点击显示 收起
1 From the Department of Nutrition and Food Science and the Biometrics Program, Department of Animal and Avian Sciences, University of Maryland, College Park.
2 Supported by the University of Maryland and the Centers for Disease Control and Prevention (purchase order 0990-0115). 3 Address reprint requests to PB Moser-Veillon, Department of Nutrition and Food Science, University of Maryland, College Park, MD 20742. 4 Address correspondence to PR Pehrsson, Nutrient Data Lab, ARS/USDA, 10300 Baltimore Boulevard, Building 005, Room 309A, Beltsville, MD 20705. E-mail: ppehrsson{at}rbhnrc.usda.gov.
ABSTRACT
Background: Iron deficiency, a pervasive problem among low-income women of childbearing age, threatens maternal health and pregnancy outcomes. The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) was designed to alleviate health problems and provides supplemental nutritious foods, nutrition education, and health care referrals.
Objectives: The aim of this study was to examine the benefits associated with participation in WIC in terms of biochemical tests of postpartum iron status in nonlactating women.
Design: WIC participants (n = 57) and eligible nonparticipants (n = 53), matched by race and age, were followed bimonthly over 6 mo postpartum. Finger stick blood samples (500 µL) were collected for measurement of plasma ferritin, transferrin receptor (TfR), and hemoglobin (Hb).
Results: The mean (±SE) Hb concentration of participants exceeded that of nonparticipants from months 2 through 6. At 6 mo, the mean Hb concentration of participants was significantly higher than that of nonparticipants (8.01 ± 0.12 and 7.63 ± 0.12 mmol/L, respectively; P < 0.05) and the prevalence of anemia was significantly lower (17% and 51%, respectively; P < 0.05). TfR and ferritin concentrations (consistently within the reference ranges) and dietary iron intakes did not differ significantly between participants and nonparticipants and were not correlated with Hb concentrations.
Conclusions: Our results suggest that WIC participants were significantly less likely to become anemic if uninterrupted postpartum participation lasted for 6 mo. The lack of correlation among iron status indicators suggests that the lower mean Hb concentration in nonparticipants at 6 mo may not have been related to improved iron status in participants but to other nutrient deficiencies or differences in access to health care and health and nutrition education.
Key Words: Hemoglobin • transferrin receptor • ferritin • anemia • women • WIC • Special Supplemental Nutrition Program for Women • Infants • Children
INTRODUCTION
The health of a woman during her reproductive years dramatically influences her health later in life, her fertility, her recovery from pregnancy and childbirth, and the health of the children she bears. Although many women are in a nutritionally depleted state before and after pregnancy (1), maternal and child health services in our society focus specifically on fetal health (2). This is a serious problem among many lower-income women, who are especially prone to health problems such as nutritional deficiencies and obesity. The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), a federally funded program, was instituted to alleviate some of the most prevalent nutritional problems. WIC serves 7 million people, primarily pregnant women and infants, through distribution of nutritious supplemental food, nutrition education, and assistance with health care referrals.
A priority system was developed to identify the most critical applicants who met established poverty income and nutritional risk criteria (3); postpartum nonlactating women fall sixth in the 7-point priority system, behind pregnant and lactating women, infants, and children. One of the most pervasive nutritional risks in this population is iron deficiency.
Iron shortages are associated with adverse changes in many components of the immune system, including resistance to infection (4, 5) and changes in the function of the immune system, but the exact consequences have not yet been determined (6). Although there is no conclusive evidence that iron deficiency directly causes a higher incidence of infections, iron deficiency anemia and infections are prevalent in low-income populations (7). Anemia, usually confirmed by a subnormal hemoglobin (Hb) concentration or hematocrit (for age and sex), may be caused by infection and inflammation; laboratory changes are measurable even when the illness or infection is mild (8). Under these conditions, the release of stored iron is blocked, often resulting in abnormally elevated ferritin and impaired delivery of iron to bone marrow for synthesis of heme (iron-deficient erythropoiesis) (9). However, in some cases iron stores may remain unchanged even if ferritin is elevated, and a misleading assessment of iron status may thus occur.
Numerous studies showed a positive effect of WIC participation on the iron status of pregnant and lactating women and on the pregnancy outcomes of low-income women (10–14). However, evidence of the effect of participation on the iron status of postpartum nonlactating mothers is lacking. The 1999 Review of the Nutritional Status of WIC Participants reported that nonlactating postpartum participants continued to report low energy and iron intakes [30% and 84% of the recommended dietary allowance (RDA), respectively] (15). This research was designed to assess the effect of WIC participation on the postpartum iron status of nonlactating women through measurement of Hb, transferrin receptor (TfR), serum ferritin, and iron intake over the 6 mo after delivery.
SUBJECTS AND METHODS
The Maryland WIC program served 89673 clients in financial year 1995 at a cost of $44 million (16). At the time of the study (1994–1995), major cutbacks in federal funding required that the Maryland WIC program admit postpartum women in selected counties only if they were of high priority (eg, anemic). Risk profiles and criteria for WIC eligibility were consistent among counties, enabling recruitment of the control group for this study. The study was approved by the Human Subjects Review Boards of the State of Maryland Department of Health and Mental Hygiene and the University of Maryland. Gift certificates to a nationally recognized children's store were awarded to the subjects on completion of each visit.
On the basis of existing acceptance policies for postpartum nonlactating applicants, 110 WIC-eligible women were recruited for this study from Maryland WIC agencies. The study group consisted of 57 WIC participants recruited from select Baltimore City sites, which accommodated all nonlactating applicants (Figure 1). The control group consisted of 52 WIC-eligible nonparticipants recruited from Prince Georges County and 1 nonparticipant recruited from Montgomery County. Postpartum risk assessment for low-risk applicants in Prince Georges and Montgomery County sites was based on the applicants' risk codes during pregnancy. Because most English-speaking WIC participants (both counties) were African American, only white and African American women were recruited.
FIGURE 1. . Study design. Personal history includes information on breast-feeding status; smoking status; maternal age; age at onset of menstruation; race; date of first prenatal visit; delivery weight; gestational duration and delivery date; number and ages of other children; number and dates of stillbirths, abortions, and miscarriages; participation in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) during pregnancy; participation in the Food Stamp Program; and use of vitamin and mineral supplements.
All subjects were free-living in an urban geographic location. Eligibility criteria for inclusion in the study or control group were 1) delivery of a full-term infant within the previous month, 2) absence of breast-feeding, 3) freedom from major medical problems likely to affect the results (eg, heart disease and diabetes), and 4) qualification as a low-risk nonlactating postpartum applicant or lack of adequate screening at the point of postpartum recertification but previous certification (during pregnancy) with low-risk criteria codes. Women aged <19 y were excluded because of the potential for the physiologic effects of growth to confound the results.
The first visit (baseline) was conducted within 30 d of delivery; each subject was interviewed to assess health and reproductive history, past WIC participation, and characteristics of the household. Height and weight were measured for each subject and blood samples were drawn for the determination of Hb, plasma TfR, and ferritin. Except for height, all measurements were repeated at 2, 4, and 6 mo postpartum (second, third, and fourth visits) and timed with recertification (first visit) and voucher pickups (maternal-infant or infant) to maximize compliance (Figure 1). Subjects with only one measurement were eliminated from the analysis.
Much of the demographic and health background information was verified with WIC screening forms that had been completed during WIC recertification by a WIC staff member (Table 1). Self-reported weight at the first prenatal visit and weight at delivery were recorded for each subject. Information on smoking status and use of vitamin and mineral supplements was documented because of the potential effect of these practices on iron status.
View this table:
TABLE 1.. Demographic, health, and reproductive characteristics of nonlactating postpartum WIC participants and nonparticipants within 30 d of delivery1
Hemoglobin, transferrin receptor, and ferritin analysis
Hb was measured in duplicate with a photometer system (HemoCue, Ängelholm, Sweden; 17); the Hb cutoff for anemia classification was 120 g/L or 7.45 mmol/L, which is the generally accepted concentration for nonpregnant women (18–20). The instrument automatically reset itself to zero after each measurement and checked the intensity of the light source and the operation of the photocell. The instrument was calibrated with the control cuvette before each set of measurements.
The precision, accuracy, and reliability of capillary Hb measurements can be complicated by significant within-subject variability, over both time and site to site, leading to potential misclassification of iron status (21). For this reason, duplicate Hb sampling was performed; if duplicates were highly variable or one measurement was in the anemic range and one in the normal range, a third measurement was taken. For measurement of TfR and ferritin, 500 µL whole blood was collected from each subject into EDTA-treated vials from single finger sticks at each of the visits. The samples were stored on ice and transported to the University of Maryland. The blood was centrifuged for 10 min at 680 x g and 4°C and the plasma separated by pipette into 2 labeled vials. Plasma samples were stored at -70°C until analyzed.
Plasma TfR was measured with the enzyme-linked immunosorbent assay (ELISA; Ramco Laboratories, Inc, Houston). A best-fit straight line was generated from the standards by using regression techniques. The cutoff of 8.5 mg/L was used to identify subjects with iron deficient erythropoiesis or stage 2 iron deficiency (22). Standards were measured during the study across 15 standard curves: the 0 and SD of the normal control standard was 5.9 ± 0.6 mg/L (company-specified range: 5.6–10.4 mg/L); the CV was 10.8%. The 0 and SD for the high standards, measured across 18 standard curves, was 11.2 ± 2.3 mg/L (company-specified range: 10.4–14.8 mg/L); the CV was 20.4%.
Plasma ferritin was measured by using an ELISA (Ramco Laboratories, Inc) (19, 21). A cutoff of 12.0 µg/L (23) was used to identify subjects with depleted iron stores or stage 1 iron deficiency. A ferritin standard (Food and Agriculture Organization standard 80/578; World Health Organization International Lab for Biological Standards, Hertfordshire, United Kingdom) was measured in duplicate at concentrations of 9.1 and 91.0 µg/L. The lower concentration measured 9.2 µg/L, or 101% of the true concentration; the higher concentration measured 86.0 µg/L, or 95% of the true concentration.
Dietary intake
Assessment of supportive data on consumption of foods that affect iron status was performed with a low-literacy self-administered food-frequency questionnaire (FFQ) designed for and tested on women participating in WIC (24–26). The FFQ was used to estimate the consumption of foods contributing iron and vitamin C. The US Department of Agriculture (Agricultural Research Service) Nutrient Data Bank was used to support the software for determining nutrient intakes. We assumed that the 1989 RDA (6) for nonpregnant women was appropriate for nonlactating women postpartum, primarily because no other standards have been adequately documented and because the postpartum nonlactating state most closely matches the nonpregnant state.
Statistical analysis
Data were analyzed with use of SAS, version 6.04 (27). Student's t tests were conducted to assess differences between participants and nonparticipants in time intervals between delivery and visits. Student's t and chi-square tests were performed to assess differences at baseline and at each visit between the 2 groups in smoking status, race, use of iron supplements, participation in the Food Stamp Program, WIC participation, gestation, household size, and dietary iron intakes. Analysis of variance was conducted to assess differences in all iron-status indicators when smoking status, participation in the Food Stamp Program, use of iron supplements, and the length of the interpregnancy interval were considered. Because only a few women in either group smoked 10 cigarettes/d, heavy and moderate smokers were combined. Partial correlation coefficients were determined between Hb, TfR, and ferritin measurements; univariate analysis of Hb was conducted to evaluate the normality of the distribution, identify potential outliers, and determine means. To allow for the effect of missed visits on differences in Hb concentrations between the groups, regression determinations of the slopes of the lines for Hb measurements over time were computed for each subject and these slopes were compared. Significant differences in slopes were considered indicative of a group x time interaction. This is comparable with a repeated-measures analysis of variance except it is more powerful in that we assume that time is continuous and that the change in Hb concentration with time is linear. Significance was set at P < 0.05.
RESULTS
Of the 110 subjects who completed the first visit, 94 (85%; 47 participants and 47 nonparticipants) completed at least the first and final visits. There were no significant differences between groups in demographic, health, and reproductive characteristics (Table 1). Eighty-four percent of the participants (48 of 57) had received WIC vouchers during at least the last trimester of their pregnancy; almost all of the nonparticipants (98%) had received WIC vouchers during their pregnancy. Because all but 2 nonparticipating subjects had participated in WIC during pregnancy, the effect of vouchers received during pregnancy on postpartum iron status could not be determined. Because of higher participation rates during pregnancy, nonparticipants received significantly more nutrition education before delivery than did participants (96% compared with 84%; P < 0.05). There were no significant differences in reproductive history (age of onset of menses and gynecologic age) or in contraceptive practices between participants and nonparticipants (Table 1).
Most of the nonparticipants had been terminated from WIC at the point of postpartum recertification because they had been coded as low risk during pregnancy and had had uneventful deliveries with no indication of anemia. Ninety-one percent of the participants were coded as having inadequate diets (eg, deficient in specific nutrients or important food groups), the most common criterion used for recertifying nonlactating applicants (28).
At the first visit, almost half of each group (47% of participants and 42% of nonparticipants) used iron supplements, primarily the remaining prenatal vitamins prescribed during pregnancy (Table 2). These percentages dropped dramatically to 15% of participants and 9% of nonparticipants by 6 mo postpartum (no significant difference). Almost all supplement users were consuming nonprescription multivitamins 6 mo postpartum. There were no significant differences in supplement use between the groups at any point during the study (Table 2).
View this table:
TABLE 2.. Iron supplementation in nonlactating postpartum WIC participants and nonparticipants1
Iron measurements
The mean Hb concentrations of participants and nonparticipants by visit are shown in Table 3. The mean Hb concentration was the same between groups at the initial visit; over time, the mean Hb concentration of participants tended to be greater than that of nonparticipants (P < 0.05). In the 2 groups, the difference increased with time; the maximum difference was significant 6 mo postpartum (8.01 ± 0.12 compared with 7.63 ± 0.12 mmol/L; P < 0.05). As shown in Table 3, both means were above the Centers for Disease Control and Prevention cutoff for anemia of 120 g/L (7.45 mmol/L) (17) at 6 mo; however, a significantly higher percentage of nonparticipants than participants had concentrations below the cutoff (51% and 17%, respectively; P < 0.05; Figure 2).
View this table:
TABLE 3.. Iron status of WIC participants and nonparticipants over the first 6 mo postpartum1
FIGURE 2. . Comparison of the prevalence of postpartum anemia (hemoglobin < 120 g/L or 7.45 mmol/L) in nonlactating participants and nonparticipants in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC). *Significantly different from participants, P < 0.05. Additionally, there was a significant time x group interaction.
One subject had an abnormally high mean Hb concentration of 10.06 mmol/L at 4 mo (6); another subject who had hemorrhaged soon after delivery had an initial mean Hb concentration of 3.66 mmol/L. These values were removed from the analysis but the subjects continued to participate in the study.
Some evidence suggests that Hb concentrations are normally lower in African American than in white persons (29–31). Most of the subjects in this study were African American; Hb data for those subjects were also analyzed separately. Because no universally accepted race-specific cutoffs for anemia have been established (29–31), the cutoff for anemia proposed by the Committee on the Prevention, Detection, and Management of Iron Deficiency Anemia Among US Children and Women of Childbearing Age of 112 g/d (31), or 6.95 mmol/L, was used. The mean Hb concentration for African American subjects in this study tended to be lower (but not significantly lower) than the mean Hb concentration of the white subjects. The percentage of African American WIC participants who were iron deficient tended to be lower (but not significantly so) than that of African American nonparticipants 6 mo postpartum (9% compared with 20%).
In both participants and nonparticipants, the mean Hb concentration was significantly higher in smokers 4 mo postpartum, consistent with established adjustments for Hb concentration (18). This strongly suggests that smoking had a strong influence in the assessment of iron deficiency in the population under study.
No significant differences were found in mean TfR between participants and nonparticipants through 6 mo postpartum. The mean TfR concentration for both groups did not change significantly over time (Table 3) and remained well within the established reference range of 3.0–8.5 mg/L (22). There were no significant differences between participants and nonparticipants in the number of women with TfR concentrations greater than the iron deficiency cutoff of 8.5 mg/L. TfR concentrations were unaffected by the length of the interpregnancy interval, participation in the Food Stamp Program, and smoking status.
No significant differences in plasma ferritin were found between WIC participants and nonparticipants through 6 mo postpartum (Table 3). Values for both groups fell from an average at baseline of 57 µg/L to 36 µg/L over 6 mo. Although the decrease may have been influenced by the dramatic decrease in use of iron supplements and changes from blood volume normalization, the mean ferritin values were always within the reference range of 20–150 µg/L for nonpregnant women. There were no significant differences in the number of subjects in each of the 2 groups with ferritin concentrations below the 12-µg/L cutoff for detection of iron deficiency.
Through the first 4 mo postpartum, and for both WIC participants and nonparticipants, substantially more subjects were identified as being iron deficient by Hb concentration than by the other 2 measurements (plasma TfR > 8.5 mg/L or ferritin < 12 µg/L). The ferritin assay almost always had the lowest level of detection of iron deficiency across time and groups. The occurrence of iron deficiency indicated by the TfR and ferritin methods was much lower than the occurrence indicated by Hb < 7.45 mmol/L. For nonparticipants 6 mo postpartum, Hb concentrations identified 51% as anemic, plasma ferritin concentrations identified 9% as iron deficient, and TfR concentrations identified 15% as iron deficient. At 6 mo, 17% of participants were identified by Hb concentration as anemic, 11% had a ferritin concentration indicative of iron deficiency, and 13% had a TfR concentration indicative of iron deficiency. The ferritin assay consistently identified fewer subjects as being iron deficient than did the TfR assay. The proportion of subjects determined to be iron deficient on the basis of the ferritin assay had increased by 6 mo postpartum, whereas the proportion determined to be iron deficient on the basis of the TfR assay had plateaued or decreased. Although the mean ferritin concentration decreased, the number of subjects with ferritin concentrations <12 µg/L did not increase.
The correlation among the 3 iron measurements was not significant and was extremely low (Hb and TfR, r = -0.07; Hb and plasma ferritin, r = 0.02; and TfR and plasma ferritin, r = 0.01). The correlation coefficient was adjusted for means between groups (P < 0.05).
Self-reported dietary intakes of iron did not differ significantly between the groups at any visit. The range across groups and time was 9.2–11.1 g/d and intake was never >74% of the RDA of 15 mg/d (6). No significant differences were found between groups at any visit for dietary intakes of vitamin C; in most cases, the intakes were 50% higher than the RDA of 60 mg/d (6).
DISCUSSION
In this study, nonlactating women participating in WIC appeared to benefit significantly from 6 mo of uninterrupted participation postpartum. To support this observation, and to account for a pattern of missing observations on Hb status, data were analyzed for missing visits. The results showed that the iron status of the participants improved with each visit, emphasizing the effect of continued WIC participation. A slight downward trend in mean Hb concentration over time was apparent for nonparticipants, whereas Hb increased slightly over time for participants (subjects with only a single measurement were eliminated from the analysis). Because the slope for Hb over time was not significantly different from zero for nonparticipants, prediction of a continuing trend for nonparticipants would have to be substantiated with continued intervention and long-term data collection.
Analysis of the rate of change (mean of individual slopes for Hb by group) showed that the mean Hb concentration by 6 mo postpartum for WIC participants plateaued and the mean Hb concentration for nonparticipants continued to decrease slightly. These data suggest a positive effect of WIC participation over time on reduction of anemia in nonlactating women postpartum. The difference between groups in mean Hb concentration and the prevalence of anemia at 6 mo underscore the need for WIC benefits to exist for a sufficiently long period to realize the program's effect.
The results of this study do not allow identification of the cause of the increased rate of anemia in nonparticipants. They do, however, support the results of several recent studies (32–36) in which anemia was attributed to iron deficiency with concurrent infection or inflammation, or both; the presence of anemia 6 mo postpartum suggests the development of this complicated health problem in nonparticipants. With concurrent infection and iron deficiency, Hb measurements may be abnormally low while TfR and ferritin measurements are in the normal range.
The higher prevalence of anemia in nonparticipants may also have been related to other nutritional deficiences and compromised health care. The occurrence of iron deficiency may be influenced by the consumption of fortified foods and adequate dietary intakes of ascorbic acid with nonheme sources of iron. In addition, women who were denied WIC participation were not receiving referrals for health care and this may have contributed to an increase in the rate of infection. On the basis of these considerations, WIC participation may be unrelated to iron status in relatively iron replete women who are free of nutritional risk.
The variability of the samples may have been underestimated in these results because samples were not necessarily collected and analyzed on the same day (37). Samples from each subject were not always analyzed in the same batch and samples from participants and nonparticipants (matched for visit) were not always analyzed together, which could have affected the estimates of variability.
Even though the subjects in this study did not have a history of anemia, an undeniable and significant reduction in the prevalence of anemia was seen with postpartum WIC participation. Because the TfR and ferritin values were not correlated with Hb concentration, a biological significance of the difference in Hb between groups was not established. The lower prevalence of anemia in participants may have been due to the secondary benefits of WIC participation: nutrition and health education that could have been associated with lower infection rates, better access to and participation in health care, or more purchasing power for health care associated with lower food expenses.
An elevated ratio of TfR to ferritin indicates reduced iron stores (decreased ferritin) or reduced delivery of iron to the tissues (increased TfR), or both, and is highly dependent on the method used for analysis of TfR. Researchers using the ELISA technique described by Flowers et al (22) for determination of TfR suggested a mass ratio of TfR to ferritin > 100 as an indication of inadequate iron stores (32). Neither group in this study had ratios of TfR to ferritin > 100. The serum TfR measurement was shown previously to be a highly reliable test of iron status except under conditions relating to dramatically increased erythropoiesis (38).
Correlations between the 3 iron measurements were consistent with correlations between iron-status measurements in a 1992 study by Dickinson (39) based on data from the second National Health and Nutrition Examination Survey for 8 measures of iron status. The lack of correlation between indexes and the low number of subjects at any visit who met all 3 criteria for iron deficiency again indicate the potential for anemia other than iron deficiency. Because the improvements in Hb were not accompanied by improvements in the results of any other tests of iron status, the reduction in the prevalence of anemia in this relatively iron-replete population may have been due to secondary causes.
For future studies, we recommend the use of biochemical indicators of infection and inflammation (eg, white blood cell count, C-reactive protein, and 1 acid glycoprotein). Documented medical information about each subject's menstrual history and presence of infection, illness, inflammation, and race-specific diseases should be collected at each visit, as should information on use of oral contraceptives and the reestablishment of menses.
National surveys showed average dietary iron intake of American women aged 15–44 y to be 10–11 mg/d (40–42), which is consistent with the findings of the present study. Although the FFQ results indicated no significant differences in iron intake between participants and nonparticipants, mean dietary iron intakes were below the RDA throughout the 6-mo period in both groups. Although the intakes were low, the relatively high consumption of heme iron (meats) and vitamin C may have contributed to iron sufficiency in most subjects through 4 mo postpartum. Because the FFQ was designed to rank intakes relative to those of other subjects (and not to measure actual intakes) and its sensitivity was low (15.5%), it was not appropriate for precise identification of iron deficiency.
We conclude from the laboratory results of iron status in this study that nonlactating women who were eligible for WIC benefits according to WIC criteria for nutritional risks benefited from continued WIC participation for 6 mo postpartum. The benefit may have been due to WIC participation under conditions of concurrent iron deficiency and infection or inflammation or due to secondary benefits such as improved nutrition education and health care. WIC participation contributes to improved health in low-income women at risk of poor nutritional status postpartum.
REFERENCES