Infectious diseases and sarcopenia-related phenotypes: A phenome-wide Mendelian randomization study with exploratory proteomic analysis.
Key claim: Genetic liability to tonsillitis and tuberculosis sequelae is associated with lower appendicular lean mass, postzoster neuralgia with higher ALM, while grip strength shows no FDR-significant infection or disease associations in this MR-PheWAS.
Abstract
Sarcopenia-related traits may be associated with infection-related disease liabilities, but broad phenome-wide genetic screening remains limited. We conducted a two-sample MR-PheWAS of more than 2000 FinnGen R12 phenome-wide endpoints against appendicular lean mass (ALM) and left-hand grip strength (HGS). Multiple testing was controlled using Benjamini-Hochberg false discovery rate (BH-FDR) correction applied separately to 2317 ALM and 2307 HGS primary-estimator tests. Twenty-one ALM endpoints and two HGS endpoints were FDR-significant; the two HGS endpoints were height and weight quantitative traits, and no disease phenotype or infection-related phenotype was FDR-significant for HGS. Among the FDR-significant ALM endpoints, four were infection-related. Rickettsioses were supported by a single instrument and were reported separately, whereas three multi-instrument phenotypes were prioritized for detailed follow-up: postzoster neuralgia, tonsillitis, and sequelae of tuberculosis. On the normalized ALM scale, postzoster neuralgia was associated with higher ALM (IVW beta = 0.01217, P = 1.17 × 10^-4), whereas tonsillitis (beta = -0.02580, P = 1.91 × 10^-4) and sequelae of tuberculosis (beta = -0.01247, P = 2.15 × 10^-4) were associated with lower ALM. Candidate-specific HGS analyses and reverse MR did not show statistically significant associations for these three phenotypes. An exploratory ALM-HGS intersection identified 23 nominally associated disease phenotypes, including five infection-related phenotypes. Exploratory proteomic analyses identified candidate proteins and pathways but did not establish mediation. Overall, the findings indicate phenotype-specific heterogeneity in genetic associations with muscle-related traits and require independent replication.
Why this matters
Mendelian randomization moves the long-standing infection–sarcopenia observational association one rung up the inferential ladder by using genetic liability as an exposure proxy across >2000 endpoints, and the phenotype split — lean mass hit, grip strength clean — argues that chronic/latent infection exposures act preferentially on muscle quantity rather than neuromuscular output. The scan does not identify a mediating immune mechanism or deliver an intervention, so it sits alongside prior longitudinal infection-burden cohort data as human-scale evidence that specific infections (not infection generically) are candidate contributors to the inflammaging–sarcopenia bridge.