We read with interest the study by Goh and Sinn [1], which demonstrated the potential chemopreventive effects of aspirin and statin in hepatocellular carcinoma (HCC) through reducing inflammation, proliferation and promoting apoptosis. Similarly, dipeptidyl peptidase-4 inhibitors (DPP4i) were suggested to have potential in anticancer treatment, in addition to their effect on diabetes mellitus (DM). DPP4/CD26 expression in HCC is upregulated and associated with poorer survival [2]. Through metabolic modulation and improving intrahepatic inflammatory cell infiltration, DPP4i may suppress carcinogenesis and enhance response to immunotherapy, the new standard of care in advanced HCC [3]. The effects of oral hypoglycemic agents such as metformin on HCC survival have been explored previously [4], but studies on DPP4i use in advanced HCC remain limited. Thus, we investigated the impact of DPP4i on the survival of patients who had DM and advanced HCC undergoing immunotherapy in a real-world setting.

We identified 451 patients (mean age 62.3 years, 82.7% males) with advanced HCC who had received at least 1 dose of immunotherapy from July 1, 2014, to December 31, 2022, in Hong Kong and Royal Prince Alfred Hospital, Australia (Supplementary Table 1). The main etiology of HCC was chronic viral hepatitis (n=331, 78.1%). 169 (37.5%) patients had DM and of which, 38 (22.5%) were treated with DPP4i. During a median follow-up of 26.8 (interquartile range [IQR] 23.8–29.9) months, 299 patients (66.3%) died. All-cause mortality of non-DM patients, non-DPP4i DM patients and DPP4i patients were 181/282 (64.2%), 95/131 (72.5%) and 23/38 (60.5%). Cancer was the leading cause of death in all subgroups, with cancer-related mortality of 143/282 (50.7%), 72/131 (55.0%) and 17/38 (44.7%) in non-DM patients, non-DPP4i DM patients and DPP4i patients.

Compared to non-DPP4i DM patients, non-DM patients (hazard ratio [HR]=0.77 [95% confidence interval 0.60–0.99], P=0.041) and users of DPP4i (HR=0.62 [0.40–0.99], P=0.043) were associated with improved survival. After adjusting for patient demographics (age, sex) and pathologic variables (HCC etiology, albumin-bilirubin [ALBI]-grade, alpha fetoprotein [AFP], platelet, alanine aminotransferase insulin use), the survival advantage of DDP4i users remained similar (HR=0.62 [0.38–1.00], P=0.049). However, the use of metformin (HR=0.82 [0.57–1.18], P=0.823) or statin (HR=0.87 [0.60–1.26], P=0.473) was not associated with a better prognosis (Table 1). The median overall survival (OS) of non-DM patients, non-DPP4i DM patients, and DPP4i users were 11.6 (IQR: 7.9–15.4), 8.1 (6.0–10.2), 14.6 (9.1–20.1) months, respectively (P=0.046). The similar HR and median OS of non-DM patients and DPP4i users suggest the groups had comparable survival.

Our results may be explained by the survival disadvantage of DM patients from their multiple comorbidities and immune dysregulation. Non-DPP4i DM patients had poorer OS when compared to other subgroups in our analyses, which aligns with the meta-analysis conducted by Wang et al. [5] DM promotes carcinogenesis through multiple mechanisms. Hyperinsulinemia and hyperglycemia lead to chronic inflammation and oxidative stress, which subsequently contribute to genomic instability and carcinogenesis [6]. Moreover, DM is associated with immune dysfunction and reduced efficacy of immunotherapy. Furthermore, DM patients with metastatic non-small cell lung cancer receiving pembrolizumab are associated with poorer survival [7]. Via reducing obesity-induced inflammation and insulin resistance, DPP4i may act as a metabolic modulator that attenuates the progression of HCC [2].

DPP4 inhibition modulates the tumor microenvironment and is shown to augment immunotherapy response in mouse melanoma models by preserving the cognate receptor C-X-C motif chemokine receptor 3 and C-X-C motif chemokine 10 (CXCR3-CXCL10) axis for lymphocyte chemotaxis [8]. This is further supported by Huang et al. [9] who demonstrated that HCC tissues from patients who received DPP4i had higher CD8+ T cell infiltration. Additionally, DPP4i inhibits angiogenesis and exerts anti-tumor effect in a dose-dependent manner in murine HCC models [2]. Through acting on CXCL12, DPP4i also has the potential to regulate tumor growth and metastasis [2]. Use of DPP4i has also been reported to improve survival in colorectal and lung cancer patients [10]. In our clinical validation, the anti-tumor and immunomodulating effects of DPP4i are possibly canceled out by the carcinogenic effects of diabetes, resulting in similar survival between non-DM patients and DPP4i users.

There are several limitations in our study. First, the cohort’s retrospective nature predisposes to residual confounding. The impact of DPP4i is confounded by exposure to other anti-diabetic agents because DPP4i are usually administered as an add-on therapy for better glycemic control. Due to the limited sample size of the cohort, subgroup analysis on DPP4i monotherapy and combination therapy could not be performed. Second, data on cancer staging and response rates were unavailable, hence ALBI grading and serum AFP that may reflect disease severity were used as covariates in multivariable analysis. Third, the dosage and the concomitant or sequential use of DPP4i, insulin, and other oral hypoglycaemic agents were not well controlled.

The use of DPP4i is associated with improved survival in diabetic HCC patients receiving immunotherapy. DPP4i may overcome the survival disadvantage of HCC patients with diabetes and offer comparable survival to non-DM patients. The findings should be further validated in prospective studies with larger sample sizes.