Early Feeding Practices and Weight Status at One Year of Age: A Comparison of Hispanic Immigrant Mother–Infant Dyads with Participants of the Infant Feeding Practices Study II

Introduction

Hispanic children are at disproportionate risk for obesity and resulting chronic health problems.^1–3 Using the benchmark of BMI for age ≥85th percentile, a recent national survey found rates among Hispanic children in all age groups between 2 and 11 years were higher than any other racial/ethnic group. ⁴ This finding is particularly significant in light of reports that nearly one in four children in the United States is Hispanic, a proportion that has more than doubled over the last three decades. Furthermore, while 90% of Hispanic children were themselves born in the United States, more than half of the children have at least one parent who is an immigrant to the United States. ⁵

To address the overall problem of childhood obesity, the Institute of Medicine has recommended early obesity prevention activities that include exclusive breastfeeding for the first half of infancy with the introduction of solid foods around 6 months. ⁶ These recommendations are consistent with the American Academy of Pediatrics (AAP) guidelines for healthy infant feeding practices. ⁷ The Infant Feeding Practices Study II (IFPS II), the nation's largest longitudinal study of infant feeding practices, provides extensive data for assessing adherence to these guidelines. The IFPS II followed mother–infant dyads throughout the first year of life and provides data about national prevalence for breastfeeding, bottle feeding, solid food introduction, and infant weight status up to 12 months of age. ⁸ A frequently cited limitation of the IFPS II is the underrepresentation of Hispanic mothers and infants (6.2%). ⁹ Data about mothers’ country of origin are not included in the IFPS II.

Given that a mother's life experiences may influence infant care practices and that a substantial proportion of US infants have mothers who are Hispanic immigrants, it is important to examine early feeding practices and infant weight status for this subset of our population. A growing body of evidence confirms that obesity has its origins in early life exposures.^10–12 A retrospective study of obese preteen children found that more than half of the children were overweight before 2 years of age and 25% became so before 3 months of age, leading to the conclusion that early infancy is the “tipping point” for preventing childhood obesity. ¹³

Researchers who have examined early infancy predictors of childhood obesity have consistently recommended that further study is needed with more socioeconomically disadvantaged and diverse racial/ethnic groups.^14–16 Therefore, the purpose of this study was to compare similarities and differences in early infant feeding practices and subsequent weight status at 1 year of age between infants of low-income, Hispanic immigrant mothers, and the largely white and higher income participants of the national IFPS II.

Methods

Comparative analyses of early infant feeding practices were conducted using longitudinal data from a study of infant feeding practices among low-income, Hispanic immigrant mothers, entitled Proyecto de Bebés Hispanos Saludables (PBHS) and the national IFPS II. The PBHS data were compiled from the electronic medical records of a large primary care pediatric clinic in northern Virginia that serves low-income families from geographically and culturally diverse backgrounds. Mothers and infants who receive care at the clinic have Medicaid or no health insurance and are eligible for the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC). Babies from the PBHS sample were born in one of five local hospitals, each with access to lactation consultants during the newborn period.

Data from both the newborn hospital records and well child visits at the clinic were used to collect feeding and anthropometric information. Included records were those of mothers who self-identified as Hispanic immigrants to the United States and who received well child care at the study setting up to 12 months infant age between 2010 and 2012. Consistent with inclusion criteria for the IFPS II, mothers in the PBHS data set were also at least 18 years of age and gave birth to a singleton baby weighing at least 5 pounds, at least 35 weeks gestation, not requiring hospitalization for more than 3 days after birth, and who had no medical conditions that would affect feeding or growth.

The IFPS II data were collected by the USDA using a nationally distributed panel survey. The majority participants (58.1%) reported income levels ≥185% of the federal poverty level and self-identified their race as white (84.4%). Questionnaires were mailed approximately monthly throughout infancy between 2005 and 2007 and included items for maternal report of current infant feeding practices and recall of weight and length measurements from the most recent well child visit. Participants in the IFPS II had to be English speaking and have a literacy level high enough to complete the questionnaires. ¹⁷

The two data files were merged yielding a sample of 2052 participants of which 550 (26.8%) were PBHS mother–infant dyads and 1502 (73.2%) were from the IFPS II. For bivariate analyses comparing PBHS and IFPS II sample characteristics, missing data ranged from 0% to 4.0%. Missing data for the mode of milk delivery measures were from 0.5% to 8.0%. Weight-for-age (WFA) percentile at 6 months data was missing for 7.1% of the PBHS sample and 20.8% of the IFPS II sample. Multivariable analysis was performed on only those records with no missing data, which meant that the PBHS analytic sample consisted of 515 cases, while the IFPS II included 1209 records.

To identify factors associated with early infant feeding practices, three mutually exclusive modes of milk delivery were used: breastfed only, breastfed and bottle fed, and bottle fed only. We calculated mode of milk delivery during the first half of infancy using age specifications from prior IFPS II research for 2 months (7–10 weeks), 4 months (15–18 weeks), and 6 months (24–28 weeks) infant age. ¹⁸ Introduction to solid foods was also assessed and dichotomized as <4 or ≥4 months of infant age as has been done in prior research.^19,20

To make comparisons between PBHS and IFPS II on weight status at 12 months infant age, WFA was assessed. For PBHS infants, Certified Medical Assistants with training in pediatric anthropometric measurement had used the Midmark Pediatric 409 Healthometer to measure weight and entered their findings in the infant's electronic medical record, which was the source for the PBHS data set. The IFPS II infant weight data were collected using maternal recall from the most recent well child visit, as reported in the mailed questionnaire.

To determine WFA at birth, 6 and 12 months, we used the WHO Child Growth Standards. ²¹ These standards have been recommended for use in clinical practice since 2010 and were developed based on an international reference population of healthy breastfed infants. ²² WFA status at 12 months was dichotomized as </≥85th percentile. The 85th percentile approximates 1 standard deviation (SD) unit above the median of the WHO growth curve population and allows for comparison of our findings with those of prior research using a similar parameter. ²³ Furthermore, because infants who attend well child visits are not precisely 6 and 12 months of age when measured, calculations of age at time of measurement were based on the actual number of days between birth and age at measurement. WFA was calculated using Epi Info software. ²⁴

Maternal characteristics included age, education, and parity, and infant characteristics of gender and birth weight were categorized to allow for comparison to prior research using the IFPS II. Bivariate analyses using the chi-square statistic were performed to assess any significant associations between PBHS and IFPS II samples in regard to these maternal and infant characteristics as well as mode of milk delivery at birth, 2, 4, and 6 months infant age, and timing for the introduction to solid food. T-tests were conducted to examine any significant differences between the two groups for WFA percentiles at birth, 6, and 12 months.

We conducted separate logistic regression analyses for PBHS and IFPS II data sets to determine any significant effects of mode of milk delivery at 6 months on WFA percentiles at 12 months. These analyses were adjusted for maternal and infant characteristics as well as the timing for introduction to solids. All analyses were performed using SPSS (IBM SPSS Statistics for Windows, 2013). The study was approved by the institutional review boards at the clinical site and the authors’ university.

Results

Table 1 presents a comparison of descriptive information about the PBHS and IFPS II groups. PBHS mothers were more likely to be between 18 and 24 years old (25% vs. 15%), while a higher proportion of IFPS II mothers were at least 35 years of age (20% vs. 15%). The PBHS mothers were more likely to report their educational level as high school or less (89% vs. 17%). Indeed, the 40% of PBHS mothers reported having less than 9 years in school. Furthermore, parity differed between the two groups with a greater proportion of PBHS mothers having 4 or more children compared to their IFPS II counterparts (14% vs. 9%). There was no significant difference in infant gender between the two groups. The average birth weight was significantly different, but clinically comparable for PBHS and IFPS II babies (7.5 vs. 7.7 lbs).

A comparison of early infant feeding practices between PBHS and IFPS II mothers also revealed similarities and differences, as presented in Table 2. The mode of milk delivery from birth through 6 months was far more likely to be a mixed method of both breastfeeding and bottle feeding for the PBHS group than the IFPS II group. Throughout the first 6 months of life, over 90% of PBHS mothers were bottle feeding (mixed breast and bottle fed or bottle fed only) their infants. By comparison, the proportion of the IFPS II mothers who were bottle feeding (mixed breast and bottle fed or bottle fed only) ranged from 54% to 59% in the first 4 months, but increased sharply to 73% by 6 months of age. There was no significant difference in the age for introduction to solid foods, with ∼40% of mothers in each group reporting that this infant feeding practice began before their babies were 4 months old.

Table 3 displays differences in WFA percentiles between infants born to PBHS and IFPS II mothers. Infants born to PBHS mothers were significantly more likely to be at or above the WFA 85th percentile at 6 months (34.2% vs. 19.8%) and 12 months (39.1% vs. 25.4%) than their IFPS II counterparts.

Logistic regression analyses were performed separately for each group to identify any significant effects that mode of milk delivery at 6 months may have had on the likelihood of high WFA percentiles at 12 months. Table 4 presents unadjusted and adjusted odds ratios (ORs) and their associated confidence intervals for WFA ≥85th percentile. Unadjusted ORs show that infants born to PBHS mothers were 2.28 (1.07, 4.81) and 2.41 (1.14, 5.11) times more likely to be ≥85th percentile if the mode of milk delivery was mixed (breast and bottle fed) or bottle fed only at 6 months infant age, compared to those who were breastfed only. For the IFPS II group, infants were 1.81 (1.31, 2.51) times more likely to be at the ≥85th percentile if the mode of milk delivery at 6 months was bottle only.

Upon adjusting for mothers’ age, education, parity, infants’ gender, birth weight, and age for introduction to solid foods, infants born to PBHS mothers were 2.86 (1.28, 6.38) times more likely to be ≥85th percentile at 12 months if their mode of milk delivery at 6 months was bottle only compared to those who were breastfed only. Similarly, among the IFPS II group, infants who were only bottle fed were 1.97 (1.35, 2.87) times more likely to be so. However, for both groups, no significant effects of mixed feeding (breast and bottle fed) at 6 months on likelihood for being ≥85th percentile at 12 months of age were found in the adjusted models.

Discussion

In this comparative study of early infant feeding practices and subsequent weight status, the low-income, Hispanic immigrant mothers who comprised the PBHS sample were more likely than their counterparts in the IFPS II to use a mixed mode of milk delivery (breastfeeding and bottle feeding) rather than the recommended practice of exclusive breastfeeding during early infancy. This finding compares with a CDC study of breastfeeding practices for US-born children in 2012 that reported that 51.4% of Hispanic mothers breastfed at 6 months with 20.8% doing so exclusively. In our study, 52.5% of PBHS mothers breastfed their babies at 6 months, but only 9.0% did so exclusively. The CDC report did not specify immigration status of the mothers, which may help explain the disparity. ²⁵

No difference was found between PBHS and IFPS II mothers in the timing for solid food introduction with ∼40% of each group initiating this practice before 4 months infant age. A study of racial/ethnic differences in early life risk factors for obesity found a similar rate of early introduction to solid foods for Hispanic mothers. However, the same study found a substantially lower proportion of non-Hispanic mothers following this practice than we found with IFPS II participants. ²⁶ The difference for IFPS II mothers may be explained by the fact that the AAP guidelines for delayed introduction solid food introduction until at least 4 months, but preferably 6 months infant age, were distributed after the period of data collection for the IFPS II. ²⁷

In the current study, the early introduction of solid food was not found to be a predictor of weight status at 12 months of age for either group. While the introduction of solid foods before the age of 4 months has been associated with later obesity in some studies, researchers who conducted a systematic review of 24 studies found no clear association between the age for solid food introduction and subsequent infancy or childhood weight status.^28–30 A recent study of IFPS II participants found the timing of solid food introduction was not related to childhood obesity at age 6 years. ³¹ Differences across studies may be linked to variation in methods used to measure the timing of solid food introduction, concomitant mode of milk feedings, and weight outcomes.

Differences in weight status were found with PBHS infants’ WFA percentiles becoming substantially higher by 6 and 12 months of age. The predictive models we tested do not completely explain this differential finding. After controlling for variables that have been used in prior research, including maternal age, education, parity, infant gender, birth weight, and age for introduction of solid food, the practice of bottle feeding exclusively was the sole predictor of weight ≥85th percentile. In this respect, both groups of participants were alike, with exclusive bottle feeding being a stronger predictor of elevated weight status for the PBHS babies than those in the IFPS II. This finding is consistent with studies that have pointed to problems with infant self-regulation of feeding when using a bottle instead of breastfeeding.^32,33

Prior research has also identified additional reasons that low-income Hispanic mothers may provide nonexclusive breastfeeding for their young infants. One IFPS II study found the perception that “breast milk alone did not satisfy my baby” was cited more often by Hispanic mothers and those with household incomes below 350% of the federal poverty level than white mothers and those with higher household incomes. ³⁴ However, mothers’ country of origin was not identified in this study. A qualitative study of low-income, Hispanic immigrant mothers from Mexico found that participants did believe breastfeeding was healthy for their infants, but that the addition of formula to the baby's diet was better because it offered the benefits of both breast milk and the added nutrients and vitamins of formula. ³⁵ Other reasons given by Hispanic mothers for nonexclusive breastfeeding in early infancy included the perception of insufficient supply of breast milk, belief that the combination of breastfeeding and formula keeps babies full longer and reduces crying, and that the addition of formula facilitates plans for a return to work. ³⁶

The practice of breastfeeding in conjunction with bottle feeding among low-income, Hispanic immigrant mothers may also be considered within the context of the life course perspective that a mother's infant feeding practices are shaped by her beliefs and cultural experiences. ³⁷ The PBHS mothers were immigrants from several countries of origin, including El Salvador (43.1%), Guatemala (11.3%), Honduras (16.4%), and Mexico (9.1%). Most were from countries identified by the World Bank classification of lower-middle income economies, defined as having a gross national income per capita (GNI) between $1,045 and $4,125. This compares with the GNI classification of high-income economy for the United States with a GNI >$12,736. ³⁸ Their average length of time in the United States before the birth of their infant was 6.17 years (SD 3.53). For 72.5% of PBHS mothers, Spanish was the only language spoken in their homes. Since many PBHS mothers grew up in geographical regions where food scarcity has impact on child development, the mixed method of both breast and bottle feeding may support a notion that good mothering is linked with the concept that “a chubby baby is a healthy baby.” In one qualitative study, low-income Hispanic mothers from Mexico reported that family members advised supplementation with formula to ensure a baby was gordito (affectionate term for a baby's overweight body habitus). ³⁵ Prior studies have found that Hispanic mothers endorse the concept that a “chubby baby is a healthy baby.”^39,40 However, one recent study found that Hispanic mothers largely did not endorse this notion. Researchers suggested that either a social desirability response bias or a recent shift in weight perception norms secondary to widespread publicity on the dangers of childhood obesity may explain this finding that is inconsistent with past studies. ⁴¹ Current maternal economic circumstances also may shape infant feeding practices. One resource available to PBHS mothers in this study was formula provided by the Supplemental Nutrition Program for WIC program. ⁴² Prior research has shown an increased consumption of formula at 6 months of age among WIC participants. ⁴³

A limitation of the present study was that the PBHS component did not have the national geographical representation of the IFPS II. Since all mother–infant participants were patients from the same clinical site, our results may not be generalizable to low-income, Hispanic immigrant mothers in other areas of the United States. A second limitation is that we were not able to quantify the amount of milk consumed by bottle and amount of solid food provided to the infant, as medical records did not include this level of detail. Future prospective studies may be able to address timing for introduction and amount of daily solid food consumption in conjunction with mode of milk delivery (breastfed only, mixed, bottle fed only) to more fully examine the impact of these feeding practices on subsequent weight status. Prior research with the IFPS II found that infants who consumed “human milk by bottle only” and “nonhuman milk by bottle only” gained more weight than those fed at breast only, but this “bottle effect” was not observed for infants who consumed “human and nonhuman milk by bottle.” The researchers concluded that bottle feeding as a mode of milk delivery might be an independent factor for infant weight gain. ²³ We were not able to make a comparison to this finding with the PBHS dyads as the content of the bottle was not specified in the medical record, only the mode of milk delivery.

Furthermore, comparisons of weight status at 12 months were based on measures of WFA, as has been done in prior research. ⁴⁴ Weight-for-length comparisons also would have contributed to the findings, but the IFPS II infant length data, gathered by means of maternal recall from the most recent well child visit, have been reported to be unreliable, so comparisons of weight-for-length were not feasible in the present study. ⁴⁵ Future studies may also examine rapid weight gain by 6 and 12 months of age as these have been found to be risk factors for later childhood obesity with other demographic groups.^46,47 In addition, we used WHO 2006 Child Growth Standards rather than the CDC 2000 growth charts to calculate weight status. The CDC now recommends that clinicians use this approach for children under 24 months of age. ²² While no comparable recommendations have been made for pediatric research, some studies have demonstrated similarities in the prevalence and prediction of early childhood overweight using CDC and WHO charts for US children.^48,49 A strength of using WHO growth percentile is the potential for clinical application of findings, as they are congruent with the screening for abnormalities in infant weight status that is routinely used by pediatric clinicians. ⁵⁰

Conclusions

There is growing evidence that the origins of childhood obesity begin early in the life course. Our findings demonstrate that infants of low-income, Hispanic immigrant mothers were more likely to breastfeed in early infancy, but less likely to do so exclusively than mothers from the national IFPS II. Subsequently, their infants were of higher weight status by 1 year of age. Additional studies are needed to more completely understand early life obesity risk factors for this growing proportion of the US population to design evidence-based clinical and public health interventions that reduce racial/ethnic health disparities in childhood obesity.