In this longitudinal cohort study, we examined the prescribed medications, chronic diseases, and physical and psychological status of individuals to identify whether polypharmacy is associated with an increased risk of sarcopenia development over a 9-year period among community-dwelling older adults with no long-term care needs. After adjusting for multifaceted confounders including time-varying factors, polypharmacy combined with the use of STOPP-J-listed PIMs or potentially muscle-wasting drugs, but not polypharmacy alone, was significantly associated with an increased OR of sarcopenia onset.
To the best of our knowledge, our study provides the first evidence of the possibility of a longitudinal association between polypharmacy combined with PIMs and sarcopenia development among community-dwelling older adults. Similarly, a previous cross-sectional study reported that polypharmacy (≥ 5 medications) in older adults was associated with a 2.2-fold higher prevalence of sarcopenia based on low appendicular lean mass [7]. Our results are consistent with the findings from previous observational studies in other populations showing associations between polypharmacy and adverse geriatric outcomes, such as frailty, sarcopenia, and mortality [5, 7–9, 21].
Several possible explanations exist for the longitudinal association between polypharmacy and new-onset sarcopenia. First, older adults have an increased risk of adverse drug reactions (ADRs) because of age-related changes in pharmacokinetics, such as absorption, bioavailability, distribution, metabolism, and excretion [22]. For example, physical frailty was associated with poor kidney function in patients with chronic kidney disease [23]. Thus, poor pharmacokinetics may enhance the adverse outcomes of polypharmacy and long-term dosing.
Second, polypharmacy is strongly associated with PIM use [24], prescribing cascades [25], and low adherence [26]. A prescribing cascade occurs when nonspecific signs and symptoms of ADRs are misinterpreted as disease, and further medications are prescribed to treat the misinterpreted conditions. Unfortunately, because the symptoms caused by polypharmacy usually mimic those of aging, such as depressive symptoms [27], cognitive decline [28], and frailty [8, 9], the probability of prescribing cascades is higher among older adults exhibiting frailty. Thus, prescribing cascades can lead to the prescription of PIMs [25]. Our results also showed that individuals with polypharmacy were also prescribed PIMs or potentially muscle-wasting drugs, suggesting that polypharmacy may be linked to overmedication. Participants with new-onset sarcopenia were not only older at baseline but also exhibited lower physical and cognitive function, as well as more depressive symptoms. It is conceivable that psychological characteristics contributed to the lower adherence. These negative influences can explain why polypharmacy is associated with preventable and unplanned hospitalization [5, 29] and ADRs [4, 5, 30]. Our results showed that the risks of sarcopenia did not increase until ≥ 5 medications were used, supporting previous findings that older Japanese adults taking ≥ 6 medications were more likely to experience ADRs [31]. Therefore, in individuals with polypharmacy, the high risk of ADRs associated with age-related changes in pharmacokinetics, PIMs, and low adherence to prescribed medications could explain the increased risks of adverse outcomes.
Furthermore, we found that PIMs or potentially muscle-wasting drug use tended to be associated with increased sarcopenia risk. In addition to medication number and possible inter-drug interactions, the types of drugs used may directly or indirectly affect sarcopenia. For example, some drugs may negatively affect muscles, inducing eating disorders, or may lead to weight loss, which influences body composition [10, 11]. Potentially muscle-wasting drugs, such as statins, can cause muscle toxicity, whereas other PIMs (e.g., glucocorticoids, β blockers, and NSAIDs) may have detrimental metabolic effects. Statins may also induce serious adverse effects, including elevated liver enzymes and skeletal muscle abnormalities, ranging from benign myalgias to severe rhabdomyolysis [10, 11], potentially derived from mitochondrial dysfunction, alterations in apoptosis-related gene expression, protein degradation, and genetic predisposition. Such mechanisms are also involved in sarcopenia pathogenesis and may underlie the association between statins and sarcopenia. Furthermore, the use of PIMs such as benzodiazepine constitutes a risk factor for adverse health outcomes and may be contraindicated for long-term use [32–34]. Therefore, despite the benefits of appropriate medications for older individuals, the prescription of too many medications, including PIMs, may negatively influence geriatric endpoints such as sarcopenia. Consistently, after adjustment for potential confounders, including time-varying factors, we observed that polypharmacy with PIMs was significantly associated with sarcopenia development, thereby explaining the longitudinal association between polypharmacy combined with PIMs and increased sarcopenia risk over a 9-year period.
This study has some limitations. First, although we adjusted for several confounding factors including time-varying variables and used a GEE with repeated measures when calculating the statistical values, some biases may have remained owing to unobserved medical conditions or altered renal or liver function. Second, information on the duration of drug treatment or medication adherence was lacking. As such, the exact doses of the prescribed medications were unknown. Third, ADR occurrence, medication underuse, and medication duplication were not assessed. Fourth, further prospective randomized control trials are necessary to evaluate the potential benefits of de-prescribing on sarcopenia risk. Notably, this study included a sample of community-dwelling older adults who participated through randomization, for whom health awareness is expected to be high, and the population representativeness is relatively high; consequently, the generalizability of the results is high. However, further research is required to determine whether similar results would be obtained in countries with different healthcare supply systems.