AR: CT may yield biased emphysema results

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon

Socioeconomic status was associated with healthier lung function but returned higher degrees of emphysema among individuals with higher levels of income and education, leading investigators to point to bias in CT breath holds as confounding emphysema exam indicators, according to a study published in the February issue of Academic Radiology.

"A strong socioeconomic gradient has also been observed for subclinical respiratory health measures such as lung function," with lower socioeconomic status having been significantly associated with increased levels of respiratory mortality, according to Gina S. Lovasi, PhD, MPH, of the department of epidemiology, Mailman School of Public Health at Columbia University in Manhattan, N.Y.

Pulling data from the Multi-Ethnic Study of Atherosclerosis (MESA), 3,706 participants without clinical cardiovascular disease were assessed for subclinical emphysema on CT and tested for lung function, which is associated with—but distinct from—subclinical emphysema. Percent emphysema was defined based as the proportion of pixels below an attenuation threshold of 910 HU from lung windows of cardiac CT scans.

"Contrary to our expectations, participants with higher SES [socioeconomic status] were observed to have more emphysema-like patterns on their CT scans, as indicated by higher percent emphysema," Lovasi and colleagues found. "However, higher SES did have the expected association with lung function consistent with prior studies."

The authors found that each additional increment of education (no high school degree, high school degree, some college, college degree and graduate degree) was associated with a 1.1 percent higher emphysema value on CT. On the other hand, all statistically significant associations between socioeconomic status indicators and lung function measures were positive.

Lovasi and colleagues conjectured that the disparity between percent emphysema and lung function was due bias from effort dependence. "All participants were coached to achieve full inspiration during spirometry by a certified spirometry technician and during CT scanning by a certified CT technician; however, it is likely that coaching of inspiration was not as intense for CT as for spirometry because inspiration has been long appreciated as a key component of valid spirometry but is only one of a number of components of CT scoring." As a result, Lovasi and colleagues concluded that suboptimal inspiration among participants with lower socioeconomic indicators would make this cohort "appear healthier than they are."

The authors concluded that "[a]vailable measures of effort may be serving only as a crude proxy for the true confounder (e.g. intrinsic motivation to comply with test instructions in a clinical research setting; rapport between technician and participant), and future research is needed to assess this potential source of bias."