The correlations between the per capita niacin consumption and the diabetes prevalence in the US. A: The trends in the niacin consumption from 1909 to 2004 (Ref. ) and in the diabetes prevalence in the US in 1958-2008 (Ref. ). Note: the data on the prevalence of diabetes in 1969-1972, 1974 and 1977 are not available. B: The 26-year lag-regression plot using the data in A. C, E and G: The trends in the niacin consumption and in the diabetes prevalence in different age groups (Ref. ). D, F and H: The lag-regression plots using the data in C, E and G, respectively. The lag time is 24, 26 and 26 years for 65-74, 45-64, and 0-44 years age groups, respectively.