Myosteatosis as an independent predictor for all-cause and cardiac mortality in initial-dialysis patients: a multicenter, retrospective cohort study

Shi-mei Hou^1,2Min Li¹Jing-yuan Cao³Yao Wang⁴Min Yang¹Li Tian¹Yan-hong Ni¹Bi-xia Yang¹Li-rong Hao⁵Jian-bing Hao⁵Chun-bo Zou³Shu-yan Zhang⁶A-feng Miao³Min Yang⁴Chun-hong Hu⁴Li Yuan⁴Jing Zheng⁷Jing-jie Xiao⁸Jian Xu^9*Bin Wang^10*

• ¹Department of Nephrology, The Third Affiliated Hospital of Soochow University, Changzhou, China
• ²Institute of Nephrology, Zhong Da Hospital, School of Medicine, Southeast University, Nanjing, China
• ³Institute of Nephrology, Taizhou People’s Hospital, Taizhou, Taizhou, China
• ⁴Institute of Nephrology, Yangzhou First People’s Hospital, Yangzhou, China
• ⁵Department of Nephrology, Southern University of Science and Technology Hospital, Shenzhen, China
• ⁶Nantong University School of Medicine, Nantong, China
• ⁷Institute of Geriatrics, Zhong Da Hospital, School of Medicine, Southeast University, Nanjing, China
• ⁸Covenant Health Palliative Institute, R416 St Marguerite Health Services Center, Edmonton, Canada
• ⁹Department of intensive care unit, Geriatric Hospital Of Nanjing Medical University, Nanjing, China
• ¹⁰Institute of Nephrology, Zhongda Hospital, Southeast University, Nanjing, China

Background: Myosteatosis is associated with adverse prognosis in diseases. We aimed to establish thresholds for myosteatosis, assess its association with all-cause and cardiac mortality in initial dialysis patients, and construct a myosteatosis-based survival nomogram. Methods: This multicentric retrospective study included 383 initial-dialysis patients (1/2014–12/2019). Endpoints include all-cause and cardiac mortality. Skeletal Muscle Index (SMI, cm2/m2) and Skeletal Muscle Density (SMD, Hounsfield Units [HU]) were measured at the third lumbar vertebra (L3) level by computed tomography (CT). Sex-specific SMI and SMD thresholds predicted all-cause mortality through the receiver operating characteristic (ROC) curves. The Cox models assessed myosteatosis-associated mortality risks. Survival prediction models were built using univariate and multivariate Cox proportional hazards regression in the training cohort, followed by internal and external validation. Results: Patients were predominantly aged 18-65 years (n=298, 77.81%), with males comprising 60.84% (n=233). All-cause mortality was 22.72% (n=87), of which 52.87% (n=46) were attributed to cardiac causes. Sex-specific SMD cutoffs for predicting all-cause mortality were 32.46 HU (AUC=0.707) in males and 34.58 HU (AUC=0.690) in females (both P<0.05). Myosteatosis was associated with higher all-cause (36.7%) and cardiac mortality (19.8%) (both P<0.001), and independently predicted both outcomes (all-cause mortality: HR=3.203, 95%CI:1.937-5.296; cardiac mortality: HR=3.418, 95%CI:1.718-6.802). The myosteatosis-based nomogram achieved a C-index of 0.761, validated in real-world data. Conclusions: Sex-specific myosteatosis thresholds (males: SMD ≤32.46 HU; females: ≤34.58 HU) derived from L3-CT independently predicted all-cause and cardiac mortality in initial-dialysis patients. The myosteatosis-based survival nomogram demonstrated moderate-to-good predictive accuracy and potential clinical utility.