High fish consumption is characteristic of Japanese-American men of the Honolulu Heart Program (HHP). Analyses of data from the Atherosclerosis Risk in Communities (ARIC) study suggest high fish intake protects the lung against smoking damage. Measurements of forced expiratory volume in 1 s (FEV1) and smoking status in the HHP cohort were done at the first examination in 1965-68. Among 8,006 men, 45 to 68 yr, 6,346 had acceptable spirograms. Within current smokers, 1,545 men consumed fish less than twice a week, and 1,264 ate fish twice a week or more. Controlling for cigarettes/d, age, height, and daily calories, separate regression models indicated an average decrease of -10.1 ml for each additional yr of smoking (95% Confidence Interval [CI]: -13.6, -6.5) at low levels of fish intake, and a decrease of -4.4 ml (95% CI: -8.2, -0.6) at high levels. The coefficients were significantly different (p = 0.03). These differences reflect a predicted FEV1 144 ml (95% CI: 62, 227) higher in the high fish group at > or = 40 yr of smoking, but no difference at < or = 35 yr. Similar analyses were conducted for cigarettes/d. On average, the FEV1 decline for each additional cigarette/d was not significantly different among subjects with low versus high fish intake. However, the predicted FEV1 at < or = 30 cigarettes/d was 52 ml (95% CI: 17, 87) higher in the high fish consumption group. No significant difference in FEV1 was noted between groups at > 30 cigarettes/d. These findings suggest that the protective role of fish is "saturated" at higher "doses" of cigarette smoking.