OK, many thanks to Lynne who managed to get the pdf. Extra Brownie Points for guessing that this was the paper I wanted, despite to my copy paste accident on the link!
It's talking about Japanese-Americans. You have to bear in mind that the information is limited by dietary preference questionnaires and the vagarities of deciding if someone has really had a heart attack. The group over 55 years is not shown as nothing much seems to make any difference in this group. Marmot discusses all of this in his paper. This is the relevant graph.
Now, you have to read this carefully. The pattern is the same, so I'll go through the top graph only. Looking at the left hand pair of vertical bars, the white one is the people who had a traditional Japanese up bringing and had a traditional Japanese diet preference. They had 2.5 heart attacks per hundred people.
Move over to the right hand side of the chart. Again the white bar is those people with a traditional Japanese up bringing but these are the ones who are now eating to what was the SAD back in the 1950s and 1960s. They had 0.4 heart attacks per 100 people. That is a relative risk of about 0.20, ie an American diet preference, compared to a traditional diet preference, appears to be HUGELY protective against CHD. Both of these groups are of a traditional Japanese style up bringing.
An American up bringing essentially doubles your risk of CHD, irrespective of whether you have a traditional Japanese dietary preference or an American dietary preference. It's worth noting that the combination of an American diet preference with an American upbringing is pretty well indistinguishable (3.0 heart attacks per hundred) from a traditional Japanese diet with traditional Japanese upbringing (2.5 heart attacks per hundred).
The protective effect of a traditional Japanese upbringing allows you to question any "dietary" intervention if it also adds in stress management, relaxation and exercise while asking you to shovel 1500kcal of rice a day down your throat!
Here's what Marmot, who had the misfortune to base his PhD on these data, had to say:
As the shot messenger I'd like to limp away now.