Impact of metformin on statin‐associated myopathy risks in dyslipidemia patients

Abstract

A growing number of patients with metabolic disorders are receiving statin and antidiabetic therapies as comedications. A signal of increased risk of myotoxicity due to potential interactions between antidiabetics and statins has been detected in previous studies. To investigate the effects of metformin on myopathy risks when added to preexisting statin therapy in dyslipidemia patients, we performed a retrospective cohort study using the Korean national health insurance data in statin‐treated dyslipidemia patients with or without concomitant metformin use. We compared the risk of myopathy in statin + metformin users against statin‐only users. Hazard ratios (HRs) and 95% confidence intervals (CIs) have been calculated following propensity score (PS) matching between study groups and subsequent stratification per patient factors. We included 4092 and 8161 patients in PS‐matched statin + metformin and statin‐only groups, respectively. The risk of myopathy decreased when metformin was used together with statins (adjusted HR 0.84; 95% CI 0.71–0.99). In subgroup analyses per individual statin agent and in stratified risk analyses, no specific statin agents or patient factors were associated with statistically significant myopathy risk. This study found that a comedication with metformin was associated with decreased myopathy risk in statin‐treated dyslipidemia patients compared to statin‐only users. Our findings suggest that metformin may provide protective effects on potential muscle toxicities induced by statin therapy.

Abbreviations

AHA

American Heart Association

AMPK

adenosine monophosphate–activated protein kinase

ASCVD

atherosclerotic cardiovascular disease

CI

confidence interval

DPP‐4

dipeptidyl peptidase‐4

EMA

European Medicines Agency

FOXO

Forkhead box protein

HIRA

Health Insurance Review & Assessment Service

HMG CoA

3‐hydroxy‐3‐methylgultaryl coenzyme A

HR

hazard ratio

ICD

International Classification of Disease

KM

Kaplan–Meier

LDL‐C

low‐density lipoprotein cholesterol

NHI

National Health Insurance

NPS

National Patients Sample

PGC‐1 α

peroxisome proliferator–activated receptor gamma coactivator 1‐alpha

PS

propensity score

SGLT‐2

sodium‐glucose cotransport‐2

T2DM

type 2 diabetes mellitus

Ub

ubiquitin

1. INTRODUCTION

The incidence of dyslipidemia has been increasing, with the prevalence rate in adults aged 20 years and above averaging at over 40% in 2020; of note is that about 90% of diabetic patients also have dyslipidemia as comorbid conditions. ¹ , ² , ³ The integral components of disease management plans for dyslipidemia involve lifestyle changes, such as physical activities and dietary control, and pharmaceutical therapy. Statins have long been used to treat lipid abnormalities, ⁴ , ⁵ , ⁶ and its usage rates have been steadily growing worldwide. ⁷ Study findings show that statins, by inhibiting 3‐hydroxy‐3‐methylgultaryl coenzyme A (HMG CoA), are effective in lowering serum cholesterol levels and decreasing risks of heart attacks or strokes compared to non‐users. ⁸ , ⁹ HMG CoA reductase inhibitors or statins are indeed among the top medications most widely prescribed in the world. ¹⁰ Although generally well tolerated, the most common adverse events associated with statins are skeletal muscle toxicities, encompassing mild to moderate muscle complaints with normal to elevated creatine kinase levels and severe rhabdomyolysis. ¹¹ , ¹² A series of studies have reported that the overall incidence of statin‐associated myopathy, ranging from muscle weakness, fatigue, or mild‐to‐moderate pain to rhabdomyolysis, was from 10% up to 27.8% in patients receiving statin therapy. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ A recent prospective cohort study also reported that, of different statin agents and doses, simvastatin 40 mg was associated with the highest incidence of myopathy (50%) whereas fluvastatin XL 80 mg and rosuvastatin 10 mg with the lowest myopathy risk (8% and 10.8%, respectively). ¹³ Statin discontinuation rates owing to statin‐associated muscle toxicities remain high, which could contribute to uncontrolled cardiovascular risks in high‐risk patients with metabolic dysfunctions. ¹⁵ , ¹⁸ , ¹⁹ Its underlying mechanism has not been fully elucidated, but it may be induced by statins suppressing not only the mevalonate synthesis pathway but also the insulin‐like growth factor 1 pathway, which leads to the overexpression of atrogin‐1 genes associated with muscle damage, ranging from myopathy, muscle cell damage, to rhabdomyolysis. ²⁰

A series of prior studies on the efficacy and safety of statins for cardiovascular disease prevention and treatment suggest that statin use may increase the risk of new onset type 2 diabetes mellitus (T2DM). ²¹ , ²² , ²³ As the number of patients with metabolic disorders is on the rise, more patients are treated for comorbid conditions of dyslipidemia and diabetes mellitus. Hence, more statin users tend to receive antidiabetic therapy as comedication, which highlights the need to evaluate the safety of the combined use of statins and antidiabetic medications. In recent clinical studies and case reports, potential adverse effects on skeletal muscles have been linked to glucose‐lowering therapies, particularly with sodium‐glucose cotransport‐2 (SGLT‐2) inhibitors ²⁴ and dipeptidyl peptidase‐4 (DPP‐4) inhibitors. ²⁵ A signal of elevated myotoxicity and sarcopenia risk has been detected with SGLT‐2 inhibitors in several studies, ²⁴ , ²⁶ , ²⁷ , ²⁸ and a case study also reported elevated rosuvastatin myotoxicity following canagliflozin initiation. ²⁹ Meanwhile, the European Medicines Agency (EMA) released safety warnings for muscular toxicity risks associated with DPP‐4 inhibitors, especially when used concomitantly with statins. ²⁵ , ³⁰ Mixed results have been reported in spontaneous reporting analyses, and subsequent large‐scale pharmacovigilance analyses based on international databases failed to identify a signal of drug interactions between DPP‐4 inhibitors and statins. ³¹ , ³²

As metformin is the first‐line glucose‐lowering therapy recommended by the treatment guidelines for T2DM and by far the most frequently used antidiabetic agent in prescription volumes, either as monotherapy or as combination therapy, ² a substantial proportion of these patient populations with metabolic dysfunctions are likely to receive metformin for their glycemic control along with statin therapy for their lipid management. Although potential effects of metformin and its mechanism to prevent statin‐associated muscle symptoms have been proposed in previous studies, ³³ , ³⁴ , ³⁵ clinical studies that investigate its effects on myotoxicity risks in dyslipidemia patients on statin therapy are still limited. Hence, there is a strong need for the verification of its safety in terms of muscular toxicities when used together with statins in metabolic disorder patients. In this study, we aim to investigate differential risks of myopathy in statin‐treated dyslipidemia patients depending on their exposure to metformin.

2. MATERIALS AND METHODS

2.1. Study population

A retrospective cohort study in dyslipidemia patients treated with statins, in combination with or without metformin, was carried out using the Korean Health Insurance Review & Assessment Service‐National Patients Sample‐2019 (HIRA‐NPS‐2019). The HIRA database stores National Health Insurance (NHI) claims data linked to healthcare services offered to the entire beneficiaries in South Korea, including patient demographics, International Classification of Disease 10th Revision (ICD‐10)‐based diagnosis codes, procedures, inpatient and outpatient medical utilization data, and comprehensive prescription data. The Initial samples of 2% of the entire national populations in 2019 had been selected based on stratified randomized sampling methods to ensure national representativeness of the NPS data. Those patients, aged 20 years or above, with healthcare encounters in 2019, who were prescribed statins under the diagnosis of dyslipidemia were first screened for the eligibility for the study cohort. To include first‐time users or those with resumed therapy post adequate intervals, the patients with a history of statin use prior to March 31, 2019 were excluded. We also excluded those with a history of myopathy prior to study entry. The study protocol was approved by the Institutional Review Board of Ajou University (202201‐HB‐EX‐004). Informed consent from study patients was waived and no further ethics approval was required as the authors are authorized by the HIRA to use the de‐identified patient data for research purposes.

2.2. Study medications and variables

Dyslipidemia patients treated with statins were assigned to the statin + metformin user group versus the statin‐only user group depending on a metformin comedication history. Seven statin substances were included in this study: atorvastatin, fluvastatin, lovastatin, pitavastatin, pravastatin, rosuvastatin, and simvastatin. Entire eligible prescription data were extracted based on a total of 127 statin substance codes and 74 metformin codes with different formulations and strengths listed in the 2019 national formulary. Use of other antidiabetic therapy, both oral and injectable agents, were permitted in both groups. Prespecified variables included patient demographics (age and sex), and myopathy‐related risk factors in terms of comorbidities and comedication patterns. The comorbid conditions that could affect the risk of myopathy were identified during the baseline period per ICD‐10 code: alcoholism (K70, G72.1), chronic kidney disease (N18), chronic hepatic failure (K72.1), hypertension (I10‐15), and hypothyroidism (E02, E03), which was adopted from a nested case–control study on DPP‐4 inhibitors' effects on statin‐associated muscular injury. ³⁶ A history of fibrate use was also identified in each patient as it has been associated with increased risk of myopathy. ³⁷ The index date was defined as the date when statin therapy was initiated or the date when metformin was first added to preexisting statin therapy. The baseline period for comorbidity and comedication was 3 months prior to the index date. The grace period was determined as 50% of the days‐supply per prior prescription; hence, it was considered as consecutive therapy when a study medication was re‐prescribed before the therapy end date (the days‐supply added to the prior prescription date).

2.3. Study outcomes

The primary outcome was the incidence and risk of myopathy in dyslipidemia patients treated with statins depending on their exposure to metformin. Myopathy assessed in this study include drug‐induced myopathy (G72.0), other specified myopathy (G72.8), unspecified myopathy (G72.9), other myositis (M60.8), unspecified myositis (M60.9), and myalgia (M79.1), ³⁸ which was adopted from a real world data‐based study on statin‐associated myopathy. ³⁹ The risk analysis was further stratified by age group, sex, comorbidities, and fibrate use in order to account for differential effects of patient factors on statin‐associated myopathy risk. The endpoint event was identified via healthcare encounters for muscle symptoms, and subsequently differential risks of muscle toxicities were assessed between groups. Only those outcome events that occurred post the index date were assessed valid and incorporated into study analyses. Subgroup analyses were designed to comprehensively evaluate differential effects per substance and per patient factor, such as sex, gender, comorbidities, and fibrate comedication history. The outcome date was determined as the earliest date a patient encountered a given outcome event. The follow‐up period started on the index date and lasted until the earliest occurrence of any of the following: an outcome event, therapy discontinuation, death, or the end of study period (December 31, 2019).

2.4. Statistical methods

In this retrospective cohort study, we designed two comparison groups: statin + metformin versus statin‐only groups. To minimize between‐group confounders, propensity score (PS) matching was performed with the nearest neighbor matching method at a ratio of 1:2. A multivariable logistic regression model was used to estimate the PS for each patient, which predicts the probability of patient exposure to statin + metformin versus statin‐only therapy given prespecified baseline variables. The baseline variables incorporated in a multivariable logistic regression as covariates were age groups, sex, comorbidities, and fibrate use history. Incidence rates per 1000 person‐years and 95% confidence intervals (CIs) of outcome events in each study group were calculated. The primary endpoint analyses were conducted with a Cox proportional‐hazards model. Hazard ratios (HRs) with 95% CIs for endpoint events were calculated by comparing the statin + metformin group against the statin‐only group (reference). The HRs and 95% CIs were then adjusted for relevant patient factors, comorbid conditions, and comedication history. Risk analyses were first performed for incident endpoint events irrespective of individual statin substances, patient comorbidities, and comedication history for myopathy risk assessment, and subsequently performed in each stratum per statin substance as well as per patient factor, such as age, gender, comorbid conditions, and fibrate use history. A p‐value of <.05 was assessed statistically significant. Kaplan–Meier (KM) curves were plotted for the cumulative incidence of primary end‐point events in the PS‐matched study cohorts. Statistical analyses were carried out with SAS 9.4 Software (SAS Institute Inc.).

3. RESULTS

3.1. Characteristics of study patients

Of the total number of Korean patients registered in HIRA database in 2019, the number of patients included in the initial patient sample was 991 189 (about 2% of the entire national health insurance beneficiaries). Of those, patients with records of statin use prior to March 2019, those aged below 20 years at study entry, and those diagnosed with myopathy prior to index date were excluded. Resultantly, patients aged 20 years or above with a history of dyslipidemia who received statin therapy but with no prior statin‐associated muscle symptoms were selected, hence leading to a total of 18 894 adult patients eligible for study entry (4098 statin + metformin users versus 14  796 statin‐only users). Baseline characteristics of the initial cohort are summarized in Table 1. Then, 1:2 PS matching was carried out to minimize possible confounding effects due to disparities across baseline variables between groups. As a result, 4092 and 8161 patients in the statin + metformin and statin‐only groups were identified, respectively. The baseline characteristics of the PS‐matched groups are also described in Table 1. After PS matching, there were no significant differences between groups per standardized difference values, with regard to patient age, sex, comorbidities, and fibrate comedication patterns. Female patients were about 42% in both PS‐matched groups, and patients aged from 40 to 79 years accounted for approximately 90% of study patients in each group.

TABLE 1.

Baseline characteristics of study patients before and after propensity score matching.

	Before propensity score matching			After propensity score matching
	Statin alone, n (%)	Statin + Metformin, n (%)	Standardized difference	Statin alone, n (%)	Statin +Metformin, n (%)	Standardized difference
Total	14 796	4098	—	8161	4092	—
Sex
Male	6793 (45.91)	2360 (57.59)	0.235	4702 (57.62)	2357 (57.6)	<0.001
Female	8003 (54.09)	1738 (42.41)		3459 (42.38)	1735 (42.4)
Age (years)
20–39	800 (5.41)	214 (5.22)	0.059	422 (5.17)	213 (5.21)	0.003
40–59	6746 (45.59)	1765 (43.07)		3507 (42.97)	1762 (43.06)
60–79	6442 (43.54)	1904 (46.46)		3803 (46.6)	1902 (46.48)
≥80	808 (5.46)	215 (5.25)		429 (5.26)	215 (5.25)
Comorbidities
Alcoholism	53 (0.36)	21 (0.51)	0.023	32 (0.39)	18 (0.44)	0.007
Chronic kidney disease	61 (0.41)	13 (0.32)	−0.016	24 (0.29)	12 (0.29)	<0.001
Chronic hepatic failure	0 (0.00)	1 (0.02)	0.022	0 (0.00)	0 (0.00)	<0.001
Hypertension	1692 (11.44)	381 (9.3)	−0.07	762 (9.34)	381 (9.31)	−0.001
Hypothyroidism	219 (1.48)	43 (1.05)	−0.039	80 (0.98)	42 (1.03)	0.005
Comedication
Fibrate	143 (0.97)	68 (1.66)	0.061	115 (1.41)	65 (1.59)	0.015

3.2. Study outcomes

The incidence and risk of myopathy in statin + metformin users versus statin‐only users were analyzed in the PS‐matched groups, and the results are summarized in Table 2. In the statin + metformin group, a total of 263 (6.4%) patients encountered the end‐point event whereas 623 (7.6%) patients did in the statin‐only group. Per 1000 person‐years, the incidence of outcome events was 201.70 (95% CI of 188.04–216.36) in the statin‐only group versus 184.33 (95% CI of 165.27–205.59) in the statin + metformin group. In the final multivariable Cox model, the HRs and 95% CIs were adjusted for patient gender, age, alcoholism, chronic renal failure, chronic hepatic failure, hypertension, hypothyroidism, and fibrate comedication history. The risk of myopathy was assessed lower when statins were used together with metformin with an adjusted HR with a 95% CI of 0.84 (0.71–0.99). These results have been also confirmed with a KM curve. The risk of statin‐associated muscle symptoms was substantially lower with the statin + metformin combination as compared to the statin‐alone therapy (Figure 1).

TABLE 2.

Myopathy risk post propensity score matching.

	Number of events	Person‐years	Incidence rate per 1000 person‐years (95% CI)	Crude HR (95% CI)	Adjusted HR (95% CI)
Statin	623	3088.68	201.70 (188.04–216.36)	1.00 (Ref)	1.00 (Ref)
Statin + Metformin	263	1426.80	184.33 (165.27–205.59)	0.84 (0.71–0.99)	0.84 (0.71–0.99)

Note: HRs were adjusted for gender, age, alcoholism, chronic renal failure, chronic hepatic failure, hypertension, hypothyroidism, and fibrate use.

Abbreviations: CI, confidence interval; HR, hazard ratio.

FIGURE 1.

Kaplan–Meier curves post propensity score matching for cumulative hazard of a myopathy endpoint.

3.3. Myopathy risk per statin agent

To investigate potential differential effects associated with each statin agent, subgroup analyses per statin substance were performed (Table 3). Of different statin agents, atorvastatin and rosuvastatin were the top two substances that were most frequently prescribed in study patients. Although the rate of myopathy was more frequent with atorvastatin, pitavastatin, pravastatin, rosuvastatin, and simvastatin when compared to each corresponding metformin combination group, the between‐group difference in overall myopathy risk was not associated with statistical significance. In the atorvastatin + metformin group, a total of 96 (6.1%) patients developed the outcome event while 255 (7.8%) patients did in the atorvastatin‐only group, with an HR (95%) of 0.86 (0.68–1.09) which was approaching statistical significance. The risk analyses of fluvastatin and lovastatin were not performed due to the small sample size and the limited number of outcome events.

TABLE 3.

Myopathy risk per statin agent.

	Person‐years	Number of patients	Number of events	Incidence rate per 1000 person‐years (95% CI)	Crude HR (95% CI)	Adjusted HR (95% CI)
Atorvastatin	1269.42	3274	255	200.88 (180.00–224.17)	1.00 (Ref)	1.00 (Ref)
Atorvastatin + Metformin	554.62	1561	96	173.09 (144.30–207.62)	0.86 (0.68–1.09)	0.87 (0.68–1.10)
Fluvastatin	4.50	9	0	—	1.00 (Ref)	1.00 (Ref)
Fluvastatin + Metformin	4.17	9	1	239.66 (43.39–1323.83)	—	—
Lovastatin	1.90	3	0	—	1.00 (Ref)	1.00 (Ref)
Lovastatin + Metformin	1.68	3	0	—	—	—
Pitavastatin	224.68	594	44	195.84 (150.25–255.25)	1.00 (Ref)	1.00 (Ref)
Pitavastatin + Metformin	116.54	322	14	120.13 (73.50–196.36)	0.61 (0.34–1.12)	0.61 (0.33–1.12)
Pravastatin	75.08	190	16	213.12 (137.99–329.14)	1.00 (Ref)	1.00 (Ref)
Pravastatin + Metformin	41.07	103	12	292.19 (181.52–470.32)	1.36 (0.64–2.88)	1.29 (0.60–2.77)
Rosuvastatin	1409.35	3817	278	197.25 (177.54–219.16)	1.00 (Ref)	1.00 (Ref)
Rosuvastatin + Metformin	660.15	1962	127	192.38 (164.54–224.93)	0.97 (0.79–1.20)	0.97 (0.79–1.20)
Simvastatin	103.76	274	30	289.12 (213.82–390.94)	1.00 (Ref)	1.00 (Ref)
Simvastatin + Metformin	48.57	130	13	267.64 (168.08–426.18)	0.92 (0.48–1.77)	0.88 (0.46–1.69)

Note: HRs were adjusted for gender, age, alcoholism, chronic renal failure, chronic hepatic failure, hypertension, hypothyroidism, and fibrate use.

Abbreviations: CI, confidence interval; HR, hazard ratio.

3.4. Stratified myopathy risk analyses

To further explore potential risks associated with patient factors, additional subgroup analyses were carried out using stratification methods in the PS‐matched groups (Figure 2). No signal of increased myopathy risk was observed with any of the patient factors (sex, age, comorbidities, and fibrate comedication) when statin + metformin users were compared against statin‐only users. The risk of muscle toxicity approached a statistical significance in male patients, albeit not statistically significant, with an HR (95% CI) of 0.84 (0.68–1.04), suggesting that metformin is safe to be used together with statin therapy and not likely to increase statin‐associated muscle toxicities. Possibly due to the limited sample size and the low frequency of outcome events, none of the comorbid conditions and comedication factors were associated with a statistically significant risk of muscle toxicities in subgroup analyses.

FIGURE 2.

Propensity score‐matched hazard analysis for a myopathy endpoint. HRs were adjusted for gender, age, alcoholism, chronic renal failure, chronic hepatic failure, hypertension, hypothyroidism, and fibrate use. CI, confidence interval; HR, hazard ratio. *p‐value for interaction, **p‐value for trend.

4. DISCUSSION

In this study, we performed a retrospective cohort study to investigate myopathy risk in statin users when metformin was added to the preexisting statin therapy in dyslipidemia patients. We found that the risk of myopathy decreased when metformin was used together with statins (adjusted HR 0.84; 95% CI 0.71–0.99). However, in subgroup analyses per individual statin substance, none of the statins were associated with elevated myopathy risk when compared against the corresponding statin substance + metformin group. Stratified risk analyses also found no increased risk of myopathy in every stratum per patient factor including sex, age, comorbidity, and fibrate use history. To the best of our knowledge, this is the first study in Korea that investigated metformin's effects on myopathy risk in statin users. A growing number of patients are receiving antidiabetic therapy together with statins as the prevalence of metabolic disorders or dyslipidemia‐diabetes comorbid conditions are on the rise. ⁴⁰ , ⁴¹ Our study findings may benefit those patients at risk for T2DM treated with a statin but worried about muscle symptoms and those with diabetes already on metformin who are to be initiated on statin therapy. A decrease in muscle symptoms can also lead to improved statin compliance and subsequently more effective prevention of cardiovascular complications in those high risk patients.

A signal of elevated myotoxicity risk has been reported with other antidiabetic medications previously, particularly with SGLT‐2 inhibitors ²⁶ , ²⁷ , ²⁸ , ²⁹ and DPP‐4 inhibitors. ²⁵ , ³⁰ In clinical trials on Japanese T2DM patients, a decrease in muscle mass was observed with the use of ipragliflozin, an SGLT‐2 inhibitor, although it was not reported as a serious adverse event, such as sarcopenia. ²⁶ , ²⁷ Long‐term use of SGLT‐2 inhibitors may contribute to aggravation of diabetes‐associated sarcopenia. ²⁸ In a case report, an elderly Japanese woman with T2DM receiving SGLT‐2 inhibitor dapagliflozin developed sarcopenia and experienced a 25% decline in her bodyweight over 12 months. ²⁴ A potential drug interaction between SGLT‐2 inhibitor canagliflozin and rosuvastatin has been suggested in a case report on an elderly patient who complained of statin‐induced muscle pain after newly being started on SGLT‐2 inhibitor therapy. ²⁹ SGLT‐2 inhibitors have been thus far believed to have a low probability of drug inter‐actions, and the pathogenesis and underlying mechanisms of its potential interactions with statins have not been fully investigated. Interactions between the two drug classes can be potentially associated with permeability glycoprotein and organic anion‐transporting polypeptides or membrane transport proteins involved in the transport of drugs. ⁴² With regard to DPP‐4 inhibitors' potential adverse effect of myotoxicity, the EMA issued safety warnings based upon a series of case reports concerning patients treated with both DPP‐4 inhibitor and statin therapies concomitantly. ²⁵ , ³⁰ A DPP‐4 inhibitors' interaction with statins may play a role in the antidiabetic's possible link to this muscular side effect, but the underlying mechanism is not clearly understood as DPP‐4 inhibitors are not known to induce or inhibit cytochrome P450 enzymes. ⁴³ , ⁴⁴ , ⁴⁵ Skeletal muscles are one of the major tissues that express DPP‐4 extensively along with the exocrine pancreas, heart, kidney, blood vessels, and lymph nodes, and possible off‐target effects of DPP‐4 inhibitors might have contributed to some adverse reactions in muscles. ⁴⁶ , ⁴⁷ In 2020, a nested case–control study in T2DM patients on statins or fibrates was conducted based on a national French health insurance data and showed that muscular injury risks were not elevated with DPP‐4 inhibitor use among statin‐treated patients with a comorbid condition of T2DM. ³⁶ Studies thus far provided conflicting results, and hence further studies are required to verify the potential link between these antidiabetic agents, statins, and potential adverse effects on skeletal muscles.

Although metformin demonstrated protective effects in terms of statin‐associated muscular adverse effects in our study when added to the preexisting statin therapy, uncertainty remains about the underlying mechanism by which metformin‐statin interactions affect myotoxicity risk in statin users. It is known that statin‐induce muscle toxicities are mediated by functional impairments of mitochondria and activation of atrogin‐1. ²⁰ Although statins are among the most widely prescribed pharmaceutical agents in the world, statin muscle pain often limits their use, and there has been no validated therapy for statin‐associated muscle symptoms thus far. In an observational study, Carris et al. investigated the effects of metformin plus statin combination on the prevalence of myopathy. ³⁵ The authors found that metformin + statin combination did not increase the myopathy prevalence, moreover this combination decreased muscle symptoms: multivariable regression analyses found that metformin was linked to a statistically significant risk reduction in muscle symptoms (a 22% risk decrease in muscle cramps [p = .049] and a 29% risk decrease in leg pains while walking [p = .01]), ³⁵ which is consistent with our study findings.

In a separate study, Carris et al. proposed possible mechanisms behind this phenomenon, suggesting metformin's pleiotropic effects of rescuing statin‐associated muscle symptoms. ³⁴ There are overlapping mechanisms underlying statin‐associated muscle toxicities and statin‐induced insulin resistance. ³⁴ It has been established that metformin is effective in decreasing insulin resistance and preventing incident diabetes mellitus, ⁴⁸ , ⁴⁹ but its effects on statin‐associated muscle symptoms have not been elucidated yet. Considering the overlapping mechanisms behind the aforementioned conditions, metformin use may provide beneficial effects on not only incident T2DM development but also muscle symptoms, the two common side effects of statins. ³⁴ In line with the proposed hypothesis, metformin indeed showed a risk reduction in muscle toxicities among statin users in our real world data‐based cohort study. Additionally, metformin may alleviate statin's adverse effects on muscles by activating peroxisome proliferator–activated receptor gamma coactivator 1‐alpha (PGC‐1α) which stimulates mitochondrial production and by downregulating atrogin‐1, a muscle specific E3 ubiquitin (Ub) ligase via adenosine monophosphate–activated protein kinase (AMPK) activation. ⁵⁰ Metformin also helps lower cholesterol levels by stimulating AMPK. ³³ These findings suggest that metformin may be effective in lowering statin‐associated muscle toxicities either through promoting mitochondrial function or through the atrogin‐1 pathway. ³³

Several studies demonstrated that statins contribute to the development of new‐onset diabetes mellitus, ⁵¹ , ⁵² , ⁵³ potentially via affecting Forkhead box protein (FOXO) group pathway, which is also related to the mechanism by which statins may induce muscle toxicities. ³³ Elsaid et al. conducted a systematic search of literature to investigate the association between statins, metformin and statin‐induced muscle toxicities and to evaluate the hypothesis that metformin could mitigate statin‐associated muscle complaints. ³³ Studies have shown that metformin can protect skeletal muscles from cardiotoxin‐induced injury through AMPK activation and calcium hemostasis ⁵⁴ and also minimize muscle cachexia by modulating skeletal muscle wasting effects induced by tumors and abnormal muscle protein metabolism. ⁵⁵ Additionally, statin‐associated muscle symptoms are possibly elicited by mitochondrial disorders and oxidative stress, altering type II glycolytic myofibers in particular since they have a low number of mitochondria. ⁵⁶ Metformin can therefore improve the activity of mitochondrial oxidative enzyme in skeletal muscles via AMPK activation and subsequently alleviate stain‐induced myotoxicities.

The American Heart Association (AHA) guidelines recommend that diabetes patients aged 40–75 years should receive moderate‐intensity statin therapy regardless of 10‐year atherosclerotic cardiovascular disease (ASCVD) risk scores and that diabetic patients with multiple ASCVD risk factors should be treated with high‐intensity statin therapy to decrease low‐density lipoprotein cholesterol (LDL‐C) by ≥50%. ⁵⁷ Metformin is also recommended for diabetes prevention in high‐risk patients albeit rarely used for that indication. ⁵⁸ , ⁵⁹ , ⁶⁰ Our study findings demonstrated the safety of metformin use in statin‐treated dyslipidemia patients with respect to myotoxicity risks, a potential side effect of statins. This study also gives a further assurance to both patients and clinicians as well as to pharmaceutical companies making combination products of statin + metformin that it is safe to promote both therapies in patients with multiple metabolic disorders as comorbidities and that metformin may even provide protective effects on statin‐associated muscle toxicities. The relevance of this relationship is augmented in consideration of the increasing number of patient populations with metabolic dysfunctions worldwide and the tendency for them to receive statins alongside glucose‐lowering therapy.

Despite several studies suggest a protective effect of metformin on statin‐associated muscle symptoms, clinical trials or observational studies that support the potential relation have been limited. Therefore, further studies are required to investigate the potential link between metformin, statins, and skeletal muscle toxicities. It is known that metformin enhances AMPK activation, indicating positive effects on controlling cholesterol levels, and also has strong glucose‐lowering effects, which are particularly beneficial to statin‐receiving patients as they are at increased risk for incident diabetes mellitus. In these patients with metabolic risk factors, combined use of statin and metformin is advised to confer protective effects on both cardiovascular and metabolic outcomes.

4.1. Limitations

This study is subject to several limitations inherent to a retrospective cohort study based on national health insurance claims data. First, patient comorbidities and outcome events were identified based on ICD‐10 diagnosis codes; hence, inaccurate documentation of diagnostic codes in health insurance data can results in over‐ or under‐estimation of the distribution of comorbidities and the frequency of endpoint events. Second, the study data were extracted from a single‐year cohort from the HIRA database. A follow‐up study with a larger sample size and a longer study period is desired to verify our study findings and to investigate long‐term effects of statin and metformin comedication. We were not able to extract relevant laboratory values, such as creatine kinase, lipid, and glycemic control levels, not available in the HIRA database. We included only those patients who have a recorded dyslipidemia diagnosis during the patient recruitment period, but we did not have complete information on lipid panels at study entry. Although PS‐matching and an adjusted multiple logistic regression model were utilized, residual confounding effects associated with unidentified covariates are still possible. For instance, exposure to any substances that inhibit or induce cytochrome P450 enzymes could have affected statin metabolism; however, such effects were not accounted for in this study. In addition, differential effects on myotoxicity risks could have been exerted not only by different statin substances but by different potencies of statin therapy. In this study, however, statin dose was not adjusted in the final multivariable Cox model because we have focused on evaluating differential myotoxicity risks by statin substance rather than by the potency of statin therapy, depending on metformin exposure. Therefore, caution is advised when interpreting the study results. Lastly, it was assumed that prescriptions were all filled and dispensed, and study patients completed their entire medication regimens, even though patient adherence can be suboptimal. Despite these limitations, our study results show the beneficial effects of metformin use in statin‐receiving patients in terms of reducing muscle toxicity risk.

5. CONCLUSIONS

This study found that, when metformin was used together with statin therapy in dyslipidemia patients, myopathy risk was lower than in statin‐only users. However, differential risks associated with specific statin agents were not detected. These findings suggest that metformin is safe to be used concomitantly with statins in patient populations with metabolic disorders and that it may also provide protective effects on potential muscle toxicities induced by statin therapy.

AUTHOR CONTRIBUTIONS

Keunhyeong Bak and Sooyoung Shin conceptualized the project and performed data extraction, curation, and analysis. Keunhyeong Bak, Suhyeon Moon, Minjung Ko, Yeo Jin Choi, and Sooyoung Shin interpreted the analyses and participated in original draft preparation of the main text. Yeo Jin Choi and Sooyoung Shin supervised and validated data analyses. Yeo Jin Choi and Sooyoung Shin edited, reviewed, and approved the final edition of the manuscript.

CONFLICT OF INTEREST STATEMENT

The authors report no conflicts of interest in this research.

ETHICAL APPROVAL STATEMENT

The study protocol was approved by the Institutional Review Board of Ajou University (202201‐HB‐EX‐004). Informed consent from study patients was waived and no further ethics approval was required as the authors are authorized by the HIRA to use the de‐identified patient data for research purposes.

Bak K, Moon S, Ko M, Choi YJ, Shin S. Impact of metformin on statin‐associated myopathy risks in dyslipidemia patients. Pharmacol Res Perspect. 2023;11:e01114. doi: 10.1002/prp2.1114

DATA AVAILABILITY STATEMENT

The data that this study was based upon are available for purchase from the HIRA. Restrictions may apply to the availability of the HIRA data, which were used with permission for this study.