To the Editor.
—The article "What If Americans Ate Less Fat?" by Browner et al1 addressed an interesting and important topic. However, the authors may have substantially underestimated the reductions in mortality rates that may follow a reduction in fat intakes to 30% of energy by overlooking the effects of random error in serum cholesterol measurement, and especially in dietary fat measurement, in projecting corresponding CHD and cancer RR estimates.Specifically, their CHD RR estimates (Table 1, p3287) for a 20 mg/dL increase in serum cholesterol levels used data from Multiple Risk Factor Intervention Trial screenees. The random measurement error associated with a single serum cholesterol measurement is such that corresponding log RR estimates are likely to be underestimated by a factor of about 1.5 (Richard Peto, Oxford University, oral communication, June 1986). This correction increases the projected impact of fat reduction on CHD mortality by more than 50%.
Sheppard L, Prentice R. What If Americans Ate Less Fat?. JAMA. 1992;267(3):361-362. doi:10.1001/jama.1992.03480030039019