点击显示 收起
Department of Nutrition and Food Science
0112 Skinner Building
University of Maryland
College Park, MD 20742
E-mail: nsahyoun{at}umd.edu
Jean Mayer USDA Human Nutrition Research Center on Aging
Tufts University
Boston, MA
Center for Nutrition Policy and Promotion
US Department of Agriculture
Arlington, VA
Dear Sir:
We thank Drs Esmaillzadeh and Azadbakht and Drs Esposito and Giugliano for their interest in our study and for their comments. Prospective studies that used detailed dietary assessment, long-term follow-up, and control of multiple confounding factors found that whole grains are associated with lower risk of cardiovascular disease (CVD; 1, 2), type 2 diabetes mellitus (3, 4), and prevalence of metabolic syndrome (5, 6). To our knowledge, our study was the first to examine whole-grain intake in a sample of healthy elderly and to relate that intake to metabolic risk factors and mortality (7).
Esmaillzadeh and Azadbakht expressed concern that our subjects were not randomly selected. We share this concern, and we acknowledged in our report that our study population was not representative of the elderly and stated that, therefore, further studies in this age-group were needed to confirm our findings. Esmaillzadeh and Azabakht also indicated a concern about an increase in underreporting of energy intake with age. The extent to which there is underreporting of calories among older US adults is not known because of paucity of such data. The work that Esmaillzadeh and Azadbakht cite as evidence of increased underreporting of energy with age is based on a food-frequency questionnaire used in a Norwegian population (8). We used 3-d diet records in the current study, so it is not clear that this finding would pertain to our study. Moreover, we made an attempt in our study to reduce the potential for underreporting by having a trained dietitian instruct participants in keeping a food record and follow up with each subject in a face-to-face interview to probe for food items that may have been missed and to obtain brand names of foods.
In response to the question about categorization of foods as whole or refined grain, we, as stated in our report, used the US Department of Agriculture Pyramid Servings Database, which provides information on food-group classification. The total number of grain servings per 100 g food was based on the 3-d food record. That total number of grain servings was classified into whole-grain servings and non-whole-grain servings according to the proportion of the grain ingredients in the food that were whole grain and nonwhole grain as determined in the database. In our study population, whole-wheat, rye, and pumpernickel breads contributed 53%, cold cereals contributed 19%, hot cereals contributed 16%, and crackers and snacks contributed 6% of the whole-grain intake.
Esmaillzadeh and Azadbakht inquired about the possibility of an interaction between sex and whole-grain intake and confounding by other foods. In separate models, 2-way interactions between sex and whole- and refined-grain intakes were tested in regression models for all metabolic risk factors, metabolic syndrome, and CVD mortality. Because none of these interactions were significant, we combined the data for men and women to increase the statistical power of the study. Moreover, as we reported, we controlled for other dietary factors, such as fruit and vegetables and dairy products, but the results remained essentially unchanged. Therefore, we did not retain those factors in our final statistical models.
Esmaillzadeh and Azadbakht stated that the use of the 3-d diet records is a major weakness of our study. Clearly, the debate over the best dietary assessment method to use in observational studies is ongoing (9). All methods of collecting dietary data have inherent problems. Although no data are available from which to compare the relative capacity of different dietary methods to capture whole-grain food intake, we believe that the diet records may present an advantage over the food-frequency questionnaire in measurement of whole-grain foods. Because the open-ended format of the diet record includes a greater amount of detail, its use can capture sources of whole grains that are not listed or that may be grouped with refined-grain foods on a food-frequency questionnaire. In addition, the 3-d diet record was more quantitative, because portion sizes were weighed or measured by using household utensils provided by the study.
Finally, Esmaillzadeh and Azadbakht questioned our finding that elderly men are consuming more whole grains than are women. This finding is likely a consequence of the use of the diet records rather than a food-frequency questionnaire for assessing whole-grain intake, because the former can capture intake in a more quantitative manner. According to Table 1 in our article, the fourth quartile of both whole- and refined-grain distribution includes fewer women; that difference is a consequence of the higher energy intake among men. The whole-grain to refined-grain intake ratio across the quartiles of whole grain is higher in women (ie, 0.11, 0.37, 0.74, and 2.12) than in men (ie, 0.07, 0.24, 0.56, and 1.33).
We agree with Esposito and Giugliano that waist circumference is a better measure for defining abdominal adiposity than is BMI, but unfortunately, waist circumference was not measured in these elderly persons. In our study, 17 of 179 men (9.5%) and 125 of 356 women (35.1%) had elevated BMI. The use of BMI rather than waist circumference may have underestimated the number of people with abdominal adiposity, particularly men, which potentially led to a modest misclassification of the metabolic syndrome. In that case, the true association may have been stronger than the one we observed.
We thank Esposito and Giugliano for highlighting the potential inflammatory role that diets rich in whole-grain foods may play in reducing CVD risk. Indeed, diets rich in whole grains have been hypothesized to play a role in preventing CVD through various potential mechanisms, including body weight, glycemic control, plasma lipids, and inflammation. It is interesting that only recently has whole-grain intake been examined in relation to markers of inflammation (10, 11). A cross-sectional study in 938 healthy middle-aged persons found no significant association between whole-grain intake and several markers of inflammation, including concentrations of C-reactive protein, fibrinogen, and interleukin 6 (10). In contrast, whole-grain intake was significantly associated with lower concentrations of C-reactive protein and tumor necrosis factor- receptor 2 in 902 diabetic women (11). At present, the effect of whole-grain foods on inflammation should be established in different populations and also in clinical studies.
Our study of whole grains, metabolic risk factors, and CVD mortality was conducted in self-selected elderly volunteers more than 2 decades ago (7). As the first study to relate whole-grain intake to metabolic risk factors and mortality in healthy elderly, it provided findings that gave potentially valuable information on the relation of whole grains to CVD risk in older persons. Clearly, further research is needed to evaluate the metabolic consequences in older adults of consuming a diet high in whole grains.
ACKNOWLEDGMENTS
None of the authors had any personal or financial conflict of interest.
REFERENCES