Article Text

Download PDFPDF

Associations between emphysema-like lung on CT and incident airflow limitation: a general population-based cohort study
  1. Elizabeth C Oelsner1,2,
  2. Benjamin M Smith1,3,
  3. Eric A Hoffman4,
  4. Aaron R Folsom5,
  5. Steven M Kawut6,
  6. Joel D Kaufman7,
  7. Ani Manichaikul8,
  8. David J Lederer1,
  9. Joseph E Schwartz1,
  10. Karol E Watson9,
  11. Paul L Enright10,
  12. John H M Austin1,
  13. Joao A C Lima11,
  14. Steven J Shea1,2,
  15. Robert G Barr1,2
  1. 1 Department of Medicine, Columbia University College of Physicians and Surgeons, New York City, New York, USA
  2. 2 Department of Epidemiology, Columbia University Mailman School of Public Health, New York City, New York, USA
  3. 3 Respiratory Division, McGill University, Montreal, Quebec, Canada
  4. 4 Department of Radiology, University of Iowa, Iowa City, Iowa, USA
  5. 5 Division of Epidemiology and Community Health, University of Minnesota, Minneapolis, Minnesota, USA
  6. 6 School of Medicine, Pulmonary, Allergy and Critical Care Division, University of Pennsylvania, Philadelphia, Pennsylvania, USA
  7. 7 Department of Environmental and Occupational Health Sciences, University of Washington, Seattle, Washington, USA
  8. 8 Department of Public Health Sciences, Division of Biostatistics and Epidemiology, University of Virginia, Charlottesville, Virginia, USA
  9. 9 Division of Cardiology, UCLA School of Medicine, Los Angeles, California, USA
  10. 10 Department of Medicine, University of Arizona, Tucson, Arizona, USA
  11. 11 Department of Cardiology, Johns Hopkins University, Baltimore, Maryland, USA
  1. Correspondence to Dr Elizabeth C Oelsner, Department of Medicine, Columbia University Medical Center, Presbyterian Hospital, New York, NY 10032, USA; eco7{at}


Emphysema on CT is associated with accelerated lung function decline in heavy smokers and patients with COPD; however, in the general population, it is not known whether greater emphysema-like lung on CT is associated with incident COPD. We used data from 2045 adult participants without initial prebronchodilator airflow limitation, classified by FEV1/FVC<0.70, in the Multi-Ethnic Study of Atherosclerosis. Emphysema-like lung on baseline cardiac CT, defined as per cent low attenuation areas<—950HU>upper limit of normal, was associated with increased odds of incident airflow limitation at 5-year follow-up on both prebronchodilator (adjusted OR 2.62, 95% CI 1.47 to 4.67) and postbronchodilator (adjusted OR 4.38, 95% CI 1.63 to 11.74) spirometry, independent of smoking history. These results support investigation into whether emphysema-like lung could be informative for COPD risk stratification.

  • COPD epidemiology
  • emphysema
  • imaging/CT MRI

Statistics from


COPD risk assessment is important to developing and targeting primary prevention and precision medicine approaches to COPD, the fourth leading cause of death. Emphysema, defined pathologically by destruction of alveolar walls, is frequently observed in COPD; nonetheless, emphysema may occur in the absence of the airflow limitation that defines COPD, and, in postmortem series, has been observed in up to one-tenth of never smokers.1

Emphysema-like lung on CT—which has been previously validated against pathological samples of emphysema,2 3 but which may also correspond to hyperinflation—is associated cross-sectionally with COPD, and has been associated longitudinally with respiratory and all-cause mortality in both patients with COPD and in persons with normal lung function.4–6 Among smokers with COPD, areas of normal-appearing lung have been found to be mechanistically influenced by areas of emphysema-like lung, and the extent of emphysema-adjacent ‘at risk’ lung was associated with lung function decline.7 Nonetheless, whether otherwise healthy individuals with greater proportions of emphysema-like lung are at increased risk of developing incident COPD is unknown.

Using information from the US general population-based Multi-Ethnic Study of Atherosclerosis (MESA), we identified 2045 middle-aged and older adults without prebronchodilator (BD) airflow limitation (FEV1/FVC<0.70) on initial spirometry in 2004–2006. In this group, we tested whether the percentage of emphysema-like lung (voxels <−950 HU (Hounsfield units), or ‘percent emphysema’) on cardiac CT at cohort baseline (2000–2002) was associated with incident airflow limitation on pre-BD and post-BD spirometry in 2010–2012. Methods for MESA, statistical analyses, and selection and characteristics of participants are described in detail in the online supplementary tables S1–S4 and figure S1.


Mean age was 58.4 years at baseline, 54.0% were female, and 75.5% were non-White. Half of participants were never smokers. Among smokers, 47.7% reported <10 pack-years, and 11.3% reported more than 30 pack-years.

Percent emphysema

Median per cent emphysema was 2.65% (IQR, 1.17, 4.99). For each participant, the upper limit of normal (ULN) for per cent emphysema was defined by previously published reference equations derived in MESA accounting for age, gender, race/ethnicity, height, weight, current smoking and scanner parameters.8 The prevalence of per cent emphysema>ULN was 5.4%.

Airflow limitation

Eleven per cent (n=224) developed incident pre-BD airflow limitation. Affected participants were older, weighed less and had lower initial lung function compared with those unaffected; 12.1% had a prior diagnosis of asthma, and 43.8% were never smokers.

Associations between per cent emphysema and incident airflow limitation

In fully adjusted models, per cent emphysema>ULN was associated with increased odds of incident airflow limitation both pre-BD (OR 2.62, 95% CI 1.47 to 4.67) and post-BD (OR 4.38, 95% CI 1.63 to 11.74). Results were similar in unadjusted models and robust to adjustment for initial FEV1/FVC ratio. Associations were consistent for the continuous measure of per cent emphysema (table 1) as well as for a fixed threshold (>5% per cent emphysema) (see online supplementary table S5). There was no evidence of effect measure modification by age, sex or race/ethnicity (p>0.2).

Supplementary file 1

Table 1

Emphysema-like lung on CT and incidence of airflow limitation at 5-year follow-up

Sensitivity analyses

Results were similar in never smokers and ever smokers, and multiplicative interaction terms with smoking status were not statistically significant (see online supplementary table S6). Among ever smokers with 30+ pack-years, per cent emphysema was associated with adjusted ORs of 2.98 (per SD, 95% CI 1.63 to 5.41) and 5.45 (per SD, 95% CI 1.59 to 18.69) for incident pre-BD and post-BD airflow limitation, respectively.

Associations were not substantially altered by adjustment for education, site, self-reported emphysema and asthma, second-hand smoke, exposure to fumes, coronary artery calcium score, initial FEV1/FVC, or fine particulate matter, nor by exclusion of participants with baseline asthma or respiratory symptoms (see online supplementary figure S2).

Supplementary file 3

These findings were robust to redefinition of pre-BD airflow limitation by lower limit of normal criteria; classification of incident COPD as post-BD airflow limitation with respiratory symptoms; and multiple imputation of missing post-BD spirometry (see online supplementary figure S2).

Additional results are provided in online supplement.


Greater per cent emphysema on CT was associated with incident pre-BD and post-BD airflow limitation at 5-year follow-up in a contemporary, multiethnic, population-based cohort of middle-aged and older adults. These results were independent of lung function and standard clinical risk factors, including smoking.

This is the first work, to our knowledge, to show that the extent of emphysema-like lung is associated with the development of airflow limitation in otherwise healthy adults, including in never smokers. Our results are consistent with mounting evidence that measures of emphysema on CT may be associated with clinical outcomes in the absence of spirometry-defined COPD. We previously demonstrated in this cohort that per cent emphysema was associated with all-cause mortality in persons without airflow limitation or COPD, mostly from respiratory mortality.5 6 The current work indicates that the adverse prognosis of per cent emphysema on mortality may be attributable, in part, to an elevated risk of developing airflow limitation.

Although we used a definition of per cent emphysema that has been validated against tissue specimens of emphysema in multiple studies,2 3 9 the pathological significance of greater emphysema-like lung in healthy never smokers is not completely known. Hyperinflation and gas trapping could contribute to lower lung density on CT; however, exclusion of participants with asthma did not alter our results. Whatever the pathophysiological basis, our findings suggest that per cent emphysema>ULN may be associated with increased risk of developing airflow limitation meeting spirometric criteria for COPD.10

Per cent emphysema was calculated from low-dose cardiac CT scans, which imaged only the lower two-thirds of the lungs, yet these measures have been shown to correlate closely with measures from full-lung scans.11 Older CT technology relied upon in this study may also have introduced random measurement error; hence, measuring per cent emphysema on the contemporary full-lung CTs that are currently indicated for lung cancer screening in heavy smokers—among whom we demonstrated a strong association—would be expected to yield more precise results.

Follow-up was limited to two spirometry exams and there was a relatively low incidence of airflow limitation over 5-year follow-up. Post-BD spirometry was also missing for a substantial number of participants with incident airflow limitation. Nonetheless, our results were consistent in both adjusted and unadjusted models, and similar in sensitivity analyses using multiple imputation.

In conclusion, emphysema-like lung on CT was associated with incident airflow limitation in a general population-based cohort, independent of smoking and other major known risk factors for COPD. These results support investigation into whether per cent emphysema could be informative for COPD risk stratification.

Supplementary file 2


The authors thank the other investigators, the staff and the participants of the MESA study for their valuable contributions. A full list of participating MESA investigators and institutions can be found at



  • Contributors All authors meet the International Committee of Medical Journal Editors (ICMJE) criteria for authorship, having all made substantial contributions to: study conception, design, and/or data collection; interpretation of data; drafting and/or critical revision of the manuscript; and final manuscript review. All coauthors agree to be accountable for all aspects of the work and will ensure any questions regarding the accuracy or integrity of any part of the work are appropriately investigated and resolved. Specific contributions of each author are enumerated as follows: ECO: literature search, funding, study design, data analysis, data interpretation, drafting of manuscript, preparation of tables and figures. BMS, SMK, AM, DJL, JES, PLE: study design, data interpretation, manuscript review. EAH, ARF, SJS, JDK, KEW, JACL: study design, data collection, data interpretation, manuscript review. JHMA: emphysema measurement, data interpretation, manuscript review. RGB: study design, funding, data collection, data interpretation, drafting and review of manuscript.

  • Funding The MESA Lung Study is funded by R01-HL077612, R01-HL093081 and RC1-HL100543. Additional funding is provided by K23-HL-130627, R21-HL129924 and R01-HL130506. MESA was supported by contracts N01-HC-95159, N01-HC-95160, N01-HC-95161, N01-HC-95162, N01-HC-95163, N01-HC-95164, N01-HC-95165, N01-HC-95166, N01-HC-95167, N01-HC-95168 and N01-HC-95169 from the National Heart, Lung, and Blood Institute and by grants UL1-TR-000040 and UL1-RR-025005 from NCRR. This publication was also developed under a STAR research assistance agreement, No. RD831697 (MESA Air), awarded by the US Environmental Protection Agency. It has not been formally reviewed by the EPA. The views expressed in this document are solely those of the authors and the EPA does not endorse any products or commercial services mentioned in this publication.

  • Competing interests The authors do not report any competing interests, besides funding from the NIH, with the following exceptions: BMS reports training/salary awards from the Quebec Health Research Fund, and grants from McGill University; EAH is a founder and shareholder of VIDA Diagnostics, a company commercialising lung image analysis software developed, in part, at the University of Iowa; SMK has received non-financial support from the ATS, personal fees from the European Respiratory Journal for serving on an editorial board, and the University of Pennsylvania has received grants from Actelion, grants from United Therapeutics, grants from Gilead, grants from Lung Biotech, and grants from Bayer for CME courses; RGB reports grants from the Alpha-1 Foundation, COPD Foundation, and Foundation for the NIH, personal fees from UpToDate, and travel reimbursement from the COPD Foundation.

  • Ethics approval The study was approved by the NHLBI as well as the institutional review boards of all collaborating institutions.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.