Abstract Background and Aims There is growing evidence that muscle mass (MM) loss, observed in long term conditions (LTCs), including chronic kidney disease (CKD), is associated with mortality. This study aimed to explore the direct association between MM and mortality, having controlled for the combined influence of ageing, long term conditions (LTCs), inflammation, renal function and body weight. Furthermore, we sought to explore whether this association differed between the biological sexes. Method An observational study was conducted using the UK Biobank cohort. Adults were recruited between 2007-2010 at 22 UK centres. MM was estimated using bioimpedance defined whole body fat-free mass (BI-FFM: Tanita BC418). Primary outcome was all-cause mortality (ACM), with the cause / date of death being obtained from the NHS Information Centre (England and Wales) and NHS Central Registry (Scotland). Follow up was defined as the period between the first visit date to either date of death / the latest date for central registry downloads (December 2021). Estimated glomerular filtration rate using creatinine (eGFRCreat) or cystatin-C (eGFRcystatin-C) were calculated (CKD-EPI 2009 equation without ethnicity correction). Participants were excluded if they were on dialysis, died from SARS-CoV-2 infection or withdrew consent. Sequences of regressions analyses (SoRA), a graphical Markov modelling technique which assesses the complexity of direct and indirect pathways of association between blocks of variables which are ordered a-priori to reflect postulated directions of association, were fitted for both sexes. Results There were 500,589 participants eligible for analysis (272,623 females / 227,966 males). There were 33,755 deaths over a median follow up period of 12.58 (IQR 11.85-13.30) yrs. The male model revealed per 5 kg-higher BI-FFM (holding weight constant) was associated with lower odds of ACM in the eGFRCreat (OR 0.827, 95% CI 0.781-0.875) and eGFRCystatin-C (OR 0.425, 95% CI 0.301-0.599) sub-models, whereas a 5 kg-higher weight (holding BI-FFM constant) was associated with higher odds of ACM in the eGFRCreat (OR 1.070, 95% CI 1.027-1.115) and eGFRCystatin-C (OR 1.395, 95% CI 1.170-1.663) sub-models. In contrast, the female model revealed a 5 kg-higher BI-FFM (holding weight constant) was associated with higher odds of ACM in the eGFRCreat (OR 2.016, 95% CI 1.650-2.463) and eGFRCystatin-C (OR 1.729, 95% CI 1.391-2.151: Fig. 1) sub-models, whereas a 5 kg-higher body weight (holding BI-FFM constant) was associated with lower odds of ACM in the eGFRCystatin-C (OR 0.935, 95% CI 0.918-0.952) but not the eGFRCreat (OR 1.020, 0.970-1.072) sub-model. Conclusion Our analysis demonstrates sex-divergent associations between BI-FFM and/or body weight with ACM when estimated kidney function is accounted for. These associations suggest sex-specific explanations as to why MM loss leads to adverse outcomes in multimorbid populations and further characterises the sex-stratified nature of the “obesity paradox” in LTCs.
Read full abstract