We performed BAL and analyzed cellular components in 20 healthy volunteers (HV) and in 98 patients. In 52 patients the following diagnoses were established; idiopathic pulmonary fibrosis (IPF, 17 cases), sarcoidosis (SAR, 15 cases), collagen vascular disease (CVD, 8 cases), pneumoconiosis (PC, 7 cases) and non-summer type hypersensitivity pneumonitis (ns-HP, 5 cases). The other 46 patients were diagnosed as having other pulmonary diseases.
There were significant differences in the proportions of pulmonary alveolar macrophage (PAM) and lymphocyte (LY) between non-smokers and smokers of HV (p<0.05 for PAM and p<0.01 for LY). However there were no significant differences in the proportion of polymorphonuclear leukocyte (PMN) and the ratio of OKT4/OKT8 between them.
Non-smoking patients were classified as having normal or increased proportions of LY and PMN and as having low, normal or high OKT4/OKT8 ratios, taking the mean±SD of non-smokers of HV as the upper and lower limit of the normal range. Smoking patients were also classified in the same way, taking the mean±SD of HV smokers as the upper and lower limit of normal.
The most common BALF cell profile in IPF showed a PMN increase without any increase of LY and a normal ratio of OKT4/OKT8. There were a slight or moderate increase of LY without any increase of PMN and a high ratio of OKT4/OKT8 for SAR. A normal range of LY and PMN and a low OKT4/OKT8 ratio were found in CVD. A marked increase of LY without any increase of PMN and a high OKT4/OKT8 ratio were observed in ns-HP. There were no characteristic findings in the BALF cell profile of PC.
The sensitivity, specificity and positive predictive value of the BALF cell profile characteristic of each disease were calculated. Although specificity was high (92-99%), sensitivity and positive predictive value differed according to diseases; the former was 80% for ns-HP, 53% for SAR, 38% for CVD and 24% for IPF, the latter was 80% for ns-HP, 53% for SAR and less than 50% for IPF and CVD.
The positive predictive value, which refers to the proportion of patients with a particular BALF cell profile was not very high. The major reason for this was that respiratory infections due to viruses, mycoplasma and other pathogens expressed similar BALF cell profiles to ns-HP and SAR and because the BALF cell profiles of the other diseases varied widely. The predictive value of ns-HP and SAR markedly increased when the probability of the diseases was considered in conjunction with other clinical informations.
It was considered that BALF cell analysis had no definitive value in the diagnosis of pulmonary diseases, whereas it did have significance in cases clinically suspected of particular diseases such as ns-HP or SAR.
View full abstract