Glycated Hemoglobin A1c; FPG, Fasting Plasma Glucose; HCC, hepatocellular carcinoma; HBV, hepatitis B virus; HR, hazard ratio; IPTW, inverse probability of treatment weighting; IQR, interquartile range; OGTT, oral glucose tolerance test; SGLT-2is, sodium-glucose cotransporter-2 inhibitors; TBIL, total bilirubin; TG, triglyceride; T2DM, type 2 diabetes mellitus; 95% CI, 95% confidence interval; CPH, Cox proportional hazards regression models; VIF, variance inflation factor; SMD, standardized mean difference; AMU, antiviral medication use.
Data Sharing Statement
The data that support the findings of this study are available in the main manuscript. Additional data are available from the corresponding author upon reasonable request.
Institutional Review Board Statement
The protocol for this research project was approved by the Ethics Committee of Changzhou Third People’s Hospital (approval code 02A-A2024033) and conforms to the provisions of the Declaration of Helsinki. Due to the retrospective nature of the study and the use of anonymized clinical data, the requirement for written informed consent was waived by the Ethics Committee.
Author Contributions
All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.
Disclosure
The authors report no conflicts of interest in this work.
References
- 1.Jeng WJ, Papatheodoridis GV, Lok ASF. Hepatitis B. Lancet. 2023;401(10381):1039–11. doi: 10.1016/S0140-6736(22)01468-4 [DOI] [PubMed] [Google Scholar]
- 2.Watson AG, Mulay AS, Gill US. Chronic hepatitis B in 2025: diagnosis, treatment and future directions. Clin Med. 2025;25(6):100527. doi: 10.1016/j.clinme.2025.100527 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Mauro E, de Castro T, Zeitlhoefler M, et al. Hepatocellular carcinoma: epidemiology, diagnosis and treatment. JHEP Rep. 2025;7(12):101571. doi: 10.1016/j.jhepr.2025.101571 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Harreiter J, Roden M. Diabetes mellitus – definition, Klassifikation, Diagnose, Screening und Prävention (Update 2023) [Diabetes mellitus: definition, classification, diagnosis, screening and prevention (Update 2023)]. Wien Klin Wochenschr. 2023;135(suppl 1):7–17. doi: 10.1007/s00508-022-02122-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Tan Y, Wei S, Zhang W, Yang J, Yang J, Yan L. Type 2 diabetes mellitus increases the risk of hepatocellular carcinoma in subjects with chronic hepatitis B virus infection: a meta-analysis and systematic review. Cancer Manag Res. 2019;11:705–713. doi: 10.2147/CMAR.S188238 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Li MY, Li TT, Li KJ, Zhou C. Type 2 diabetes mellitus characteristics affect hepatocellular carcinoma development in chronic hepatitis B patients with cirrhosis. World J Clin Cases. 2023;11(5):1009–1018. doi: 10.12998/wjcc.v11.i5.1009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kunutsor SK, Balasubramanian VG, Zaccardi F, et al. Glycaemic control and macrovascular and microvascular outcomes: a systematic review and meta-analysis of trials investigating intensive glucose-lowering strategies in people with type 2 diabetes. Diabetes Obes Metab. 2024;26(6):2069–2081. doi: 10.1111/dom.15511 [DOI] [PubMed] [Google Scholar]
- 8.Luo Y, Li R, Kang J, et al. Chronic hepatitis B with type 2 diabetes mellitus: association between glycemic control and liver fibrosis. World J Diabetes. 2025;16(7):107256. doi: 10.4239/wjd.v16.i7.107256 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Singh S, Singh PP, Singh AG, Murad MH, Sanchez W. Anti-diabetic medications and the risk of hepatocellular cancer: a systematic review and meta-analysis. Am J Gastroenterol. 2013;108(6):881–892. doi: 10.1038/ajg.2013.5 [DOI] [PubMed] [Google Scholar]
- 10.Yen FS, Lai JN, Wei JC, et al. Is insulin the preferred treatment in persons with type 2 diabetes and liver cirrhosis? BMC Gastroenterol. 2021;21(1):263. doi: 10.1186/s12876-021-01773-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lee JY, Jang SY, Nam CM, Kang ES. Incident hepatocellular carcinoma risk in patients treated with a sulfonylurea: a nationwide, nested, case-control study. Sci Rep. 2019;9(1):8532. doi: 10.1038/s41598-019-44447-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Simon TG, King LY, Chong DQ, et al. Diabetes, metabolic comorbidities, and risk of hepatocellular carcinoma: results from two prospective cohort studies. Hepatology. 2018;67(5):1797–1806. doi: 10.1002/hep.29660 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Plaz Torres MC, Jaffe A, Perry R, Marabotto E, Strazzabosco M, Giannini EG. Diabetes medications and risk of HCC. Hepatology. 2022;76(6):1880–1897. doi: 10.1002/hep.32439 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.ElSayed NA, Aleppo G, Bannuru RR. American diabetes association professional practice committee. 2. diagnosis and classification of diabetes: standards of care in Diabetes—2024. Diabetes Care. 2024;47(suppl 1):S20–S42. doi: 10.2337/dc24-S002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Singal AG, Llovet JM, Yarchoan M, et al. AASLD Practice Guidance on prevention, diagnosis, and treatment of hepatocellular carcinoma. Hepatology. 2023;78(6):1922–1965. doi: 10.1097/HEP.0000000000000466 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Arvind A, Memel ZN, Philpotts LL, Zheng H, Corey KE, Simon TG. Thiazolidinediones, alpha-glucosidase inhibitors, meglitinides, sulfonylureas, and hepatocellular carcinoma risk: a meta-analysis. Metabolism. 2021;120:154780. doi: 10.1016/j.metabol.2021.154780 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Zhou YY, Zhu GQ, Liu T, et al. Systematic review with network meta-analysis: antidiabetic medication and risk of hepatocellular carcinoma. Sci Rep. 2016;6:33743. doi: 10.1038/srep33743 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Chao LT, Wu CF, Sung FY, et al. Insulin, glucose and hepatocellular carcinoma risk in male hepatitis B carriers: results from 17-year follow-up of a population-based cohort. Carcinogenesis. 2011;32(6):876–881. doi: 10.1093/carcin/bgr058 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Chettouh H, Lequoy M, Fartoux L, Vigouroux C, Desbois-Mouthon C. Hyperinsulinaemia and insulin signalling in the pathogenesis and the clinical course of hepatocellular carcinoma. Liver Int. 2015;35(10):2203–2217. doi: 10.1111/liv.12903 [DOI] [PubMed] [Google Scholar]
- 20.Liu J, Zhi Q, Liu Y, et al. Insulin promotes hepatocarcinoma tumorigenesis by up-regulating PKM2 expression. Exp Cell Res. 2021;408(2):112872. doi: 10.1016/j.yexcr.2021.112872 [DOI] [PubMed] [Google Scholar]
- 21.Qi H, Ning X, Yu C, et al. Succinylation-dependent mitochondrial translocation of PKM2 promotes cell survival in response to nutritional stress. Cell Death Dis. 2019;10(3):170. doi: 10.1038/s41419-018-1271-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hou PP, Luo LJ, Chen HZ, et al. Ectosomal PKM2 Promotes HCC by inducing macrophage differentiation and remodeling the tumor microenvironment. Mol Cell. 2020;78(6):1192–1206.e10. doi: 10.1016/j.molcel.2020.05.004 [DOI] [PubMed] [Google Scholar]
- 23.Kramer JR, Natarajan Y, Dai J, et al. Effect of diabetes medications and glycemic control on risk of hepatocellular cancer in patients with nonalcoholic fatty liver disease. Hepatology. 2022;75(6):1420–1428. doi: 10.1002/hep.32244 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Chen Y, Mushashi F, Son S, Bhatti P, Dummer T, Murphy RA. Diabetes medications and cancer risk associations: a systematic review and meta-analysis of evidence over the past 10 years. Sci Rep. 2023;13(1):11844. doi: 10.1038/s41598-023-38431-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Nishina S, Yamauchi A, Kawaguchi T, et al. Dipeptidyl peptidase 4 inhibitors reduce hepatocellular carcinoma by activating lymphocyte chemotaxis in mice. Cell Mol Gastroenterol Hepatol. 2018;7(1):115–134. doi: 10.1016/j.jcmgh.2018.08.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Kawaguchi T, Nakano D, Koga H, Torimura T. Effects of a DPP4 Inhibitor on Progression of NASH-related HCC and the p62/Keap1/Nrf2-Pentose Phosphate Pathway in a Mouse Model. Liver Cancer. 2019;8(5):359–372. doi: 10.1159/000491763 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Yang X, Liu Y, Li M, et al. Predictive and preventive significance of AMPK activation on hepatocarcinogenesis in patients with liver cirrhosis. Cell Death Dis. 2018;9(3):264. doi: 10.1038/s41419-018-0308-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Ko M, Kim J, Lazim R, et al. The anticancer effect of metformin targets VDAC1 via ER-mitochondria interactions-mediated autophagy in HCC. Exp Mol Med. 2024;56(12):2714–2725. doi: 10.1038/s12276-024-01357-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Shen Z, Zhou H, Li A, et al. Metformin inhibits hepatocellular carcinoma development by inducing apoptosis and pyroptosis through regulating FOXO3. Aging. 2021;13(18):22120–22133. doi: 10.18632/aging.203464 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Giorda CB, Picariello R, Tartaglino B, et al. Hepatocellular carcinoma in a large cohort of type 2 diabetes patients. Diabet Res Clin Pract. 2023;200:110684. doi: 10.1016/j.diabres.2023.110684 [DOI] [PubMed] [Google Scholar]
- 31.Wang L, Berger NA, Kaelber DC, Xu R. Association of GLP-1 Receptor Agonists and Hepatocellular Carcinoma Incidence and Hepatic Decompensation in Patients With Type 2 Diabetes. Gastroenterology. 2024;167(4):689–703. doi: 10.1053/j.gastro.2024.04.029 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Kojima M, Takahashi H, Kuwashiro T, et al. Glucagon-like peptide-1 receptor agonist prevented the progression of hepatocellular carcinoma in a mouse model of nonalcoholic steatohepatitis. Int J Mol Sci. 2020;21(16):5722. doi: 10.3390/ijms21165722 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Chou OHI, Ning J, Chan RNC, et al. Lower risks of new-onset hepatocellular carcinoma in patients with type 2 diabetes mellitus treated with sglt2 inhibitors versus DPP4 Inhibitors. J Natl Compr Canc Netw. 2024;22(2 D):e237118. doi: 10.6004/jnccn.2023.7118 [DOI] [PubMed] [Google Scholar]
- 34.Ahmed H, Gomaa S, Abdulrazzak E, Alabdul Razzak I, Kovalovich KK. Therapeutic Potential of Sodium-Glucose Cotransporter 2 (SGLT2) Inhibitors in Liver Disease: focus on Cirrhosis. Cureus. 2025;17(5):e84768. doi: 10.7759/cureus.84768 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Shimizu I, Kohno N, Tamaki K, et al. Female hepatology: favorable role of estrogen in chronic liver disease with hepatitis B virus infection. World J Gastroenterol. 2007;13(32):4295–4305. doi: 10.3748/wjg.v13.i32.4295 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Li H, Lu D, Chen J, et al. Post-transplant hepatitis B virus reactivation impacts the prognosis of patients with hepatitis B-related hepatocellular carcinoma: a dual-centre retrospective cohort study in China. Int J Surg. 2024;110(4):2263–2274. doi: 10.1097/JS9.0000000000001141 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Du Z, Lai Z, Huang Y, et al. HBV reactivation in patients with hepatocellular carcinoma treated with pd-1/-l1 antibodies and concurrent antiviral prophylaxis agents: a prospective observational study. Clin Cancer Res. 2026;32(4):705–714. doi: 10.1158/1078-0432.CCR-25-2859 [DOI] [PubMed] [Google Scholar]
- 38.Sun LY, Wang NY, Diao YK, et al. Comparison between models for detecting hepatocellular carcinoma in patients with chronic liver diseases of various etiologies: ASAP score versus GALAD score. Hepatobiliary Pancreat Dis Int. 2025;24(4):412–422. doi: 10.1016/j.hbpd.2023.12.004 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available in the main manuscript. Additional data are available from the corresponding author upon reasonable request.