Medical records of the 15,924 twin-pairs in the National Academy of Sciences-National Research Council (NAS-NRC) twin registry were collected for an additional 16 years through 1994 when the surviving twins were aged 67 to 77 years. Compared with earlier analyses (Hrubec, Z, and Omenn, G. S., Alcohol. Clin. Exp. Res., 5:207-215, 1981), when subjects were aged 51 to 61, there were 23% more diagnoses of alcoholism (34.4 per 1,000 prevalence), 32% more diagnoses of alcoholic psychosis (5.4 per 1,000), and 25% more twins with liver cirrhosis (17.7 per 1,000). Overall, 5.3% of the cohort had at least one of the diagnoses related to alcoholism. Probandwise concordance rates (%) were: alcoholism-26.7 monozygotic (MZ), 12.2 dizygotic (DZ) (p < 0.0001); alcoholic psychosis-17.3 MZ, 4.8 DZ (p < 0.05); and cirrhosis-16.9 MZ, 5.3 DZ (p < 0.001). Concordance for any diagnosis related to alcoholism was 30.2 MZ, 13.9 DZ (p < 0.0001). Maximum-likelihood modeling indicated that approximately 50% of the overall variance was due to additive genetic effects; in all diagnosis categories, a totally environmental model gave a significantly poorer fit to the data. Bivariate and trivariate genetic analyses indicated most of the genetic liability for the organ-specific endpoints of psychosis and cirrhosis was due to the shared genetic liability for alcoholism. Once the shared variance with alcoholism was considered, there was no further shared genetic liability for psychosis and cirrhosis. Our results confirm Hrubec and Omenn's conclusion that there was significantly greater concordance in MZ twins-pairs for alcoholic psychosis and cirrhosis in the NAS-NRC twins, and concordance rates remained similar to those reported 16 years earlier. In contrast, we found most of the genetic liability to organ-specific complications of alcoholism was shared with the genetic liability for alcoholism per se; only a small portion of the genetic variance of the individual complications was independent of the genetic predisposition for alcoholism.