The diagnostic utility of transbronchial lung biopsy (TBB) is partly a function of its size. However, objective parameters that reflect biopsy specimen size have not yet been well-defined. We studied clinical records and histopathologic lung tissue slides of 116 patients who underwent diagnostic TBB, aiming to define the possible significance of association between seven parameters and three categories of pathologic diagnoses. Three of the seven parameters were clinical: age, sex, and chest roentgenographic infiltrates (localized vs diffuse). The remaining four parameters were histopathologic and morphometric: total number of tissue fragments, total number of alveoli (per biopsy specimen), total tissue area (alveolated plus nonalveolated), and lung total area (alveolated tissue alone). The three categories of pathologic diagnoses were as follows: infection, tumor, and nonspecific diagnoses. The nonspecific diagnoses included diagnoses of fibrosis and/or chronic inflammation. The alveoli were microscopically counted by one of us (S.D.G.). The number of biopsy fragments, the total tissue area, and the total lung area were measured in square millimeters by a computer-assisted digitizing system using specific (Bio-Quant) software (R and M Biometrics Inc). The significance of the associations between the seven parameters and the three diagnostic categories were assessed by the chi2 test for association. Overall, the following four possible associations were found to be statistically significant: (1) age--a lower percentage of patients with infection was found among patients with increasing age (p less than 0.001); (2) roentgenographic findings--a greater percentage of tumor diagnoses were found in patients with localized infiltrates (p = 0.006); (3) number of biopsy fragments--a greater percentage of patients with diagnoses of infection was identified among patients whose biopsy specimens contained the highest number of tissue fragments (p = 0.04); and (4) number of alveoli--a greater percentage of diagnosis of infection was made in patients whose biopsy specimens contained greater than or equal to 20 alveoli (p = 0.01). Our findings support the notion that the diagnostic utility of TBB is related to its size. However, this relationship between TBB size and diagnostic utility was apparent only for diagnoses of infection and not for diagnoses of tumor. We conclude that TBB specimens containing 20 or more alveoli may (1) be declared to be adequate for diagnosis, (2) in the appropriate clinical setting, they will be most likely to yield a diagnosis of infection, and (3) the number of alveoli does not appear to be associated to the diagnosis of tumor.