To examine the contribution of environmental and genetic risk factors to familial aggregation in chronic obstructive pulmonary disease (COPD), 325 first-degree (1d) relatives and 56 spouses of 150 COPD patients were compared with 222 1d relatives and 49 spouses of 107 nonpulmonary patient controls for the prevalence of two clinical outcomes: 1) airways obstruction (AO; 1-sec forced expiratory volume less than 68% of forced vital capacity) and 2) chronic bronchitis (CB; cough and sputum for 3+ months per year for 2+ years). The loglinear model was used to study direct and indirect (ie, those mediated by other risk factors) components of familial aggregation. Three risk factors were found to be independently associated with CB and/or AO: alpha 1-antitrypsin deficiency (PiZ allele), personal cigarette smoking, and parental cigarette smoking. Because 1d relatives of COPD patients were more likely to have a PiZ allele, be heavy smokers (1+ packs per day), and be exposed to parental smoking than 1d relatives of controls, these three factors also constitute indirect components of familial aggregation. However, after controlling for the three factors, 1d relatives of COPD patients were more likely to have AO and CB than 1d relatives of controls (direct component). This direct component might have a genetic basis, because no such association was found when spouses instead of 1d relatives were compared. Thus, both shared environmental factors (personal and passive smoking) and shared genetic factors (alpha 1-antitrypsin and a possible direct genetic component) contribute to familial aggregation in COPD. The loglinear model provides a useful tool for analyzing familial aggregation in diseases of multifactorial etiology.