Errors in Meta-Analysis, Scientific Frustration and Nutrition Guidance
Wait for it….Scientists call for new media coverage to correct misinformation generated from a March 17 Annals of Internal Medicine paper “Association of Dietary, Circulating, and Supplement Fatty Acids with Coronary Risk: A systematic Review and Meta-analysis”. [Link to corrected paper]. But don’t count on new headlines on the subject. While scientists may be outraged by shoddy methodology, poor reviewing, and calling for retraction, it is unlikely this scientific correction will reach consumers.
While perspectives on the original paper are accessible on TalkingNutrition, (Mar 17 and Mar 18), readers have to cross a paywall to see the corrected paper, the supplemental data and the original version published on Mar 17. Basically, the addition of the final sentence “Nutritional guidelines on fatty acids and cardiovascular guidelines may require reappraisal to reflect the current evidence” was based on changes to the manuscript. Inclusion criteria for the meta-analysis were redefined in the new analysis with respect to sudden cardiac death. This increased the numbers to 105,085 participants with 6,229 incident coronary outcomes rather than the 103,052 participants and 5,726 coronary outcomes in the original version.
Today, the Annals of Internal Medicine has a link beside the abstract to a 2013 paper published by co-author Mozaffarian on ω-3 long-chain fatty acids and mortality. Is this an effort by the American College of Physicians to remedy an apparently weak review process and misleading headlines? Maybe because Mozaffarian and colleagues reported a 27% lower total mortality with higher total ω-3 LCPUFA concentrations. Individuals over 65y in the highest quintile of phospholipid ω-3 LCPUFA lived an average 2.2 years longer (vs lowest quintile). Clearly a different message than ‘no supportive evidence’.
Dr Walter Willett, Professor from Harvard University, says meta-analyses are often not appropriate for nutrition studies. Indeed, Dr Robert Heaney, Professor at Creighton University, has published 6 guidelines to follow when conducting meta-analyses of nutrition studies. Hopefully, researchers will apply these guidelines, editorial boards will review manuscripts with diligence, and journalists and other communicators will follow the International Food Information Council Foundation and Harvard School of Public Health guidelines when communicating science.
Chowdhury R, Warnakula S, Kunutsor S, Crowe F, Ward HA, Johnson L, Franco OH, Butterworth AS, Forouhi NG, Thompson SG, Khaw K-T, Mozaffarian D, Danesh J, Di Angelantonio E. Association of dietary, circulating, and supplemental fatty acids with coronary risk: A systematic review and meta-analysis. 2014 Ann Intern Med doi:10.7326/M13-1788 (corrected paper).
Kupferschmidt K. Scientists fix errors in controversial paper about saturated fats. 2014 Science Insider. Posted Mar 24, 2104 at 3:15pm.
Mozaffarian D, Lemaitre RN, King IB, Song X, Huang H, Sacks FM, Rimm EB, Wang M, Siscovick DS. Plasma phospholipid long-chain ω-3 fatty acids and total and cause-specific mortality in older adults: A cohort study. 2013 Ann Intern Med doi: 10.7326/0006-4819-158-201304020-00003
Heaney RP. Guidelines for optimizing design and analysis of clinical studies of nutrient effects. 2013 Nutr Rev doi: 10.1111/nure.12090
Fineberg HV, Rowe S. Improving public understanding: Guidelines for communicating emerging science on nutrition, food safety, and health. 1998 JNCI doi: 10.1093/jnci/90.3.194