We examined bronchoalveolar lavage (BAL) cell data from 19 patients with a lung disorder presenting clinical, radiologic, functional, and course characteristics of drug-associated interstitial pneumonitis. In each of them, one of 13 different drugs was incriminated and no other cause was found. In one case due to bleomycin, a neutrophil and eosinophil alveolitis was present. In the other 18, the common denominator was a lymphocyte alveolitis, either pure (n = 6) or associated with neutrophilia (n = 5), eosinophilia (n = 3), or neutrophilia and eosinophilia (n = 4). In addition, in all patients, an inverted CD4/CD8 lymphocyte ratio was observed. In eight patients who underwent another BAL, lymphocyte alveolitis decreased but was persistent in two of them two to four months after cessation of treatment with the drug incriminated, whereas interstitial pneumonitis had resolved clinically. In five patients, after resolution of pneumonitis and after an almost normal BAL cell profile was obtained, resumption of treatment with the suspected drug for two to four weeks induced a rise in lymphocyte population in a third BAL. In conclusion, apart from one case of bleomycin lung, the most striking feature of drug-associated alveolitis in this series was expansion of lymphocyte population and imbalance in lymphocyte subsets. When a provocation test was performed, variations in alveolar lymphocyte levels paralleled withdrawal and readministration of the drug responsible for alveolitis. These data could be of value in diagnosing and understanding drug-induced lung disorders.