Lung structure and longitudinal change in cardiac structure and function: the MESA COPD Study

Abstract

Background: Lung structure and cardiac structure and function are associated cross-sectionally. The classic literature suggests relationships of airways disease to cor pulmonale and emphysema to reduced cardiac output (CO) but longitudinal data are lacking.

Methods: The Multi-Ethnic Study of Atherosclerosis (MESA) COPD Study was a multicentre longitudinal COPD case-control study of participants 50-79 years with ≥10 pack-years smoking without clinical cardiovascular disease. Segmental airway wall area (WA) and percent emphysema were measured on computed tomography. Right and left ventricle parameters were assessed on cardiac magnetic resonance imaging (cMRI) in exams 6 years apart. Longitudinal and period cross-sectional associations were evaluated with mixed models adjusted for demographics, body size and smoking.

Results: The 187 participants with repeated cMRI were 67±7 years old; 42% had COPD; 22% currently smoked; and the race/ethnicity distribution was 54% White, 30% Black, 14% Hispanic and 3% Asian. Greater WA at enrolment was associated with longitudinal increase in right ventricular (RV) mass (3.5 (95% CI 1.1-5.9) g per 10 mm² WA). Greater percent emphysema was associated with stably lower left ventricular (LV) end-diastolic volume (-7.8 (95% CI -10.3- -3.0) mL per 5% emphysema) and CO (-0.2 (95% CI -0.4- -0.1) L·min^-1 per 5% emphysema).

Conclusion: Cardiac associations varied by lung structure over 6 years in this multi-ethnic study. Greater WA at enrolment was associated with longitudinal increases in RV mass, whereas greater percent emphysema was associated with stable decrements in LV filling and CO.

Figure 1. Six-year follow-up of right ventricular mass by quintile of segmental airway wall area at enrollment.

Model includes time between exams, the interaction of time between exams and WA at enrollment, age, race/ethnicity, sex, educational attainment, weight, height, body surface area, smoking status, pack-years, systolic blood pressure, hypertensive medication use, diuretic use, diabetes, low density lipoprotein, statin use, oxygen saturation, study site, and LV mass, weighted by inverse probability of selection into the cohort. Categorical covariates were fixed at their proportion in the sample; continuous variables were set at the sample mean. Abbreviations: RV, right ventricle; g, gram; LV, left ventricle; Q1, first quintile; Q2, second quintile; Q3, third quintile; Q4, fourth quintile; Q5, fifth quintile; CT, computed tomography.

Figure 2. Association of segmental airway wall area at enrollment and longitudinal change in right ventricular mass over 6 years.

Models included age, race/ethnicity, sex, educational attainment, weight, height, body surface area, smoking status, pack-years, systolic blood pressure, antihypertensive medication use, diabetes, low density lipoprotein, statin use, diuretic use, oxygen saturation, study site, left ventricle mass, and lumen area. Models were weighted by the inverse of the probability of selection into the study. * Excludes one extreme outlier with WA and lumen area measures > 2 SD above mean. Abbreviations: RV, right ventricle; g, gram; mm, millimeter; CT, computed tomography; SD, standard deviation.

Figure 3. Six-year follow-up of left ventricular end diastolic volume by quintiles of percent emphysema at enrollment.

Period cross-sectional differences, or the average cross-sectional differences at enrollment and follow-up visits, were obtained from linear mixed models with time-varying measures of percent emphysema and LV end diastolic volume. Model includes time between exams, the interaction of time between exams and percent emphysema at enrollment, age, race/ethnicity, sex, educational attainment, weight, height, body surface area, smoking status, pack-years, systolic blood pressure, hypertensive medication use, diuretic use, diabetes, low density lipoprotein, statin use, oxygen saturation, and study site, weighted by inverse probability of selection into the cohort. Categorical covariates were fixed at their proportion in the sample; continuous variables were set at the sample mean. Abbreviations: LV, left ventricle; mL, milliliter; Q1, first quintile; Q2, second quintile; Q3, third quintile; Q4, fourth quintile; Q5, fifth quintile; CT, computed tomography.