Effect of muscle mass and strength on depression: does Mendelian randomization address causality?

Sarcopenia is a progressive and generalized skeletal muscle disorder that was first described as an age-related decline in lean body mass. The definition of sarcopenia by the European Working Group on Sarcopenia in Older People (EWGSOP) now includes low muscle strength and muscle mass or quality in the concept of sarcopenia. This definition is supported by the Asian Working Group on Sarcopenia, whereas the Sarcopenia Definitions and Outcomes Consortium (SDOC) in the United States defines sarcopenia as low muscle strength and low gait speed. These definitions emphasize the new focus on muscle function, usually defined by muscle strength, muscle power, or physical performance.(1)

This paradigm shift from low muscle mass to function as the defining feature of sarcopenia was precipitated by studies showing muscle function to be a more powerful predictor of clinically relevant outcomes than muscle mass alone.(1) Although muscle function outcomes such as fractures and falls have been included in predictive validity tests of sarcopenia definitions, a substantial proportion of the validity tests investigated global health outcomes such as mortality or hospitalization. This is in stark contrast to other medical diagnoses for which clear disease-specific outcomes have been defined for validation purposes,(2) and is reflected in the fact that an operational definition of sarcopenia still lacks consensus. Importantly, knowledge of sarcopenia has evolved to encompass not just a muscle-centered perspective but also the role played by the nervous system in the regulation of muscular contraction, and their reciprocal functions. Thus, skeletal muscle mass, strength, and function have been linked to common mental disorders such as depression. The prevalence of depression in sarcopenia is estimated to be 28% and the coexistence of low muscle mass and strength elevates the risk for depressive symptoms with muscle function showing a stronger association.(3)

Identifying risk factors and mechanisms that cause psychiatric diseases poses a challenge because of the numerous confounders. The potential for reverse causality between the risk factors and these diseases that are generally multifactorial, compounds the issue. Mendelian randomization (MR) is an analytic tool that overcomes these drawbacks of confounding and reverse causation by assigning individuals to groups based whether or not they carry a particular genetic variant that is linked to an exposure. Randomization occurs by natural distribution of the genetic variant in the population. The analysis is designed to determine whether an observational association between the exposure and health outcome is compatible with a causal relationship between them. The assumptions of MR are presented in Figure 1. The gene variant, exposure, and health outcome can be derived from the same population (single sample randomization) or the gene variant and exposure could be from one population and the gene variant and health outcome could be from another population (double sample randomization).

Figure 1.

Graphic example of Mendelian Randomization representing the assumptions (a) the genetic variant must be associated with the risk factor or exposure; (b) the genetic variant must not be associated with any confounder of the exposure-outcome association; and (c) the genetic variant must be associated with the outcome only through the exposure. ALM: appendicular lean mass DSM: Diagnostic and Statistical Manual of Mental Disorders; HPA: hypothalamic pituitary axis; ICD: International Classification of Diseases; SNP: single nucleotide polymorphism. Figure created in Biorender.com

Since MR uses genetic variants as proxies for risk factors (which in general, are present before fully syndromal psychiatric disease), reverse causality is unlikely. To reduce the risk of population heterogeneity in samples of psychiatric patients, the Diagnostic and Statistical Manual of Mental Disorders (DSM) or International Classification of Diseases (ICD) diagnoses rather than self-assessment questionnaires or clinical scores are preferred. In this issue, Zhang et al investigated whether sarcopenia modifies the risk of depression in a double sample MR study. The methods included use of genetic instruments from genome-wide association studies of non-overlapping datasets from the UK and FinnGen cohort biobanks, a diagnosis of depression based on ICD, and appropriate methods to assess for adherence to the key assumptions of MR. Appendicular lean mass (ALM, which represents total amount of muscle mass in the arms and legs), hand grip strength, and walking pace were the sarcopenia traits whose causal association with depression were evaluated. Bioelectrical impedance analysis (BIA) which was used to assess ALM is a validated measure of body composition, but it has limited reliability particularly in individuals with BMI>34 kg/m². The results showed that only ALM was negatively correlated with the risk of depression after accounting for hand grip strength, walking pace, and body mass index (BMI). Although low grip strength and slow walking pace increased the risk for depression, the causality dissipated when adjusted for the other sarcopenia traits suggesting a predominant causal effect of low ALM on the risk for depression.

In contrast, a study using MR to determine whether non-fat mass measured by BIA was a risk factor for depression, found no causal relationship. A caveat to this study is that patients were selected based on a relatively liberal definition of depression including self-reporting, clinical assessment, and examination of medical records.(4) In another MR study, using depression as the exposure rather than the outcome, a causal relationship between depression and both muscle mass and strength was demonstrated using genetic variants linked to depression and ALM and hand-grip strength as outcome indicators.(5)

Several large epidemiologic studies have also shown that low muscle strength is associated with depressive symptoms without a significant correlation with muscle mass. However, randomized controlled trials are singularly lacking. A small randomized controlled trial (N = 20) showed that resistance training reduced depressive symptoms which was accompanied by an increase in muscle strength and performance.(6) The discordant findings in the study by Zhang et al may be explained by inherent limitations of MR analysis. For instance, MR provides causality of the risk factor with onset of disease, and by selecting on disease incidence does not address the question of the causal relationship of muscle strength with disease progression. Additionally, muscle strength displays age-related decline in its trajectory pattern, and MR, which provides estimates of lifetime risk, may not be sensitive to a critical window. This is particularly relevant since decline and not baseline muscle strength is associated with all-cause mortality.(7) The relative importance of muscle mass and muscle strength in causing depression is a nuanced question that will need to be tested in a large randomized controlled trial. Nevertheless, results from genome-wide association studies provide an excellent source of large and precise data to establish causality using MR, and emerging methods may provide solutions to overcome its limitations.

Data Statement

The data has not been previously presented orally or by poster at scientific meetings.