Quotation bias in reviews of the diet-heart idea.

Criticism of the diet-heart idea is often met with the argument that consensus committees have settled the issue unanimously. To see how these committees have explained discordant results, quotations from papers with such findings were sought in three recent authoritative reviews. Only two of twelve groups of controversial papers were quoted correctly, and only in one of the reviews. About half of the papers were ignored. The rest were quoted irrelevantly; or insignificant findings in favour of the hypothesis were inflated; or unsupportive results were quoted as if they were supportive. Only one of six randomized cholesterol-lowering trials with a negative outcome were cited and only in one of the reviews. In contrast, each review cited two, four, and six non-randomized trials with a positive outcome, respectively. It appears as if fundamental parts of the diet-heart idea are based on biased quotation.