Computed tomography–derived muscle quantity and attenuation phenotypes and all-cause mortality in patients receiving maintenance haemodialysis: a retrospective cohort study

Reduced muscle mass and myosteatosis are common in patients undergoing maintenance haemodialysis; however, the prognostic relevance of combined computed tomography (CT)-defined muscle quantity–attenuation phenotypes remains uncertain. We retrospectively studied data from 390 adults receiving maintenance haemodialysis who underwent non-contrast abdominal CT. The skeletal muscle index (SMI) and muscle radiation attenuation (MRA), an imaging marker of myosteatosis, were measured at the third lumbar vertebral level. The participants were classified into four phenotypes using cohort median values: high SMI/high MRA, low SMI/high MRA, high SMI/low MRA, and low SMI/low MRA. Cox and Fine–Gray models were used to evaluate all-cause mortality and incident fractures, respectively. During a median follow-up of 1,800 days, 151 deaths and 69 incident fractures occurred. The adjusted overall association across the four CT-defined phenotypes was not statistically significant (global P = 0.071). In exploratory pairwise comparisons with the high SMI/high MRA phenotype, mortality estimates were higher for the high SMI/low MRA phenotype (hazard ratio [HR], 1.86; 95% confidence interval [CI], 1.14–3.06) and the low SMI/low MRA phenotype (HR: 1.72; 95% CI: 1.04–2.85), but not for the low SMI/high MRA phenotype. In a supporting continuous analysis, each 1-SD lower MRA was associated with higher all-cause mortality (HR, 1.40; 95% CI, 1.14–1.74). No statistically significant overall association with incident fracture was detected, although the estimates were imprecise. The overall association of the four CT-defined muscle quantity–attenuation phenotypes with all-cause mortality was not statistically significant in the primary adjusted global test, and the association with incident fracture remained inconclusive. Exploratory analyses suggested an association between lower MRA and higher mortality in both the continuous analysis and high SMI/low MRA pairwise comparison. These findings are hypothesis-generating and require external validation.