INTRODUCTION

Cardiovascular disease remains a leading cause of morbidity and mortality worldwide. Statins, which are inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, are fundamental in the prevention of atherosclerotic cardiovascular disease (ASCVD) [1,2]. Statin usage has increased substantially due to aging populations and clinical guidelines that emphasize early intervention to prevent ASCVD [3]. Although statins are generally safe and effective, they have been associated with a modest increase in the risk of type 2 diabetes mellitus (T2DM), which itself represents a risk factor for ASCVD [4–7]. Consequently, it should be a public health priority to understand and optimize statin therapy in the context of diabetes risk.

The association between statin use and the risk of T2DM has been demonstrated in multiple studies, including large-scale prospective cohort studies and randomized controlled trials [4,6–16]. In the JUPITER trial, rosuvastatin at a dose of 20 mg was associated with an approximately 30% increased risk of developing T2DM compared to placebo (odds ratio [OR], 1.28; 99% confidence interval [CI], 1.07 to 1.54) [4]. A meta-analysis of randomized controlled trials similarly indicated that statin therapy is associated with an approximately 10% increased risk of new-onset T2DM [7]. In the METSIM cohort study, the use of statins in participants without baseline diabetes was linked to a 46% higher risk of T2DM over 6 years (hazard ratio [HR], 1.46; 95% CI, 1.22 to 1.74). The T2DM risk increases in proportion to the statin dose, and statin use is also associated with reduced insulin sensitivity [13,17,18].

The mechanisms underlying the diabetogenic effects of statins are not fully understood. However, recent genetic studies have highlighted the role of genetic susceptibility in the increased T2DM risk associated with statin use [19,20]. Certain alleles associated with lowering low-density lipoprotein cholesterol (LDL-C), particularly in genes such as HMGCR and NPC1L1, have been linked to an elevated risk of T2DM. The inhibition of the gene encoding HMGCR, a primary target of statins, has been connected to impaired glucose tolerance and an increased risk of T2DM. This may stem from direct or indirect effects on β-cell calcium channels. Each additional allele of HMGCR rs17238484-G is associated with a 0.06 mmol/L decrease in LDL-C and a 2% increase in T2DM risk, accompanied by increases in body weight, waist circumference, plasma insulin, and glucose levels [21]. Similarly, the HMGCR rs12916-T allele is associated with a 6% increased risk of T2DM and has comparable effects on LDL-C, body weight, and waist circumference [22]. Therefore, understanding whether genetic susceptibility modifies the diabetogenic effect of statins is crucial for precision prevention. This knowledge may assist clinicians in balancing the benefits of ASCVD prevention against potential diabetes risk.

Polygenic risk scores (PRS), which aggregate genetic risk across multiple loci to quantify cumulative genetic predisposition, have the potential to support a tailored assessment of T2DM risk in statin users. This study aimed to evaluate the collective influence of genetic predisposition, as quantified by T2DM PRS, and regular statin use on T2DM risk. Using a propensity score-matched approach to control for factors affecting statin use, we investigated the interaction between statin use and T2DM PRS in determining T2DM risk.

METHODS

Study Population

UK Biobank

This research was conducted using the UK Biobank Resource under Application No. 87860.

The UK Biobank is a large prospective cohort study that collects clinical, genetic, and lifestyle data from over 500 000 individuals aged 40–69 years. These participants were recruited from 22 assessment centers across the United Kingdom between 2006 and 2010 [23]. At the time of enrollment, participants completed a touchscreen questionnaire detailing their medical history and current medications. This was followed by an interview with trained health professionals. Medications were categorized by type, such as statins, insulin, metformin, and aspirin. Anthropometric measurements were taken, including body mass index (BMI) determined via bioimpedance. Blood samples were also collected for biochemical analyses, including LDL-C, high-density lipoprotein cholesterol (HDL-C), glucose, and hemoglobin A1c (HbA1c) levels. Health outcome data were linked to electronic health records and included information from primary care, hospital inpatient, and death registry data.

Exclusion criteria

Participants with baseline diagnoses of major cardiovascular diseases, as indicated by International Classification of Diseases, 10th revision (ICD-10) codes, were excluded from the analysis (n=35 030). These conditions encompassed ischemic heart diseases (I20–I25), cerebrovascular diseases (I63–I66), and transient ischemic attacks (G45–G46). Individuals diagnosed with diabetes (n=26 304), participants with HbA1c levels above 48 mmol/mol (6.5%, n=3071), and those reporting the use of insulin or biguanides (n=287) were also omitted. Furthermore, the study excluded individuals with missing data for key covariates, including BMI (n=6191), polygenic scores for T2DM (n=3968), Townsend deprivation index, and HDL-C levels (Figure 1).

Ultimately, the study population comprised 333 840 participants who did not have baseline T2DM or ASCVD and for whom complete data on relevant covariates were available. Propensity score matching (PSM) was performed to facilitate subsequent analyses (Table 1).

Definition of Study Variables

Disease status

The disease status of participants was established using the UK Biobank’s First Occurrence data, which integrates self-reported diagnoses, electronic health records (including inpatient and primary care data), and death registry information to ascertain the initial diagnosis date by ICD-10 code. Baseline diabetes status, categorized as type 1, type 2, or unclassified, was documented. Incident T2DM was identified using the first recorded date of non–insulin-dependent diabetes mellitus (ICD-10 code E11; data fields 130708, 130709). The analysis utilized the time to first occurrence of T2DM. Baseline dyslipidemia status was determined using the first recorded date of disorders of lipoprotein metabolism and other lipidemias (ICD-10 code E78; data field 130814). Baseline hypertension status was defined using a composite of essential hypertension, hypertensive heart disease, hypertensive renal disease, hypertensive heart and renal disease, and secondary hypertension (ICD-10 codes I10, I11, I12, I13, I15; data fields 131286, 131288, 131290, 131292, 131294). Chronic kidney disease status was determined based on the first diagnosed date of chronic renal failure (ICD-10 code N18; data field 132032).

Medication status

Medication use was evaluated at the initial assessment center visit. Information on prescribed statins and other medications was coded according to Read code version 2, British National Formulary, and dictionary of medicines and devices standards and mapped to the UK Biobank lookup table to ensure consistency. Participants who reported using statins at baseline were classified as regular statin users.

Polygenic scores

We utilized pre-calculated PRS values for T2DM, provided by Genomics PLC [24]. The standard PRS information, available upon request from the UK Biobank, is derived from external genome-wide association studies (GWAS) to ensure independence from the UK Biobank’s own cohort data. The PRS calculation method involved conducting a fixed-effect inverse variance meta-analysis of external GWAS. The standard PRS set includes scores for all UK Biobank participants, promoting broad applicability. For T2DM, the PRS demonstrated an area under the curve of approximately 0.62, indicating moderate predictive performance in this population. Further details regarding PRS are documented in the referenced source [24]. Access to the UK Biobank PRS Release is available through an application to the UK Biobank’s Research Analysis Platform ( https://biobank.ndph.ox.ac.uk/ukb/label.cgi?id=300).

Confounders

Potential confounders were selected based on evidence from the existing literature to consider factors associated with both regular statin use and the risk of T2DM. Demographic and socio-demographic characteristics included age, sex, ethnicity, annual household income, and education level. Lifestyle factors consisted of smoking status and alcohol use at baseline. Clinical and biomarker variables included BMI, HbA1c, LDL-C, HDL-C, and triglycerides. The baseline comorbid conditions considered were dyslipidemia, hypertension, chronic kidney disease, and a derived Charlson comorbidity index. Medication use was incorporated as the total number of medications taken at baseline, as well as aspirin use. Additionally, prediabetes was evaluated as a candidate covariate, defined as an HbA1c level exceeding 40 mmol/mol but less than 50 mmol/mol [25–28].

To identify potential predictors, we initially assessed the univariate association of covariates with baseline statin use. Subsequent multivariable logistic regression modeling was conducted, with feature selection informed by F scores using the Python Scikitlearn SelectKBest package. Pairwise correlations between variables were evaluated, and variables with a correlation coefficient around 0.7 were considered highly correlated. Only 1 variable from each highly correlated pair was retained for the model. Specifically, LDL-C and total cholesterol, systolic blood pressure and diastolic blood pressure, and the Charlson comorbidity index and baseline age exhibited high correlations, leading to the inclusion of only 1 representative variable from each pair in the final multivariable analysis.

Statistical Analysis

To ensure balanced baseline characteristics between statin users and non-users, we applied PSM. The variables incorporated into the matching model included age, sex, ethnicity, education, Townsend deprivation index, BMI, HDL-C, LDL-C, triglycerides, diastolic blood pressure, baseline hypertension, baseline dyslipidemia, prediabetes status, and the number of medications taken at baseline. PSM was executed at a 1:5 ratio, employing a caliper width of 0.2 pooled standard deviations using the MatchIt package in R (R Foundation for Statistical Computing, Vienna, Austria) [29].

The demographic characteristics of statin users and non-users were compared both before and after matching (Table 1). The crude incidence rates of T2DM, expressed as events per 1000 person-years across PRS strata, were calculated and visualized (Figure 2A).

In the matched cohort, a multivariable Cox proportional hazards regression model was employed to assess the relationship between statin use, PRS, and the risk of T2DM. This model incorporated weights derived from the variable ratio matching process (Table 2). Right-censoring was applied at the earliest occurrence of loss to follow-up, the end of the follow-up period (July 1, 2023), or death. The proportional hazards assumption was tested using Schoenfeld residuals (Supplemental Material 1). For the stratified analyses, separate models were fitted for the subgroups of statin users and non-users, allowing us to examine how the association between T2DM PRS and T2DM risk varies across these populations. These models were adjusted for potential confounders, including age, sex, ethnicity, HbA1c level, and BMI. PRS was categorized into 3 groups: <30, 30–70, and >70%. The middle category (30–70%) served as the reference group for estimating ordinal effects. The significance of these interactions was evaluated using likelihood ratio tests. In the combined model, the interaction between statin use and T2DM was assessed by including terms representing the combinations of PRS categories and statin use as joint variables. The model was additionally adjusted for age, sex, ethnicity (white vs. non-white), BMI, and HbA1c level. Furthermore, we validated the interaction effect using unmatched data and through propensity score-matched analyses. The results of the propensity-score-adjusted Cox proportional hazard model and a Cox proportional hazard model that adjusted for all variables used in the PSM are presented in Supplemental Materials 2–4.

Ethics Statement

This study was approved by the Institutional Review Board of Seoul National University (IRB No. E2309/002-013).

RESULTS

After PSM, we identified a cohort comprising 12 366 statin users and 30 797 non-users, with an estimated effective sample size of 18 822.1 (Table 1). The baseline characteristics of the matched cohort, stratified by statin use, were generally well-balanced across key variables. The mean age of statin users was 60.05 years, which closely matched that of non-users at 60.31 years. The sex distribution was similar between groups, with male participants comprising 54.0% of statin users and 55.1% of non-users. Socioeconomic indicators, including Townsend deprivation index and education level, displayed no significant differences between groups. Body composition, as measured by BMI, was also comparable (28.35 kg/m² for statin users vs. 28.45 kg/m² for non-users). Most participants in both groups identified as white (statin users, 94.5%; non-users, 94.5%). The prevalence of hypertension was slightly lower among statin users (50.5%) than non-users (53.5%). The prevalence of prediabetes was nearly identical between groups, at 14.3% for statin users and 14.3% for non-users. However, statin users exhibited a moderately higher prevalence of dyslipidemia (53.8%) compared to non-users (47.8%). The covariate balance, assessed through absolute standardized mean differences, indicated that most variables were well matched. Detailed balance metrics are provided in Supplemental Material 5.

Figure 2B presents survival curves depicting the cumulative probability of developing new-onset T2DM over time. These curves are stratified by regular statin use and T2DM PRS group, defined as low, intermediate, and high (<30, 30–70, and >70%, respectively), within the matched population. The curves were generated using a weighted Cox proportional hazards model adjusted for age, sex, ethnicity, BMI, and HbA1c level to evaluate differences in T2DM risk across subgroups. The findings reveal a clear gradient in T2DM risk according to PRS group and statin use. Specifically, in the high-PRS group, statin users displayed a significantly higher risk of T2DM compared to non-users (HR, 1.22; 95% CI, 1.02 to 1.46; p=0.031). Among those with intermediate PRS, statin users had a lower hazard of developing T2DM than non-users. In contrast, no significant interaction was observed in the low-PRS group (HR, 0.88; 95% CI, 0.70 to 1.10, p=0.258).

Table 2 presents the results of the Cox proportional hazards models assessing the interaction between regular statin use and T2DM PRS group on the hazard of developing T2DM. In the combined model, statin users with a high PRS faced the highest hazard of T2DM, exhibiting a 56% increased risk compared to the reference group, which consisted of statin non-users with an intermediate PRS (HR, 1.56; 95% CI, 1.37 to 1.78; p<0.001). Statin non-users with a high PRS also displayed a significantly elevated hazard (HR, 1.30; 95% CI, 1.13 to 1.49; p<0.001). In the intermediate-PRS group, regular statin use did not significantly impact T2DM risk (HR, 1.00; 95% CI, 0.88 to 1.14; p=0.984). Conversely, statin non-users in the low-PRS group had a reduced hazard (HR, 0.75; 95% CI, 0.65 to 0.87; p<0.001). Regular statin users in the low-PRS group demonstrated the greatest reduction in T2DM hazard, with a 37% lower risk compared to the reference group (HR, 0.63; 95% CI, 0.53 to 0.74; p<0.001).

In the stratified analyses, statin users with a high PRS exhibited a significantly higher hazard of T2DM compared to their counterparts with an intermediate PRS (HR, 1.17; 95% CI, 1.03 to 1.33; p=0.016). However, statin users in the low-PRS group exhibited no significant difference (HR, 0.86; 95% CI, 0.71 to 1.05; p=0.140). Among individuals not using statins, the T2DM hazard was highest in the high-PRS group (HR, 1.30; 95% CI, 1.13 to 1.49; p<0.001). Conversely, those in the low-PRS group had a significantly lower hazard compared to the intermediate-PRS group (HR, 0.75; 95% CI, 0.65 to 0.87; p<0.001).

DISCUSSION

In this study, leveraging genetic data from the large-scale UK Biobank cohort, we applied PSM and Cox proportional hazards modeling to explore the interaction between T2DM PRS and regular statin use regarding the risk of T2DM. Regular statin use was linked to an increased hazard of T2DM exclusively in individuals with a high T2DM PRS. Thus, genetic predisposition may amplify the diabetogenic effects of statins.

Our findings align with previous research on the diabetogenic impact of statins and the modifying role of genetic predisposition. Mendelian randomization studies have consistently shown that genetic proxies for the low-density lipoprotein (LDL)-lowering effects of statins are linked to an increased risk of T2DM [21,22,30–32]. Notably, a study by Swerdlow et al. [21] found that genetic variants in the HMGCR gene, which encodes the enzyme targeted by statins, are associated with a 0.22 mmol/L reduction in LDL-C and a 12% increase in T2DM risk per allele (OR, 1.12; 95% CI, 1.04 to 1.20). Ference et al. [31] employed genetic risk scores that simulate the effects of statins to investigate the causal impacts of lower LDL-C levels on cardiovascular and diabetes risk, potentially mediated by variants in HMGCR and PCSK9. Their analysis revealed that HMGCR and PCSK9 variants were associated with 13% (OR, 1.13; 95% CI, 1.06 to 1.20) and 11% (OR, 1.11; 95% CI, 1.04 to 1.19) increased risks of T2DM, respectively. Despite these diabetogenic effects, each variant reduced cardiovascular events by 19% (OR, 0.81 for both) for every 0.26 mmol/L decrease in LDL-C, with additive benefits when present together. Importantly, the diabetogenic risk was more pronounced in individuals with impaired fasting glucose, although it was still modest relative to the substantial reduction in cardiovascular risk. Liu et al. [33] investigated DNA methylation patterns and identified significant associations between statin use and DNA methylation at specific CpG sites within the ABCG1 gene, particularly at cg06500161. These methylation changes were also correlated with elevated fasting glucose and insulin levels, suggesting that disruptions in lipid and glucose metabolism represent key pathways.

The biological mechanism by which genetic predisposition modifies the diabetogenic effect of statins remains largely unknown. Pathways involving pancreatic beta cells, skeletal muscle, and liver metabolism have been proposed [34]. One experimental study reported that increased cholesterol accumulation in pancreatic beta cells can impair insulin secretion, implicating cholesterol metabolism in statin-induced diabetogenic effects. Statins inhibit HMG-CoA reductase, a key enzyme in cholesterol synthesis, thus reducing hepatic cholesterol production and indirectly increasing insulin resistance and glucose production. This effect is thought to be exerted largely through the inhibition of HMGCR. HMGCR inhibition has been demonstrated to disrupt signaling involving AMP-activated protein kinase (AMPK), an essential regulator of insulin sensitivity and glucose metabolism. Disruption of AMPK signaling is associated with insulin resistance and impaired glucose tolerance, especially in those genetically predisposed to T2DM. This association is supported by studies linking HMGCR inhibition with insulin resistance [35,36]. However, direct experimental evidence to substantiate this link is currently lacking.

This study had several limitations. First, we utilized retrospective observational data, which carries the inherent risk of unmeasured residual confounding despite efforts to adjust for known covariates. Second, the UK Biobank cohort predominantly includes individuals of European descent, limiting the generalizability of our findings to other populations. Future validation studies in more diverse ethnic groups are warranted to increase the applicability of the results. Nevertheless, the large sample size and robust analytical framework of this study may provide a foundation for similar investigations in other populations. Third, UK Biobank participants generally exhibit higher socioeconomic status and healthier lifestyles than the general UK population, potentially introducing selection bias [37]. Nevertheless, such biases are likely to attenuate observed associations rather than exaggerate them, as healthier cohorts may display lower baseline risks for adverse outcomes. Therefore, our findings are likely conservative estimates and remain relevant for populations with higher baseline risks. Fourth, while our study examined the class effect of statins, we did not differentiate between specific types of statins, nor did we account for their dose, duration of use, or adherence. Variations in these factors could confound the observed associations and should be systematically investigated in future research [38]. Addressing medication adherence, specific statin types, and dose-response effects in future studies will more thoroughly clarify the relationship between statin use and T2DM risk. Fifth, we conducted a complete case analysis while assuming that missing data were missing completely at random. This assumption might not fully capture underlying dependencies or patterns in missingness, potentially biasing the results. This issue should be explored in subsequent studies to confirm the robustness of our findings.

Our study leveraged the UK Biobank, a large population-based cohort with extensive genetic and phenotypic data. This enabled robust PSM and survival analysis. Unlike previous studies that relied on genetic proxies to gauge the LDL-lowering effects of statins, our method reduces confounding by using PSM and provides a direct estimate of the interaction between regular statin use and T2DM PRS. The findings indicate that incorporating genetic data can deepen our understanding of traditional risk factors and refine our management of the balance between cardiovascular and T2DM risk when prescribing statins. To our knowledge, this is the first study to examine the interaction between statin use and T2DM PRS within a propensity score-matched cohort.

In conclusion, genetic susceptibility, as quantified by T2DM PRS, may increase the risk of T2DM associated with statin use. This result suggests that incorporating PRS into clinical assessments could aid in identifying individuals at elevated risk of T2DM when considering statin therapy, thus facilitating more personalized and preventive strategies. Further research is needed to validate these results, considering dose-response relationships, drug interactions, and lifestyle factors, to elucidate effective T2DM prevention strategies for long-term statin users.

Supplemental Materials

Notes

Conflict of Interest

The authors have no conflicts of interest associated with the material presented in this paper.

Funding

This work was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (2021R1C1C2011327).

Author Contributions

Conceptualization: Park JH, Sung J. Data curation: Park JH, Lee J. Formal analysis: Park JH, Lim KT, Gil Y. Funding acquisition: Sung J, Lee J. Methodology: Park JH, Lim KT, Sung J. Project administration: Sung J, Lee J, Park JH. Visualization: Park JH, Lim KT. Writing – original draft: Park JH, Lim KT, Sung J. Writing – review & editing: Sung J, Park JH, Lee J, Lim KT, Gil Y.

Acknowledgements

Figure 1

Eligible population selection process. Participants diagnosed with diabetes of any type (ICD-10 codes E10–E14) or major atherosclerotic cardiovascular disease at baseline were excluded. As a result, 333 840 complete cases were extracted from a total population of 502 236. Using these 333 840 complete cases, propensity scores were calculated, and up to five controls were matched for each treated individual with a caliper of 0.2 pooled standard deviations. HbA1c, hemoglobin A1c; BMI, body mass index; T2DM, type 2 diabetes mellitus; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; ICD-10, International Classification of Diseases, 10th revision. ¹The numbers of missing data categories include duplicates.

Figure 2

(A) Incidence rate of T2DM by PRS groups and regular statin use status in the matched population, shown in units of 1000 person-years. (B) Cumulative incidence plot for T2DM risk stratified by regular statin use status. Error bars and shaded areas represent the 95% confidence intervals, highlighting statistically significant differences between groups. T2DM, type 2 diabetes mellitus; PRS, polygenic risk score.

REFERENCES

• 1. Arnett DK, Blumenthal RS, Albert MA, Buroker AB, Goldberger ZD, Hahn EJ, et al. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Circulation 2019;140(11):e563-e595. https://doi.org/10.1161/CIR.0000000000000677ArticlePubMedPMC
• 2. Zhao W, Zhao SP. Different effects of statins on induction of diabetes mellitus: an experimental study. Drug Des Devel Ther 2015;9: 6211-23. https://doi.org/10.2147/DDDT.S87979ArticlePubMedPMC
• 3. Blais JE, Wei Y, Yap KK, Alwafi H, Ma TT, Brauer R, et al. Trends in lipid-modifying agent use in 83 countries. Atherosclerosis 2021;328: 44-51. https://doi.org/10.1016/j.atherosclerosis.2021.05.016ArticlePubMed
• 4. Ridker PM, Pradhan A, MacFadyen JG, Libby P, Glynn RJ. Cardiovascular benefits and diabetes risks of statin therapy in primary prevention: an analysis from the JUPITER trial. Lancet 2012;380(9841):565-571. https://doi.org/10.1016/S0140-6736(12)61190-8ArticlePubMedPMC
• 5. Newman CB, Preiss D, Tobert JA, Jacobson TA, Page RL 2nd, Goldstein LB, et al. Statin safety and associated adverse events: a scientific statement from the American Heart Association. Arterioscler Thromb Vasc Biol 2019;39(2):e38-e81. https://doi.org/10.1161/ATV.0000000000000073ArticlePubMed
• 6. Ruscica M, Macchi C, Morlotti B, Sirtori CR, Magni P. Statin therapy and related risk of new-onset type 2 diabetes mellitus. Eur J Intern Med 2014;25(5):401-406. https://doi.org/10.1016/j.ejim.2014.03.003ArticlePubMed
• 7. Sattar N, Preiss D, Murray HM, Welsh P, Buckley BM, de Craen AJ, et al. Statins and risk of incident diabetes: a collaborative meta-analysis of randomised statin trials. Lancet 2010;375(9716):735-742. https://doi.org/10.1016/S0140-6736(09)61965-6ArticlePubMed
• 8. Macedo AF, Douglas I, Smeeth L, Forbes H, Ebrahim S. Statins and the risk of type 2 diabetes mellitus: cohort study using the UK clinical practice pesearch datalink. BMC Cardiovasc Disord 2014;14: 85. https://doi.org/10.1186/1471-2261-14-85ArticlePubMedPMC
• 9. Kennedy RL, Vangaveti V, Malabu UH. Statin treatment and diabetes. J Am Coll Cardiol 2013;61(18):1933-1934. https://doi.org/10.1016/j.jacc.2013.01.071ArticlePubMed
• 10. Preiss D, Seshasai SR, Welsh P, Murphy SA, Ho JE, Waters DD, et al. Risk of incident diabetes with intensive-dose compared with moderate-dose statin therapy: a meta-analysis. JAMA 2011;305(24):2556-2564. https://doi.org/10.1001/jama.2011.860ArticlePubMed
• 11. Byrne CD, Wild SH. Increased risk of glucose intolerance and type 2 diabetes with statins. BMJ 2011;343: d5004. https://doi.org/10.1136/bmj.d5004ArticlePubMed
• 12. Rajpathak SN, Kumbhani DJ, Crandall J, Barzilai N, Alderman M, Ridker PM. Statin therapy and risk of developing type 2 diabetes: a meta-analysis. Diabetes Care 2009;32(10):1924-1929. https://doi.org/10.2337/dc09-0738ArticlePubMedPMC
• 13. Cederberg H, Stančáková A, Yaluri N, Modi S, Kuusisto J, Laakso M. Increased risk of diabetes with statin treatment is associated with impaired insulin sensitivity and insulin secretion: a 6 year follow-up study of the METSIM cohort. Diabetologia 2015;58(5):1109-1117. https://doi.org/10.1007/s00125-015-3528-5ArticlePubMed
• 14. Barkas F, Liamis G, Elisaf M, Anastasiou G, Liberopoulos E. Metabolically healthy obesity and risk of incident type 2 diabetes in 1077 statin-treated individuals: a six-year retrospective study. Eur J Prev Cardiol 2020;27(19):2341-2344. https://doi.org/10.1177/2047487319899191ArticlePubMed
• 15. Na E, Cho S, Kim DJ, Choi J, Han E. Time-varying and dose-dependent effect of long-term statin use on risk of type 2 diabetes: a retrospective cohort study. Cardiovasc Diabetol 2020;19(1):67. https://doi.org/10.1186/s12933-020-01037-0ArticlePubMedPMC
• 16. Vallejo-Vaz AJ, Kondapally Seshasai SR, Kurogi K, Michishita I, Nozue T, Sugiyama S, et al. Effect of pitavastatin on glucose, HbA1c and incident diabetes: a meta-analysis of randomized controlled clinical trials in individuals without diabetes. Atherosclerosis 2015;241(2):409-418. https://doi.org/10.1016/j.atherosclerosis.2015.06.001ArticlePubMed
• 17. Mansi IA, Chansard M, Lingvay I, Zhang S, Halm EA, Alvarez CA. Association of statin therapy initiation with diabetes progression: a retrospective matched-cohort study. JAMA Intern Med 2021;181(12):1562-1574. https://doi.org/10.1001/jamainternmed.2021.5714ArticlePubMed
• 18. Kokkinos P, Nylen E, Faselis C, Pittaras A, Samuel IB, Lavie C, et al. Progression to insulin therapy in patients with type 2 diabetes according to cardiorespiratory fitness, body mass index, and statin therapy. Mayo Clin Proc 2024;99(2):249-259. https://doi.org/10.1016/j.mayocp.2023.05.005ArticlePubMed
• 19. Laakso M, Fernandes Silva L. Statins and risk of type 2 diabetes: mechanism and clinical implications. Front Endocrinol (Lausanne) 2023;14: 1239335. https://doi.org/10.3389/fendo.2023.1239335ArticlePubMedPMC
• 20. Ruscica M, Ferri N, Banach M, Sirtori CR, Corsini A. Side effects of statins: from pathophysiology and epidemiology to diagnostic and therapeutic implications. Cardiovasc Res 2023;118(17):3288-3304. https://doi.org/10.1093/cvr/cvac020ArticlePubMed
• 21. Swerdlow DI, Preiss D, Kuchenbaecker KB, Holmes MV, Engmann JE, Shah T, et al. HMG-coenzyme A reductase inhibition, type 2 diabetes, and bodyweight: evidence from genetic analysis and randomised trials. Lancet 2015;385(9965):351-361. https://doi.org/10.1016/S0140-6736(14)61183-1ArticlePubMedPMC
• 22. Yang G, Schooling CM. Statins, type 2 diabetes, and body mass index: a univariable and multivariable mendelian randomization study. J Clin Endocrinol Metab 2023;108(2):385-396. https://doi.org/10.1210/clinem/dgac562ArticlePubMed
• 23. Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, et al. UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med 2015;12(3):e1001779. https://doi.org/10.1371/journal.pmed.1001779ArticlePubMedPMC
• 24. Thompson DJ, Wells D, Selzam S, Peneva I, Moore R, Sharp K, et al. A systematic evaluation of the performance and properties of the UK Biobank Polygenic Risk Score (PRS) Release. PLoS One 2024;19(9):e0307270. https://doi.org/10.1371/journal.pone.0307270ArticlePubMedPMC
• 25. Han SJ, Kim KH. Adjusting for confounders in outcome studies using the Korea National Health Insurance Claim Database: a review of methods and applications. J Prev Med Public Health 2024;57(1):1-7. https://doi.org/10.3961/jpmph.23.250ArticlePubMed
• 26. Vell MS, Loomba R, Krishnan A, Wangensteen KJ, Trebicka J, Creasy KT, et al. Association of statin use with risk of liver disease, hepatocellular carcinoma, and liver-related mortality. JAMA Netw Open 2023;6(6):e2320222. https://doi.org/10.1001/jamanetworkopen.2023.20222ArticlePubMedPMC
• 27. Kim J, Kennedy Neary MT, Aschard H, Palakkamanil MM, Do R, Wiggs JL, et al. Statin use in relation to intraocular pressure, glaucoma, and ocular coherence tomography parameters in the UK Biobank. Invest Ophthalmol Vis Sci 2022;63(5):31. https://doi.org/10.1167/iovs.63.5.31Article
• 28. Ding J, Davis-Plourde KL, Sedaghat S, Tully PJ, Wang W, Phillips C, et al. Antihypertensive medications and risk for incident dementia and Alzheimer’s disease: a meta-analysis of individual participant data from prospective cohort studies. Lancet Neurol 2020;19(1):61-70. https://doi.org/10.1016/S1474-4422(19)30393-XArticlePubMed
• 29. Ho D, Imai K, King G, Stuart EA. MatchIt: nonparametric preprocessing for parametric causal inference. J Stat Softw 2011;42(8):1-28. https://doi.org/10.18637/jss.v042.i08Article
• 30. Wu P, Moon JY, Daghlas I, Franco G, Porneala BC, Ahmadizar F, et al. Obesity partially mediates the diabetogenic effect of lowering LDL cholesterol. Diabetes Care 2022;45(1):232-240. https://doi.org/10.2337/dc21-1284ArticlePubMed
• 31. Ference BA, Robinson JG, Brook RD, Catapano AL, Chapman MJ, Neff DR, et al. Variation in PCSK9 and HMGCR and risk of cardiovascular disease and diabetes. N Engl J Med 2016;375(22):2144-2153. https://doi.org/10.1056/NEJMoa1604304ArticlePubMed
• 32. Smit RA, Trompet S, Leong A, Goodarzi MO, Postmus I, Warren H, et al. Statin-induced LDL cholesterol response and type 2 diabetes: a bidirectional two-sample Mendelian randomization study. Pharmacogenomics J 2020;20(3):462-470. https://doi.org/10.1038/s41397-019-0125-xArticlePubMed
• 33. Liu Y, Shen Y, Guo T, Parnell LD, Westerman KE, Smith CE, et al. Statin use associates with risk of type 2 diabetes via epigenetic patterns at ABCG1. Front Genet 2020;11: 622. https://doi.org/10.3389/fgene.2020.00622ArticlePubMedPMC
• 34. Galicia-Garcia U, Jebari S, Larrea-Sebal A, Uribe KB, Siddiqi H, Ostolaza H, et al. Statin treatment-induced development of type 2 diabetes: from clinical evidence to mechanistic insights. Int J Mol Sci 2020;21(13):4725. https://doi.org/10.3390/ijms21134725ArticlePubMedPMC
• 35. Ye J. Mechanism of insulin resistance in obesity: a role of ATP. Front Med 2021;15(3):372-382. https://doi.org/10.1007/s11684-021-0862-5ArticlePubMed
• 36. Johnson JD. On the causal relationships between hyperinsulinaemia, insulin resistance, obesity and dysglycaemia in type 2 diabetes. Diabetologia 2021;64(10):2138-2146. https://doi.org/10.1007/s00125-021-05505-4ArticlePubMed
• 37. Schoeler T, Speed D, Porcu E, Pirastu N, Pingault JB, Kutalik Z. Participation bias in the UK Biobank distorts genetic associations and downstream analyses. Nat Hum Behav 2023;7(7):1216-1227. https://doi.org/10.1038/s41562-023-01579-9ArticlePubMedPMC
• 38. Ghang H, Lee J. Multilevel analysis of the relationship between prescribing institutions and medication adherence among patients with hypertension and diabetes in Korea. J Prev Med Public Health 2023;56(6):504-514. https://doi.org/10.3961/jpmph.23.252ArticlePubMedPMC

Figure & Data

References

Citations

Citations to this article as recorded by