Abstract
Iron status and dietary correlates of iron status have not been well described in Hispanic older adults of Caribbean origin. The aim of this study was to evaluate iron status and describe dietary components and correlates of iron status in Hispanic older adults and in a neighborhood-based comparison group of non-Hispanic white older adults. Six hundred four Hispanic and non-Hispanic white adults (59-91 y of age) from the Massachusetts Hispanic Elders Study were included in the analysis. We examined physiological markers of iron status as well as dietary factors in relation to iron status. Dietary intake was assessed by FFQ. Our results revealed that Hispanics had significantly lower geometric mean serum ferritin (74.1 μg/L vs. 100 μg/L; P < 0.001), lower hemoglobin concentrations (137 ± 13 vs. 140 ± 12 g/L; P < 0.01), higher prevalence of anemia (11.5 vs. 7.3%; P < 0.05), and suboptimal hemoglobin concentrations (<125 g/L) for this age group (21.4 vs. 13.3%; P < 0.05). Iron deficiency anemia was higher (7.2% vs. 2.3%; P < 0.05) in Hispanic women. Hispanics had lower mean intakes of total iron, vitamin C, supplemental vitamin C, and total calcium than did non-Hispanic whites. After adjusting for age, sex, BMI, alcohol use, smoking, total energy intake, inflammation, diabetes, and liver disease, intake of heme iron from red meat was positively associated and dietary calcium was negatively associated with serum ferritin. This population of Hispanic older adults was significantly more likely than their non-Hispanic white neighbors to suffer from anemia and poor iron status, particularly among women. Cultural variation in dietary patterns may influence iron availability and body iron stores and contribute to an increased risk for iron deficiency anemia among some Hispanic older adults.
Introduction
Iron is an essential mineral that functions to bind oxygen as part of heme in hemoglobin and myoglobin. In addition, iron is needed in various heme- and nonheme-containing metalloenzymes in the body. Iron can also participate in the generation of free radicals, which can damage DNA, proteins, and cellular components, such as cell membranes. Thus, availability of free iron in the body is tightly controlled during transport and metabolism by binding to various proteins. In the cell, ferritin protein sequesters free iron and the synthesis of ferritin is increased in response to an increase in iron availability. Serum ferritin has been shown to be a good biomarker of body iron stores. Very high body iron stores have been associated with various pathologies, including cardiac failure, liver cirrhosis, hepatic cancer, and diabetes (1-5). On the other hand, iron deficiency anemia (IDA)6 remains an important problem.
In older adults, anemia is prevalent and is associated with increased morbidity and mortality (6,7). In one study, anemia was observed in 11% of individuals >65 y of age, and in >20% of those ≥85 y (8). Onset is insidious, so anemia can often go undetected (9). Approximately 80% of cases have a known etiology, with >34% linked to nutrition, and therefore modifiable (10,11). In contrast to younger age groups, the majority of cases in older adults are not typically due to iron deficiency (12). Rather, many cases are due to chronic infectious or inflammatory processes called “anemia of chronic disease.” We previously reported that although 10% of elderly participants in the Framingham Heart Study cohort had anemia, the majority were due to anemia of chronic disease and not to iron deficiency (12). Furthermore, elevated serum ferritin concentrations were 4 times more prevalent than IDA (13).
Dietary composition is an important determinant of iron bioavailability and iron status (13-17). Several studies demonstrated associations between heme iron from meat or iron supplements and serum ferritin (15,16). We previously reported associations between dietary factors and risk of elevated serum ferritin in the Framingham Heart Study (18). However, that population is almost exclusively non-Hispanic white, and results cannot be generalized to other ethnic groups with differing dietary patterns. The Hispanic population is the fastest-growing ethnic minority in the United States (19). A traditional Hispanic dietary pattern is more plant-based than the typical American diet (20,21), and plant-based diets are associated with low iron absorption and poor iron status.
Iron absorption occurs predominantly in the duodenum and upper jejunum (22) and absorption is influenced by the presence of gastric acid, which enhances iron uptake, along with dietary factors, such as ascorbate and citrate. However, iron absorption can also be inhibited by calcium, plant phytates, and tannins. Studies have evaluated iron bioavailability of foods from various Hispanic diets, but these do not provide information on the long-term effects of such dietary patterns on iron status (20,23-25). The objectives of this study were, therefore, to describe the iron nutriture and status of Hispanic older adults and to identify dietary correlates associated with biological markers of iron status in this population.
Methods
Study population
This analysis was conducted with data from the Massachusetts Hispanic Elders Study (MAHES), a statewide epidemiologic survey conducted between 1993 and 1997 to examine health, nutrition, diet, and disability status among elderly Hispanics residing in the state of Massachusetts. A 2-stage cluster sampling method, previously described (26), was utilized to obtain a representative sample of Hispanic older adults from neighborhoods in Massachusetts and non-Hispanic white older adults from the same neighborhoods to use as a comparison group. Briefly, counties were randomly selected based on the proportional representation of the Hispanic population >55 y of age within the county according to 1990 census data. Within these selected counties, census blocks that contained at least 2 Hispanics ≥55 y of age were randomly sampled. Through door-to-door enumeration, both Hispanic and non-Hispanic white older adults were simultaneously located within the same census blocks.
Of those invited, 83% of Hispanics (n = 779) and 54% of non-Hispanic whites (n = 251) participated (26).The non-Hispanic white group was randomly selected, from the same neighborhoods as Hispanic participants, to control for socioeconomic indicators representative of the Hispanic group (21). The final sample included 618 Puerto Ricans and Dominicans, and 161 “other Hispanics” comprised of Cubans, Mexicans, Central Americans, or South Americans. The “other Hispanic” group was excluded from this analysis due to their heterogeneity. Subjects lacking serum ferritin (n 233) or with missing or implausible total energy intake (>16,736 kJ or <2510 kJ) (n = 31) were also excluded. Therefore, the current analysis included 604 subjects aged 59-91 y; 451 Hispanic older adults from Puerto Rico and the Dominican Republic, and 153 non-Hispanic whites. All study procedures were approved by the Institutional Review Board of the Tufts New England Medical Center in Boston, MA.
Dietary intake
Dietary intake data were collected using a FFQ specifically designed and validated for Hispanic adults residing in the northeastern United States (13). Trained bilingual interviewers collected dietary information from study subjects using 3-dimensional models and measurement tools to aid in portion size estimation. Nutrient data were assessed with the Nutrient Data System, version 4.06 (NDS, University of Minnesota).
Biochemical measurements
Fasting blood samples were collected by venipuncture into evacuated EDTA-containing tubes during the home visit. All blood samples were taken to the Jean Mayer Human Nutrition Research Center on Aging (HNRCA), at Tufts University, in a cooler of ice, within 2 h of collection, and clinical chemistries were determined. White blood cell count (WBC), red blood cell count (RBC), and mean cell volume (MCV) were measured in whole blood specimens with a system 9000 Diff Model Automated Cell Counter (Serono-Baker Diagnostics). Hemoglobin was measured with the cyanmethemoglobin procedure. Hematocrit was calculated from the measured variables as (MCV) × (RBC)/10.
Physiologic iron stores were estimated on the basis of serum ferritin concentrations. Serum ferritin was measured using the Magic Ferritin 125I radioimmunoassay (Ciba Corning). As a means of quality control, we conducted the assay of the WHO International Ferritin Standard 80/578 with the Magic Ferritin radioimmunoassay. Mean ferritin values were within 5-10% of the stated concentration of this quality control. Plasma C-reactive protein (CRP), an inflammatory marker, was measured with a Behring Nephelometer (Behring Diagnostics). Inflammation was defined as plasma CRP ≥10 mg/L (27).
Socio-demographic and health status
Information on age, education, income, physical activity, and alcohol and tobacco use, as well as medical history, was obtained by a questionnaire adapted from questions used in the Hispanic Health and Nutrition Examination Survey. All methods were pretested for clarity with subjects similar to those in the target population before the beginning of formal data collection. These factors represent nonnutritional determinants of serum ferritin and were analyzed for their possible confounding effects on body iron status.
Criteria for evaluating iron status
Ferritin is an acute-phase protein whose synthesis and secretion by hepatic cells is increased by inflammatory cytokines (12,13,28-30). Thus, whereas serum ferritin typically reflects body iron stores (31,32), it is an unreliable indicator of iron status in the presence of various acute and/or chronic disease conditions such as inflammation, infection, liver disease, and malignancy as its concentration becomes disproportionately elevated relative to actual iron stores. We attempted to control for the possible confounding effects of these conditions by defining various disease conditions and adjusting for these in our models.
To evaluate the potential confounding effects of anemia of chronic disease the following guidelines were established. Plasma CRP concentrations ≥10.0 mg/L was used to define inflammation (n = 82, 18% of Hispanics; n = 21, 13% of non-Hispanic whites) (13,27). Infection status was based on sex-specific cutoff values for WBC count (13). Men with WBC concentrations <3.9 × 103 or >10.6 × 103/mm3 and women with concentrations <3.5 × 103 or >11.0 × 103/mm3 were categorized as having an infection. However, infection was not included in the final analysis, as it did not appear to affect serum ferritin. Diabetes and self-reported physician-diagnosed liver disease were also considered for their confounding effects. Subjects with one or more of the following were defined as having diabetes: fasting blood glucose >6.99 mmol/L, random (nonfasting) blood glucose >11.10 mmol/L, or pharmacological treatment for diabetes (diabetes related drugs or insulin). To increase comparability, we followed similar analysis procedures to those performed by Fleming et al. (13) in the Framingham Heart Study.
Hereditary hemochromatosis is an autosomal recessive disorder characterized by elevated iron absorption that affects ∼1 in 250 persons in populations of northern European descent and is characterized by pathologic elevations in blood ferritin concentrations (33). These elevations reflect actual increases in body iron stores. Given the probability that individuals within this study population may be homozygous for hemochromatosis is very small, we did not exclude subjects at risk for this mutation.
Iron stores
Generally acceptable cutoff points for iron status measures were selected. Given the age of our population and the evidence supporting a higher serum ferritin threshold for depleted iron stores in the elderly, we selected a conservative cutoff of <15 μg/L for evaluation of iron deficiency (34,35). Based on data from NHANES II (19), the 95th percentile values for serum ferritin concentration in women aged 18-64 y and in men aged 18-64 y were 193 and 299 μg/L, respectively. Therefore, we selected sex-specific cutoffs to define elevations of serum ferritin >200 μg/L for women and >300 μg/L for men, as used previously (12).
Definition of anemia
The WHO criteria for anemia are estimated on the basis of hemoglobin or hematocrit concentrations (12,36,37). Prevalence estimates of anemia are based on the NHANES III cutoff criteria for an abnormal hemoglobin concentration (<119 g/L for women and <124 g/L for men). We further evaluated the prevalence of suboptimal hemoglobin concentrations <125 g/L (in both sexes) in this population, as a preponderance of recent research indicates that current guidelines are not clinically optimal. Our evaluation of “suboptimal” hemoglobin concentrations allows us to evaluate iron status using cutoffs that may be functionally relevant (38,39). We also evaluated the prevalence of abnormally low hematocrit concentrations, using the WHO criteria (men <0.40; women <0.37).
We attempted to differentiate IDA from other types of anemia, after an evaluation of disease states. We defined IDA as serum ferritin concentrations <15 μg/L, or hemoglobin <119 g/L for women and <124 g/L for men, and MCV <80 μm3/cell (40).
Statistical analysis
All statistical procedures were performed with SAS for Windows, version 8.1 (SAS Institute). Descriptive comparisons by ethnicity were examined by t test and chisquare analyses. Prevalence estimates for low hemoglobin concentrations, suboptimal hemoglobin, and anemia were obtained by sex and ethnic group. Serum ferritin values were assessed for both the total sample and the majority subset without inflammation (defined here as CRP ≥10 mg/L). Variable distributions were assessed and logarithmic transformations applied when necessary.
For the Hispanics only, the logged ferritin variable was regressed onto dietary intake variables, adjusting for age, sex, BMI, total energy intake, and additional potential confounders. In addition to age, sex, and BMI, nondietary factors previously shown to be associated with serum ferritin include age, sex, BMI, smoking status, and aspirin use (13). Total energy intake was included in the regression analyses to adjust for differences in energy intakes and for possible systematic over- or underreporting of dietary intake (13,41). Alcohol intake (no intake, moderate, or heavy) was also included, as it is known to be positively associated with serum ferritin (42). Moderate intake was defined as ≤13.2 g/d for women and ≤26.4 g/d for men and heavy intake was defined as >13.2 g/d and >26.4 g/d for women and men, respectively. Use of medications, including aspirin, nonsteroidal anti-inflammatory agents, antiplatelet medications, antiulcer medications, and anticoagulants, were examined for possible confounding effects but were excluded from the regression analysis when the variables did not appear to affect serum ferritin concentrations in this sample. We also adjusted for presence of inflammation (CRP ≥10 mg/L, yes/no), diabetes (yes/no), and liver disease (yes/no) as these may confound the diet-ferritin associations. Repeating the analyses excluding those with inflammation did not alter the results, therefore the full sample is presented. Values presented in the text are means ± SD.
The 1st set of dietary factors included total intake of iron, vitamin C, calcium, dietary fiber, and caffeine (from all sources). To examine the impact of various nutrient intakes on iron status, the 2nd set of dietary factors divided total nutrient intakes into their dietary and supplement components. Dietary iron was further separated into heme and nonheme iron. Because white meat has considerably less heme iron, we examined heme iron from red meat and heme iron from white meat separately. Nonheme iron was defined as the difference between total dietary iron and heme iron. Supplemental intake of iron, vitamin C, and calcium were entered as categorical (yes/no) variables.
The 3rd set of dietary factors included general food and beverage groups composed of items collapsed from the FFQ. Food groups were created by first combining total daily intake (g) of specific foods. The total was then divided by a median-defined portion size and then multiplied by 7 to reach the total servings per week. Food groups included milk (skim, low fat, or whole), fruit (including all fruit and fruit juices), starchy vegetables (including plantains, potatoes, and other root crops), and all other vegetables (including tomatoes). The legumes group included dried beans and peas. White meats included chicken, turkey, and fish. The red meat group was composed of liver, beef, pork, and all processed meats. Fortified and refined grain products were classified as different groups based on the form of iron used for fortification, i.e., the bread and pasta group, which consisted of items fortified primarily with iron sulfate, and the sweet baked goods group, which consisted of items fortified with elemental iron powder produced by hydrogen reduction (i.e., reduced iron). Breads were white or wheat, corn bread, corn tortillas, and sweet baked goods. Cold breakfast cereals were divided into refined or whole-grain cereals on the basis of the content of whole grain or bran.
Results
Demographic and health measures
The sample included 451 Hispanics and 153 non-Hispanic whites (Table 1). Our Hispanic group was slightly younger than our non-Hispanic white group. Compared with non-Hispanic whites, fewer Hispanics were heavy alcohol users or current smokers. Type 2 diabetes was more prevalent in the Hispanic group (P < 0.01), as previously reported (26). The prevalence of inflammation (CRP ≥10 mg/L), BMI, or total energy intake between Hispanics and non-Hispanic whites did not differ.
TABLE 1.
Descriptive characteristics of Hispanics and non-Hispanic whites in the Massachusetts Hispanic Elders Study, 1993-19971
| Hispanic | Non-Hispanic white | |
|---|---|---|
| n | 451 | 153 |
| Age, y | 68.7 ± 7.1*** | 71.1 ± 7.2 |
| Female, % | 58.4 | 60.3 |
| BMI, kg/m2 | 28.4 ± 5.8 | 28.1 ± 6.8 |
| Energy intake,2 kJ/d | 7,183 ± 2,717 | 7,233 ± 2,746 |
| Alcohol use,3 % | ||
| Moderate intake | 31.4 | 53.2 |
| Heavy intake | 1.3*** | 10.3 |
| Current smoker, % | 15.8*** | 30.1 |
| Medication use,4 % | 42.3 | 39.0 |
| Inflammation,5 % | 18.0 | 13.0 |
| Elevated serum ferritin,6 % | ||
| Total sample | 8.2 | 13.7 |
| Excluding those with inflammation5 | 8.3 | 12.1 |
| Diabetes,7 % | 36.3** | 23.1 |
| Liver disease,8 % | 4.0 | 2.6 |
Values are means ± SD or %. Asterisks indicate different from non-Hispanic whites,
P < 0.01,
P < 0.001 by chi-square or t test analysis.
1 kcal = 4.184 kJ.
Moderate alcohol intake is defined as ≤13.2 g/d for women and ≤26.4 g/d for men. Heavy alcohol intake is defined as >13.2 g/d for women and >26.4 g/d for men.
Includes aspirin, nonsteroidal anti-inflammatory medications, antiplatelet medications, antiulcer medications, and anticoagulants.
CRP ≥10.0 mg/L.
Elevated serum ferritin >200 μg/L for women and >300 μg/L for men.
Defined as: 1) fasting blood glucose >6.99 mmol/L; 2) random (nonfasting) blood glucose >11.10 mmol/L; or 3) on medication for diabetes (diabetes related drugs or insulin).
Self-reported.
Dietary intake
The dietary intake of Hispanics included greater amounts (P < 0.05) of starchy vegetables, rice, milk, and legumes than the non-Hispanic white group (Table 2). Mean intakes of red meat, other vegetables, breads, cereals, coffee, and tea were lower among Hispanics than among non-Hispanic whites.
TABLE 2.
Mean and median servings/wk of food groups by Hispanics and non-Hispanic whites1
| Hispanic, n = 451 |
Non-Hispanic whites, n = 153 |
|||
|---|---|---|---|---|
| Food Intake | Mean ± SD | Median | Mean ± SD | Median |
| Milk | 8.5±7.7* | 7.0 | 7.2±6.3 | 5.9 |
| Fruit | 12.5±10.2 | 9.9 | 12.8±10.1 | 10.5 |
| Rice | 7.9±5.3*** | 7.3 | 1.1±1.6 | 0.5 |
| Starchy vegetables | 7.4±6.8*** | 5.6 | 4.1±4.8 | 2.9 |
| Legumes | 3.3±2.8*** | 2.7 | 1.8±3.1 | 0.9 |
| Other vegetables | 9.6±8.4** | 7.8 | 11.9±8.9 | 9.3 |
| White meats | 4.9±3.7 | 4.0 | 4.2±3.3 | 3.4 |
| Red meats | 2.1±2.6*** | 1.3 | 3.7±3.2 | 2.9 |
| Breads | 6.0±5.7*** | 5.0 | 11.2±9.2 | 8.1 |
| Cereals | 6.5±7.9** | 3.9 | 8.8±8.6 | 7.0 |
| Coffee and tea | 13.1±1.7*** | 10.5 | 23.6±17.32,3 | 21.7 |
Values are means±SD or medians. Asterisks indicate different from non-Hispanic whites, t test,
P < 0.05;
P < 0.01;
P < 0.001.
Excluding one subject with value >100 servings/wk.
Excluding one subject with value >600 servings/wk.
Hispanics had a 43% lower mean total dietary iron intake than non-Hispanic whites (Table 3). When total dietary iron was broken into heme and nonheme iron, estimated intake of heme iron from meat was the same in the 2 groups (0.7 mg heme iron/d); however, the Hispanic group consumed less heme from red meat and more from white meat (P < 0.001).
TABLE 3.
Mean and median daily intake for total nutrients and dietary components by Hispanics and non-Hispanic whites1
| Nutrient intake, mg | Hispanics, n = 451 |
Non-Hispanic whites, n = 153 |
||
|---|---|---|---|---|
| Mean±SD | Median | Mean±SD | Median | |
| Total iron, mg | 15.8±22.0* | 11.6 | 27.7±64.4 | 14.5 |
| Heme iron from red meat, mg | 0.3±0.3*** | 0.2 | 0.4±0.3 | 0.3 |
| Heme iron from white meat, mg | 0.4±0.3*** | 0.4 | 0.3±0.2 | 0.3 |
| Nonheme iron, mg | 10.9±4.6*** | 10.1 | 14.4±9.8 | 11.5 |
| Supplemental iron,2 mg | 28.0±49.4 | 18.0 | 51.8±123 | 18.0 |
| Total vitamin C, mg | 167±203** | 121 | 242±332 | 133 |
| Dietary vitamin C, mg | 126±77.1 | 113 | 130±129 | 111 |
| Supplemental vitamin C,2 mg | 187±356.5* | 60.0 | 324±437 | 60.0 |
| Total calcium, mg | 742±456** | 661 | 855±451 | 789 |
| Dietary calcium, mg | 692±382* | 618 | 765±364 | 672 |
| Supplemental calcium,2 mg | 273±540 | 162 | 251±249 | 162 |
| Dietary fiber, g | 16.9±7.7 | 15.5 | 18.1±9.4 | 15.9 |
| Caffeine, mg | 182±172** | 140 | 374±832 | 281 |
Values are means±SD or medians. Asterisks indicate different from non-Hispanic whites, t test,
P < 0.05;
P < 0.01;
P < 0.001.
Supplement users only:14.9% Hispanics and 24.4% non-Hispanic whites used supplemental iron; 21.6% Hispanics and 34.6% non-Hispanic whites used supplemental vitamin C; 18.5% Hispanics and 35.9% non-Hispanic whites used supplemental calcium.
As iron bioavailability is known to be inhibited by calcium and enhanced by vitamin C intake, it was of interest to examine the dietary intake of these 2 nutrients. Total intake of vitamin C was 31% lower for Hispanics than non-Hispanic whites (P < 0.001). When supplemental intake was excluded, however, mean intake of vitamin C from food was similar across groups. Mean calcium intake was 13% lower for Hispanics than non-Hispanic whites (P < 0.01), with 19% of Hispanics and 36% of non-Hispanic whites using calcium supplements. Mean dietary calcium intake, excluding supplements, was also lower among Hispanics than non-Hispanic whites (P < 0.05) despite the finding that Hispanics had higher milk intake than non-Hispanic whites. Mean daily caffeine intake was lower in Hispanics than in non-Hispanic whites (P < 0.01). Intake of dietary fiber was similar in both ethnic groups.
Biological markers of iron status
Hispanic older adults had lower mean hemoglobin concentrations (P < 0.01), and higher prevalence of suboptimal hemoglobin concentrations (P < 0.05) than non-Hispanic whites (Table 4). When examined by sex, these differences appeared to be due mainly to particularly low hemoglobin status in Hispanic women. Greater than 28% of Hispanic women had suboptimal hemoglobin compared with 12% of men. Microcytosis (MCV < 80 μm3) was ~ 10 times more prevalent in the Hispanic group than in the non-Hispanic whites (P < 0.001). Again, this was more apparent among Hispanic women. More than 11% of Hispanic women had microcytosis compared with <1% of non-Hispanic white women. Macrocytosis (MCV > 90 μm3), an indicator of folate or vitamin B-12 deficiency, was not present in either group. The geometric mean serum ferritin concentration in Hispanics was significantly lower than that of non-Hispanic whites; however, the prevalence of elevated serum ferritin did not differ significantly between groups. Results for serum ferritin were similar when we excluded subjects with inflammation [geometric mean 73.6 μg/L in Hispanics (n = 366) vs. 95.5 μg/L in non-Hispanic whites (n = 132); P < 0.05] and prevalence of elevated ferritin (8.3% in Hispanics vs. 12.1% in non-Hispanic whites; P > 0.05). There was a trend toward higher prevalence of IDA in Hispanics than in non-Hispanic whites (P = 0.08). This was significant (P < 0.05) for Hispanic women, >7% of whom had IDA, compared with 2% of non-Hispanic white women.
TABLE 4.
Biological markers of iron status and prevalence of abnormal values by sex and by Hispanic or non-Hispanic origins1
| Total |
Men |
Women |
||||
|---|---|---|---|---|---|---|
| Blood values | Hispanic | Non-Hispanic whites | Hispanic | Non-Hispanic whites | Hispanic | Non-Hispanic whites |
| n | 443 | 153 | 184 | 61 | 259 | 90 |
| Serum ferritin,1,2 μg/L | 74.1 ± 6.3*** | 100 ± 7.2 | 85.1 ± 2.6 | 107 ± 2.6 | 69.2 ± 2.5** | 95.5 ± 2.1 |
| Hemoglobin,1 g/L | 137 ± 13** | 140 ± 12 | 143 ± 15 | 145 ± 13 | 130 ± 11*** | 136 ± 12 |
| Low, % | 11.5* | 7.3 | 12.0 | 8.2 | 11.2* | 6.7 |
| Suboptimal,3% | 21.4* | 13.3 | 12.0 | 8.2 | 28.2* | 16.7 |
| Hematocrit1 | 0.40 ± 0.04* | 0.41 ± 0.04 | 0.42 ± 0.05 | 0.42 ± 0.04 | 0.39 ± 0.03** | 0.40 ± 0.04 |
| Low hematocrit, % | 27.5 | 20.5 | 26.1 | 21.3 | 28.6 | 20.0 |
| MCV <80 μm/cell, % | 6.6*** | 0.7 | 6.0 | 3.3 | 11.2*** | 0.1 |
| IDA,4 % | 6.5 | 3.3 | 6.4 | 4.9 | 7.2* | 2.3 |
Values are means ± SD, adjusted for age, sex, and inflammation (CRP ≥10.0 mg/L). Asterisks indicate different from non-Hispanic whites,
P < 0.05;
P < 0.01;
P < 0.001.
Elevated serum ferritin: > 200 μg/L for women and > 300 μg/L for men.
Suboptimal hemoglobin concentrations (<125 g/L).
Serum ferritin <15 μg/L or hemoglobin <119 g/L (women) or <124 g/L (men) and MCV <80 μm3 per cell.
Determinants of serum ferritin in Hispanic and non-Hispanic white older adults
The relation between serum ferritin and ethnicity was assessed by adjusting sequentially for age and sex, nondietary factors, dietary factors, expanded nutrient factors, and food and beverage factors. Hispanics consistently had significantly lower serum ferritin than non-Hispanic whites after adjusting for dietary variables. The inclusion of food and beverages reduced the significance of ethnicity to a greater extent than did nutrient intake variables.
After controlling for age, sex, BMI, alcohol intake, smoking status, energy intake, inflammation, diabetes, and liver disease, neither total iron intake nor intake of other nutrients tested were significantly associated with serum ferritin among Hispanics (Table 5). However, when total iron, vitamin C, and calcium were expanded into their component dietary sources, intakes of heme iron derived from red meat and caffeine were positively associated, and dietary calcium negatively associated with serum ferritin. Each mg of heme iron consumed from red meat was associated with a 46% greater log serum ferritin concentration (P < 0.02). Every 100 mg of caffeine consumed was associated with a 5% greater log serum ferritin (P < 0.04), and each 100 mg of dietary calcium consumed was associated with a 4% lower log serum ferritin concentration (P < 0.01).
TABLE 5.
Association of dietary factors and serum ferritin among Hispanics only1
| Nutrient intakes, unit/d | Serum ferritin2 | P-value |
|---|---|---|
| Total iron, mg | -0.0002 ± 0.0023 | 0.92 |
| Total vitamin C, mg | 0.0001 ± 0.0003 | 0.64 |
| Total calcium, mg | -0.0002 ± 0.0001 | 0.12 |
| Dietary fiber, g | 0.0006 ± 0.009 | 0.94 |
| Caffeine, mg | 0.0005 ± 0.0003 | 0.08 |
| Expanded nutrient intakes | ||
| Heme iron from red meat, mg | 0.46 ± 0.20 | 0.02 |
| Heme iron from white meat, mg | -0.28 ± 0.18 | 0.12 |
| Nonheme iron, mg | 0.0038 ± 0.021 | 0.85 |
| Supplemental iron,5 mg | 0.077 ± 0.19 | 0.68 |
| Dietary vitamin C, mg | 0.0001 ± 0.0007 | 0.86 |
| Supplemental vitamin C,5 mg | -0.072 ± 0.16 | 0.65 |
| Dietary calcium, mg | -0.0004 ± 0.0002 | 0.01 |
| Supplemental calcium,3 mg | 0.13 ± 0.16 | 0.41 |
| Dietary fiber, g | -0.0028 ± 0.012 | 0.81 |
| Caffeine, mg | 0.00054 ± 0.00026 | 0.04 |
Values are β-coefficient ± SE. Intakes were adjusted for age, sex, BMI (kg/m2), alcohol intake (none/light /moderate/heavy), smoking status (current yes/no), total energy intake/d (kJ), inflammation (yes/no), diabetes (yes/no), liver disease (yes/no). R2 = 0.091 (nutrient intakes); R2 = 0.12 (expanded nutrient intakes).
Serum ferritin (μg/L) was log transformed.
Supplement use (yes/no)
Finally, we assessed the association between food components and serum ferritin among Hispanics. After adjusting for age, sex, BMI, alcohol intake, smoking status, total energy intake, inflammation, diabetes, liver disease, and supplemental use of iron and calcium, milk intake was negatively associated, and red meat positively associated, with serum ferritin concentrations (Table 6). Each serving of milk was associated with a 2% lower log serum ferritin concentration, whereas each serving of red meat was associated with a 5% greater log serum ferritin concentration.
TABLE 6.
Food components of diet (servings/wk) and serum ferritin among Hispanics only1
| Food variable,2,3servings/wk | Regression coefficient ± SE | P-value |
|---|---|---|
| Milk | -0.017 ± 0.0072 | 0.02 |
| Fruit | -0.0063 ± 0.0053 | 0.23 |
| Rice | 0.0021 ± 0.013 | 0.87 |
| Starchy vegetables | -0.0031 ± 0.0075 | 0.69 |
| Legumes | -0.020 ± 0.019 | 0.28 |
| Other vegetables | 0.0084 ± 0.0059 | 0.15 |
| White meat | -0.016 ± 0.014 | 0.25 |
| Red meat | 0.049 ± 0.021 | 0.02 |
| Bread | -0.016 ± 0.0095 | 0.10 |
| Cereal | -0.0073 ± 0.0065 | 0.26 |
| Coffee and tea | 0.0060 ± 0.0039 | 0.13 |
Serum ferritin (μg/L) was log transformed.
Adjusted for age, sex, BMI (kg/m2), alcohol intake (none/light /moderate/heavy), smoking status (current yes/no), total energy intake (kJ), inflammation (CRP ≥10.0 mg/L yes/no), diabetes (yes/no), liver disease (yes/no).
R2 = 0.13.
Discussion
We have shown previously that Hispanic older adults living in the northeastern U.S. had functional limitations in excess of a non-Hispanic white comparison group (26). Further research is needed to determine whether low iron status is a contributing factor. To our knowledge, this is the first study to assess iron status and dietary determinants of iron status in Hispanic older adults of Caribbean origin. This representative group of older Hispanics in Massachusetts was more likely to be anemic than non-Hispanic whites living in the same neighborhood (6.5% vs. 3.3%). The Hispanics in our sample had a 60% higher prevalence of suboptimal hemoglobin (<125 g/L) than their neighborhood non-Hispanic white counterparts. These findings are notable because of the possible etiologic implications for the excess burden of chronic disease within this population. A pre-ponderance of recent research has indicated that current WHO criteria for classification of anemia (<119 g/L for women and <124 g/L for men) are not clinically optimal in an elderly population and should be revised (38,39). Studies have shown that individuals with hemoglobin concentrations in the low to normal range of ≤125 g/L were as much as 1.5 times as likely to have difficulty performing daily tasks, whereas women with higher hemoglobin concentrations performed significantly better on tests of mobility (38,43). This notion is further supported by the Women’s Health and Aging Studies I and II that demonstrated the prevalence of mobility difficulties was inversely related to hemoglobin concentration even within a hemoglobin range considered normal.
In contrast to the present findings in Hispanic older adults, we previously showed that IDA was uncommon (1.2% prevalence) in a free-living, white, elderly Framingham Heart Study cohort (12). Interestingly, although significantly lower than the Hispanic population (6.5%), the prevalence of anemia for our generally low-income non-Hispanic white comparison population (3.3%) was also higher than that of the Framingham older adults. The reason for these ethnic and socio-economic level differences in iron status is unknown, but may reflect, at least in part, differences in iron intake and the contribution of dietary factors that are known to influence iron bioavailability.
We found significant ethnic differences in nutrient intake and intake of specific food items. Despite living in the same neighborhoods, these Caribbean Hispanics differed from non-Hispanic whites in both demographics and dietary intake patterns. Bartholomew et al. (44) showed that ethnicity and therefore ethnically related food choices may affect the health status of a population. We have previously reported that the intake of heme, supplemental iron, dietary vitamin C, and ethanol were significant positive predictors of iron stores (serum ferritin) in white, elderly persons in the Framingham Heart Study cohort (13). We also found that heme intake was positively associated with serum ferritin among these elderly Hispanics. The relation of heme iron and meat, specifically red meat, has consistently been observed in studies across varying populations (13,45,46). The association of meat with serum ferritin has been attributed to the high concentration of highly bioavailable heme iron found exclusively in animal products (13). In addition, the presence of red meat in the diet may enhance dietary iron absorption from nonheme sources. A study examining iron absorption from Latin American diets indicated that variation in both heme and nonheme iron absorption among Hispanics was due in large part to a variation in meat intake (23). Hallberg et al. (24) showed that the addition of xsmeat to a typical cereal-based Latin American meal effectively elevated rates of iron absorption. Our current findings of elderly Hispanics living in Massachusetts support the importance of red meat intake as an important determinant of body iron stores (47). Vitamin C can promote the absorption of iron and we have previously reported that dietary vitamin C intake was positively associated with iron stores in the elderly Framingham Heart Study cohort (13). Total vitamin C intake was about one-third lower in our Hispanic compared with non-Hispanic older adults and may have been an additional factor that contributed to lower iron stores in this group.
Ethanol consumption is associated with higher serum ferritin concentrations in the elderly (18). Hispanics in our study were much less likely to consume significant amounts of alcohol, which may have influenced their iron stores. On the other hand, some dietary factors have an inhibitory effect on iron bioavailability. It has been shown that calcium given in the form of cheese, milk, and calcium chloride reduces iron absorption (47,48). Calcium can interfere with absorption of both heme and nonheme iron (47,49), although the negative influence of calcium can be less significant in mixed diets (48,50). We found that dietary calcium intake and milk intake had a negative association with serum ferritin, supporting the potential practical effect of habitual dietary calcium on iron bioavailability and iron status. The higher milk intake noted in Hispanics compared with non-Hispanic whites may contribute to lower body iron stores in this group.
This study examined dietary intake and iron status in a representative group of Hispanic older adults residing in Massachusetts and in a non-Hispanic white neighborhood comparison group. Unlike other regions of the country where Mexican Americans are more prevalent, Puerto Ricans and Dominicans make up the majority of the Hispanic population in the Northeast. The differing cultural backgrounds and dietary patterns of these Hispanic groups may lead to group-specific effects on iron status. Therefore, these results may not be generalized to Mexican Americans or other Hispanic groups.
Our results demonstrate that Hispanic older adults in Massachusetts, particularly Hispanic women, are more likely than their non-Hispanic white neighbors to suffer from IDA. The cause of the relatively high rates of IDA in this Hispanic elderly population is uncertain. However, differences in dietary patterns are evident and may play an important role in the determination of iron status and the increased risk for anemia in these Hispanic older adults. Such dietary patterns could be modified to help alleviate ethnic differences in nutritional status.
Acknowledgments
We thank Janice Maras and Peter Bakun at the HNRCA for their assistance with the dietary data and Gayle Petty of the Nutrition Evaluation Laboratory at the HNRCA for laboratory analyses.
Footnotes
Supported by the USDA, Agricultural Research Service, under agreement 58-1950-9-001 and by NIH/NIA grants AG10425 and AG023394.
- CRP
- C-reactive protein
- IDA
- iron deficiency anemia
- MCV
- mean cell volume
- WBC
- white blood cell count.
Literature Cited
- 1.Salonen JT, Nyyssonen K, Korpela H, Tuomilehto J, Seppanen R, Salonen R. High stored iron levels are associated with excess risk of myocardial infarction in eastern Finnish men. Circulation. 1992;86:803–11. doi: 10.1161/01.cir.86.3.803. [DOI] [PubMed] [Google Scholar]
- 2.Stevens RG, Jones DY, Micozzi MS, Taylor PR. Body iron stores and the risk of cancer. N Engl J Med. 1988;319:1047–52. doi: 10.1056/NEJM198810203191603. [DOI] [PubMed] [Google Scholar]
- 3.Tuomainen TP, Punnonen K, Nyyssonen K, Salonen JT. Association between body iron stores and the risk of acute myocardial infarction in men. Circulation. 1998;97:1461–6. doi: 10.1161/01.cir.97.15.1461. [DOI] [PubMed] [Google Scholar]
- 4.Tuomainen TP, Nyyssonen K, Salonen R, Tervahauta A, Korpela H, Lakka T, Kaplan GA, Salonen JT. Body iron stores are associated with serum insulin and blood glucose concentrations. Population study in 1,013 eastern Finnish men. Diabetes Care. 1997;20:426–8. doi: 10.2337/diacare.20.3.426. [DOI] [PubMed] [Google Scholar]
- 5.Salonen JT, Tuomainen TP, Nyyssonen K, Lakka HM, Punnonen K. Relation between iron stores and non-insulin dependent diabetes in men: case-control study. BMJ. 1998;317:727. doi: 10.1136/bmj.317.7160.727. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Olivares M, Hertrampf E, Capurro MT, Wegner D. Prevalence of anemia in elderly subjects living at home: role of micronutrient deficiency and inflammation. Eur J Clin Nutr. 2000;54:834–9. doi: 10.1038/sj.ejcn.1601099. [DOI] [PubMed] [Google Scholar]
- 7.Izaks G, Westendorp RG, Knook DL. The definition of anemia in older persons. JAMA. 1999;281:1714–17. doi: 10.1001/jama.281.18.1714. [DOI] [PubMed] [Google Scholar]
- 8.Penninx BW, Pahor M, Cesari M, Corsi AM, Woodman RC, Bandinelli S, Guralnik JM, Ferrucci L. Anemia is associated with disability and decreased physical performance and muscle strength in the elderly. J Am Geriatr Soc. 2004;52:719–24. doi: 10.1111/j.1532-5415.2004.52208.x. [DOI] [PubMed] [Google Scholar]
- 9.Sheth TN, Choudhry NK, Bowes M, Detsky AS. The relation of conjunctival pallor to the presence of anemia. J Gen Intern Med. 1997;12:102–6. doi: 10.1046/j.1525-1497.1997.00014.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Woodman R, Ferrucci L, Guralnik J. Anemia in older adults. Curr Opin Hematol. 2005;12:123–8. doi: 10.1097/01.moh.0000154030.13020.85. [DOI] [PubMed] [Google Scholar]
- 11.American Heart Association [cited Dec 2005]; Available from: www.americanheart.org.
- 12.Fleming DJ, Jacques PF, Tucker KL, Massaro JM, D’Agostino RB, Sr., Wilson PW, Wood RJ. Iron status of the free-living, elderly Framingham Heart Study cohort: an iron-replete population with a high prevalence of elevated iron stores. Am J Clin Nutr. 2001;73:638–46. doi: 10.1093/ajcn/73.3.638. [DOI] [PubMed] [Google Scholar]
- 13.Fleming DJ, Jacques PF, Dallal GE, Tucker KL, Wilson PW, Wood RJ. Dietary determinants of iron stores in a free-living elderly population: the Framingham heart study. Am J Clin Nutr. 1998;67:722–33. doi: 10.1093/ajcn/67.4.722. [DOI] [PubMed] [Google Scholar]
- 14.Payette H, Gray-Donald K. Dietary intake and biochemical indices of nutritional status in an elderly population, with estimates of the precision of the 7-d food record. Am J Clin Nutr. 1991;54:478–88. doi: 10.1093/ajcn/54.3.478. [DOI] [PubMed] [Google Scholar]
- 15.Garry PJ, Hunt WC, Baumgartner RN. Effects of iron intake on iron stores in elderly men and women: longitudinal and cross-sectional results. J Am Coll Nutr. 2000;19:262–9. doi: 10.1080/07315724.2000.10718925. [DOI] [PubMed] [Google Scholar]
- 16.Preziosi P, Hercberg S, Galan P, Devanlay M, Cherouvrier F, Dupin H. Iron status of a healthy French population: factors determining biochemical markers. Ann Nutr Metab. 1994;38:192–202. doi: 10.1159/000177811. [DOI] [PubMed] [Google Scholar]
- 17.Cook JD, Reddy MB. Effect of ascorbic acid intake on nonhemeiron absorption from a complete diet. Am J Clin Nutr. 2001;73:93–8. doi: 10.1093/ajcn/73.1.93. [DOI] [PubMed] [Google Scholar]
- 18.Fleming DJ, Tucker KL, Jacques PF, Dallal GE, Wilson PW, Wood RJ. Dietary factors associated with the risk of high iron stores in the elderly Framingham heart study cohort. Am J Clin Nutr. 2002;76:1375–84. doi: 10.1093/ajcn/76.6.1375. [DOI] [PubMed] [Google Scholar]
- 19.Hajat A, Lucas JB, Kington R. Health outcomes among Hispanic subgroups: data from the national health interview survey, 1992-95. Adv Data. 2000;310:1–14. [PubMed] [Google Scholar]
- 20.Rosado JL, Lopez P, Morales M, Munoz E, Allen LH. Bioavailability of energy, nitrogen, fat, zinc, iron and calcium from rural and urban Mexican diets. Br J Nutr. 1992;68:45–58. doi: 10.1079/bjn19920065. [DOI] [PubMed] [Google Scholar]
- 21.Bermudez OI, Falcon LM, Tucker KL. Intake and food sources of macronutrients among older Hispanic adults: association with ethnicity, acculturation, and length of residence in the United States. J Am Diet Assoc. 2000;100:665–73. doi: 10.1016/s0002-8223(00)00195-4. [DOI] [PubMed] [Google Scholar]
- 22.Muir A, Hopfer U. Regional specificity of iron uptake by small intestinal brush-border membranes from normal and iron-deficient mice. Am J Physiol. 1985;248:G376–9. doi: 10.1152/ajpgi.1985.248.3.G376. [DOI] [PubMed] [Google Scholar]
- 23.Acosta A, Amar M, Cornbluth-Szarfarc SC, Dillman E, Fosil M, Biachi RG, Grebe G, Hertrampf E, Kremenchuzky S, et al. Iron absorption from typical Latin American diets. Am J Clin Nutr. 1984;39:953–62. doi: 10.1093/ajcn/39.6.953. [DOI] [PubMed] [Google Scholar]
- 24.Hallberg L, Rossander L. Improvement of iron nutrition in developing countries: comparison of adding meat, soy protein, ascorbic acid, citric acid, and ferrous sulphate on iron absorption from a simple Latin American-type of meal. Am J Clin Nutr. 1984;39:577–83. doi: 10.1093/ajcn/39.4.577. [DOI] [PubMed] [Google Scholar]
- 25.Taylor PG, Mendez-Castellanos H, Martinez-Torres C, Jaffe W, Lopez de Blanco M, Landaeta-Jimenez M, Leets I, Tropper E, Ramirez J, et al. Iron bioavailability from diets consumed by different socioeconomic strata of the Venezuelan population. J Nutr. 1995;125:1860–8. doi: 10.1093/jn/125.7.1860. [DOI] [PubMed] [Google Scholar]
- 26.Tucker KL, Falcon LM, Bianchi LA, Cacho E, Bermudez OI. Self-reported prevalence and health correlates of functional limitation among Massachusetts elderly Puerto Ricans, Dominicans, and non-Hispanic white neighborhood comparison group. J Gerontol A Biol Sci Med Sci. 2000;55:M90–7. doi: 10.1093/gerona/55.2.m90. [DOI] [PubMed] [Google Scholar]
- 27.Tietz N. Clinical guide to laboratory tests. Harcourt Brace Company; Philadelphia: 1995. [Google Scholar]
- 28.Cook JD, Skikne BS, Lynch SR, Reusser ME. Estimates of iron sufficiency in the US population. Blood. 1986;68:726–31. [PubMed] [Google Scholar]
- 29.Johnson MA, Fischer JG, Bowman BA, Gunter EW. Iron nutriture in elderly individuals. FASEB J. 1994;8:609–21. doi: 10.1096/fasebj.8.9.8005389. [DOI] [PubMed] [Google Scholar]
- 30.Qvist I, Norden A, Olofsson T. Serum ferritin in the elderly. Scand J Clin Lab Invest. 1980;40:609–13. doi: 10.3109/00365518009091971. [DOI] [PubMed] [Google Scholar]
- 31.Cook JD, Lipschitz DA, Miles LE, Finch CA. Serum ferritin as a measure of iron stores in normal subjects. Am J Clin Nutr. 1974;27:681–7. doi: 10.1093/ajcn/27.7.681. [DOI] [PubMed] [Google Scholar]
- 32.Worwood M. Serum ferritin. Clin Sci (Lond) 1986;70:215–20. doi: 10.1042/cs0700215. [DOI] [PubMed] [Google Scholar]
- 33.Edwards C, Griffen LM, Goldgar D, Drummond C, Skolnick MH, Kushner JP. Prevalence of hemochromatosis among 11,065 presumably healthy blood donors. N Engl J Med. 1988;318:1355–62. doi: 10.1056/NEJM198805263182103. [DOI] [PubMed] [Google Scholar]
- 34.Holyoake TL, Stott DJ, McKay PJ, Hendry A, MacDonald JB, Lucie NP. Use of plasma ferritin concentration to diagnose iron deficiency in elderly patients. J Clin Pathol. 1993;46:857–60. doi: 10.1136/jcp.46.9.857. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Guyatt GH, Patterson C, Ali M, Singer J, Levine M, Turpie I, Meyer R. Diagnosis of iron-deficiency anemia in the elderly. Am J Med. 1990;88:205–9. doi: 10.1016/0002-9343(90)90143-2. [DOI] [PubMed] [Google Scholar]
- 36.World Health Organization . Nutritional anemias. World Health Organization; Geneva: 1972. [Google Scholar]
- 37.Khusun H, Yip R, Schultink W, Dillon DHS. World Health Organization hemoglobin cut-off points for the detection of anemia are valid for an Indonesian population. J Nutr. 1999;129:1669–74. doi: 10.1093/jn/129.9.1669. [DOI] [PubMed] [Google Scholar]
- 38.Chaves PH, Ashar B, Guralnik JM, Fried LP. Looking at the relationship between hemoglobin concentration and prevalent mobility difficulty in older women. Should the criteria currently used to define anemia in older people be reevaluated? J Am Geriatr Soc. 2002;50:1257–64. doi: 10.1046/j.1532-5415.2002.50313.x. [DOI] [PubMed] [Google Scholar]
- 39.Cesari M, Penninx B, Woodman R. Hemoglobin levels and skeletal muscle: results from the In CHIANTI study. J Gerontol A Biol Sci Med Sci. 2004;59:249–54. doi: 10.1093/gerona/59.3.m249. [DOI] [PubMed] [Google Scholar]
- 40.Smith DL. Anemia in the elderly. Am Fam Physician. 2000. 2000;62:1565–72. [PubMed] [Google Scholar]
- 41.Willett WC. Nutrition and chronic disease. Public Health Rev. 1998;26:9–10. [PubMed] [Google Scholar]
- 42.Whitfield JB, Zhu G, Heath AC, Powell LW, Martin NG. Effects of alcohol consumption on indices of iron stores and of iron stores on alcohol intake markers. Alcohol Clin Exp Res. 2001;25:1037–45. [PubMed] [Google Scholar]
- 43.Chaves P, Ashar B, Guralnik J. What constitutes normal hemoglobin concentration in community-dwelling disabled older women? J Am Geriatr Soc. 2004;52:1811–16. doi: 10.1111/j.1532-5415.2004.52502.x. [DOI] [PubMed] [Google Scholar]
- 44.Bartholomew AM, Young EA, Martin HW, Hazuda HP. Food frequency intakes and sociodemographic factors of elderly Mexican Americans and non-Hispanic whites. J Am Diet Assoc. 1990;90:1693–6. [PubMed] [Google Scholar]
- 45.Bothwell TH. Overview and mechanisms of iron regulation. Nutr Rev. 1995;53:237–45. doi: 10.1111/j.1753-4887.1995.tb05480.x. [DOI] [PubMed] [Google Scholar]
- 46.Brussaard JH, Brants HA, Bouman M, Lowik MR. Iron intake and iron status among adults in the Netherlands. Eur J Clin Nutr. 1997;51(Suppl 3):S51–8. [PubMed] [Google Scholar]
- 47.Hallberg L, Brune M, Erlandsson M, Sandberg AS, Rossander-Hulten L. Calcium: effect of different amounts on nonheme- and heme-iron absorption in humans. Am J Clin Nutr. 1991;53:112–9. doi: 10.1093/ajcn/53.1.112. [DOI] [PubMed] [Google Scholar]
- 48.Cook JD, Dassenko SA, Lynch SR. Assessment of the role of nonheme-iron availability in iron balance. Am J Clin Nutr. 1991;54:717–22. doi: 10.1093/ajcn/54.4.717. [DOI] [PubMed] [Google Scholar]
- 49.Hallberg L, Rossander-Hulten L, Brune M, Gleerup A. Calcium and iron absorption: mechanism of action and nutritional importance. Eur J Clin Nutr. 1992;46:317–27. [PubMed] [Google Scholar]
- 50.Whiting SJ. The inhibitory effect of dietary calcium on iron bioavailability: a cause for concern? Nutr Rev. 1995;53:77–80. doi: 10.1111/j.1753-4887.1995.tb01507.x. [DOI] [PubMed] [Google Scholar]