Did you read the New England paper? They specifically measured serum levels, did a very thorough sub group analysis (including ethnicity and many other possible factors), broke down the supplementation by dosage.

Could you cite the paper you mean?