Vascular Pruning on CT and Interstitial Lung Abnormalities in the Framingham Heart Study

Abstract

Background: Pulmonary vascular disease is associated with poor outcomes in individuals affected by interstitial lung disease. The pulmonary vessels can be quantified with noninvasive imaging, but whether radiographic indicators of vasculopathy are associated with early interstitial changes is not known.

Research question: Are pulmonary vascular volumes, quantified from CT scans, associated with interstitial lung abnormalities (ILA) in a community-based sample with a low burden of lung disease?

Study design and methods: In 2,386 participants of the Framingham Heart Study, we used CT imaging to calculate pulmonary vascular volumes, including the small vessel fraction (a surrogate of vascular pruning). We constructed multivariable logistic regression models to investigate associations of vascular volumes with ILA, progression of ILA, and restrictive pattern on spirometry. In secondary analyses, we additionally adjusted for diffusing capacity and emphysema, and performed a sensitivity analysis restricted to participants with normal FVC and diffusing capacity.

Results: In adjusted models, we found that lower pulmonary vascular volumes on CT were associated with greater odds of ILA, antecedent ILA progression, and restrictive pattern on spirometry. For example, each SD lower small vessel fraction was associated with 1.81-fold greater odds of ILA (95% CI, 1.41-2.31; P < .0001), and 1.63-fold greater odds of restriction on spirometry (95% CI, 1.18-2.24; P = .003). Similar patterns were seen after adjustment for diffusing capacity for carbon monoxide, emphysema, and among participants with normal lung function.

Interpretation: In this cohort of community-dwelling adults not selected on the basis of lung disease, more severe vascular pruning on CT was associated with greater odds of ILA, ILA progression, and restrictive pattern on spirometry. Pruning on CT may be an indicator of early pulmonary vasculopathy associated with interstitial lung disease.

Keywords: epidemiology (pulmonary); imaging; interstitial lung disease; pulmonary circulation.

Figure 1

A, Axial CT sections from a Framingham Heart Study participant with normal pulmonary parenchyma (left) and interstitial lung abnormalities (right). B, Three-dimensional volumetric reconstructions of the pulmonary vascular tree, generated from the CTs in A, demonstrating relative loss of the smallest pulmonary vessels (ie, vascular pruning) in the participant with interstitial lung abnormality on CT. Vessels are color-coded by size.

Figure 2

Associations of vascular pruning on CT (BV5/TBV) with odds of ILA, ILA progression, and restrictive pattern on spirometry (expressed per SD lower BV5/TBV). ∗Results expressed per –SD (SD lower) CT vascular volume parameter. Results of multivariable logistic and multinomial logistic regression models with adjustment for age, sex, height, weight, smoking status, total pack-years of cigarette exposure, personal educational attainment, occupation category, median household income, and Framingham. Heart Study cohort (offspring vs third generation). BV5 = blood vessel volume of pulmonary vessels with cross-sectional area < 5 mm²; ILA = interstitial lung abnormality; TBV = total blood vessel volume of all pulmonary vessels.