Reduction of alcohol-related mortality is a national goal for health promotion and disease prevention. We conducted this analysis to determine whether trends in New Mexico's Hispanics, non-Hispanic Whites, and American Indians were consistent with national trends in alcohol-related mortality, and whether differences in drinking patterns could account for racial and ethnic differences in rates. Age-adjusted, race-specific, and ethnic-specific alcohol-related mortality rates and 95% confidence intervals were calculated for 5-year periods for 1958-1991 using New Mexico vital statistics data. We estimated the prevalence of acute and chronic at-risk drinking behaviors and abstinence from data collected by the Behavioral Risk Factor Surveillance System (BRFSS) for the period 1986-1992. We found that alcohol-related mortality rates varied substantially by race, ethnicity, sex, age, and calendar period. American Indians had the highest rates for both sexes. Rates increased sharply from the period 1958-1962 until the late 1970s and the early 1980s, and then began to decrease rapidly. However, during the most recent decade, the rates have followed contrasting trends in the three ethnic and racial groups. Although rates have continued to decline among non-Hispanic Whites, rates for Hispanics and American Indians have not declined, and still remain substantially higher than rates during the 1958-1962 period. Differences in at-risk drinking behaviors reported to the BRFSS do not explain the contrast in race-specific and ethnic-specific mortality rates. Although progress has been made in reducing national per capita alcohol consumption and alcohol-related mortality, certain high-risk racial and ethnic groups may not be sharing in the progress.