Mendelian randomization revealing the protective effect of sodium-glucose cotransporter 2 inhibition on prostate cancer with verified evidence from electronic healthcare and biological data

Abstract

AbstractBackgroundObservational studies indicated a decreased risk of prostate cancer by SGLT2 inhibitors, but high-quality evidence is lacking to make a clear conclusion. We evaluated the effect of SGLT2 inhibition on prostate cancer risk by triangulating evidence from three methods.MethodsGenetic variants associated with HbA1clevels (P<5×10-8) in the genomic region of the target gene,SLC5A2, were used to proxy SGLT2 inhibition. In discovery, Mendelian randomization (MR) was applied to estimate effects of genetically proxied SGLT2 inhibition on risk of prostate cancer and its subtypes (79,148 cases and 61,106 controls). In a validation using electronic healthcare data, the association of incidence of prostate cancer between 24,155 new users of SGLT2 inhibitors and 24,155 new users of the active comparator, dipeptidyl peptidase 4 inhibitors, was estimated using electronic health-care data. In a biological validation, the differential gene expression ofSLC5A2between normal prostate tissue and tumour tissue were estimated in 691 prostate cancer patients. To validate the influence of glucose, the association between HbA1clevels and incident prostate cancer during 10-years of follow-up were estimated.FindingsFor genetic evidence, genetically proxied SGLT2 inhibition reduced the risk of overall (odds ratio=0.56, 95%CI=0.38 to 0.82), advanced (OR=0.52, 95%CI=0.27 to 0.99) and early-onset (OR=0.27, 95%CI=0.11 to 0.72) prostate cancer. For electronic healthcare evidence, usage of SGLT2 inhibitor was associated with a 23% reduced risk of prostate cancer (hazard ratio=0.77, 95%CI=0.61 to 0.99) in males with diabetes. For biological evidence, expression levels of theSLC5A2gene in tumour prostate tissue was 2.02-fold higher than that in normal tissue (P=0.006). Genetically proxied HbA1cand observed HbA1cprovided little evidence to support an association with total/incident prostate cancer, implies a non-glycemic effect of SGLT2 inhibition on prostate cancer.InterpretationThis study provides genetic, electronic healthcare, and biological evidence to support a beneficial effect of SGLT2 inhibition on reducing prostate cancer risk. Future trials are warranted to investigate whether SLGT2 inhibitors can been recommended for diabetic individuals with high risk of prostate cancer or considered as an anti-prostate cancer therapy.FundingAMS, MRC and NSFC.Graphical abstractResearch in contextEvidence before this studyWe searched PubMed, Embase andclinicaltrials.govdatabases from inception up to July 11, 2023 using the search terms: “SGLT2 inhibitor”, “canagliflozin”, “dapagliflozin”, or “empagliflozin” and “prostate cancer” and “clinical trials”, without language restrictions. Some functionals studies provided evidence of SGLT2 inhibition on reduce the viability of prostate cancer cells but lack of human-based evidence. Only one clinical trial study is investigating the role of SGLT2 inhibitors on prostate cancer in individuals with diabetes. Other 46 trials of SGLT2 inhibitors set prostate cancer as secondary outcome, the prostate cancer cases were limited for these studies, power issues have prevented clear causal inference. Little has been done to establish the causal role of SGLT2 inhibition on total and incident prostate cancer.Added value of this studyIn this study, Mendelian randomization (MR) was applied in 140,254 men (79,148 with prostate cancer), and suggested that genetically proxied SGLT2 inhibition showed an effect on 44%, 48% and 73% reduced risk of total-, advanced- and early-onset prostate cancer in the general male population. Validation analysis using electronic health-care record data (81,122 men with diabetes) suggested that usage of SGLT2 inhibitor was associated with a 23% reduced risk of prostate cancer in males with diabetes. The differential expression analysis in 639 men with prostate cancer showed that the expression of SGLT2 was 2.02 folds higher in prostate cancer tissues compared with that in surrounding normal prostate tissues. As a benchmark, MR and observational analyses showed little evidence to support an effect of HbA1c on prostate cancer, which suggests a potential non-HbA1ceffect of SGLT2 inhibition on prostate cancer.Implication of all the available evidenceThere were multiple sources of evidence to support a protective role of genetically proxied SGLT2 inhibition and usage of SGLT2 inhibitors on risk of prostate cancer in men with and without diabetes and/or prostate cancer. Future clinical trials should be prioritised for investigation of the long-term use of SGLT2 inhibitors in the prevention and treatment of prostate cancer.