Cryptogenic fibrosing alveolitis: clinical features and their influence on survival

Abstract

A retrospective analysis of 220 cases fulfilling criteria for cryptogenic fibrosing alveolitis (CFA) attending the Brompton Hospital between 1955 and 1973 has been carried out and patients have been followed for between four and 21 years. The frequency of various clinical features confirms previous reports. The 2: 1 male preponderance was similar in all age groups. The mean age at presentation was 54 years ± 12 SD; 202 (92%) of the patients presented with dyspnoea, the severity of which was related to the reduction in vital capacity (p<0·003) and to the radiographic profusion score (p<0·01) but not to its duration. Twenty-one per cent of the 220 had joint symptoms, 10% having clinical rheumatoid arthritis. Eleven per cent had other types of connective tissue disorder. In all, 30% had polyarthritis or other immunological disorders and 70% lone CFA. Apart from an increase in rheumatoid factor in those with polyarthritis, there were no other clinical or survival differences between those with and without associated immunological disease. Forty-five per cent of 205 subjects had antinuclear antibody, and this occurred equally in those with and without associated connective tissue disorders. One hundred and forty-seven initial radiographs were available for reclassification using the ILO/UC system and only three were normal. Small rounded opacities were seen in 16% and small irregular opacities in 84%. Pleural changes were uncommon. Histological confirmation had been obtained in 118 patients and material was still available for review using a semiquantitative analysis in 68 (biopsy 42 and necropsy 26). Of the biopsies one could be classified as desquamative interstitial pneumonia (DIP) and 17 as endstage fibrosis; the other 24 showed a mixed cellular and fibrotic pattern. The necropsy material showed much greater fibrosis and less acute inflammatory cellularity in spite of an interval between onset of symptoms and death of less than four years in 21 of 26 patients. One hundred and fifty-six patients have died (mean survival 3·2 years). Eleven (5%) are believed to be alive but have been lost to follow-up. Fifty-five per cent of deaths were attributable directly or indirectly to CFA. There was also an excess of deaths from cardiovascular disease and lung cancer. Using a life-table analysis and a log rank test, longer survival was seen in younger patients (p<0·001) and women (p<0·01). After correction for age and sex, lesser grades of dyspnoea (p<0·03) and lesser radiographic abnormality (p<0·001), absence of right axis deviation (p<0·001), and a higher Pao2 (p<0·01) also related to longer survival. Subjects with more cellular histology also survived longer (p<0·02). Factors having no influence on survival included duration of dyspnoea before presentation, degree of reduction of FEV1, FVC, and TLC, the presence of “connective tissue” disorders, autoantibodies, smoking history, cough, sputum, crackles, clubbing, ESR, or immunoglobulins.