Human Milk Feeding and Ultraprocessed Food Intake in Toddlerhood

Abstract

Objective:

The objective of the study was to estimate associations between early-life human milk feeding and ultraprocessed food (UPF) intake at two timepoints during toddlerhood among children born at <35 weeks’ gestation.

Study Design:

Children were enrolled in the Omega Tots trial (2012–2017, Ohio) at 10–17 months’ corrected age after having discontinued human milk and formula feeding. Caregivers reported children’s human milk feeding history at baseline and past month diet through a food frequency questionnaire at baseline and follow-up (180 days later). We used the NOVA classification system to estimate UPF intake. We estimated covariate-adjusted associations between human milk feeding (ever and duration) and UPF intake at baseline and follow-up using linear and logistic regression.

Results:

Nearly 89% (n = 295) of 333 toddlers had received human milk but only 4.2% (n = 14) were fed exclusively human milk to 6 months of age. UPFs represented 37.7 (standard deviation [SD] = 13.2)% and 43.4 (SD = 11.3)% of total calories at the two timepoints. Human milk feeding (exclusive or otherwise) was unassociated with UPF intake in toddlerhood (e.g., months of exclusive human milk feeding with the number of daily servings of UPFs at follow-up: β = −0.09, 95% confidence interval [CI]: −0.26, 0.08).

Conclusion:

In this sample of toddlers born preterm, any exposure to as well as the duration of human milk feeding was unassociated with UPF intake during the second year of life. These results require replication in larger samples given the small number of children in some human milk feeding categories.

Introduction

Recent mounting evidence for the harmful health effects of consuming ultraprocessed foods (UPFs) has directed nutrition research attention toward evaluating diet on the basis of industrial processing. The NOVA (not an acronym) classification system—which classifies foods and beverages into four groups based on the nature and extent of processing—defines UPFs as formulations of ingredients made mostly or entirely from industrial substances, containing additives and little to no unprocessed/intact foods.^1,2 UPFs currently represent 58.2% and 66.2% of total energy intake for U.S. preschoolers and school-aged children, respectively.³ Higher UPF intake has been associated with greater adiposity,^4–9 as well as adverse lipid profiles¹⁰ in pediatric populations.

The benefits of breastfeeding are well-established with recently updated American Academy of Pediatrics (AAP) guidance recommending exclusive breastfeeding for 6 months and continued breastfeeding until 2 years or beyond.¹¹ Although this recommendation is based on the robust evidence for reduced risk of adverse infant and maternal health outcomes, a growing body of research also suggests that human milk feeding could be beneficial in promoting diet quality later in childhood.^12–21 This has been explained by two potential mechanisms. First, repeated exposure to mother’s breast milk introduces infants to a higher diversity of flavor and taste experiences than formula, which later influences their behavioral response (acceptance) of similar stimuli.^12,22 Second, breastfeeding promotes parent feeding styles that are responsive to infant cues enhancing the infant’s self-regulation of intake.²³ Thus, it could be hypothesized that children who were fed formula rather than human milk as well as those bottle fed (where caregivers are more likely to overfeed by promoting bottle emptying) may gravitate toward UPF food products, which largely are highly palatable and energy dense.

Among the studies examining the association between human milk feeding and later child diet, few have characterized diet in terms of levels of processing (UPF intake).^17,19–21 Children in many of these studies tended to be older (≥4 years), although improving diet quality as early as toddlerhood is a public health priority.^24,25 Importantly, human milk feeding typically was operationalized only in terms of duration, although other characteristics such as exclusivity and mode (both feeding at the breast and expressed milk) may be meaningful.

Children born preterm are a particularly high-risk group for breastfeeding problems because of their developmental and neurological immaturity. Moreover, infants born preterm are at increased risk of nutrition-related chronic diseases such as cardiovascular disease, hypertension, and diabetes later in life,²⁶ indicating that minimizing their risk factors (including dietary) across the lifespan is critical. Yet, studies on the link between early-life feeding and diet in childhood specifically among preterm children are scarce. The aims of the present study are to evaluate the association of multiple aspects of early-life human milk feeding with the UPF content of diet in toddlerhood in a sample of children born preterm.

Methods

Study sample

This study utilized data from the Omega Tots trial, a double-blind, randomized, placebo-controlled trial conducted at Nationwide Children’s Hospital in 2012–2017 (NCT01576783) designed to test the efficacy of dietary polyunsaturated omega fatty acid supplementation on the cognitive development of toddlers born preterm.²⁷ Children born at <35 weeks’ gestation were enrolled at 10 to 17 months’ of age (corrected for prematurity) (n = 377) and randomized to receive either a placebo or fatty acid supplement for 180 days. All children had a history of having been admitted to a neonatal intensive care unit (NICU) in Columbus, OH postbirth or of a neonatology clinic visit at Nationwide Children’s Hospital. Further inclusion criteria were discontinuation of human milk or formula consumption, weighing between 5th and 95th percentile for corrected age and sex, and the family’s primary language being English. Children were excluded if they consumed fatty acid supplements, fatty fish, or nutritional-support beverages containing DHA > 2 times per week. A legal guardian for each child provided written informed consent for participation. The study was reviewed and approved by the Institutional Review Board at Nationwide Children’s Hospital. Details about the trial have been published previously.²⁷ The original trial was registered on ClinicalTrials.gov (NCT01576783).

Exposures

Early-life feeding exposures were evaluated retrospectively through an interviewer-assisted questionnaire completed by caregivers at the baseline study visit. As children were <3 years of age at study entry, the recall period met recommendations for reliable assessment of breastfeeding history.²⁸ We defined human milk exposure as direct (at the breast) or indirect (expressed or pumped) feeding of human milk, whether mother’s own milk or donor milk. Exclusive human milk feeding was defined as not having been fed other foods, beverages, or formula during the time the child was fed human milk. We determined if toddlers ever received human milk (binary variable), the duration of human milk feeding and exclusive human milk feeding (continuous variables), and whether they consumed exclusively human milk up to 2, 4, and 6 months (binary variables).

Outcomes

A standard, semiquantitative food frequency questionnaire (FFQ) was used to estimate toddlers’ dietary intake.^29,30 Because the intervention tested in the trial was a supplement and not a food, it was not captured in the FFQ. On the date of enrollment and at 180 days postrandomization, the primary caregiver completed the FFQ with a 1-month recall period.

The NOVA classification system was used to determine which FFQ items were UPFs.³¹ Due to ambiguity about preparation methods for some FFQ items, three researchers (S.Y., C.R., K.E.) independently assigned each item on the FFQ to one of 4 NOVA food processing categories: (1) unprocessed and minimally processed foods, (2) processed culinary ingredients, (3) processed foods, and (4) ultraprocessed foods. After making their independent assignments, the researchers discussed any items that were discordant and came to a consensus. Items identified as UPFs are listed in Supplementary Table S1.

We quantified the average daily amount of UPFs in the diet in three ways: (1) servings of UPFs, (2) grams of UPFs, and (3) % energy from UPF. A single serving of each food was defined by the FFQ and the average servings of UPFs per day were calculated. Grams of UPF were determined by assigning each food item a weight per serving and multiplying the servings/day of that food by the weight in grams. The percent energy of UPFs was determined by assigning each food item an amount of energy and multiplying the energy of one serving by the number of servings per day. This was then divided by the total energy consumed by the child to get the percent energy from UPFs. Weight and energy assignments for FFQ items were determined by the developers of the FFQ at the Nutrition Questionnaire General Service Center at the Harvard T.H. Chan School of Public health and estimates from the U.S. Department of Agriculture (USDA) FoodData Central database.³²

Covariate data

Potential confounders of the association between early-life feeding exposures with diet in toddlerhood were selected a priori based on previous literature. Measured at baseline, these covariates included child’s sex, race, and ethnicity (because of large race-based disparities in the prevalence of preterm birth and in the outcomes of preterm birth), age in months (corrected for prematurity), gestational age at birth, as well as maternal age, education, marital status, annual household income, participation in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), and treatment arm.

Statistical analyses

We compared descriptive baseline characteristics of the analytic sample by exposure to any human milk feeding (ever fed human milk vs. never fed human milk) using t tests and chi-square tests for continuous and categorical variables, respectively.

Associations between human milk feeding variables and continuous variables reflecting UPF intake (servings/day, grams/day, and % energy) were estimated using linear regression after adjustment for potential confounders.

For consistency with previous literature on this topic,^17,19,21 we also examined an alternative measure of UPF intake categorized into tertiles (tertile 1: 0.22–2.96 servings/day, tertile 2: 2.98–5.27 servings/day, tertile 3: 5.28–13.05 servings/day). Covariate-adjusted associations of human milk feeding with tertiles of UPF intake were examined using logistic regression with the lowest tertile of UPF intake as the reference group. The median value of each tertile was fit as a continuous variable in the applicable regression models for tests of linear trend (p-trend).

Because toddlerhood is a time of rapid dietary change, we also examined whether human milk feeding exposures were associated with changes in UPF intake from baseline to follow-up using linear mixed models with an interaction of time with exposure variables. Since ∼16% of children were multiples (twins, triplets), we accounted for their statistical nonindependence using a random effect for family in the linear mixed models. All analyses were performed using SAS 9.4 software (SAS Institute, Cary, NC).

Sensitivity analysis

Given that UPF intake or level of food processing is a relatively new construct in nutritional science, we also examined associations of human milk feeding with the Toddler Diet Quality Index (TDQI), a food group and nutrient-based indicator of diet quality to compare the direction and magnitude of associations. The TDQI applies a scoring system (range: 0–100) to the toddler-specific diet recommendations under the 2020–2025 Dietary Guidelines for Americans (DGA), with higher scores representing better diet quality.³³

Results

The analytic sample included toddlers who had data on human milk feeding history as well as plausible dietary data at enrollment (n = 333) and follow-up (n = 284). Nearly 89% (n = 295) of the 333 were ever fed human milk, but only 60 (18.1%), 24 (7.2%), and 14 (4.2%) were exclusively human milk fed to 2, 4, and 6 months of age, respectively. Distributions of feeding categories combining duration, mode, and exclusivity are presented in Supplementary Table S2.

Characteristics of the analytic sample at baseline and comparisons among those who were ever fed human milk and those who were not are presented in Table 1. Most demographic and socioeconomic characteristics did not differ between exposure groups, although mothers with higher education appeared more likely (although not statistically significantly so) to have fed their child human milk than mothers with lower educational attainment (p = 0.07). Caregivers who fed their child human milk similarly appeared to be more likely to be married or living with their partner (70%), as compared with those who did not feed their child human milk (55%) (p = 0.07). Compared with children who were never fed human milk, those who were fed any human milk were born at a lower average birthweight (1692.2 [572] vs. 1888.1 [419.1], p = 0.04) and earlier gestational age (31.1 [2.9] vs. 32.5 [1.7], p = 0.09).

Table 1.

Baseline Characteristics of Toddlers Born Preterm by Exposure to Human Milk, Omega Tots Trial (2012–2017, n = 333, Columbus, Ohio)

Characteristics	Never fed human milk (n = 38)	Fed human milk (n = 295)	Full sample (n = 333)	p ^a
Child’s sex,%
Male	16 (42)	156 (53)	172 (52)	0.211
Female	22 (58)	139 (47)	161 (48)
Child’s race,%				0.515
White alone	21 (55)	191 (65)	212 (64)
Black alone	12 (32)	75 (25)	87 (26)
Other	5 (13)	29 (10)	34 (10)
Child’s ethnicity,%				0.462
Not Hispanic/Latino	37 (97)	279 (95)	316 (95)
Hispanic/Latino	1 (3)	16 (5)	17 (5)
Child’s birthweight, grams	1888.1 (419.1)	1692.2 (572)	1714.6 (559.6)	0.042
WIC participation, %				0.408
No	17 (45)	153 (52)	170 (51)
Yes	21 (55)	142 (48)	163 (49)
Maternal education, %				0.073
Less than high school	5 (14)	11 (4)	16 (5)
High school or equivalent	10 (28)	60 (20)	70 (21)
Some college	7 (19)	78 (27)	85 (26)
Associates degree	3 (8)	25 (9)	28 (9)
Bachelor’s degree	9 (25)	77 (26)	86 (26)
Postgraduate degree	2 (6)	42 (14)	44 (13)
Marital status, %				0.072
Not married or living with partner	17 (45)	89 (30)	106 (32)
Married or living with partner	21 (55)	205 (70)	226 (68)
Annual household income, USD				0.427
<20,000	12 (32)	60 (20)	72 (22)
25,000–35,000	9 (24)	70 (24)	79 (24)
35,000–45,468	7 (18)	59 (20)	66 (20)
>45,468	10 (26)	104 (35)	114 (34)
Treatment arm, %				0.775
Placebo	20 (53)	148 (50)	168 (50)
Fatty acid supplementation	18 (47)	147 (50)	165 (50)
Child’s gestational age at birth, weeks	32.5 (1.7)	31.1 (2.9)	31.2 (2.8)	0.003
Maternal age, years	32.5 (8.7)	30.6 (6.3)	30.8 (6.6)	0.093

Values represent mean (standard deviation) or number (percent). Missing data: Maternal education (n = 4), marital status (n = 1), annual household income (n = 2). Other race includes Asian/Pacific Islander, other race, and multiple race.

p value for mean or group differences by exposure to human milk.

Mean (standard deviation) UPF intake at baseline and follow-up were 4.4 (2.4) servings/day, 334.3 (249.6) grams/day, 37.7 (13.2)% of total energy, and 4.9 (2.2) servings/day, 376.1 (284.2) grams/day, 43.4 (11.3)% of total energy, respectively. Mean (standard deviation) TDQI scores at baseline and follow-up were 50.4 (9.8) and 48.1 (9.9), respectively.

Table 2 shows the associations of human milk feeding with continuous UPF intake (and TDQI) at baseline. Associations with diet measured at follow-up (180 days) are presented in Table 3. Effect estimates for associations of human milk feeding with UPF variables had confidence intervals that included the null, although some imprecise associations suggested slightly reduced percent energy from UPFs at baseline (β = −2.50 [95% CI: −7.10, 2.10]) and follow-up (β = −2.82 [95% CI: −7.04, 1.40]) among those ever exposed to human milk (compared with those never exposed). The TDQI, however, was strongly positively associated with any exposure to human milk (β = 5.09 [95% CI: 1.95, 8.23]) at baseline and at follow-up (β = 3.89 ([95% CI: 0.31, 7.46]). Longer duration of human milk feeding was also associated with slightly higher TDQI scores indicating better diet quality (β per month of feeding = 0.25 [95% CI: 0.003, 0.50)] at baseline and β per month of feeding = 0.29 [95% CI: 0.02, 0.57]).

Table 2.

Adjusted Associations for Human Milk Feeding and Baseline Ultraprocessed Food Intake in Toddlerhood, Omega Tots Trial (2012–2017, n = 333, Columbus, Ohio)

	β (95% CI)
	UPF servings/day	UPF grams/day	UPF % energy	Total_TDQI
Ever fed human milk
no (n = 38)	referent	referent	referent	referent
yes (n = 295)	0.08 (−0.77, 0.94)	−29.06 (−115.88, 57.75)	−2.50 (−7.10, 2.10)	5.09 (1.95, 8.23)
Exclusively human milk fed for 6 months	referent	referent	referent	referent
no (n = 319)	−0.06 (−1.38, 1.26)	19.71 (−114.84, 154.26)	−3.54 (−10.62, 3.54)	0.85 (−4.12, 5.82)
yes (n = 14)
Exclusively human milk fed for 4 months	referent	referent	referent	referent
no (n = 309)	−0.54 (−1.54, 0.45)	−4.59 (−106.60, 96.70)	−3.04 (−8.38, 2.31)	1.98 (−1.77, 5.73)
yes (n = 24)
Exclusively human milk fed for 2 months	referent	referent	referent	referent
no (n = 273)	−0.23 (−0.94, 0.48)	31.37 (−41.25, 103.99)	−0.04 (−3.87, 3.79)	1.54 (−1.14, 4.22)
yes (n = 60)
Duration of exclusive human milk feeding (months)	−0.07 (−0.25, 0.10)	6.25 (−11.27, 23.77)	−0.25 (−1.18, 0.68)	0.32 (−0.33, 0.96)
mean (SD) = 0.8 (1.6)
Duration of human milk feeding (months)	−0.0003 (−0.07, 0.07)	1.37 (−5.39, 8.14)	−0.10 (−0.46, 0.26)	0.25 (0.003, 0.50)
mean (SD) = 4.6 (4.6)

Associations at baseline with toddlers corrected age at 10–17 months. Covariates, measured at baseline, were child’s sex, race, ethnicity, prematurity corrected age, gestational age at birth, as well as maternal age, education, marital status, annual household income, WIC participation, and treatment arm.

UPF, ultraprocessed food; TDQI, toddler diet quality index; WIC, women, infants, and children.

Table 3.

Adjusted Associations for Human Milk Feeding and Follow-up Ultraprocessed Food Intake in Toddlerhood, Omega Tots Trial (2012–2017, n = 284, Columbus, Ohio)

	β (95% CI)
	UPF servings/day	UPF grams/day	UPF % energy	Total_TDQI
Ever fed human milk
no (n = 35)	referent	referent	referent	referent
yes (n = 245)	0.24 (−0.61, 1.08)	8.69 (−96.14, 113.52)	−2.82 (−7.04, 1.40)	3.89 (0.31, 7.46)
Exclusively human milk fed for 6 months
no (n = 272)	referent	referent	referent	Referent
yes (n = 12)	−0.45 (−1.76, 0.87)	−62.00 (−224.81, 100.81)	−1.72 (−8.30, 4.86)	−0.93 (−6.53, 4.68)
Exclusively human milk fed for 4 months
no (n = 263)	referent	referent	referent	referent
yes (n = 21)	−0.64 (−1.64, 0.36)	−63.57 (−187.17, 60.03)	−2.81 (−7.80, 2.18)	0.42 (−3.84, 4.68)
Exclusively human milk fed for 2 months
no (n = 230)	referent	referent	referent	referent
yes (n = 54)	−0.32 (−1.04, 0.41)	−33.05 (−122.95, 56.85)	−1.68 (−5.31, 1.95)	1.53 (−1.56, 4.62)
Duration of exclusive human milk feeding (months)
mean (SD) = 0.9 (1.6)	−0.09 (−0.26, 0.08)	−11.27 (−32.73, 10.19)	−0.53 (−1.39, 0.34)	0.28 (−0.46, 1.02)
Duration of human milk feeding (months)
mean (SD) = 4.9 (4.8)	−0.02 (−0.09, 0.04)	−2.13 (−10.17, 5.9)	−0.16 (−0.48, 0.16)	0.29 (0.02, 0.57)

Associations at baseline with toddlers corrected age at 16–22 months. Covariates, measured at baseline, were child’s sex, race, ethnicity, prematurity corrected age, gestational age at birth, as well as maternal age, education, marital status, annual household income, WIC participation, and treatment arm. UPF, ultraprocessed food; TDQI, toddler diet quality index; WIC, women, infants, and children.

Associations of human milk feeding with tertiles of dietary outcomes at baseline and follow-up showed similar results as continuous dietary outcomes (data not shown). All associations with UPF outcomes were null, but toddlers who were ever fed human milk had much higher odds of being in the highest tertile of TDQI score (tertile 1: 16.72–46.24, tertile 2: 46.24–54.23, tertile 3 54.23–76.22) at baseline (tertile 3 vs. 1: OR = 4.88 [95% CI: 1.47, 16.19], p-trend = 0.01). These associations were weaker with the passage of time; however, each additional month of human milk feeding was associated with only slightly higher odds of being in the highest tertile of TDQI score at follow-up (tertile 3 vs. 1: OR = 1.08 [95% CI: 1.00, 1.17], p-trend = 0.07).

Changes in UPF and TDQI by categorical human milk feeding variables are presented in Table 4. There were no significant interaction p-values, indicating that changes in dietary outcomes from baseline to follow-up did not differ by any exposure to human milk as well as by having been exclusively fed human milk up to 2, 4, and 6 months. For example, mean (standard error) change in UPF percent energy among toddlers who were ever fed human milk was 5.59 ± 0.87, whereas it was 6.72 ± 2.38 among their never fed human milk counterparts (p-for-interaction = 0.65). For the same groups, mean (standard error) change in TDQI was −2.87 ± 0.54 vs. −1.25 ± 1.47 (p-for-interaction =0.30). Similarly, changes in dietary outcomes by continuous human milk feeding variables were null (data not shown, all interaction p-values not significant).

Table 4.

Change from Baseline to 180-Days for Dietary Outcomes by Categorical Human Milk Feeding Exposures; Omega Tots Trial (2012–2017, Columbus, Ohio)

	Baseline	Day 180	Mean change (baseline to 180-days)	p-value*
UPF servings/day
Ever fed human milk				0.853
No	4.65 ± 0.50	5.07 ± 0.50	0.42 ± 0.45
Yes	4.57 ± 0.31	5.07 ± 0.31	0.51 ± 0.16
Exclusively human milk fed for 6 months				0.874
No	4.53 ± 0.31	5.02 ± 0.31	0.49 ± 0.16
Yes	5.13 ± 0.62	5.50 ± 0.61	0.37 ± 0.74
Exclusively human milk fed for 4 months				0.7977
No	4.57 ± 0.31	5.05 ± 0.31	0.48 ± 0.16
Yes	4.59 ± 0.52	5.22 ± 0.52	0.63 ± 0.56
Exclusively human milk fed for 2 months				0.774
No	4.59 ± 0.32	5.06 ± 0.32	0.47 ± 0.17
Yes	4.53 ± 0.40	5.11 ± 0.40	0.58 ± 0.35
UPF grams/day
Ever fed human milk				0.7164
No	355.34 ± 52.91	381.62 ± 58.42	26.28 ± 54.50
Yes	305.34 ± 33.04	352.71 ± 34.86	47.38 ± 19.98
Exclusively human milk fed for 6 months				0.5642
No	304.86 ± 30.89	351.86 ± 32.91	46.99 ± 19.77
Yes	393.45 ± 66.49	385.37 ± 81.73	−8.08 ± 93.24
Exclusively human milk fed for 4 months				0.6504
No	302.70 ± 33.01	348.97 ± 34.77	46.27 ± 19.51
Yes	391.78 ± 52.87	405.71 ± 62.48	13.93 ± 68.58
Exclusively human milk fed for 2 months				0.3845
No	296.87 ± 33.55	349.00 ± 35.34	52.13 ± 20.86
Yes	355.21 ± 41.49	365.91 ± 46.86	10.70 ± 42.74
UPF percent energy/day
Ever fed human milk				0.6534
No	40.24 ± 2.56	46.96 ± 2.63	6.72 ± 2.38
Yes	36.96 ± 1.59	42.55 ± 1.63	5.59 ± 0.87
Exclusively human milk fed for 6 months				0.5247
No	37.62 ± 1.58	43.22 ± 1.61	5.59 ± 0.83
Yes	34.21 ± 3.10	42.35 ± 3.22	8.14 ± 3.91
Exclusively human milk fed for 4 months				0.6601
No	37.64 ± 1.57	43.24 ± 1.60	5.61 ± 0.85
Yes	34.77 ± 2.63	41.74 ± 2.73	6.97 ± 2.98
Exclusively human milk fed for 2 months				0.4724
No	37.34 ± 1.61	43.32 ± 1.65	5.98 ± 0.91
Yes	37.56 ± 2.02	42.06 ± 2.12	4.49 ± 1.86
TOTAL TDQI
Ever fed human milk				0.3021
No	44.96 ± 1.89	43.71 ± 2.01	−1.25 ± 1.47
Yes	50.21 ± 1.22	47.34 ± 1.26	−2.87 ± 0.54
Exclusively human milk fed for 6 months				0.5774
No	49.65 ± 1.23	47.01 ± 1.26	−2.64 ± 0.52
Yes	49.35 ± 2.16	45.33 ± 2.47	−4.02 ± 2.41
Exclusively human milk fed for 4 months				0.3089
No	49.50 ± 1.23	46.95 ± 1.26	−2.56 ± 0.53
Yes	50.67 ± 1.90	46.16 ± 2.11	−4.51 ± 1.85
Exclusively human milk fed for 2 months				0.5827
No	49.56 ± 1.26	46.99 ± 1.29	−2.56 ± 0.57
Yes	49.76 ± 1.52	46.49 ± 1.64	−3.27 ± 1.15

Values indicate mean ± standard error unless indicated otherwise. UPF, ultraprocessed food; TDQI, toddler diet quality index.

p value for the interaction of time with exposure variable.

Discussion

In this secondary analysis of a nutrient-supplementation trial of toddlers that were born preterm, we found that UPF intake represented a high proportion of energy intake (37.7 [13.2]% and 43.4 [11.3]% at 10–15 and 16–22 months of age), highlighting that diets this early in childhood have room for improvement in terms of levels of processing. Early human milk feeding history exposure variables were generally unassociated with UPF intake at baseline and follow-up as well as the change in UPF intake between these timepoints. However, those who were ever fed human milk had notably higher diet quality as measured by the TDQI compared with those who were never fed human milk at both timepoints (β = 5.09 [95% CI: 1.95, 8.23] at baseline; β = 3.89 [95% CI: 0.31, 7.46] at follow-up), and each additional month of human milk feeding was associated with slightly higher TDQI scores.

Previous research on early-life feeding history and UPF intake in early childhood, all carried out in populations outside of the United States, have reported some favorable associations with human milk feeding. Fronseca et al. 2019 found that among Brazilian children 4–7 years of age, duration of exclusive breastfeeding (per month) was inversely associated with percent energy from UPFs (β: −0.7; 95% CI: −1.3 to 0.1; p = 0.038).¹⁷ In a sample of 6–24-month-old Brazilian children, Soares et al. 2021 reported that the absence of breastfeeding was associated with higher intake of processed and ultraprocessed foods.¹⁹ Similarly, Spanieol et al., 2020 reported that breast milk intake was associated with lower odds of consuming UPF (OR: 0.26; CI 95%: 0.09–0.74) among Brazilian children less than 2 years old.²⁰ Lastly, Garcia-Blanco et al. 2022 reported that among 4–5-year-old children in Spain, duration of breastfeeding exposure (<6 months, 6–12 months, >12 months) was associated with incrementally higher UPF intake compared no breastfeeding.²¹ Caution must be taken when comparing our results to previous studies because of the use of different dietary assessment methods (FFQ vs. 24-hour recall and food records), which can affect the categorization of UPFs. Furthermore, the geographic location of the sample could also affect the findings because maternal diet, which influences breast milk content, breastfeeding practices (mode, duration, exclusivity), and food items categorized as UPF (including their natural or added flavors) can all vary by region. Lastly, our study includes only children born preterm whereas these previous studies either excluded those born preterm¹⁹ or likely included few to none (inferred based on the distribution of birthweight when available). Parrerias et al. 2021 included a sample of mostly preterm children, although differences in the diversity of UPF consumed periodically were not significantly different by the use of complementary feeding before 6 months, and no further findings on adjusted associations were reported.

A mechanism through which longer (and any) human milk feeding is expected to improve diet is through exposing the infant to diverse taste profiles in human milk that will later facilitate acceptance of similar stimuli by the toddler. In our study, human milk feeding was not associated with higher UPF intake, but it was associated with higher diet quality as measured by the TDQI. Of note, the TDQI and percent energy from UPFs in our sample were significantly inversely correlated, but only moderately so. For example, while most refined carbohydrates would be considered UPFs and also count negatively on the TDQI score, several FFQ items that are categorized under healthy components of the TDQI (such as the total vegetables and dairy components) were identified as UPFs. It is plausible that since UPF items in the modern food environment have a wide range of flavors (including flavors of traditionally “healthy” food groups), this broad category of a food processing framework may not be a dietary measure where associations with human milk feeding would have a strong magnitude of association.

Our study has some limitations. First, maternal diet during human milk feeding (lactation) was not accounted for as maternal diet data were unavailable. As breast milk content and flavor exposures will be directly influenced by maternal diet, future studies designed to account for this covariate are needed. Second, FFQs were not designed to provide an adequate level of information about food preparation to unambiguously determine levels of food processing. However, with the NOVA classification system being a relatively new paradigm, all traditional measures of dietary assessment are subject to this limitation to some degree. A strength of our analysis is that we determined UPF items by independent assignment from three researchers and adjudication of discordant assignments. Lastly, the imprecise associations (wide confidence intervals) of UPF intake with some categorical exposures suggest that our analysis may have been underpowered. However, our total analytic sample was fairly large, detecting significant associations with the TDQI even in exposure categories with small sample sizes. Moreover, this study is the first on this topic with two timepoints in toddlerhood and exclusively based on a U.S. preterm sample, which is a novel contribution.

Conclusion

In conclusion, in a sample of U.S. toddlers born preterm, several measures of early-life human milk feeding (exclusive or otherwise) were not associated with UPF intake at two time points of toddlerhood. At the outset, these results may appear reassuring for caregivers of preterm infants whose early feeding decisions and ability to meet breastfeeding goals are often undermined by neonatal medical complications. However, the present findings must be balanced with known benefits of human milk feeding for short- and long-term health outcomes as well as positive associations of human milk feeding with other constructs and measures of healthy diets, including diet quality scores like the TDQI, which we found to be strongly associated with any human milk feeding in our sample. The small sample size in certain categories of human milk feeding must also be considered, highlighting the need for future studies in much larger samples.

Footnotes

Acknowledgments

The authors thank the families who participated in our research, Kelly M. Boone, MA; Mark A. Klebanoff, MD; Abigail Norris Turner, PhD; Joseph Rausch, PhD; Mary Ann Nelin, MD; Lynette K. Rogers, PhD; Keith Owen Yeates, PhD; Leif Nelin, MD; and Kelly W. Sheppard, PhD; Seanceray Bellinger, BS; Holly Blei, BS; Ashlea Braun, MS, RDN, LD; Anne Brown, BS; Lautaro Cabrera, BA; Chelsea Dillon, BS; Eva Fabian, MPH; Connor Grannis, BA; Rachel Haeuptle; Nathan Hanna, BS, MSN, RN; Chenali Jayadeva, BS; Sarah Landry, MPH; Julia Less, MPH; Cara Lucke, BS; Melissa Kwitowski, BA; Joseph Macklin, BA; Krista McManus, BA, MS; Emily Messick, BS; Yvette Noah, BS; Grace Pelak, BS; Evan Plunkett, MD; John Rissell, BS; Rachel Ronau, DO; Ashley Ronay, MPH; Katie Smith, BS; Sarah Snyder, BS; Reena Oza-Frank, PhD; Kamma Smith; and Justin Jackson for their contributions.

Authors’ Contributions

S.F.Y.: conceptualization (equal); data curation (equal); writing—original draft (lead); formal analysis (lead); and writing—review and editing (equal). C.A.R.: writing—original draft (supporting); data curation (equal); and writing—review and editing (equal). K.M.E.: writing—original draft (supporting); data curation (equal); and writing—review and editing (equal). S.A.K.: conceptualization (equal); data curation (equal); funding acquisition (lead); project administration (lead); supervision (lead); formal analysis (supporting); and writing—review and editing (equal).

Disclosure Statement

The study sponsors had no role in the study design, the collection, analysis and interpretation of data, the writing of the report, and the decision to submit the article for publication. The authors have no relevant conflicts of interest to disclose. Samrawit Yisahak wrote the first draft of the article.

Funding Information

This project was supported by the Health Resources and Services Administration (HRSA) of the U.S. Department of Health and Human Services (HHS) under Grant Number R40MC28316 and title “Nutrition as Intervention: Trial of PUFA to Boost Development of Preterm Toddlers” for grant amount $296,181. This information or content and conclusions are those of the author and should not be construed as the official position or policy of, nor should any endorsements be inferred by HRSA, HHS, or the U.S. Government. This project was also supported by grant R01HD100493 from the National Institutes of Health, 12-FY14-171 from the March of Dimes, a grant from the Allen Foundation, a grant from Cures Within Reach, Grant Number UL1TR00170 from the National Center for Advancing Translational Sciences/National Institutes of Health, and internal support from The Research Institute at Nationwide Children’s Hospital.

Ethical Standards Disclosure

This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving research study participants were approved by the Nationwide Children’s Hospital Institutional Review Board. Written informed consent was obtained from all subjects/patients.

Supplementary Material

References

Monteiro

, Cannon

, Levy

, et al. NOVA. The star shines bright. World Nutrition, 2016; 7(1–3):28–38.

Monteiro

, Cannon

, Moubarac

J-C

, et al. Nutrition and health. The issue is not food, nor nutrients, so much as processing. Public Health Nutr, 2009; 21(3):497–501; doi: 10.1017/S1368980009005291

Neri

, Martinez‐Steele

, Monteiro

, et al. Consumption of ultra‐processed foods and its association with added sugar content in the diets of US children, NHANES 2009‐2014. Pediatr Obes, 2019; 14(12):e12563; doi: 10.1111/ijpo.12563

Bawaked

, Fernández‐Barrés

, Navarrete‐Muñoz

, et al. Impact of lifestyle behaviors in early childhood on obesity and cardiometabolic risk in children: Results from the Spanish INMA birth cohort study. Pediatr Obes, 2020; 15(3):e12590; doi: 10.1111/ijpo.12590

Costa

, Assunção

MCF

, Loret de Mola

, et al. Role of ultra-processed food in fat mass index between 6 and 11 years of age: A cohort study. Int J Epidemiol, 2021; 50(1):256–265; doi: 10.1093/ije/dyaa141

Vedovato

, Vilela

, Severo

, et al. Ultra-processed food consumption, appetitive traits and BMI in children: A prospective study. Br J Nutr, 2021; 125(12):1427–1436; doi: 10.1017/S0007114520003712

Khandpur

, Neri

, Monteiro

, et al. Ultra-processed food consumption among the paediatric population: An overview and call to action from the European childhood obesity group. Ann Nutr Metab, 2020; 76(2):109–113; doi: 10.1159/000507840

Costa

, Rauber

, Leffa

, et al. Ultra-processed food consumption and its effects on anthropometric and glucose profile: A longitudinal study during childhood. Nutr Metab Cardiovasc Dis, 2019; 29(2):177–184; doi: 10.1016/j.numecd.2018.11.003

Chang

, Khandpur

, Neri

, et al. Association between childhood consumption of ultraprocessed food and adiposity trajectories in the avon longitudinal study of parents and children birth cohort. JAMA Pediatr, 2021; 175(9):e211573-e211573; doi: 10.1001/jamapediatrics.2021.1573

10.

Rauber

, Campagnolo

PDB

, Hoffman

, et al. Consumption of ultra-processed food products and its effects on children’s lipid profiles: A longitudinal study. Nutr Metab Cardiovasc Dis, 2015; 25(1):116–122; doi: 10.1016/j.numecd.2014.08.001

11.

Meek

, Noble

, Section on Breastfeeding. Policy statement: Breastfeeding and the use of human milk. Pediatrics, 2022; 150(1); doi: 10.1542/peds.2022-057988

12.

Ventura

. Does breastfeeding shape food preferences links to obesity. Ann Nutr Metab, 2017; 70 Suppl 3(Suppl. 3):8–15; doi: 10.1159/000478757

13.

Burnier

, Dubois

, Girard

. Exclusive breastfeeding duration and later intake of vegetables in preschool children. Eur J Clin Nutr, 2011; 65(2):196–202; doi: 10.1038/ejcn.2010.238

14.

Jackson

, Johnson

. Does breast-feeding reduce offspring junk food consumption during childhood? Examinations by socio-economic status and race/ethnicity. Public Health Nutr, 2017; 20(8):1441–1451; doi: 10.1017/S1368980016003517

15.

Specht

, Rohde

, Olsen

, et al. Duration of exclusive breastfeeding may be related to eating behaviour and dietary intake in obesity prone normal weight young children. PLoS One, 2018; 13(7):e0200388; doi: 10.1371/journal.pone.0200388

16.

Husk

, Keim

. Breastfeeding and dietary variety among preterm children aged 1–3 years. Appetite, 2016; 99:130–137; doi: 10.1016/j.appet.2016.01.016

17.

Fonseca

, Ribeiro

SAV

, Andreoli

, et al. Association of exclusive breastfeeding duration with consumption of ultra-processed foods, fruit and vegetables in Brazilian children. Eur J Nutr, 2019; 58(7):2887–2894; doi: 10.1007/s00394-018-1840-9

18.

ParreirasI

EEF

, LimaI

, PereiraI

RCSS

, et al. Breastfeeding practices, consumption of ultra-processed foods in complementary feeding and associated factors in premature children. Ciência e Natura, 2021; 43:e20–e20; doi: 10.5902/2179460X43692

19.

Soares

, Juvanhol

, Ribeiro

SAV

, et al. Prevalence of processed and ultra-processed food intake in Brazilian children (6–24 months) is associated with maternal consumption and breastfeeding practices. Int J Food Sci Nutr, 2021; 72(7):978–988; doi: 10.1080/09637486.2021.1885016

20.

Spaniol

, Da Costa

THM

, Bortolini

, et al. Breastfeeding reduces ultra-processed foods and sweetened beverages consumption among children under two years old. BMC Public Health, 2020; 20(1):330–339; doi: 10.1186/s12889-020-8405-6

21.

García-Blanco

, de la O Pascual

, Berasaluce

, et al. Individual and family predictors of ultra-processed food consumption in Spanish children. The SENDO project. Public Health Nutr, 2022:1–22; doi: 10.1017/S136898002200132X

22.

Forestell

, Mennella

. Early determinants of fruit and vegetable acceptance. Pediatrics, 2007; 120(6):1247–1254; doi: 10.1542/peds.2007-0858

23.

Taveras

, Scanlon

, Birch

, et al. Association of breastfeeding with maternal control of infant feeding at age 1 year. Pediatrics, 2004; 114(5):e577–e583; doi: 10.1542/peds.2004-0801

24.

Stoody

, Spahn

, Casavale

. The Pregnancy and Birth to 24 Months Project: A series of systematic reviews on diet and health. Am J Clin Nutr, 2019; 109(Suppl_7):685S–697S; doi: 10.1093/ajcn/nqy375

25.

Oliver Olid

, de la

, Bueso

, et al. Longer Breastfeeding Duration is Associated With Lower Consumption of Ultraprocessed Foods in a Sample of Spanish Preschoolers: The SENDO Project. J Acad Nutr Diet, 2023; 123(11):1555–1567; doi: 10.1016/j.jand.2023.05.028

26.

Luu

, Katz

, Leeson

, et al. Preterm birth: Risk factor for early-onset chronic diseases. CMAJ, 2016; 188(10):736–746; doi: 10.1503/cmaj.150450

27.

Keim

, Boone

, Klebanoff

, et al. Effect of docosahexaenoic acid supplementation vs placebo on developmental outcomes of toddlers born preterm: A randomized clinical trial. JAMA Pediatr, 2018; 172(12):1126–1134; doi: 10.1001/jamapediatrics.2018.3082

28.

, Scanlon

, Serdula

. The validity and reliability of maternal recall of breastfeeding practice. Nutr Rev, 2005; 63(4):103–110; doi: 10.1111/j.1753-4887.2005.tb00128.x

29.

Willett

, Sampson

, Stampfer

, et al. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol, 1985 1985;122(1):51–65; doi: 10.1093/oxfordjournals.aje.a114086,

30.

Parrish

, Marshall

, Krebs

, et al. Validation of a food frequency questionnaire in preschool children. Epidemiology, 2003; 14(2):213–217; doi: 10.1097/01.EDE.0000041256.12192.23

31.

Monteiro

, Cannon

, Levy

, et al. Ultra-processed foods: What they are and how to identify them. Public Health Nutr, 2019; 22(5):936–941; doi: 10.1017/S1368980018003762

32.

U.S. Department of Agriculture. FoodData Central (fdc.nal.usda.gov). 2019.

33.

Kay

, Duffy

, Harnack

, et al. Development and Application of a Total Diet Quality Index for Toddlers. Nutrients, 2021; 13(6):1943; doi: 10.3390/nu13061943

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.39 MB