We determined the prevalence of respiratory symptoms and lung dysfunction in a large sample of habitual smokers of freebase cocaine ("crack") alone and in combination with tobacco and/or marijuana. In addition, we compared these findings with those in an age- and race-matched sample of nonusers of crack who did or did not smoke tobacco and/or marijuana. A detailed respiratory and drug use questionnaire and a battery of lung function tests were administered to (1) a convenience sample of 202 habitual smokers of cocaine (cases) who denied intravenous drug abuse and (2) a reference sample of 99 nonusers of cocaine (control subjects). The cocaine smokers (85% black) included the following: 68 never-smokers of marijuana, of whom 43 currently smoked tobacco and 25 did not, and 134 ever-smokers of marijuana (42 current and 92 former), of whom 92 currently smoked tobacco and 42 did not. The control subjects (96% black) included the following: 69 never-smokers of marijuana, of whom 26 currently smoked tobacco and 43 did not, and 30 ever-smokers of marijuana (18 current and 12 former), of whom 21 currently smoked tobacco and 9 did not. Cases smoked an average of 6.5 g cocaine per week for a mean of 53 months. The median time of the most recent use of crack prior to study was 19 days (range less than 1 to 180 days). After controlling for the use of other smoked substances, frequent crack use was associated with: (1) a high prevalence of at least occasional occurrences of acute cardiorespiratory symptoms within 1 to 12 h after smoking cocaine (cough productive of black sputum [43.7%], hemoptysis [5.7%], chest pain [38.5%], usually worse with deep breathing, and cardiac palpitations [52.6%]) and (2) a mild but significant impairment in the diffusing capacity of the lung.(ABSTRACT TRUNCATED AT 250 WORDS)