Reply

Abstract

We thank Dr. Vinciguerra for his interest in our article,1 which examined associations between two dietary indices, the Healthy Eating Index-2010 (HEI-2010) and the alternate Mediterranean Diet Score (aMED), with risk of chronic liver disease (CLD) mortality and hepatocellular carcinoma (HCC) incidence in the NIH-AARP Diet and Health Study. He specifically asks about protein intake, pointing out a recently published study examining the association of protein intake with total mortality in NHANES III. In that study, the authors observed that high protein intake was associated with higher mortality among those 65 years or younger, but lower mortality among those aged 66 years or older.2 As discussed in our article, we examined overall dietary patterns because it is difficult to disentangle individual components of diet from each other. For example, individuals who like to eat red meat may be more likely to eat sweets and less likely to eat beans, fruits and vegetables, and whole grains. As we mentioned, both HEI-2010 and aMED have been associated with a variety of disease outcomes.1 Nevertheless, it is also important to examine distinct aspects of diet as well, which is why we included these data. As there may be some confusion about the terms used for each dietary component in our article, we wish to clarify here. The total protein foods component in HEI-2010 specifically refers to lean red meat and poultry, eggs, as well as beans and peas (which are included when the total protein foods standard is not met). It does not include dairy; or seafood and plant proteins (except beans and peas), which are separate components in HEI-2010. We would also like to clarify that the red and processed meat of aMED was actually associated with CLD and HCC in the opposite direction, as mentioned by Vinciguerra, with participants eating more red and processed meat having nonsignificantly higher risk of HCC incidence and CLD mortality than participants eating less. Looking across the sources of protein, we see some evidence for increased risks, with eating more red and processed meat and decreased relative risks with eating more eggs, beans, and peas. Associations for dairy were mixed for each endpoint. As is always the case, and particularly with regard to observational studies, public health recommendations should only be made based on the totality of the evidence, not on the results of individual studies. Unfortunately, relatively few studies of diet in CLD and HCC are available. Future studies of dietary patterns and individual dietary components, such as protein, are needed. Wen-Qing Li, Ph.D.1,2 Yikyung Park, Sc.D.3 Katherine A. McGlynn, Ph.D.4 Albert R. Hollenbeck, Ph.D.5 Philip R. Taylor, M.D., Sc.D.2 Alisa M. Goldstein, Ph.D.2 Neal D. Freedman, Ph.D.3 1Department of Dermatology Brown University Warren Alpert Medical School Providence, RI 2Genetic Epidemiology Branch Division of Cancer Epidemiology and Genetics National Cancer Institute National Institutes of Health Rockville, MD 3Nutritional Epidemiology Branch Division of Cancer Epidemiology and Genetics National Cancer Institute National Institutes of Health Rockville, MD 4Hormonal and Reproductive Epidemiology Branch Division of Cancer Epidemiology and Genetics National Cancer Institute National Institutes of Health Rockville, MD 5AARP (retired) Washington, DC