We thank Drs Enright, Miller and Petsonk et al for their insightful comments regarding our recently published paper.1 Their concerns can be summarised as follows: (1) use of the fixed forced expiratory volume in 1 s/forced vital capacity (FEV₁/FVC) ratio of 0.70 rather than a lower limit of normal greatly “overdiagnoses” chronic obstructive pulmonary disease (COPD) and is thus detrimental to both public health and the psychological health of patients; (2) GOLD stage I COPD does not represent disease but is, in most people, simply normal ageing; and (3) our measure of “COPD-related hospitalisations” probably included many hospitalisations not caused by COPD.

While we recognise the potential for overdiagnosis posed by use of the GOLD fixed ratio criterion, historically the greater public health problem has been the underdiagnosis (and corresponding undertreatment) of COPD, especially lung function impairment that clinicians would consider clinically relevant (GOLD stage II or more severe).2 3 We have seen many patients with misdiagnosed lung disease, and the misdiagnosis is almost always a result of either not having performed spirometric tests or having poorly done spirometric tests. We therefore believe that the lack of spirometric testing in clinical practice is the larger threat to misdiagnosis and hence that, from a public health and educational perspective, the emphasis should probably be on underdiagnosis rather than on overdiagnosis. The problem of underdiagnosis is further exacerbated by poor quality spirometry and underestimation of FVC owing to inadequate emptying of the lungs. This is not to say that we should ignore the problem of potential overdiagnosis. The debate that is going on at present about the different criteria for airflow obstruction is a healthy one that can best be informed by analyses such as ours and the subsequent discussion that such analyses generate. This is how knowledge advances, and we are happy to have stimulated a lively discussion and anticipate that it will lead to further studies and analyses.

The question of whether the GOLD stage I classification really just represents “normal” ageing in most people is another area of ongoing debate. Although we do not dispute that the FEV₁/FVC decreases with ageing and that not everyone meeting GOLD stage I criteria has “disease”, we also know that, on average, older individuals with better lung function tend to live longer and are more healthy than those with worse lung function.4 We also know that those with respiratory symptoms fare worse than do those without such symptoms. Finally, while we acknowledge the potentially negative psychological consequences of labelling someone as having a chronic disease, patients understand the concept of disease staging, and telling someone that they have potentially early stage COPD may in fact have some positive public health consequences. For instance, one study has noted that patients who underwent spirometric tests and were found to have “mild” disease were more likely than subjects with “normal” lung function to stop smoking, and that cessation rates were even higher in those who were told they have “moderate” and “severe” disease.5 Again, we anticipate that trying to address the questions surrounding normal ageing versus the development of early disease will lead to further studies and analyses of longitudinal data.

As to whether our measure of COPD-related hospitalisations was too inclusive, we agree that not all of these hospitalisations were for COPD exacerbations per se. The link between COPD and other diseases (such as pneumonia, congestive heart failure and lung cancer) has been well established, and many hospitalisations in patients with COPD are for these co-morbid conditions.6 However, lung function impairment does consistently seem to increase the risk of such hospitalisations.

In conclusion, the letters published here raise some important issues at the intersection of how we define disease (particularly in epidemiological studies) and how we treat it. In practice, while we use the GOLD criteria to classify the severity of disease in our patients with respiratory disease, we use clinical criteria such as the presence of symptoms or quality of life impairment to guide treatment. This is particularly true in mild disease where the best interventions are smoking cessation, weight loss and exercise.