点击显示 收起
1 From the Department of Nutritional Sciences, The Pennsylvania State University, University Park.
See corresponding article on page 1375.
The conceptual framework of feedback regulation of iron absorption has been one of the central pillars in the study of human iron nutrition for many decades. The notion that as iron accumulates in the body, both heme and nonheme iron are less efficiently absorbed was tested and validated in many studies. Although many of those studies were direct human trials of iron absorption that used tracers, other approaches such as field trials, in which the most iron-deficient segments of the population respond to a greater extent than do other segments, also lend support to this regulatory schema. Despite the abundance of studies regarding inhibitors and enhancers of iron bioavailability, many experts on iron would agree that iron status itself is the major factor in determining the efficiency of iron absorption (1). Cell culture and molecular studies identified coordinated responses to variations in iron uptake, iron storage, and transcytosis of iron in mucosal cells when there is a change in the iron status of enterocytes (2). The nature of the feedback signal from hepatic liver stores is being elucidated by using information from a genetic knockout mouse model (3). Thus, the pieces are gradually being put together to further clarify the nature of this feedback regulation. On the basis of their clinical and human absorption studies, Hallberg and coworkers (4, 5) approached this question in a somewhat different fashion. They reexamined their nearly 35 y of experience in iron absorption studies in which radioactive iron was administered to human volunteers in highly controlled laboratory experiments. The authors conducted a pseudo-meta-analysis in which they combined data sets from many of their studies to generate sex-specific prediction equations that related the efficiency of iron absorption to plasma ferritin concentrations. Using this combined data set with sufficient statistical power, they verified the strong negative relation between absorption efficiency and plasma ferritin concentration. Importantly, they concluded that when ferritin concentrations were > 60–80 µg/L, the efficiency of absorption was nearly identical to basal iron requirements (ie, 1 mg/d). An inference from this model is that populations would not be likely to accumulate excess iron through normal dietary intakes or, perhaps, even through moderate doses of highly bioavailable iron salts. Their analysis, however, does not exclude the possibility of clinical, genetic, or pharmacologic situations whereby > 1 mg Fe/d would pass across mucosal cells.
The concept of iron balance being achieved through a tightly controlled and efficient process of absorption was examined in an elderly cohort with the use of the cycle 20 data set from the Framingham Heart Study; the results of that examination are presented in this issue of the Journal (6). The authors probed this underlying question in a group of healthy elderly subjects by examining the relation of dietary factors (both nutrient and diet models) and nondietary factors to plasma ferritin concentrations. This question has important implications for dietary policy in the preelderly because there is a growing belief that individuals with high iron stores have an increased risk of developing pathologic processes (7). The authors showed that after adjustment for nondietary covariates, there was an association between iron supplement use ( 30 mg Fe/d), consumption of > 21 servings of fruit/wk, or consumption of 7 servings of red meat/wk and the risk of high iron stores (adjusted odds ratios of 4.32, 2.88, and 3.61, respectively). Neither total dietary iron intake nor light-meat intake was associated with the risk of elevated ferritin concentrations.
To place these findings in perspective, one could examine the report on dietary reference intakes that was recently released by the Institute of Medicine, Panel on Micronutrients, which considered iron requirements (8). In that report, the relation between dietary iron intake and iron status in the third National Health and Nutrition Examination Survey is shown in Table H-5. Among men aged > 71 y, those in the lowest quartile of iron intake (median: 8.3 mg Fe/d) had a median plasma ferritin concentration of 148 µg/L. In the highest quartile of iron intake (median: 25.1 mg Fe/d), the median plasma ferritin concentration was 134 µg/L. Among women aged > 71 y, the median plasma ferritin concentrations of those in the lowest and highest intake quartiles (10.8 and 21.1 mg Fe/d, respectively) were 107.5 and 87 µg/L, respectively. In the men, there were no significant increases in median serum ferritin concentrations between the third and seventh decades of life, whereas in the women there was a clear plateau after the fifth decade through the sixth decade. These data support the concept of tight homeostatic regulation of iron absorption in an iron-replete population during the preelderly and elderly years because there was no significant difference in median iron stores between the subjects in the lowest quartile and those in the highest quartile. The interpretation of the results from the Framingham Heart Study is of great importance because it suggests that elderly individuals with high supplement use, high red-meat intake, and high fruit consumption will have high plasma ferritin concentrations. Who exactly is at risk? There was a strong relation between the risk of high iron stores and heme-iron intake when subjects consumed > 7 servings of red meat/wk; however, this intake was nearly 3 times that of the median intake of 2.5 servings/wk, and thus the real effect may be seen only in a fringe group. A similar argument might be made for supplement use, but the use of iron supplements in most studies is hard to document. Because Fleming et al (6) went to great lengths to remove subjects with diseases or genetic conditions that might alter plasma ferritin concentrations (45% reduction in sample size), the influence of these confounders on the relation between dietary factors and the risk of high iron stores was removed. In its analysis of dietary reference intakes, the Panel on Micronutrients did not go to great lengths to control for potential covariates. Thus, the 2 approaches are not entirely comparable. Other studies, however, used models and methods similar to those used by Fleming et al in studies of diet and ferritin in the elderly. Garry et al (9) found no relation between dietary heme iron and plasma ferritin in a healthy, aging cohort from New Mexico. Considering the heated debate on the causal relation of dietary iron intake and iron accumulation with disease processes, these data are very sparse. Some studies in the elderly showed a significant adjusted odds ratio for dietary heme-iron intake and myocardial infarction but not for dietary heme-iron intake and serum ferritin (10, 11). Meta-analyses and examinations of all causes of death relative to serum ferritin concentrations failed to show a relation between ferritin concentrations and myocardial infarction (12, 13).
In summary, this most recent foray into the relation between dietary iron intake and storage iron draws into question the precision with which the homeostatic feedback loop regulates iron absorption in the elderly. It may prove to be informative to conduct the same analyses on data collected from these subjects during their younger years and compare the results with those of the present study. Perhaps other studies with careful collection of dietary, clinical, and iron-assessment data may help clarify the apparent disagreement.
REFERENCES