The use of statins are associated with an increased risk of new-onset diabetes in patients with hypertension and obstructive sleep apnoea, a longitudinal study

Abstract

Background

Statins, a kind of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors, are first-line cholesterol-lowering drugs that are widely used in the primary and secondary prevention of coronary atherosclerotic heart disease (CAD). However, the safety of statins has been in the spotlight as recent studies have shown that statins may increase the incidence of diabetes. Therefore, we conducted a study of statins use and new-onset diabetes(NODM) in people with hypertension and obstructive sleep apnea (OSA) to better understand the relationship and to provide guidance for future clinical management.

Methods

We conducted a retrospective cohort study using data from the Urumchi Hypertension Database (UHDATA), including patients aged ≥ 18 years diagnosed with hypertension and obstructive sleep apnoea treated at our Hypertension Centre between 2015 and 2019. The study was followed until November 2023 and the primary endpoint was new onset diabetes during the follow-up period. The hazard ratio (HR) and 95% confidence interval (CI) were calculated using the Cox proportional hazards model. Sensitivity analyses were performed by excluding those with pre-diabetes at baseline.

Results

8755 patients with hypertension and OSA, and 80.1% were followed up. During median follow-up of 31 months, 740 patients developed NODM. The incidence of NODM per 1000 person-years was 53.1. In Cox regression analysis, the risk of diabetes is significantly higher in patients who continue to take statins (HR = 1.77, 95% CI, 1.34–2.34, P < 0.001), and the results remain significant in sensitive analysis.

Conclusions

In patients with OSA and hypertension, continuous statins use increases the risk of diabetes and physicians should be vigilant about monitoring blood glucose levels when using statins in this patients.

Introduction

Diabetes mellitus (DM) is a metabolic disease characterized by hyperglycemia. DM, in particular type 2 diabetes mellitus, has become one of the major diseases posing a serious threat to public health worldwide [1]. With an estimated rate of nearly 52%, cardiovascular diseases (CVDs) are the leading cause of mortality in patients with type 2 diabetes [2]. Patients with diabetes have a 2- to 3-fold higher risk of developing cardiovascular disease compared to those without diabetes [3]. Many risk factors for type 2 diabetes have been demonstrated in previous studies, such as male gender, older age, being overweight, obesity, high blood pressure (HBP), sleep, alcohol, tobacco, depression and hypertriglyceridemia [4]. In recent years, studies have shown that statins may also increase the risk of developing diabetes, in addition to the aforementioned factors [5,6,7,8,9].

Recent studies have shown that some protective drugs, such as statins, may increase the risk of diabetes [5,6,7,8,9]. For instance, Cederberg H, Ooba N, Jimin Lee, Ridker and Hail Li’s studies showed that statins would increase the risk of diabetes in varying degrees [5,6,7,8, 10]. However, in West of Scotland Coronary Prevention Study, pravastatin therapy resulted in a 30% reduction of the cases of type 2 diabetes mellitus in this study [9]. Therefore, it is controversial whether statins are associated with new onset diabetes mellitus. Meanwhile, there are still some limitations in previous studies. For instance, the research population is limited to a single gender or elderly population [5, 11] or shorter follow-up period [12].

Both hypertension and obstructive sleep apnea (OSA) are independent risk factors for CVDs morbidity and mortality, and the comorbidity rate is very high [13, 14]. Patients with hypertension and or OSA often have dyslipidemia, as high as 15% [15], and are often the candidates for statins use [16, 17]. In addition, hypertension and OSA are most significant risk factors for diabetes [18, 19]. Studies have shown that in patients with OSA, intermittent hypoxia and repeated arousal can negatively affect the function of the autonomic nervous system, leading to the secretion of catecholamines. Increased secretion of adrenaline, norepinephrine, and cortisol can further lead to impaired glucose metabolism and insulin sensitivity, as well as dysfunction of beta cells caused by increased gluconeogenesis and reduced glucose uptake [20]. In addition, increased blood pressure has been found to elevate markers of inflammation associated with the insulin signaling pathway and function of beta cells, which may contribute to the development of diabetes, too [21]. However, in this specific population the association of statins use and risk of diabetes mellitus has been less studied. Therefore, we carried out an age and gender-stratified longitudinal study in clinical patients to explore the association between statins use and NODM among patients with hypertension and OSA.

Materials and methods

This study was approved by the Ethics Committee of People’s Hospital of Xinjiang Uygur Autonomous Region, and the research strictly adhered to the ethical standards of the Declaration of Helsinki. All participants provided written informed consent.

Data sources and study population

Participants were the patients from the Hypertension Database in Urumchi (UHDATA), a specialized database for hypertension in Urumqi that was created by The Key Laboratory for Hypertension Clinical Research and The National Health Committee of China, as described in our recent studies [13].

The medicines and diagnostic data used during the follow-up were mainly from UHDATA and Urumqi’s social medical insurance system, the latter includes inpatient and outpatient diagnosis information, drug prescription information, and cost information.

Baseline data were collected from patients aged ≥ 18 years with hypertension and OSA at the People’s Hospital of Xinjiang Uygur Autonomous Region, China, between January 2015 and December 2019. Inclusion criteria for this study were as follows: (1) age ≥ 18 years; (2) inclusion of hypertension and OSA in discharge diagnosis. The exclusion criteria were as follows: (1) a history of diabetes, (2) Cushing’s syndrome, (3) malignant tumors; 3)usage of steroid medication; (4) mixed or central sleep apnea (5) acute or chronic pancreatitis and (6) incomplete data on important parameters.

Data collection and definition

Baseline data were extracted from medical electronic system, including the demographics: age, sex, waist, systolic and diastolic blood pressure (SBP, DBP), body mass index (BMI), cigarette consumption (yes or no), alcohol intake (yes or no); hypertension duration and discharge diagnosis; biochemical measurements: uric acid (UA), serum creatinine, plasma aldosterone, fasting blood glucose (FBG), total cholesterol (TC), triglyceride (TG), low-density lipoprotein cholesterol (LDL-C), high density lipoprotein cholesterol (HDL-C), alanine aminotransferase (ALT), aspartate aminotransferase (AST), serum potassium and prescribed agents at discharge such as antihypertensive, anti-diabetic agent and or statins.

Exposure

As determined by the number of statins prescriptions in the social medical insurance database, an average of 4 or more statins prescriptions per year is defined as continuous users; an average of 1 to 4 statins prescriptions per year is defined as intermittent users; no statins prescriptions during follow-up is defined as none users.

Follow up and outcome

The main outcome for the current study was new-onset diabetes mellitus during follow-up. All patients’ outpatient or inpatient medical records and medical insurance prescription records were collected and organized for follow-up information. The follow-up period ended on Novmber 2023. Patients were followed from the date of study-cohort entry until study outcome occurred or the deadline is reached, whichever occurred first. Diagnosis of new onset diabetes was based on the following criteria: (1) fasting plasma glucose (FPG) ≥ 7.0 mmol/l, 2 h plasma glucose (2hPG) ≥ 11.1 mmol/l in an OGTT; (2) HbA1c ≥ 7.0% or random blood glucose ≥ 11.1 mmol/l; (3) glucose-lowering medication started during follow-up; (4) diabetes diagnosed by physician as per medical records and/or FPG ≥ 7.0 mmol/l, 2hPG ≥ 11.1 mmol/l or HbA1c ≥ 7.0 in outpatient/primary care laboratory measurements.

Statistical analysis

Participants were divided into none users, intermittent users and persistent users according to their usage of statins.

Continuous variables were presented as mean ± SD or median (interquartile range) according to the normality test and compared between groups using analysis of variance or non-parametric Kruskale-Wallis H or Students’ t-test. Categorical variables were summarized as numbers and percentages and compared between groups using Pearson’s chi-square test.

The cumulative incidence of the primary outcome was calculated using Kaplan-Meier curves, and the log-rank test was used to determine group differences. Cox proportional hazard regression models were constructed in total, gender, age, abdominal obesity and OSA severity stratified. Variables with P < 0.10 in univariate Cox regression were included in regression models, and we assessed collinearity between variables according to tolerance and variance inflation factor (VIF). Indicators with tolerances < 0.1 or VIF > 10 are considered to have collinearity. Multivariate cox proportional hazards models were used to find independent predictors of endpoint events, and their components were adjusted for known endpoint event-influencing factor.

We performed sensitivity analyses by excluding pre-diabetes, and the results still suggest that statins may increase the risk of diabetes.

Statistical analyses were performed using Statistical Package for the Social Sciences (SPSS) version 23.0 and R4.3.2.

Result

A total of 8755 patients with hypertension and OSA at baseline were initially identified, of whom 6727 the inclusion and exclusion criteria and 5390 completed follow-up were finally analyzed (Fig. 1.).

Fig. 1

Experimental flow chart

Baseline characteristics of the participants

The median age of the study population was 49.0 years, and 3821 (70.9%) patients were men. As in Table 1, participants who continue to take statins tended to have older age, obesity, long duration of hypertension and taking two or more antihypertensive drugs. In addition, There was no significant difference in triglyceride and glucose (TyG) index between the three groups. Details of the baseline characteristics across the statins groups are shown in Table 1.

Table 1 Baseline characteristics of study participants by Statins use

Relationship between usage of Statins and outcome

This section describes Table 2. During the follow-up of 13924.2 person-years (median 31 months, interquartile range: 16–48), 740 patients developed NODM, with an incidence of 53.1/1000 person-years. Per 1000 person-years, the incidence of NODM in patients not using statins, those using statins intermittently, and those continuing to use statins were 48.6, 48.9, 78.1/1000 person-years respectively. The sensitivity analysis excludes patients who were diagnosed with pre-diabetes at baseline and then analyses the incidence rate of new-onset diabetes. In sensitivity analysis, the incidence rates in these different groups are 41.5, 35.8, and 57.7/1000 person-years (Table 2).

Table 2 Incidence of events among groups

The Kaplan-Meier curve shows the cumulative risk of NODM in these three groups. The risk of diabetes increased during the follow-up period, regardless of whether statins were taken or not. Among these, the cumulative risk of diabetes is significantly higher in patients who continue to take statins than in the other two groups, followed by those who take statins intermittently, and those who do not take statins have the lowest cumulative incidence rate of diabetes. (log-rank P < 0.001, Fig. 2A、2B).

Fig. 2

A Kaplan-Meier curves of the three groups in total participates. B. Kaplan-Meier curves of the three groups in participates without pre-diabetes

This section describes Table 3. In total participants, the risk of diabetes was significantly increased in the group that continued to use statins(HR = 1.84, 95%CI: 1.42–2.39, P < 0.001), while there was no significant difference in those taking statins intermittently(HR = 1.05, 95%CI: 0.85–1.30, P = 0.663). We then further adjusted for fasting blood glucose, hypertension duration, abdominal circumference, alanine aminotransferase, uric acid, systolic blood pressure, triglyceride, high density lipoprotein-C, age, gender and anti-hypertension agents develop models 1, 2 and 3, the results were consistent with the above, which suggests that continuous statins use may increase the risk of diabetes(HR = 1.63, 95%CI: 1.26–2.12, P < 0.001, HR = 1.73, 95%CI: 1.33–2.29, P < 0.001, HR = 1.77, 95%CI: 1.34–2.34, P < 0.001). When pre-diabetic patients were excluded for sensitivity analysis, the results were consistent with the total participants, suggesting that the results are stable and reliable(HR = 1.58, 95%CI: 1.16–2.17, P < 0.001). We selected factors with p < 0.1 based on univariate cox regression to be adjusted and multivariable Cox to calculate Variance Inflation Factor (VIF) and Tolerance. (Sup Table 1.)

Table 3 Cox regression analysis for the association of Statins use and risk of incident diabetes in total and sensitivity analysis by excluding those with pre-diabetes at baseline (no Statins use as the reference, HR, 95%CI, P)

Tables 4 and 5 are described in this section. In our study, Table 4 shows the stratified analysis of the total population by gender, age, obesity status, and OSA severity, and Table 5 shows the sensitivity analysis of excluding prediabetic patients based on Table 4. In gender stratification, persistent statins users from both male and female populations also showed a significantly higher risk of NODM (HR = 1.80 95%CI: 1.30–2.49, P < 0.001; HR = 1.99, 95%CI: 1.14–3.48, P = 0.015), which is consistent with the result in total participants. In sensitivity analysis, statins were significantly associated with NODM in male (HR = 1.69,95%CI: 1.17–2.44, P = 0.006) while the results weren’t observed in female, which may be caused by the relatively small sample size. Nevertheless, intermittent users neither male nor female showed a significantly increased risk for NODM. In the age groups, people aged < 45 and 45–60 years old indicated that statins were associated with NODM (HR = 2.58 95%CI: 1.52–4.38, P < 0.001; HR = 1.74, 95%CI: 1.18–2.56, P = 0.005) while the results weren’t observed in people aged over 60 years old, which may be due to the relatively small sample size. In sensitivity analysis, aged < 45 years old showed a significantly higher risk of NODM (HR = 2.15 95%CI: 1.15–4.01, P = 0.016). Both in the overall population and in the sensitivity analysis, the risk of diabetes is significantly higher in obese patients who continue to take statins (HR = 1.77 95%CI: 1.33–2.36, P < 0.001; HR = 1.60, 95%CI: 1.16–2.22, P = 0.005). The mild, moderate and severe OSA groups showed a 1.85 (95% CI: 1.15,2.97), 1.87 (95% CI: 1.08,3.25) and 1.69 (95% CI: 1.04,2.75) fold risk of NODM, respectively, which was significant in adjusted models. After sensitivity analysis, it was still significant. More details of the Cox regression are shown in Tables 4 and 5. The above results refer to model 3 of the Tables 4 and 5.

Table 4 Stratified Cox regression analysis for the association of Statins use and risk of incident diabetes by gender, age, obesity, and OSA severity (no Statins use as the reference, HR, 95%CI, P)

Table 5 Sensitivity analysis for stratified Cox regression for the association of Statins use and risk of incident diabetes in participants without pre-diabetes (no Statins use as the reference, HR, 95%CI, P)

Discussion

Statins are a kind of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors. They are currently the drugs of choice in most cases of reducing low-density lipoprotein cholesterol [22] and statins have been widely utilized in the primary and secondary prevention of CVDs [23, 24].

Although statins are considered to be good safety profile in general, its diabetogenic effects, inducing diabetes and increase in hemoglobin A1c or fasting glucose level has got considerable attention [8].However, previous study results are conflicting and studies are scarce in co-existent patients with hypertension and OSA, vulnerable patient groups to dyslipidemia, CVDs and statins use. Therefore, the aim of this retrospective cohort study was to investigate whether there is an increase in the risk of diabetes with statins in this spfecific patient groups.

With longitudinal design, we observed the following. First, after adjustment for confounders, statins treatment increases the risk of type 2 diabetes by more than 50%, suggesting that statins use in this particular population significantly increases the risk of diabetes. Then, in patients with hypertension and OSA, the prevalence of diabetes is higher than in the general population, and statins use further increases the risk of NODM.

The present research findings are consistent with previous studies. In another study of 25,970 participants, Mansi et al. performed a retrospective observational study suggested that statins users had a higher odds of incident diabetes (odds ratio 1.87; 95% CI 1.67–2.01) and diabetes with complications (odds ratio 2.50; 95% CI 1.88–3.32) compared to non-users [25]. In a meta-analysis by Sattar N et al., statins therapy was associated with a 9% increased risk of type 2 diabetes, based on FPG or physician-reported diagnosis of diabetes [26].

It has been suggested that statins are involved in new-onset diabetes through the following suggested mechanisms. Statins can increase insulin resistance, and the pancreas can compensate for early insulin resistance by increasing insulin secretion, however, the pancreas was unable to compensate for the long-term increase in insulin resistance, leading to a decrease in insulin secretion finally [27]. Statins inhibit the insulin secretion through reduce the production of ATP by suppressing the synthesis of ubiquinone (CoQ10) [28]. Glucose transporter 4 (SCL2A4 or GLUT4) is a facilitated transporter which responsible for peripheral insulin-mediated glucose influx. Statins have been shown to induce insulin resistance via reduction of GLUT4 translocation to the cell membrane [29]. Recently, professor Wu Yue’s team have revealed that statins induce a reduction of Clostridium-rich microbiota, affecting Clostridium sp.-bile acids axis to inhibit chenodeoxycholic acid to ursodeoxycholic acid (CDCA-to-UDCA) transformation and reduce serum UDCA and GLP-1 finally impair glucose metabolism [30]. Previous studies have shown that OSA itself is a major risk factor for diabetes. In both animal models and humans, intermittent hypoxia and circadian disruption lead to sympathetic arousal and inflammation, which have been shown to be associated with pancreatic beta cell loss and dysfunction [31,32,33].

Current studies have shown that statins are associated with NODM, especially in male, young and middle-aged, obese, moderate and severe sleep apnea participants who are at relatively high risk for diabetes, therefore, individuals taking statins, particularly for extended periods, should monitor their blood glucose levels or consider alternative lipid-lowering medications. The higher risk of type 2 diabetes found in our study suggests that previous studies may have underestimated the importance of statin-induced diabetes, especially in hypertensive patients with OSA. Although statins have a beneficial effect on the prevention of cardiovascular disease, the adverse effects of long-term statins use also require close attention by clinicians. Paul M Ridker et al. showed that in the JUPITER primary prevention trial, the cardiovascular and mortality benefits of statins treatment exceeded the risk of diabetes, including high-risk participants with diabetes risk [10]. This suggests that short-term and long-term use of statins may have different risk-benefit ratios for patients. Therefore, in clinical practice, patients who need to take statins for a long time should be treated after weighing the benefits and harms. It is recommended that fasting blood glucose and glycosylated haemoglobin levels be assessed in patients who require long-term statins use. For patients with high fasting glucose and glycosylated haemoglobin levels or impaired glucose control, it is recommended to monitor fasting glucose and glycosylated haemoglobin levels regularly and to educate patients about lifestyle interventions, such as diet and exercise, to reduce the impact of other risk factors on blood glucose and reduce the risk of new-onset diabetes.

A meta-analysis by Michael G. Silverman et al. suggested that the major vascular events for each of the 5 different established interventions (statins, diet, bile acid sequestrants, ileal bypass, and ezetimibe) that ultimately act predominantly through upregulation of LDL receptor expression, regardless of the type of treatment, the clinical benefit normalized to the magnitude of LDL-C reduction are consistent.Meanwhile, the use of statins and nonstatins therapies that act via upregulation of LDL receptor expression to reduce LDL-C were associated with similar RRs of major vascular events per change in LDL-C. Therefore, there may be additional options to reduce LDL-C while protecting cardiovascular and cerebrovascular health [34]. Proprotein convertase subtilisin–kexin type 9 (PCSK9) inhibitor is a new class of drugs that effectively lower LDL-C levels. Marc S. Sabatine et al. conducted a RCT concluded that inhibition of PCSK9 with evolocumab on a background of statins therapy can further lowered LDL cholesterol and reduced the risk of cardiovascular events [35]. And in FOURIER randomised controlled trial, they further demonstrated that evolocumab could reduce cardiovascular risk without increase the risk of new-onset diabetes and worsen glycaemia [36]. In addition, positive correlations have been found between PCSK9 levels and the platelet activation markers soluble CD40L, platelet factor-4 and P-selectin, and anti-PCSK9 reduces platelet aggregability and activation.Studies have shown that the ability of anti-PCSK9 mAb therapy to improve cardiovascular outcomes in patients may be due to inhibition of PCSK9’s direct effects on endothelial function, plaque stabilisation and platelet activation, and indirectly through its dramatic effects on lipid homeostasis [37]. It is suggested that we can reduce the incidence of diabetes while lowering LDL-C and protecting cardiovascular system by adjusting or combining with other types of lipid-lowering drugs. However, further research is still needed.

Strengths of our analysis include its longitudinal design with a large sample size and a series of adjustments for confounders, which helped to obtain a relatively stable and reliable results. Secondly, our study included a sample from a population at high risk of diabetes, and the results may contribute to the prevention of diabetes and as a reference for clinicians using statins. Meanwhile, our research has some limitations, which should be taken into account when interpreting the observational results: firstly, this is a single-centre observational study with the drawbacks of non-randomised studies. To confirm the stability of the results, we conducted sensitivity analysis and adjusted for multiple variables, we adjusted for factors such as fast blood glucose, HTN duration, Abdominal circumference, alanine aminotransferase, uric acid, cystic blood pressure, triglycerides, high-density lipoprotein-C, age, gender, anti-HTN agents. Second, the number of female participants is relatively small, which may partly explain the lack of a significant association between statins and diabetes. Thirdly, our follow-up studies lack relevant information on treatment for central OSA and CPAP, making it impossible for us to confirm other factors that might influence the results of the study.

In conclusion, in this observational study, statins therapy was significantly associated with new-onset diabetes in patients with hypertension and OSA after adjustment for multiple confounding variables. Therefore, attention should be paid to glycaemic control when using statins in this population and clinicians need to be more cautious when selecting statins.