Diet quality among pregnant women in the Navajo Birth Cohort Study

2.2. Dietary assessment

A modified HSFFQ survey was administered to NBCS women (Baer et al., 2005). The Navajo diet includes native plant‐ and game‐based foods such as dandelion greens, blue corn mush, and mutton (lamb), and transitional foodstuffs such as frybread (deep fried wheat flour), Navajo burgers (fry bread, hamburger patty, and condiments) and Navajo tacos (Kopp, 1986). The NBCS program added these Navajo foods to the HSFFQ instrument based on the guidance from Navajo community members and field staff. FFQ surveys were administered in person by trained women (predominately Navajo) who were either university employees, Navajo Division of Health Community Health representatives, or Indian Health Service employees. Staff were trained using food models and were instructed to use the models while administering the survey. FFQs were administered by trained staff during 36‐week prenatal appointments at the hospital or clinic. In some cases, FFQs were administered at the hospital within 1 week of delivery (n = 38; 17%). Participants were asked about their typical portion size and frequency of consuming each food item (never to six times per day) within the 4 weeks prior to the date of the interview. For the FFQ and other surveys administered, NBCS participants could opt not to answer any question, and this was recorded as “not reported.”

FFQs were processed, and reported consumption data were converted to nutrient intake by assigning a daily frequency weight and using the Nutrition Data System for Research (NDSR), a computer‐based software application developed at the University of Minnesota Nutrition Coordinating Center. The Nutrition Coordinating Center Food and Nutrient Database serves as the source of food composition information used by the NDSR for estimating nutrient intake based on FFQ responses (Schakel, Sievert, & Buzzard, 1988; Sievert, Schakel, & Buzzard, 1989). This database includes over 18,000 foods including 8,000 brand name products. Ingredient choices and preparation methods provide more than 160,000 food variants. Values for 165 nutrients, nutrient ratios, and other food components are generated from the database. The U.S. Department of Agriculture (USDA) Nutrient Data Laboratory is the primary source of nutrient values and nutrient composition in NDSR (Feskanich, Sielaff, Chong, & Buzzard, 1989). Representative recipes for fry bread, blue corn mush, Navajo burger, Navajo taco, and mutton‐based foods were added to the database for nutrient determination. Multivitamin supplementation (i.e., prenatal multivitamins) was excluded from our dietary intake analyses.

Pregnancy and age‐specific estimated average requirement (EAR) intakes were used as benchmarks to determine intake adequacy and the proportions of participants meeting current recommendations using the EAR cut‐point method (Institute of Medicine, 2000). Intake ratios were used as a metric to examine potential changes in nutrient intake between the Butte et al. study and the present study. Mean nutrient intake values were determined for NBCS participants (n = 242), and mean nutrient intake values reported in Butte et al. were used to calculate intake ratios for overlapping nutrients. An intake ratio <1 suggests lower estimated intake, and intake ratios >1 suggest greater estimated intake among NBCS participants compared with the Butte et al. study.

2.3. Diet quality

The Health Eating Index (HEI) is a metric developed to evaluate diet quality (Freedman, Guenther, Krebs‐Smith, & Kott, 2008; Miller et al., 2011). The HEI total score is the summation of individual component scores with a maximum total score of 100. Higher scores indicate greater consistency with recommended healthy diet guidelines. We calculated scores for 13 dietary components including total fruit, whole fruit, total vegetables, greens and beans, whole grains, dairy, total protein foods, seafood and plant proteins, fatty acid ratio, refined grains, saturated fat sodium, and added sugar. For the individual components of added sugar, sodium, saturated fat, and refined grains, a low score indicates increased dietary intake of these foods; for all other components, a low score indicates low dietary intake of those foods. We converted NDSR food servings per food group to cup equivalents to conform with HEI‐2015 scoring methods. We then calculated a population ratio HEI score for NBCS participants using the method previously outlined (Freedman et al., 2008; Guenther et al., 2014). This method obtains estimates of a population's total intake by taking the sum total of the relevant nutrient and energy across all sampled participants. Then the ratio of each nutrient to energy is calculated and converted to an HEI score. The total HEI score is the sum of HEI values for each dietary component. For comparison, we calculated two National Health and Nutrition Examination Survey (NHANES) HEI scores using a subset of NHANES 2011–2012 and 2013–2014 24‐hr dietary recall data for women aged 14–45 years, not considering pregnancy and lactation status (N = 1,492) and a subset of NHANES 2011–2012 and 2013–2014 24‐hr dietary recall data for pregnant women aged 14–45 years (N = 162) using the complex survey design population method.

2.4. Sensitivity analysis

We conducted a sensitivity analysis to investigate how daily prenatal supplements would impact the profile of the study population meeting EAR recommendations. We identified the most commonly consumed prenatal vitamin, as reported by participants, and acquired the reported micronutrient content for this brand. We randomly sampled 5% (N = 11), 25% (N = 56), 50% (N = 111), 75% (N = 167), and 95% (N = 211) of the total study participants (N = 222), for 1,000 iterations, and each time, we added the referred daily supplement values from the prenatal vitamin with the estimated micronutrient intake from the FFQ questionnaire.

2.5. Urinary iodine measurements

Iodine concentrations measured in urine samples collected in the morning, or from other spot urine collections, have been shown to adequately assess a population's iodine status (Rohner et al., 2014). During a women's 36‐week prenatal appointment, trained staff collected spot urine samples (N = 193) in sterile 50‐ml urine collection cups pre‐screened for metals that were provided by the National Center for Environmental Health's Division of Laboratory Sciences. Urine samples were stored at ‐80 °C until shipped overnight on dry ice to the Division of Laboratory Sciences and, once received, samples were stored at ≤−20°C until preparation for analysis. Urine iodine concentrations were measured using inductively coupled plasma dynamic reaction cell mass spectrometry (Centers for Disease Control and Prevention, 2011).

2.6. Statistical methods

Descriptive statistics for a continuous variable includes sample mean with standard deviation denoted by Mean (SD) and median with the interquartile range denoted by Median (IQR). For hypothesis testing of no difference between groups, the ANOVA and Kruskal–Wallis tests were performed as parametric and non‐parametric approaches for robustness of testing conclusion. For categorical variables, frequency and percentage were displayed for each category. The Fisher's exact test was used for testing for associations between stratified groups when count data are presented using a contingency table. Intakes of added sugars were expressed as contribution (%) to TEI and stratified by age and participation in food assistance programs.

For HEI calculations, we utilized the SAS macros provided by the National Cancer Institute (National Cancer Institute, n.d.). Specifically, the weighted means and variance–covariance matrix were estimated from the dataset and then used for Monte Carlo simulations to generate 10,000 independent observations. The mean of individual HEI components, the total scores, and their confidence intervals were further calculated from the simulated datasets. For NHANES data, the sampling weight was also accounted for. All the statistical analyses were performed using SAS 9.4 (SAS Institute Inc., Cary, NC) or R 3.2.1 (R Development Core Team, Vienna, Austria).

3.2. Mean nutrient intake ratios suggest some improvement in estimated intake over time

Mean intake ratios between Butte et al. (1981) and our study were calculated as a metric to evaluate potential changes in estimated dietary intakes over time (Figure ​(Figure1a).1a). Almost all nutrients examined had ratios close to or above one, indicating that estimated micronutrient intake has remained similar or improved slightly over time. Potential improvements were evident for the B vitamins (particularly folate and niacin).

Open in a separate window

Figure 1

Estimated dietary intake of micronutrients¹ among Navajo Birth Cohort Study (NBCS) women. (a) Estimated nutrient intakes of NBCS cohort compared with a 1981 Navajo study. Mean nutrient intake values were used to calculate an intake ratio during pregnancy for NBCS cohort versus 1981 study by Butte et al. (NBCS/Butte). An intake ratio <1 indicates decrease in mean estimated intake. An intake ratio >1 indicates increased mean estimated intake. (b) Prevalence of inadequate intakes in the NBCS cohort for nutrients with an Estimated Average Requirement (EAR)². (c) Sensitivity analysis based on daily prenatal vitamin intake by NBCS women. The multivitamin brand used by the majority of NBCS women was utilized to assess the percentage of NBCS women meeting the EAR for select micronutrients if a given proportion of NBCS women took this brand of multivitamin daily

3.3. High proportion of NBCS women have dietary intakes below recommended guidelines

Based on dietary intake alone, a high proportion of NBCS women (>50%) had inadequate intakes of calcium, folate, iron, magnesium, vitamin D, and vitamin E (Figure ​(Figure1b).1b). The proportion of participants with estimated intakes below the EAR did not differ significantly across subgroups stratified by age, body mass index, education level, household income, and food assistance program participation (Table S4). Vitamin D had the greatest proportion of women with inadequate intakes, followed by vitamin E. A slightly lower proportion of adolescent NBCS women (ages 14–18 years) had inadequate intakes of iron and vitamin D compared with older NBCS women. A high proportion of NBCS women reported prenatal vitamin intake (>90%), however our survey tool limited the ability to collect information on frequency of use. Because a majority of NBCS women reported using a specific brand of prenatal vitamins, we conducted a sensitivity analysis to assess the percentage of NBCS women meeting the EAR for select micronutrients if a given proportion of NBCS women took this brand of prenatal vitamin daily (Figure ​(Figure1c).1c). This analysis showed a substantial increase in the prevalence of adequate intake for vitamin D, folate, iron, and zinc among NBCS women with daily supplemental intake. The greatest effect of supplementation was on vitamin D, folate, and iron, indicated by the steep slopes for these nutrients (Figure ​(Figure1c).1c). Supplementation increased the prevalence of calcium and vitamin E adequacy, but to a much lesser extent compared with other nutrients. Even with 95% of NBCS women taking a daily prenatal vitamin, only 40% and 58% of NBCS women would achieve adequate vitamin E and calcium intake, respectively.

3.4. Urinary iodine concentrations indicate less than adequate iodine intake among NBCS women

The median urinary iodine (mUI) concentration for NBCS women was 90.8 μg/L (95% CI [80, 103.5 μg/L]; Figure ​Figure2),2), which indicates population‐level iodine insufficiency when compared with World Health Organization (WHO) sufficiency level of 150 μg/L. Additionally, the NBCS mUI was lower compared with pregnant women from several NHANES cycles (Figure ​(Figure2).2). The mUI of NBCS FFQ participants was comparable with NBCS women overall.

Open in a separate window

Figure 2

Median urinary iodine (median, interquartile range) for Navajo Birth Cohort Study (NBCS) and National Health and Nutrition Examination Survey (NHANES) groups. The dotted horizontal line indicates minimum urinary iodine (UI) concentration (150 μg/L) for sufficiency during pregnancy as defined by the World Health Organization. Spot urines were used to measure UI concentrations in NBCS and NHANES groups

3.5. Overall diet quality for NBCS women is similar to NHANES but differs for some components

Overall, average diet quality of NBCS women based on total HEI scores was similar to that reported for the general population of women of child‐bearing age and for pregnant women in NHANES (Table ​(Table2).2). However, there were notable differences among some of the individual components (Figure ​(Figure3).3). NBCS women had diets lower in dairy, seafood/plant protein, and refined grains and higher in added sugar, whole fruit, total fruit, and whole grains, compared with women of child‐bearing age in NHANES. When compared with a subset of women in NHANES who were pregnant, NBCS women had higher added sugar intake. Both NBCS women and NHANES pregnant women had diets lower in seafood and plant protein.

Open in a separate window

Figure 3

Radar plot of Healthy Eating Index‐2015 individual component scores for Navajo Birth Cohort Study (NBCS) and National Health and Nutrition Examination Survey (NHANES) groups. Each component point illustrates the percentage of total points received

4.1. Opportunities to address iodine intake

Dairy and seafood are major sources of iodine. This study is the first to identify the potential for iodine insufficiency among Navajo women. NBCS women had inadequate intake of iodine, with a median urinary iodine concentration <150 μg/L, suggesting a risk of iodine deficiency. Iodine is important for child neurodevelopment and essential for the normal production of thyroid hormones (WHO, 2013). Iodine fortification of table salt in the United States was implemented in the 1920s in response to iodine insufficiency. Although not documented in the literature, there is a range of sentiments among Navajo communities towards iodized salt. Our 30 years of experience on the ground and extensive conversations with Navajo field staff indicates that iodized salt is not sold in the primary grocery store chain on Navajo Nation. Although iodized salt can be purchased at other locations, it is unclear if iodized salt is sought out and purchased by residents. Some NBCS women live in proximity to national chain grocery stores that do sell iodize salt; however, our study design does not enable an analysis of the proportion of NBCS women that are purchasing and using iodized salt at the dinner table.

Although pregnant women are at greater risk for iodine deficiency, iodine is not required in prenatal multivitamin supplements in the United States (Leung, Braverman, & Pearce, 2012). Brand name prenatal vitamins are increasingly including iodine; however, inconsistencies remain in iodine per daily dose across brands (Lee, Stagnaro‐Green, Mackay, Wong, & Pearce, 2017). NBCS enrolment surveys indicated a majority of NBCS women reported taking a prenatal multivitamin, however they predominately used a brand of prenatal vitamins that does not list iodine content on the product label (data not shown). Contrasting our findings in NBCS women, a separate survey conducted by the New Mexico Pregnancy Risk Assessment Monitoring System (PRAMS) program found a majority of Navajo mothers reported no multivitamin/folic acid vitamin use in the month prior to pregnancy (Navajo Epidemiology Center). Together these findings support the need for efforts emphasizing education and access to iodine containing multivitamins as a part of women's health care, with special attention during reproductive age years and pregnancy (Alexander et al., 2017). Our study also highlights knowledge gaps on iodine intake among Navajo women, barriers to taking iodine containing supplements, and Navajo experiences with previous iodine fortification efforts. The New Mexico PRAMS program in partnership with the Navajo PRAMS workgroup monitors health status, behaviours, and experiences of Navajo mothers that reside in New Mexico. Partnerships with the New Mexico and Navajo PRAMS programs would present opportunities to learn more about promotion and access to iodine containing multivitamins. The PRAMS program is also available in Arizona and Utah, states that extend into Navajo Nation. Collaboration between the three states on issues related to iodine could provide a comprehensive dataset that can be used to develop evidence‐based public health strategies that are feasible and culturally relevant for Navajo communities.

The authors would like to acknowledge the tremendous efforts of the NBCS Study Team for administering FFQ surveys and engaging with NBCS participants. The NBCS Study team includes Qetarah Anderson*, Lorraine Barton, David Begay*, Delila Begay*, Francine Begay*, Mae‐Gilene Begay*, Nikki Begay*, Priscilla Begay*, Malcolm Benally*, Benita Brown*, Courtney Burnette, Miranda Cajero, Carla Chavez, Karen Cooper, LaShelly Crank*, Erica Dashner, Vanessa De La Rosa, Ruofei Du, Danielle Duarte, Esther Erdei, Adrienne Ettinger, Myra Francisco*, Joseph Hoover, Laurie Hudson, Lisa Kear, CJ Laselute*, Lynda Lasiloo, Ji‐Hyun Lee, Johnnye Lewis (PI), Li Luo, Debra MacKenzie, Chris Miller, Anita Muneta*, Teddy Nez*, Sara Nozadi, Elena O'Donald, Jennifer Ong, Tim Ozechowski, Bernadette Pacheco, Mallery Quetawki, Sandy Ramone*, Johnna Rogers*, Anna Rondon*, Diedra Sam*, Melissa Samuel*, Abigail Sanders*, Chris Shuey, Becky Smith, Charlotte Swindal, Marcia Tapaha*, Roxanne Thompson*, Doris Tsinnijinnie*, Monique Tsosie*, Shasity Tsosie*, Chris Vining*, Josephine Watson*, and Maria Welch* (* indicates a Navajo team member). Thank you to Vanessa Garcia with the UNM CTSC biochemical laboratory for her guidance on FFQ processing. We acknowledge the indigenous partners we worked with over the last three decades and from whom we learned about the lasting impact of the abandoned mines, with particular acknowledgement of those from Tachee/Blue Gap and Red Water Pond Road communities on Navajo, the 1,304 families in our original DiNEH Project, and the more than 750 families currently in the Navajo Birth Cohort Study. We also acknowledge the partnership of our research partners from Southwest Research and Information Center and their Navajo Home Environmental Assessment staff, Navajo Nation Department of Health including Mae‐Gilene Begay and the Community Health and Environmental Research Specialists, Navajo Nation Environmental Protection Agency, and faculty and staff from the University of New Mexico team, including Dr David Begay of Community Environmental Health Program for his support in understanding Navajo culture and tradition. Funding support was provided by the University of New Mexico Environmental Health Signatures Program and the Navajo Birth Cohort Study by U01 TS000135‐05, and training support provided by K12GM088021‐08 Academic Science Education and Research Training (ASERT) Program. Additional funding provided by Navajo Birth Cohort Study/ECHO UG3/UH3OD023344, NIH Office of the Director.