Age, Obesity, Sarcopenia, and Proximity to Death Explain Reduced Mean Muscle Attenuation in Patients with Advanced Cancer.

BACKGROUND

Cancer cachexia is characterized by skeletal muscle loss. A feature of muscle wasting, reduction in the mean muscle attenuation from computed tomography images is believed to reflect pathological infiltration of fat into muscle. It is a reported prognostic indicator in cancer patients.

OBJECTIVES

To develop an explanatory multivariate model of muscle attenuation of cancer patients incorporating age, sex, disease characteristics, body composition. Time to death ≤92 days was included in the model as the demarcation of end-stage disease.

DESIGN

Multivariate general linear model regression analysis of total mean muscle attenuation and change in muscle attenuation.

SETTING

Regional cancer center (Alberta, Canada).

PARTICIPANTS

Patients with gastrointestinal and respiratory tract cancers (mean age 64±11 years, 44% female).

MEASUREMENTS: Total adipose tissue and skeletal muscle cross sectional area, and mean muscle attenuation at the 3rd lumbar vertebra were assessed from baseline computed tomography (n=1719), and a subset with repeated measures (n=246 patients with a total of 871 images).

RESULTS

At baseline, muscle attenuation associated with total skeletal muscle (β 0.09; 95% CI 0.07 to 0.11; p<0.001) and adipose tissue (β -0.032; 95% CI -0.035 to -0.029; p<0.001) cross sectional areas, age (β -0.28; 95% CI -0.32 to -0.24; p<0.001), time to death ≤92 days (β -1.9; 95% CI -3.1 to -0.7; p=0.003) and male sex (β -2.3; 95% CI -3.5 to -1.1; p<0.001). Change in muscle attenuation over time associated with total adipose tissue cross sectional area (β -0.008; 95% CI -0.012 to -0.004; p<0.001) and time to death ≤92 days (β -1.6; 95% CI -3.0 to -0.2; p=0.03).

CONCLUSIONS

The radiation attenuation of skeletal muscle is lowest in individuals who are older, less muscular, have a higher fat mass and are within 92 days of death. Men had lower muscle attenuation than women when controlled for other variables.