Dietary Risk Assessment in the WIC Program (2002)

The preceding chapters have discussed the poor reliability and validity of methods used to assess the diet and physical activity in individuals. This chapter addresses the nutritional vulnerability of pregnant and postpartum women and children as groups and presents results from relevant dietary intake studies as well as relationships between income and dietary risk. A discussion of infants has been omitted since this report does not cover dietary risk for this high-risk group.

The need for food energy and the Estimated Average Requirements (EARs) for most nutrients are higher for pregnant and lactating women than they are for other women in the childbearing years (IOM, 2001). At the same time, the effects of nutrient shortfalls potentially are more serious for pregnant and lactating women than for other women. Both the woman’s health and that of the embryo, developing fetus, or infant may be affected. For example, inadequate energy intake may contribute to low gestational weight gain and fetal growth restriction. Likewise, inadequate iron intake may lead to maternal anemia and to low iron stores in the infant. The combination of higher requirements and/or higher recommended nutrient densities and more serious results of deficiency means that pregnant and lactating women are more vulnerable to nutrition problems than are other adults—both women and men. For example, iron deficiency is nearly

twice as prevalent among low-income women in their reproductive years, in comparison with those who are more advantaged (Looker et al., 1997), but the three tests required to determine iron deficiency are not a routine part of health care or of WIC services.

In addition, mounting evidence indicates that practices during pregnancy may have a long-term impact on health. For example, periconceptional intake of folic acid is important not only in the prevention of central nervous system and other birth defects but also in reducing the risk of cancer and other chronic disease in later life (Toren et al., 1996). Likewise, increasing maternal intake of omega-3 fatty acids is associated with increased gestation duration (Allen and Harris, 2001), improved fetal neurological development (Innis, 2000), and lowered maternal cardiovascular risk (Mori and Beilin, 2001).

Little attention has been paid to maternal nutrition after pregnancy, particularly among women who are not lactating, possibly because they have had a low priority for receipt of WIC services. Two studies were identified that address postpartum, nonlactating women. Caan and colleagues (1987) examined the influence of extended maternal food supplementation (5–7 months) in the interpregnancy interval compared with more limited supplementation (0–2 months). All women received WIC benefits during both the index and subsequent pregnancies. In comparison to those with limited feeding, women with extended supplementation had significantly improved outcomes in the subsequent pregnancy: birth length was increased by 0.3 cm and birth weight was 120 g higher after controlling for gestational duration and other potential confounding variables (e.g., maternal smoking and the birth weight of the prior infant). Maternal iron status was improved—hemoglobin levels were increased significantly, on average by 0.3 mg/dL with extended feeding. In addition, risk of maternal overweight and obesity (defined as > 120 percent of ideal weight in this study) was reduced twofold among women on extended supplementation (Caan et al., 1987). Pehrsson and coworkers (2001) examined three indicators of iron status among postpartum, nonlactating participants and eligible nonparticipants (women unserved because of lack of funds). Women who participated in WIC for the full 6 months were significantly less likely to become anemic than were the eligible nonparticipants.

The years prior to age 5 are a time of rapid growth and development. The results of a shortfall during the first 4 years of life can be very serious, including both stunted physical growth and cognitive deficits. Compared with children

from more affluent families, low-income children are more likely to have anemia (CDC, 1998a), to be stunted and/or overweight (CDC, 1998a), to have higher blood lead levels (NCHS, 1998) and, perhaps in consequence, to be developmentally delayed or learning disabled (Brooks-Gunne and Duncan, 1997). They also are more likely to have experienced hunger in the past year and to come from a family where the head of household reports fear of going out into the neighborhood (Brooks-Gunne and Duncan, 1997).

Over the past few decades, overweight and obesity have become more prevalent among women in the childbearing years (Flegal et al., 1998; Kuczmarski et al., 1994; Mokdad et al., 1999) and among young children in the population as a whole (Ogden et al., 1997). Overweight and obesity are more prevalent in lower- than in higher-income groups (NCHS, 1998), as well as in the subgroups of the population that give rise to many of those who are eligible for WIC (Must et al., 1999). Thus, it is reasonable to expect that the prevalence of overweight and obesity would be higher in populations served by WIC. In 1990, 19 percent of pregnant women in WIC were obese (Kim et al., 1992)¹; but by 1994, the prevalence of obesity had increased to 22 percent (Randall et al., 1995). A more recent analysis of data on pregnant women enrolled in the Ohio and Kentucky WIC programs, using current adult criteria for obesity (BMI ≥ 30 kg/m²) and overweight (BMI ≥ 25 kg/m²) (NIH, 1998; WHO, 1995), showed that during the first trimester of pregnancy, over one-fourth of all women and one-third of African-American women were obese (Whitaker et al., 1997, 2001). Few data are available on the extent to which those who enter pregnancy at a healthy weight become overweight during pregnancy or the postpartum period. However, excessive weight gain during pregnancy may be a factor that increases the risk of new postpartum overweight and obesity among young and mature gravidas in general (Gunderson et al., 2000; Scholl et al., 1996).

Among 4-year-old children in WIC, the prevalence of overweight (weight-for-height ≥ ninety-fifth percentile) increased from 8.2 percent in 1983 to 10.6 percent in 1995, a relative increase of almost one-third in just 12 years (Mei et al., 1998). Other recent analyses of data on children enrolled in the Ohio and Kentucky WIC programs (Whitaker, personal communication) in 1998 have used the 2000 Centers for Disease Control and Prevention growth charts

(Kuczmarski et al., 2000) to estimate the proportion of 4-year-olds at risk for overweight (BMI ≥ eighty-fifth and < ninety-fifth percentile) and overweight (BMI ≥ ninety-fifth percentile). Across these two states, the prevalence of at risk for overweight and overweight among children 48 to 60 months of age were 14 percent and 12 percent, respectively. Although this could be interpreted as saying that WIC is making things worse, there is no reason to believe that this increased prevalence of overweight is isolated to those enrolled in WIC or is somehow caused by WIC. Examination of nationally representative cross-sectional surveys of 4- and 5-year-olds in the United States has shown that the rate of overweight (weight-for-height ≥ ninety-fifth percentile) nearly doubled from 5.8 to 10 percent between 1971 and 1994 (Ogden et al., 1997). Similar increases are also being seen among preschool children in other developed countries (Bundred et al., 2001).

Because the problem of obesity disproportionately affects low-income women and because the prevalence of obesity has increased substantially in low-income children in recent years, it might appear that food supplementation in WIC is a counterproductive strategy for preventing or treating the problem of obesity in WIC. However, there is no evidence that enrollment in the WIC Program is a risk factor for obesity among income-eligible children. In fact, low-income children enrolled in WIC do not appear to have higher weight-for-height than those low-income children not enrolled in WIC (CDC, 1996).

There is also a poorly understood paradox that food insecurity and obesity can coexist. Food insecure women are more likely to be overweight than those who are food secure (Olson, 1999; Townsend et al., 2001), and this appears to be true even within the population of women receiving food stamps (Townsend et al., 2001). The question of whether children living in more food-insecure households are more likely to be overweight has not yet been addressed using the current U.S. Department of Agriculture Food Security Scale (Gleason et al., 2000). However, Alaimo and colleagues, in a series of analyses using data from the Third National Health and Nutrition Examination Survey, have examined several child outcomes in relation to household food insufficiency (living in a family that reports “sometimes or often not getting enough food to eat”)—a state thought to be different and more adverse than food insecurity (Carleson and Briefel, 1995). Although there was no tendency for household food insufficiency to increase the risk of overweight in 2- to 7-year-old children (Alaimo et al., 2001b), school-age children living in food-insufficient households have been shown to have poorer parent-reported health status (Alaimo et al., 2001c) and poorer outcomes on certain measures of cognitive, academic, and psychosocial development, after controlling for other socioeconomic indicators (Alaimo et al., 2001a).

Thus, the above current evidence suggests that food supplementation may decrease, rather than increase, the risk of overweight among food-insecure mothers in WIC and that it may promote varied aspects of health and well-being among food-insufficient children without adding to the risk of overweight.

Maternal overweight and obesity are associated with a reduced risk of fetal growth restriction but have serious consequences for maternal health and other aspects of fetal well-being. Risks of maternal hypertension, gestational diabetes, and cesarean section increase with increasing BMI; hospitalization expenses increase as well (Galtier-Dereure et al., 1995). Risk of late fetal death is increased twofold in women who are obese before pregnancy (BMI ≥ 30). In nulliparas, fetal death is increased threefold in the overweight and fivefold in the obese and there are trends suggesting a rise in both early neonatal death and very preterm delivery (< 32 weeks) with increasing maternal pregravid BMI (Cnattingus et al, 1998). A similar effect on mortality was observed as part of the Collaborative Perinatal Project: perinatal mortality was increased approximately twofold among overweight gravidas (BMI 25–30) and more than threefold among those who were obese (BMI > 30) (Naeye, 1990). Maternal pregravid obesity also is a risk factor for major congenital defects in the fetus (Naeye, 1990). The well-known protective effect of folic acid intake on risk of neural tube defects appears to be absent among obese women (Shaw et al., 1996; Waller et al., 1994; Werler et al., 1996). While the reasons for this effect have yet to be identified, a similar lack of benefit was observed with zinc and obesity: risk of low birth weight was decreased in normal weight, but not in obese women receiving supplemental zinc (Goldenberg et al., 1995).

Overweight children are more likely to become obese adults (Power, 1997; Serdula et al., 1993). By 6 to 9 years of age, an overweight child with an obese parent has more than a 70 percent chance of being obese in young adulthood (Whitaker et al., 1997). At any age, once obesity develops, it is very difficult to treat (Barlow and Dietz, 1998; NIH Technology Assessment Conference Panel, 1993). Even in childhood, being overweight is associated with abnormalities in cardiovascular disease risk factors such as blood pressure, serum lipid concentrations, and serum insulin concentrations (Freedman et al., 1999). There are now alarming increases in the prevalence of Type II diabetes among young adults (Mokdad et al., 2000) and adolescents (Fagot-Campagna et al., 2000) that are attributable to the problem of obesity early in life. For all these reasons, there is great interest in preventing children from ever becoming overweight.

Because WIC serves almost 6 million children under 5 years of age, the program has a unique opportunity to provide the early intervention required to prevent childhood obesity. Anthropometric measurements in WIC identify applicants at nutritional risk because of overweight or obesity. While there are no highly accurate methods to determine which normal weight infants and children in WIC are likely to become overweight, it is clear that both maternal and sibling obesity substantially (and independently) increase the risk that a newborn in WIC will become overweight by 4 years of age (Whitaker et al., 2001). This population of newborns may be a promising target group for developing obesity prevention strategies in WIC. However, these strategies will require a new nutrition counseling paradigm that takes into account the evidence that many families in WIC with already overweight children may not believe that their children are overweight (Baughcum et al., 2000; Jain et al., 2001). Furthermore, the paradigm will need to consider the important role of activity, along with diet, in obesity prevention.

No representative studies have reported on the nutrients or foods consumed by the women and children applying for WIC. Thus, indirect data must be used to examine the potential dietary risk of these groups. Relatively little information is available about dietary intake of pregnant or lactating women or postpartum women for the 6-month period after delivery, which makes the data even more indirect for these groups. This section covers dietary intake information about women and children in general, and about those served by WIC.

For most nutrients, the recommended method to assess the adequacy of nutrient intake by a population is a three-step process:

1. Obtain estimates of the usual intake distribution of the population. This requires at least two nonconsecutive days of research-quality diet recalls or records from a representative sample of the population of interest.

2. Make statistical adjustments of the data to remove within person variation in intake.

3. Determine the percentage of the population with usual intakes above or below the Estimated Average Requirement (EAR) (IOM, 2000a).

For iron, because the distribution of requirements is skewed, the probability approach should be used as described by the Institute of Medicine (IOM, 2000a, 2001; NRC, 1986).

No studies were found reporting percentages of individuals with intakes below the EAR for age or physiological status. Therefore, to identify problem nutrients, the committee used tables of percentiles for usual intakes of nutrients from food published in IOM reports on Dietary Reference Intakes (IOM, 1997, 1998, 2000b, 2001) and the EARs for young children and women by physiologic status. The dietary intake data used for those tables are from the 1994–1996 Continuing Survey of Food Intakes by Individuals (CSFII) for phosphorus and magnesium and from the 1988–1994 Third National Health and Nutrition Examination Survey (NHANES III) for the remaining nutrients. This method allowed the identification of a range for the percentage of individuals with intakes below the EAR. Three nutrients were excluded: calcium, for which an EAR has not been set; folate, for which current intakes are likely to be higher than reported because of the fortification of enriched cereal grains and the use of micrograms of dietary folate equivalents for setting the EAR (IOM, 1998); and iron, for which appropriate data are lacking.

Nutrients for which more than 5 percent of the age group has an intake less than the EAR are shown in Table 8-1. No data are available to determine the extent to which persons who are categorically and income-eligible for WIC are represented in the group with intakes below the EAR, but data covered in a later section, “Associations of Food Intake with Income,” suggest that they may be at increased risk.

In a report to the U.S. Department of Agriculture that addressed the estimation of dietary risk (and other components of nutritional risk), Sigma One Corporation (2000) used 1-day dietary intake data from Phase 1 of NHANES III (1988– 1991) for women ages 17 to 49 years and children ages 1 to 4 years. The cut points shown in Table 8-2 represent modal levels used to determine dietary adequacy obtained from 1997 WIC state plans. The cut points are similar to, but not coincident with, the recommended number of servings for each food group specified by the Food Guide Pyramid. Since only approximately 6 percent of the women studied had intakes that met or exceeded the cut point for each food group, Sigma One concluded that 94 percent of the women were at dietary risk. Reported intakes by children ages 1 to 4 years were somewhat better: 15 percent met modal intake for each food group placing 85 percent at dietary risk.

Krebs-Smith and colleagues (1997) conducted a similar but more rigorous analysis of the food group intake of adults. They determined the percentages who met energy and age-based Food Guide Pyramid recommendations (see Table 3-2) using 3 days of dietary data from the 1989–1991 CSFII. Only 0.1 percent of the women met the recommendations for all five basic food groups. Notably, the average age of the men and women who met all the recommendations was 60 years—well beyond the childbearing years. In another study of women ages 18 to 39 years, nearly 34 percent had fewer than 2.5 servings of fruits and vegetables per day (LSRO/FASEB, 1995). Munoz and coworkers (1997), in a rigorous analysis of 1989–1991 CSFII data, found that none of the children ages 2 to 5 years (n = 1,028) met the minimum recommendations for all five food groups, after allowing for smaller portion sizes of grains, fruits, vegetables, and meats. Considering the five food groups individually, recommendations for the meat group were least likely to be met—only about 13 percent of the children met them. A much smaller (n = 110) but similar study of older children (ages 7 to 14 years) conducted in Alabama found that only 5 percent and 9 percent met Pyramid recommendations for fruit and milk products, respectively, over 3 days. Moreover, the mean proportion of calories from fat was well above the recommended 30 percent for children ages 2 to 5 years (CDC, 1996), suggesting that a high percentage of preschool children do not meet the recommendation. Based on data from NHANES III (1988–1994), only about 23 percent of children ages 2 to 5 years met the dietary recommendation for total fat intake (Troiano et al., 2000).

Bowman and colleagues compared the Healthy Eating Index (HEI; see Chapter 5) of selected groups using 2 days of 1994–1996 CSFII data from subjects of all ages. (A score of 80 or more implies a good diet.) African Americans had lower HEI scores than other ethnic groups (59 compared with 64–67). Children ages 2 to 3 years had a higher mean HEI score (74) than did older children (68 for children ages 4 to 6 years) or women in the childbearing years (61–62). McCullough et al. (2000) used data from Food Frequency Questionnaires obtained in the Nurses’ Health Study to calculate HEI scores and found that only 1 to 2 percent of the 67,272 subjects (many of whom are beyond the childbearing years) had an HEI score greater than 90.

Women and children participating in WIC have intakes of energy and certain nutrients that exceed those of nonparticipants (Rush, 1988; Suitor et al., 1990). While data are lacking to determine whether the changes resulted from WIC participation, it is reasonable to infer that intakes while served by WIC are at least as high and probably higher than intakes of new applicants. Participants have consistently reported food intakes that resulted in protein intakes at or above the 1989 Recommended Dietary Allowances (Kramer-LeBlanc et al., 1999). Data from this study are not available on the percentages of participants whose intakes were below the EARs for nutrients. However, nutrients identified as problematic included calcium, iron, folic acid, zinc, and magnesium during pregnancy; vitamin C and zinc during lactation; and iron, calcium, and magnesium in nonbreastfeeding, postpartum women. The overall diet of pregnant and postpartum WIC participants tended to be low in calories (a mean of 70–89 percent of the recommended energy allowance) with a low nutrient density and 33– 37 percent of calories from fat—well above the recommended 30 percent. Mean intakes for the WIC target nutrients were fairly similar for WIC participants, income-eligible nonparticipants, and the total sample of pregnant, lactating, and postpartum women (Kramer-LeBlanc et al., 1999). In some instances (e.g., folic acid, B₆), however, their intakes were higher than those of the comparison groups of income-eligible individuals not participating in WIC or of the total sample in the same age range.

Dietary data from 332 pregnant women participating in the 1988–1994 NHANES III (Mardis and Anand, 2000) were examined for WIC participants, for income-eligible nonparticipants, and for those whose incomes exceed the threshold for WIC. On average, all groups consumed less than the recommended number of servings from the Food Guide Pyramid based upon a 2,200 kcal diet as shown in Table 3-2 (using three servings as the cut-off value from the milk group). However, each group consumed more than the recommended percentage of energy from fat and saturated fat and more than 2,400 mg of sodium. In addition, WIC participants consumed significantly fewer servings of milk than did the women with incomes greater than 185 percent of poverty.

Dietary data show an inverse relationship between income and dietary intake of energy and of certain food groups. For example, children from low-income households (< 131 percent of poverty) are less likely to meet current recommendations for the consumption of fruit and milk products than are those from more affluent families (Munoz et al., 1997). Likewise, lower-income adults (women and minorities in particular) are over-represented among the group whose diets fail to meet any of the recommendations for food consumption in

the Food Guide Pyramid (Krebs-Smith et al., 1997). The percentage of women who report the consumption of fruit during 3 days increases with increasing income level (LSRO/FASEB, 1995).

On average, children in poverty have higher fat intakes than those in more affluent families (USDA, 1999). Smaller percentages of low-income (< 131 percent of poverty) children ages 3 to 5 years met recommendations for total fat than did those whose household incomes were 131 to 350 percent and over 350 percent of poverty (21 percent, 34 percent, and 44 percent, respectively). Similarly, only about 12 percent of the lowest-income children met recommendations for saturated fat intake compared with about 25 percent of the children at 131 to 350 percent of poverty and about 31 percent of the higher-income children.

Block and Abrams (1993) used data from the second NHANES and from the CSFII to examine associations of nutrient and food intakes with income. They found that women with incomes near poverty or below poverty had lower mean intakes for every nutrient examined (protein, calcium, folic acid, iron, zinc, vitamins A, C , E, and B₆). Low-income women, in particular, ate few fruits and vegetables. About half ate no vegetables at all, including potatoes, when surveyed over four nonsuccessive days. In a random sample of mothers in a rural New York county, nearly 75 percent of the food-insecure mothers (n = 103) reported 0 to 2 servings of fruits and vegetables daily, compared with 55 percent of the food-secure mothers (n = 90) (Kendall et al., 1996).

Scores from summary measures such as the HEI improve with increased income, in part because of increased variety of intake with increased income. Based on 1994–1996 CSFII data for all ages, people with household incomes at or below 50 percent of poverty had average variety scores of 6.9 (out of a possible 10), whereas those with household incomes at or above 300 percent of poverty had average variety scores of 7.9. Those from the lowest-income households also had lower average scores for saturated fat (5.7) and sodium (6.6) than did people whose household income was more than three times the poverty level (saturated fat, 6.6; sodium, 7.9) (Bowman et al., 1998).

Findings from the Dietary Quality Index, an overall score determined by whether or not an individual’s diet met recommendations for fat (total, saturated), cholesterol, fruits and vegetables, complex carbohydrate, protein, sodium, and calcium (NRC, 1989) were similar—higher income was associated with better scores. Those whose diets were judged as “poor” had lower incomes and were 5 to 7 times more likely not to be college graduates than those with diets classified as “good” (Patterson et al., 1994).

Throughout history, it has been observed that people restrict their food selection as food prices rise (Karp and Greene, 1983). In the instance of the United States oil embargo that occurred in the 1970s, the prevalence of anemia increased among poor urban children when a major source of iron (meat) began to disappear from the family diet due to the rising cost of food (Karp and Greene, 1983). Thus, when income falls and food selection is sufficiently narrow, dietary

quality suffers and nutritional status frequently will worsen, particularly among the poor.

Box 8-1 summarizes the broad range of evidence suggesting that individuals who are both categorically and income-eligible for WIC participation generally are also at dietary risk. The need for energy and for most nutrients is increased during pregnancy and lactation; thus, the effects of low intake are more serious for pregnant and lactating women than for other women.

When there is a shortfall, the health of both mother and fetus is likely to be affected. Likewise, the years prior to age 5 are a time of rapid growth and development. The results of a shortfall during the first 4 years of life can be very serious, including stunted physical growth and cognitive deficits. The inverse relationship between quality of intake and income is well documented. The diets of many low-income women and children are of a low nutrient density, contain more fat and saturated fat than recommended by the Dietary Guidelines, and fail to meet food group recommendations specified by the Food Guide Pyramid, especially in the fruit and vegetable groups. These data suggest that essentially all women and children who are income-eligible for WIC are at dietary risk.