| dc.description.abstract | Degenerative rotator cuff tear (DCT) is a leading cause of shoulder disability, and our understanding of the etiological factors behind DCT remains limited. Recent research challenges the notion that DCTs solely result from repetitive microtrauma, suggesting intrinsic tendinous abnormalities and factors that contribute to tendinous weakness are potential risk factors. Conditions like obesity and diabetes, known to impact tendon health, may predispose individuals to rotator cuff injury. While traditional epidemiological studies link obesity and diabetes to cuff disease, most are limited in providing evidence for causality due to inconsistent definitions of cuff disease, lack of temporality between exposure and outcome, confounding and other biases. In this work, I leverage data from large electronic health record (EHR) linked biobanks such as Vanderbilt University Medical Center (VUMC) Synthetic Derivative (SD), BioVU and UK Biobank to build and validate algorithms for DCT, and to evaluate causal roles of diabetes and obesity on DTCs, by incorporating methods rooted in instrumental variable analysis – Mendelian Randomization (MR) – to overcome challenges faced by traditional epidemiologic studies. With the VUMC SD, I validated two algorithms (imaging-based and non-imaging based) to identify DCT cases and controls with over 90% predictive value. I comprehensively compared these algorithms to show non-imaging-based methods yield larger numbers and have reduced selection bias than imaging-based methods. I then applied these algorithms to the UK Biobank and BioVU to conduct genome-wide association studies (GWAS). Using a meta-analysis of these DCT GWAS, and published GWAS data on obesity and diabetes traits, I conduct MR analyses to provide evidence of positive causal associations between obesity related measures and DCT, but limited evidence between diabetes related traits and DCT. I report a potential effect of fasting insulin on DCT risk which is attenuated after adjusting for body mass index, and especially waist-to-hip ratio. Additionally, I use MVMR to investigate the potential mediating role of fasting insulin on the association between obesity and DCT and found little evidence of mediation. This work suggests obesity may be causally related to DCT, independent of diabetes, and a deeper understanding may inform approaches to prevention, mitigation, and treatment of DCT. | |
