Chest computed tomography: is it ready for major studies of chronic obstructive pulmonary disease?

Chronic obstructive pulmonary disease (COPD) requires for diagnosis the presence of chronic airflow limitation and emphysema or chronic bronchitis. In the USA, the disease causes w100,000 deaths annually, it is the fourth most important cause of death and afflicts about 16 million persons. About 1% of COPD in the USA is associated with a1-antitrypsin (a1-AT) deficiency and is conveniently referred to as a1-AT-COPD. This form of the disease occurs in younger persons, progresses more rapidly, is associated with a higher frequency of panlobular emphysema (PLE) than usual COPD, and may be treated by augmentation therapy with concentrated human a1-AT. However, there is limited evidence that a1-AT replacement modifies the rate of progression of a1-AT-COPD. How can the rate of progression be monitored? Using tools derived from usual COPD to monitor a1AT-COPD might not be correct. These diseases share pathophysiological similarities but there are also important differences. a1-AT-COPD is mostly due to the rapid development of PLE predominantly in the lower lobes; airflow obstruction and hyperinflation are mainly due to emphysema and losses of elastic recoil. In usual COPD, airflow limitation is largely secondary to airway damage and remodelling and emphysema (mainly centrilobular emphysema (CLE)) predominates in the upper lobes with a heterogeneous distribution in the lung. The extent of CLE is quite variable for a given level of airflow limitation. Moreover, SANDERS et al. [1] and SANTIS et al. [2] found that 68 – 80% of smokers showed emphysema by high-resolution computed tomography (HRCT) in the presence of normal lung function tests. These facts highlight the complexity of COPD and the relative independence and at the same time interdependence of airway remodelling and emphysema in their contribution to the development of COPD. Which end-points should then be used to assess available and future new drugs for the treatment of both forms of COPD? The rate of decline of forced expiratory volume in one second (FEV1), the traditional end-point used, requires a large number of subjects studied over a period of years [3], and does not track the presence and progression of emphysema well enough. FEV1 progression would be difficult to use in a1-AT-COPD since the number of patients that could be studied is much smaller and emphysema is the main pathological abnormality, thus it makes sense to measure emphysema and emphysema progression. The computed tomography (CT) scanner could potentially do it. In 1984, HAYHURST et al. [4] published an important study that was the base for our actual understanding of how to use CT scans for the diagnosis of emphysema. By comparing anatomical emphysema with CT in the same patients, they showed that the distribution curve of attenuated values was significantly shifted toward the lowest attenuation values in emphysematous lungs. Based on this concept, several groups have defined cut-off points in lung attenuation that could best quantify the extent of emphysema in the lungs. As is usually the case, we have ended up with multiple ways of looking at the same issue. Among these, the work by GEVENOIS and coworkers [5, 6] stands out for having carefully compared transversal CT slices and transversal anatomical lung slices showing that both macroscopic [5] and microscopic [6] emphysema are well quantitated by CT in smokers using an attenuation cut-off point of -950 Hounsfield units (HU). They also showed that normal lungs can have up to 7% of attenuation v-950 HU and this value can increase with increases in total lung capacity and older age [7]. A key question still remaining is whether evidence from CT showing that a drug has delayed destruction of lung tissue by emphysema is sufficient evidence of drug efficacy to satisfy regulatory requirements. The FEV1 is a well established measure of airflow obstruction, which has been shown convincingly to relate to patients9 functional status, prognosis and quality of life. What evidence is there that CT-demonstrated severity of emphysema is related to these clinical parameters? In this issue of the Journal, DOWSON et al. [8] attempt to answer these important questions. In a cross-sectional study of HRCT in 111 patients with a1-AT deficiency, they determined the relative area of pixels representing lung tissue with a density v-910 HU for a single upper zone slice (level of the aortic arch) and a single lower zone slice (level of the inferior pulmonary vein). HRCT was performed in both full inspiration and full expiration. In concert with previous studies, they showed moderately strong, *McGill University, Royal Victoria Hospital, Montreal, Quebec, Canada. Boston University School of Medicine, Boston, USA.