Lung Volumes in Smokers without Chronic Obstructive Pulmonary Disease: A Pointer to Disease Development?

The late lamented Nobel Prize–winning psychologist Daniel Kahneman spent much of his career identifying the biases in our thinking which unconsciously shape our lives. One of the most common of these is confirmation bias, in which the observer favors a result that fits their prior preconceptions. Medicine is not immune to this type of thinking (and neither is this editorialist), and it poses a challenge to anyone who makes an unexpected observation. This has proved to be the case when studying the transition from apparently healthy lungs to the onset of established chronic obstructive pulmonary disease (COPD).

We know that the forced expiratory volume in 1 second (FEV₁) declines with age, and this decline is accelerated in patients who develop emphysema and small airway disease due to tobacco exposure. The physiological threshold for the presence of airflow obstruction is customarily defined by a ratio of FEV₁ to forced vital capacity (FEV₁/FVC) of 0.7 or less, and there is empirical support for this (1), although a value below the population defined lower limit of normal is being increasingly used (2). Moreover, an FEV₁/FVC ratio below 0.7 is a good predictor of future disease progression in early disease (3). Disease severity is graded in the Global Initiative for Chronic Obstructive Lung Disease (GOLD) strategy (4) by the degree of FEV₁ percent predicted loss across GOLD Grades 1–4. This implies an incremental and monotonic progression from grade to grade as the disease progresses.

Less attention has been paid to the changes in lung volume that accompany those in FEV₁, in part because lung volume measurements are more time consuming to apply in large populations and show more between test variability than spirometry (5). Residual volume (RV) increases early in the natural history of COPD, although less was known about when total lung capacity (TLC) changes (6). Large longitudinal studies such as COPDGene and SPIROMICS have remedied this deficit. Thanks to the work of Arjomandi and colleagues, we know that the ratio of RV to TLC derived from supine inspiratory and expiratory computed tomography (CT) scans was the best predictor of lung function loss in tobacco-exposed unobstructed SPIROMICS study participants (7). They used routine pulmonary function data from 7,479 unobstructed smokers to show that a high RV/TLC ratio predicted the subsequent risk of developing COPD and of dying (8). In a further analysis of SPIROMICS data, they observed that a high TLC but normal RV favored progress to GOLD Stage 1 COPD, whereas subjects with a normal TLC but high RV usually progressed to GOLD Stage 2 (9). They confirmed their findings in COPDGene data, which suggested that smokers with an isolated high TLC had more emphysema and less small airway disease than those with a high RV/TLC ratio (10).

In this issue of the AnnalsATS, Arjomandi and colleagues (pp. 494–505) provide further longitudinal data once more drawn from SPIROMICS that extend their previous observations (11). They focus on the 496 smoking participants with normal initial spirometry who not only completed the initial assessments and CT scan but also completed annual follow-up visits for 3 years as well as a further CT scan, questionnaires, and spirometry tests around Year 4. They analyzed their data as continuous variables for the time to COPD onset and rate of lung function decline and in mutually exclusive subgroups with isolated high TLC (n = 115), high RV/TLC without raised TLC (n = 115), and both high TLC and high RV/TLC (n = 50).

They found that RV/TLC was the most important predictor of baseline symptoms and severe exacerbations. Moreover, compared with high TLC alone, a raised RV/TLC was associated with a greater risk of developing COPD, a more rapid rate of FEV₁ decline, and worse health status, especially if accompanied by a low or normal TLC. Finally, there were differences in lung structure between the groups that supported the observations in COPDGene that high TLC without air trapping had “more” emphysema, although the absolute amounts were quite small. Overall, there was little evidence of changes in lung structure over time in the different subgroups, but when people with an isolated high TLC developed obstruction, it was mild at GOLD Grade 1, whereas over the same follow-up period, those with a high RV/TLC moved into GOLD Grade 2.

These findings confirm that a raised RV/TLC is a marker of COPD risk and is associated with more “airway” pathology and a faster progression to GOLD Grade 2 disease. The change in symptoms over time is compatible with the modest deterioration in lung mechanics, with little impact on breathlessness or exercise performance but worsening health status and increased medication use when RV/TLC is high. The subgroups were surprisingly well balanced, although FEV₁ was lowest in the high-RV/TLC group and highest when only TLC was elevated, suggesting a “horse race effect” in terms of COPD progression. Even in a large, well-conducted study, there are inevitable limitations. Two issues stand out. First, the supine lung volume values are split by tertiles, as there are no normative data for this measurement. CT lung volume correlates well with seated plethysmographic volume, but it is consistently lower (10), making it hard to operationalize clinically. Second, it would have helped to have more data from subjects with normal or midrange RV and TLC values to better understand whether a high TLC is truly beneficial in this setting. Whether changes in lung elastance are sufficient to explain the high-TLC group or whether they reflect differences in chest wall compliance and muscle strength remains to be explored.

Nonetheless the authors are to be congratulated on their rigorous approach to furthering our understanding of this transitional state from health to disease. Many suggestions have been made to help us determine susceptible tobacco smokers (12, 13), and we need to know how early changes in RV/TLC occur, something that should be possible in younger smoking cohorts such as BEACON, where CT abnormalities have already been found (14), and now in the U.S. SOURCE study (15). So, have our biases been confirmed? Well, up to a point. Changes in residual volume are a measurable and robust early sign of COPD, but the subsequent progression of this disease is affected by how big the lungs are. Whether this is an intrinsic or acquired feature is still to be determined and opens new areas to explore in this rapidly changing field.