Letters Section Editor: Robert M. Golub, MD, Senior Editor.
In Reply: Dr Chen and colleagues note 2 points of contention with our systematic review: the validity of dietary assessments for estimating fatty acid intake and our assessment of the quality of the animal and in vitro literature.
They are correct that the studies included in our meta-analysis did not measure fatty acid composition but rather estimated intake of fatty acids based on dietary information obtained from patients. However, their assertion that the methods used to define fatty acid intake in these studies were not valid is contradicted by other studies. As evidence that dietary questionnaires and records correlate poorly with direct measurements in patients, Chen et al cited 2 studies. The first study1 reported correlation coefficients between patient-reported intakes and measured content in subcutaneous tissue of 0.49 to 0.50 (P<.001) for polyunsaturated fats and 0.47 (P<.001) for eicosapentaenoic acid. Although not perfect, these correlations cannot be considered poor. Additional studies have demonstrated the validity of dietary questionnaires for the measurement of omega-3 fatty acids.2- 4 The second study5 cited by Chen et al examined the use of biomarkers for protein and energy intake to correct for reported intake and did not assess omega-3 fatty acids. The validity of the conclusions in this study has also been questioned because its analysis did not take into account random error over time in the biomarkers.6
MacLean CH, Newberry SJ. Omega-3 Fatty Acids and Cancer Risk—Reply. JAMA. 2006;296(3):282. doi:10.1001/jama.296.3.282-b