In most clinical series of patients with cryptogenic fibrosing alveolitis (CFA), disease severity is staged using lung function indices. However, many physiologic indices are measured in routine clinical practice; the choice of variable to evaluate functional severity is contentious. Computed tomography (CT) provides a reproducible means of quantifying the morphologic extent of disease. The aim of this study was to evaluate the functional consequences of smoking-related lung damage in CFA and to identify functional measures best reflecting the extent of fibrosing alveolitis on CT. Sixty-eight patients with CFA were studied. Fourteen patients with emphysema on CT were characterized by relative preservation of FVC and TLC (p < 0.005) and relative depression of DLCO (p < 0.05) and KCO (p < 0.00005). On multivariate analysis, the extent of fibrosing alveolitis and the presence of emphysema were independent determinants of functional impairment; there was no independent relationship between smoking history and functional abnormalities. In patients without emphysema on CT, percent predicted DLCO (r = -0.68), oxygen desaturation on exercise (r = 0.64), and the physiologic component of the clinical-radiographic-physiologic (CRP) score (r = 0.62) correlated much better with the extent of disease on CT than spirometric and plethysmographic volumes. A composite functional index was generated against the extent of disease on CT, using multivariate analysis; comparison with the CRP score suggested that the relationship between morphologic disease extent and the CRP score would be improved by the inclusion of DLCO and by the use of negative weighting for depression of FEV1. These findings indicate that in CFA, the presence of concurrent emphysema on CT has a more profound influence upon functional measures than the smoking history, and underline the importance of both the measurements of DLCO and exercise testing in the assessment of the severity of CFA.