Longitudinal Association Between Muscle Loss and Mortality in Ever Smokers

Abstract

Background: Body composition measures, specifically low weight or reduced muscle mass, are associated with mortality in COPD, but the effect of longitudinal body composition changes is undefined.

Research question: Is the longitudinal loss of fat-free mass (FFM) associated with increased mortality, including in those with initially normal or elevated body composition metrics?

Study design and methods: Participants with complete data for at least one visit in the COPDGene study (n = 9,268) and the ECLIPSE study (n = 1,760) were included and monitored for 12 and 8 years, respectively. Pectoralis muscle area (PMA) was derived from thoracic CT scans and used as a proxy for FFM. A longitudinal mixed submodel for PMA and a Cox proportional hazards submodel for survival were fitted on a joint distribution, using a shared random intercept parameter and Markov chain Monte Carlo parameter estimation.

Results: Both cohorts demonstrated a left-shifted distribution of baseline FFM, not reflected in BMI, and an increase in all-cause mortality risk associated with longitudinal loss of PMA. For each 1-cm² PMA loss, mortality increased 3.1% (95% CI, 2.4%-3.7%; P < .001) in COPDGene, and 2.4% (95% CI, 0.9%-4.0%; P < .001) in ECLIPSE. Increased mortality risk was independent of enrollment values for BMI and disease severity [BODE (body mass, airflow obstruction, dyspnea, and exercise capacity) index quartiles] and was significant even in participants with initially greater than average PMA.

Interpretation: Longitudinal loss of PMA is associated with increased all-cause mortality, regardless of BMI or initial muscle mass. Consideration of novel screening tests and further research into mechanisms contributing to muscle decline may improve risk stratification and identify novel therapeutic targets in ever smokers.

Keywords: COPD; mortality; muscle wasting; sarcopenia.

Graphical abstract

Figure 1

The population percentiles of fat-free mass index (FFMI) and BMI for non-Hispanic White participants between 45 and 70 years of age in COPDGene and ECLIPSE, compared with healthy population data. In the upper panels, participant BMI reflects that of the general population, as demonstrated by the normal distribution. In contrast, in the lower panels the FFMI distributions are right-skewed, with the blue-shaded portions representing participants overrepresented in the lower percentiles of FFMI compared with their BMI. COPDGene = Genetic Epidemiology of COPD Study; ECLIPSE = Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints.

Figure 2

Effect sizes from stratified analyses in the COPDGene cohort. Stratification was based on the value of the variable (listed in the gray boxes on the right) at the time of study enrollment. Each dot represents the point estimate for the increased risk of mortality per 1-cm² loss of pectoral muscle area, and horizontal lines depict the CI. avg = average; Black/AA = Black/African American; BODE = body mass, airflow obstruction, dyspnea, and exercise capacity; COPDGene = Genetic Epidemiology of COPD Study; ECLIPSE = Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints; PMA = pectoral muscle area.

Figure 3

Kaplan-Meier survival curves for participants in COPDGene with more than one pectoral muscle area (PMA) measurement. Patients were dichotomized on the basis of whether their PMA was above or below the cohort mean at enrollment (adjusted for sex and height) as well as whether they lost more than the expected age-related muscle area. avg = average; COPDGene = Genetic Epidemiology of COPD Study.