Drug risks associated with sarcopenia: a real-world and GWAS study

Abstract

Introduction

Drug-induced sarcopenia has not received adequate attention. Meanwhile, there is growing recognition of the importance of effective pharmacovigilance in evaluating the benefits and risks of medications.

Aims

The primary aim of this study is to investigate the potential association between drug use and sarcopenia through an analysis of adverse event reports from the Food and Drug Administration (FDA) Adverse Event Reporting System (FAERS) and to evaluate the genetic factors contributing to drug-induced sarcopenia using summary-data-based Mendelian randomization (SMR).

Methods

We obtained reports of adverse drug reactions from FAERS. Primary outcomes included sarcopenia and potential sarcopenia. We calculated the Proportional reporting ratio (PRR) to assess the risk of specific adverse events associated with various drugs, applying chi-square tests for statistical significance. Additionally, we used SMR based on Genome-wide association study (GWAS) to evaluate the potential associations between drug target genes of some significant medications and sarcopenia outcomes. The outcome data for sarcopenia included metrics as hand grip strength and appendicular lean mass (ALM).

Results

A total of 55 drugs were identified as inducing potential sarcopenia, and 3 drugs were identified as inducing sarcopenia. The top 5 drugs causing a potential risk of sarcopenia were levofloxacin (PRR = 9.96, χ² = 1057), pregabalin (PRR = 7.20, χ² = 1023), atorvastatin (PRR = 4.68, χ² = 903), duloxetine (PRR = 4.76, χ² = 527) and venlafaxine (PRR = 5.56, χ² = 504), and the 3 drugs that had been proved to induced sarcopenia included metformin (PRR = 7.41, χ² = 58), aspirin (PRR = 5.93, χ² = 35), and acetaminophen (PRR = 4.73, χ² = 25). We identified electron-transfer flavoprotein dehydrogenase (ETFDH) and protein Kinase AMP-Activated Non-Catalytic Subunit Beta 1 (PRKAB1) as the primary drug target genes for metformin, while Prostaglandin-endoperoxide Synthase 1 (PTGS1) and Prostaglandin-endoperoxide Synthase 2 (PTGS2) were considered the primary action target genes for aspirin and acetaminophen according to DrugBank database. SMR showed that the expression abundance of ETFDH was negatively correlated with right hand grip strength (blood: OR = 1.01, p-value = 1.27e-02; muscle: OR = 1.01, p-value = 1.42e-02) and negatively correlated with appendicular lean mass (blood: OR = 1.03, p-value = 7.73e-08; muscle: OR = 1.03, p-value = 1.67e-07).

Conclusions

We find that metformin, aspirin, and acetaminophen are specifically noted for their potential to induce sarcopenia based on the analyses conducted. We perform signal mining for drug-associated sarcopenia events based on real-world data and provides certain guidance for the safe use of medications to prevent sarcopenia.

Background

The characteristic feature of sarcopenia is age-related decline in skeletal muscle mass and quality, and may lead to adverse outcomes such as frailty, fractures, immobility, hospitalization, and even death, especially for elderly individuals. As the prevalence of age-related diseases increases, sarcopenia may further jeopardize the physical and mental health of middle-aged and elderly individuals, increasing healthcare expenditures [1]. Furthermore, there are secondary causes of sarcopenia, such as those resulting from other diseases or drugs. They are relatively more preventable than primary sarcopenia. Currently, drug-related sarcopenia has not received much attention [2]. Some studies have reported drugs that might cause muscle-related side effects: Ganga et al. [3] conducted a systematic review and they found that the percentage of muscle-related problems was generally higher in the statin treatment group (12.7%) compared to the placebo group (12.4%, p-value = 0.06). The systematic review by Biguetti et al. [4] revealed that among 1367 patients from 5 studies using hydroxychloroquine or chloroquine, 37 patients exhibited myopathy, while 252 patients showed elevated levels of muscle enzymes (aldolase, creatine kinase, and lactate dehydrogenase). During clinical drug treatment, especially with long-term or combination therapy, drug-related muscle atrophy is often overlooked. In clinical practice, if doctors can detect drug-induced sarcopenia early and discontinue the medication promptly, it can improve the patient’s prognosis. Currently, there is still a lack of relevant research in this area [2]. Drug-induced sarcopenia remains an underrecognized phenomenon in both clinical practice and pharmacovigilance. Many studies lacking comprehensive analyses that utilize large-scale adverse event reporting systems. This gap underscores the need for research that directly investigates the relationship between specific drugs and sarcopenia outcomes [2].

Food and Drug Administration (FDA) Adverse Event Reporting System (FAERS) is the world’s largest spontaneous reporting adverse event repository, collecting and documenting adverse events related to drug therapy in the real world. It serves as an important method for identifying early safety issues with medications [5, 6]. OpenVigil is an another web-based analytical tool designed for FAERS data analysis [7]. It can perform disproportionality analysis on the data, directly applying drug vigilance results to real-life scenarios [8]. It primarily utilizes methods such as the Proportional Reporting Ratio (PRR) and Reporting Odds Ratio (ROR) to analyze adverse reaction signals.

Mendelian randomization (MR) is a novel research method that uses genetic variation as an instrumental variable (IV) to estimate the causal relationship between exposure and outcome [9]. Furthermore, MR can be used to assess the relationship between drug target genes and diseases. Mendelian randomization is considered superior to observational studies because it avoids interference from reverse causation and confounding factors [10]. Another approach is Summary data-based Mendelian randomization (SMR), which utilizes GWAS data from quantitative trait locus (QTL) studies and summary statistics to investigate the multigenic associations between gene expression levels and complex outcomes [11]. Therefore, we can use GWAS to study the potential association between our drug target genes of interest and sarcopenia, verifying whether the drug has other potential targets that lead to sarcopenia while also avoiding reverse causation.

In this study, we utilize the OpenVigil tool to analyze the risk of drug-related sarcopenia and classify and summarize drugs that may contribute to sarcopenia. Mendelian randomization study based on GWAS further elucidates the potential causal relationship between the primary drug target genes and sarcopenia. Our study aids in the early detection of adverse drug events and provides initial insights into their potential mechanisms.

Methods

Screening of adverse drug reaction reports

Supplementary Table 1 provides information about the databases and sources used in this study. All systems are accessible to all users. Due to the nature of the data sources, informed consent from the original patients was not required for this study. The human-related datasets used in this study were provided by third parties and are available for open use in any research project, with appropriate ethical approval obtained for each study where the data were utilized.

FAERS is a database that contains adverse event reports, medication error reports and product quality complaints resulting in adverse events that were submitted to FDA, which is designed to collect, analyze, and report adverse events related to drug use [5]. Adverse events and medication errors are coded using terms in the Medical Dictionary for Regulatory Activities (MedDRA) terminology. As of now, FAERS has received a total of 29,661,136 reports, including 2,683,991 death reports. Reports are submitted to FAERS by healthcare professionals, consumers, and manufacturers. The FDA receives voluntary reports directly from healthcare professionals (such as physicians, pharmacists, nurses, and others) as well as consumers (including patients, family members, lawyers, and others) [5]. Additionally, some reports in the FAERS data are duplicates, which are most commonly found in reports submitted by both healthcare professionals and consumers. Therefore, we only selected reports submitted by healthcare professionals. We used OpenVigil 2.1 to extract adverse drug reaction reports from the FAERS database. The reporting period was restricted to the first quarter of 2004 to the second quarter of 2023. We selected a specific age range for our analysis based on the epidemiology of sarcopenia, which predominantly affects middle-aged and older individuals and the age of the population was limited to those 30 years and older. Based on the Asian Working Group for Sarcopenia (AWGS) 2019 definition of sarcopenia [6], we categorized the primary outcomes into 2 types: one was sarcopenia (Search term: “sarcopenia”), and the other outcome was defined as potential sarcopenia (Search terms: “Muscle degeneration” OR “Muscle atrophy” OR “Muscle weakness” OR “Weakness muscle” OR “Muscle weakness aggravated” OR “Generalised muscle weakness” OR “Activities of daily living impaired” OR “Underactivity”). The variables we collected include, but are not limited to, patient demographics (such as age and sex), types of adverse events, and reporting regions. Finally, we calculated the proportional reporting ratio (PRR) based on the number of adverse event reports, using the formula: PRR = (A/(A + B)) / (C/(C + D)), where A represents the number of specific adverse reactions reported for the drug, B denotes the number of other adverse reactions reported for the drug, C indicates the number of specific adverse reactions reported for other drugs, and D represents the number of other adverse reactions reported for other drugs. Chi-square tests were employed to examine differences in frequency distributions.

Standardization of drug names

Based on the RxReasoner database, we obtained the generic names and classifications for each drug. After merging drugs with the same Anatomical Therapeutic Chemical (ATC) codes, we used the PRR and χ². We established filtering criteria of drug-event counts ≥ 30, PRR > 2, and χ² > 4 to further screen the drugs [12].

SMR analysis

Mendelian randomization requires 3 valid instrumental variable assumptions for effective inference: (1) Instrumental variables must be associated with the exposure; (2) Instrumental variables must be independent of any confounding factors related to the exposure-outcome relationship; and (3) Instrumental variable must be independent of the outcome, conditional on the exposure and confounding factors. We utilized SMR method to simulate the causal associations between the drug target genes of interest and the phenotypes. SMR is one of the extended methods of MR, specifically designed for summarizing GWAS and eQTL studies. It is a method for identifying polygenic associations between gene expression and complex traits and can explain all results under a polygenic model. To satisfy these three fundamental assumptions, first, the IVs in SMR analysis were significant QTLs (P < 5 × 10e-8) located within 100 kb of the drug target gene regions. Secondly, to ensure the robustness of the results, we selected loci with an r² < 0.3 and a minor allele frequency (MAF) > 0.01. We obtained the target genes of the drugs based on the DrugBank database. Since a single drug often has multiple targets, we focused exclusively on the primary targets with established pharmacological effects reported by related literature. We obtained the outcome data for sarcopenia from publicly available GWAS. We used left-hand grip strength, right-hand grip strength, and appendicular lean mass (ALM) as outcomes to assess muscle mass [13, 14]. The were all sourced from the summary GWAS of the UK Biobank cohort [15]. The eQTL data from blood and muscle tissue were respectively obtained from the eQTLGen Consortium (https://www.eqtlgen.org/) and the GTEx database (https://gtexportal.org/). They serve as proxies for drug target genes [16]. We derived estimates of the causal effect of drug targets on outcomes, typically expressed as odds ratios (OR) and calculated 95% confidence intervals for these estimates to provide a measure of precision. We reported p-values to assess the statistical significance of our findings. The HEIDI (Heterogeneity in Effect Estimates of Direct and Indirect Effects) test is a statistical method used to detect potential linkage disequilibrium in SMR results. A p-value less than 0.05 indicates that the association may be influenced by linkage disequilibrium [17].

SMR analysis was conducted using SMR software version 1.03 (https://cnsgenomics.com/software/smr/#Overview) [17].

Results

The population distribution characteristics of adverse reactions

Due to the limited number of reports directly causing sarcopenia, we have adjusted the screening threshold to PRR > 2 and χ² > 4. During the study period, a total of 4107 cases were reported to be associated with potential sarcopenia, while 33 cases were directly linked to adverse events causing sarcopenia. As shown in Table 1, we had compiled baseline information from adverse reaction reports. We could observe that these two definitions of outcomes have different population distributions: Potential sarcopenia was most commonly reported in females aged 45–60, with a predominant reporting location in North America, while sarcopenia was more frequent in males aged 60 and above, with a predominantly European population. The most common adverse outcome types were hospitalization or prolonged hospital stays.

Table 1 Population characteristics sarcopenia-related reports

Drugs with a high reporting frequency

A total of 55 drugs were identified as inducing potential sarcopenia. The detailed data for this section can be found in supplementary Table 2. Table 2 listed the top 30 drugs associated with potential sarcopenia. The top 5 drugs causing a potential risk of sarcopenia were levofloxacin (PRR = 9.96, χ² = 1057), pregabalin (PRR = 7.20, χ² = 1023), atorvastatin (PRR = 4.68, χ² = 903), duloxetine (PRR = 4.76, χ² = 527) and venlafaxine (PRR = 5.56, χ² = 504).

Table 2 Top 30 drugs most associated with potential Sarcopenia

We organized the indications, ATC codes, and PRR for these 55 drugs. The results are shown in Fig. 1. It could be observed that the drugs associated with potential sarcopenia were mainly related to the Nervous System, Musculoskeletal System, and Cardiovascular System.

Fig. 1

Sunburst chart created based on the drugs’ anatomical system subgroups, treatment subgroups, and ATC codes. The width of each slice represents the relative PRR, with larger slices indicating stronger signals

The 3 drugs reported to be associated with sarcopenia included metformin (PRR = 7.41, χ² = 58), aspirin (PRR = 5.93, χ² = 35), and acetaminophen (PRR = 4.73, χ² = 25) (Supplementary Table 3). We intersected the lists of drugs associated with causing sarcopenia and those linked to potential sarcopenia, and found that the only common drugs between the two groups were aspirin and acetaminophen. This finding suggests that these two drugs may have a higher likelihood of being associated with the risk of sarcopenia.

Drug-target SMR

We searched for the primary drug target genes of metformin, aspirin, and acetaminophen. We identified ETFDH and PRKAB1 as the primary drug target genes for metformin [18, 19]. PTGS1 and PTGS2 were considered the primary action target genes for aspirin and acetaminophen [20]. Among these, metformin is considered an inhibitor of ETFDH and an agonist of PRKAB1, while aspirin and acetaminophen are inhibitors of PTGS1 and PTGS2. SMR described the genetic associations of these drug target genes with left and right hand grip strength and ALM (Supplementary Table 4, Fig. 2). Due to the lack of eQTL data for PTGS1 and PTGS2 in muscle tissue, we only simulated the genetic effects of PRKAB1 and ETFDH in muscle tissue. The results showed that the associations of PTGS1, PTGS2, and PRKAB1 with the outcome indicators were not statistically significant (p-value > 0.05), While the expression levels of ETFDH might be associated with the outcomes to some extent. The expression abundance of ETFDH was negatively correlated with right hand grip strength (blood: OR = 1.01, p-value = 1.27e-02; muscle: OR = 1.01, p-value = 1.42e-02) and negatively correlated with appendicular lean mass (blood: OR = 1.03, p-value = 7.73e-08; muscle: OR = 1.03, p-value = 1.67e-07). HEIDI test did not find evidence that the results were affected by linkage disequilibrium. This suggested that metformin-induced sarcopenia was probably associated with its inhibition of ETFDH expression [18].

Fig. 2

Forest plot illustrating the genetic effects of ETFDH on the outcomes

Discussion

Currently, clinical staff have insufficient awareness of drug-induced sarcopenia. Evidence-based early identification and intervention of the condition through data mining will facilitate necessary treatment, improving patient prognosis, minimizing hospitalization, and reducing the burden of patient care [21]. Our study conducted disproportionality analysis based on the FAERS to assess signals related to sarcopenia. Additionally, we utilized the Mendelian randomization method to evaluate the association between drug targets and sarcopenia. This research identified certain drugs that may contribute to sarcopenia and discovered that the ETFDH target of metformin could potentially underlie its muscle-related adverse effects. Our findings hold significant clinical implications. First, by identifying drugs associated with sarcopenia, clinicians can better assess the medication risks for their patients and intervene promptly to prevent disease progression. Furthermore, these findings provide valuable evidence for drug safety monitoring, highlighting the need to enhance drug monitoring and education to raise awareness among healthcare professionals regarding drug-induced sarcopenia.

Our research findings corroborate the conclusions of some scholars. In a randomized controlled trial conducted by Parker et al. [22], subjects who orally received atorvastatin experienced a mean increase in creatine kinase of 20.8 ± 141.1 U/L (p-value < 0.0001), and there were more incidents of myalgia compared to the placebo group. A systematic review comprising 7 studies found that in adult rheumatoid arthritis patients, the use of corticosteroids was positively associated with sarcopenia (OR = 1.46) [23]. Nakayama et al. conducted a cohort study on Japanese rheumatoid arthritis patients and found an association between the use of nonsteroidal anti-inflammatory drugs (NSAIDs) and sarcopenia [24]. Evidence regarding some drugs causing sarcopenia remains unclear, necessitating further clinical research for validation. Adverse reaction reporting follows the principle of reporting when suspicion arises, which may lead to a certain degree of false positives. However, we still need to approach these conclusions with caution, as early diagnosis and treatment of the condition remain of paramount importance [25].

Our study identified 3 drugs potentially associated with sarcopenia: metformin, aspirin, and acetaminophen. Currently, many antidiabetic medications are believed to affect skeletal muscle metabolism [26]. Among them, biguanides may be involved in this process through mechanisms such as altering insulin sensitivity and intervening in the AMPK pathway [27]. ETFDH is one of the primary drug targets of metformin. Metformin inhibits mitochondrial respiratory chain complex by suppressing ETFDH [28]. Some scholars have pointed out that ETFDH is crucial for the oxidative phosphorylation efficiency of mitochondria in skeletal muscle [29, 30]. This could be one of the primary mechanisms by which metformin causes sarcopenia and is consistent with our main conclusion. There is no evidence to suggest that aspirin and acetaminophen’s primary drug target genes are factors causing muscle loss. Further research is needed to supplement this viewpoint.

The strength of this study lies in its analysis based on real-world data, utilizing the FAERS database for disproportionality analysis, which reveals potential associations between drugs and sarcopenia. Additionally, the use of Mendelian randomization analysis enhances the reliability of the results, allowing for causal inferences. This study is the first systematic assessment of drug-induced sarcopenia, laying the groundwork for further research in this area. There are still some limitations. It should be acknowledged that the methods employed in this study did not establish definite causality. Further carefully designed controlled trials are needed to establish the causal relationship between drugs and adverse events [31]. FDA reports also have certain limitations: They do not require proof of causality between the product and the event, and reports often lack sufficient detail for accurate assessment. Additionally, various factors may influence whether an event is reported, such as the time since the product’s market launch and the public attention surrounding the event, which makes it is subject to reporting bias and confounding factors. Due to the presence of duplicate reports in the FDA reporting system, it cannot be used to calculate the incidence of adverse events or medication errors, meaning there is a lack of data on overall drug exposure in the population. This limitation restricts our ability to accurately assess the true risk of outcomes following drug use. Mendelian Randomization simulated the impact of certain target genes on outcomes. On the one hand, our GWAS population consisted of individuals of European descent, so whether the conclusions can be extended to other populations remains to be studied. On the other hand, a drug may have numerous target genes, and we only selected the most significant ones. There are various methods for using loci within cis-regulatory regions as IVs, such as Top SNP approach we employed. However, there is still a lack of consensus in the literature regarding the use of different methods [32, 33]. The specific mechanisms still rely on in vivo and in vitro studies. Despite these limitations, our study remains meaningful.

Conclusions

In conclusion, our study highlights the potential risk of drug-induced sarcopenia and underscores the importance of pharmacovigilance and genetic analysis in identifying and understanding drug-related muscle adverse events. These findings have implications for clinical practice, emphasizing the need for personalized medicine approaches and tailored interventions to mitigate the risk of sarcopenia associated with certain drugs.