Development of Food Pattern Recommendations for Infants and Toddlers 6–24 Months of Age to Support the Dietary Guidelines for Americans, 2020–2025

Abstract

Background

Developing food-based dietary guidelines (FBDGs) for infants and toddlers is a complex task that few countries have attempted.

Objectives

Our objectives are to describe the process of food pattern modeling (FPM) conducted to develop FBDGs for the Dietary Guidelines for Americans, 2020–2025 for infants 6 to <12 mo and toddlers 12 to <24 mo of age, as well as the implications of the results and areas needing further work.

Methods

The US 2020 Dietary Guidelines Advisory Committee, with the support of federal staff, conducted FPM analyses using 5 steps: 1) identified energy intake targets; 2) established nutritional goals; 3) identified food groupings and expected amounts, using 3 options for the amount of energy from human milk in each age interval; 4) estimated expected nutrient intakes for each scenario, based on nutrient-dense representative foods; and 5) evaluated expected nutrient intakes against nutritional goals.

Results

For human milk–fed infants (and toddlers), example combinations of complementary foods and beverages were developed that come close to meeting almost all nutrient recommendations if iron-fortified infant cereals are included at 6 to <12 mo of age. These combinations would also be suitable for formula-fed infants. For toddlers not fed human milk, 2 patterns were developed: the Healthy US-Style Pattern and the Healthy Vegetarian Pattern (a lacto-ovo vegetarian pattern). Achieving nutrient recommendations left virtually no remaining energy for added sugars.

Conclusions

It is challenging to meet all nutrient needs during these age intervals. Added sugars should be avoided for infants and toddlers <2 y of age. Further work is needed to 1) establish a reference human milk composition profile, 2) update and strengthen the DRI values for these age groups, and 3) use optimization modeling, in combination with FPM, to identify combinations of foods that meet all nutritional goals.

Introduction

Establishing healthy dietary patterns in infancy and early childhood is crucial to support immediate needs for growth and development and to promote lifelong health by helping to reduce the risk of obesity and associated cardiometabolic disorders later in life (1, 2). However, developing evidence-based dietary guidelines for infants and toddlers is not a simple task, in part because the scientific evidence for associations of dietary intake to health outcomes is relatively scant for this age group for most topics (2) (Part D: Chapters 4–6). The Scientific Report of the 2020 Dietary Guidelines Advisory Committee (DGAC) (2) is the first DGAC report to extensively review the period from birth to age 24 mo, as mandated by federal legislation, the Agricultural Act of 2014 [aka the 2014 Farm Bill (3)]. This enabled the inclusion of food-based dietary guidelines (FBDGs) for infants and toddlers 6 to <24 mo of age in the Dietary Guidelines for Americans, 2020–2025 (4).

Several different approaches to developing FBDGs exist, including single- or multi-objective optimization modeling, food pattern modeling (FPM), and a combination of these methods. Optimization modeling is a mathematical approach that optimizes an “objective function” (e.g., minimizes cost or total energy of the diet) while complying with various constraints (e.g., nutrient requirements), usually via linear or goal programming (5., 6., 7.). FPM is a way to develop nutritionally adequate dietary patterns based on amounts of foods to be consumed from each of a variety of food groups and subgroups, usually informed by observed dietary intakes in the target population (2). Other countries have used FPM (e.g., Ireland) (8), or a combination of FPM and optimization modeling (e.g., Australia) (9), to develop FBDGs for young children.

In the United States, FPM is used to develop the USDA Food Patterns, which are aimed at meeting the DRIs and Dietary Guidelines for Americans recommendations, within energy needs, for each age-sex group. In these models, the key elements that comprise a healthy dietary pattern are fairly consistent from age 2 y onward. The time period between birth and 24 mo, however, is characterized by major changes in feeding patterns and dietary intake. Exclusive breastfeeding is recommended for about the first 6 mo (4). For infants who are not fed human milk, or are mixed-fed (i.e., both human milk and infant formula), commercial infant formula is recommended until 12 mo of age (10). The transition from sole consumption of human milk and/or infant formula to a varied diet that includes nutrient-dense complementary foods and beverages (CFBs) is recommended to begin at ∼6 mo of age (10). Thus, the 2020 DGAC concluded that USDA Food Patterns are not necessary for infants younger than age 6 mo and began modeling exercises at age 6 mo when CFBs start to become an important part of the diet (11, 12).

FPM for infants and toddlers 6 to <24 mo of age must take into account the primary milk source because human milk, infant formula, and cow milk differ widely in nutrient content and bioavailability (13, 14). Thus, the nutrients from CFBs needed to meet the DRIs vary by milk source. In addition, FPM should be conducted separately by age group (6 to <12 mo; 12 to <24 mo), for several reasons. First, 6 to <12 mo of age is a time when infants are learning to eat new foods, so the variety, amounts, and textures of CFBs increase and change substantially during those 6 mo. Second, most of the DRI values for 6 to <12 mo are Adequate Intake (AI) estimates, with RDAs established only for protein, iron, and zinc, whereas RDAs are established for most nutrients for ages 12 mo and older (15., 16., 17., 18., 19., 20., 21., 22.). Third, infant formula is not recommended after 12 mo of age, and most infants in the United States (66%) are no longer receiving human milk after 12 mo of age (23), so the predominant milk source in the second year of life is cow milk.

The complementary feeding period is important not only for providing essential nutrients, but also for introducing infants and toddlers to various types of CFBs that can be beneficial to health and development. For example, certain foods should be introduced before age 12 mo to reduce the risk of food allergies (e.g., peanut, egg) (2) (Part D: Chapter 5), and this can be taken into account during FPM. Other important aspects of complementary feeding, however, are generally not addressed during FPM, such as 1) providing foods in different textures and forms to help develop manual dexterity, hand-eye coordination, and dexterity required for chewing and swallowing; and 2) implementing responsive feeding practices, modeling of healthy eating behaviors, and bonding through food and mealtimes.

The main objective of this article is to describe the FPM conducted by the 2020 DGAC, with support of USDA and U.S. Department of Health and Human Services federal staff, to support the development of FBDGs for infants and toddlers in the United States (2) (Part D: Chapter 7). We also discuss the implications of the results and identify areas for further work.

Methods

The analyses described herein were informed by the process to establish and model USDA Food Patterns for ages 2 y and older. The methods for establishing the USDA Food Patterns for ages 2 y and older have been described in detail elsewhere (2) (Part D: Chapter 14) (24). This article highlights the modifications made to reflect the unique feeding aspects of the population 6 to <24 mo of age. The food groups and subgroups used for developing USDA Food Patterns are as follows:

• Fruits;

• Vegetables: dark green, red and orange, beans and peas, starchy, and other;

• Dairy, including calcium-fortified soy beverages [among plant-based milk alternatives, only fortified soy beverage is currently considered a dairy equivalent (4)];

• Grains: whole grains and refined grains; and

• Protein foods: meats, poultry, and eggs; seafood; nuts, seeds, and soy products.

For each food group or subgroup, nutrient profiles used in FPM are based on a consumption-weighted average of nutrient-dense forms of foods. Nutrient-dense representative foods are defined as those within each item cluster in forms with the least amounts of added sugars, sodium, and solid fats (2) (Part D: Chapter 14).

Figure 1 shows the steps in FPM for infants and toddlers 6 to <24 mo of age. The Committee established 5 energy intake levels for this age group (600–1000 kcal/d, in 100-kcal increments); established the nutritional goals for the FPM analyses; identified the food groupings and expected amounts, using 3 options for the amount of energy from human milk in each age interval; estimated the expected nutrient intakes from each food group for each scenario, based on nutrient-dense representative foods; and evaluated the expected nutrient intakes against the nutritional goals. The age intervals used were 6.0–8.9, 9.0–11.9, and 12.0–23.9 mo; for simplicity these are labeled as 6 to <9, 9 to <12, and 12 to <24 mo throughout.

Figure 1.

Steps used in food pattern modeling for infants and toddlers ages 6 to <24 mo. AI, Adequate Intake; CFB, complementary food and beverage; WWEIA, What We Eat in America.

For infants 6 to <12 mo of age, the first goal was to identify combinations of CFBs that would meet the nutrient needs of infants whose milk source is human milk (i.e., no infant formula). Infant formula is fortified, so formula-fed infants generally have higher intakes of certain nutrients (e.g., iron) than infants not receiving formula. Based on dietary intake data, achieving adequate iron and zinc intakes at ages 6 to <12 mo for infants fed human milk was identified by the Committee as a major challenge and, hence, was a key focus of the Committee's work aimed at combinations of CFBs during that age range. The CFB combinations included seafood, eggs, and nut products (in small amounts), in accordance with recommendations to introduce these foods during this age period (2) (Part D: Chapter 5). In addition, a maximum for dairy (no more than 0.5 cup eq/d) was set, given that infants at this age are receiving human milk or infant formula. After developing combinations of CFBs for infants fed human milk, the Committee then estimated the expected nutrient intakes of infants fed infant formula if they consumed the same types and combinations of CFBs.

For toddlers 12 to <24 mo of age, the Committee first conducted FPM for toddlers fed neither human milk nor infant formula (including those fed a lacto-ovo vegetarian diet), and then examined combinations of CFBs for toddlers fed human milk.

For both age intervals, the Committee was guided by the principle that CFBs should be nutrient-rich, particularly in nutrients for which potential risk of inadequacy exists, while also limiting exposures and intakes of other food components when they are of concern, such as added sugars.

Results

Infants 6 to <12 mo old fed human milk

In the first model, with food group amounts in proportion to the amounts in the 1000-kcal Pattern for ages 2 y and older (25), numerous nutrient gaps were evident for both 6 to <9 and 9 to <12 mo. Gaps existed for iron and zinc, as expected, but also for several other micronutrients. The iron content of this first model was only ∼1–2 mg/d at 6 to <9 mo and 1–4 mg/d at 9 to <12 mo (far below the RDA of 11 mg), and zinc content was 1.4–2.5 mg/d at 6 to <9 mo (below the RDA of 3 mg) and 2–4 mg/d at 9 to <12 mo (closer to the RDA of 3 mg). Thus, the second step was to replace 56 kcal of grains with 56 kcal of fortified infant cereal (0.5 oz eq). For iron, this second model included ∼8–9 mg/d at 6 to <9 mo and ∼8–11 mg/d at 9 to <12 mo. These amounts were closer to the RDA, but still below it for most energy levels and human milk proportion options. For zinc, this second model included 3–5 mg/d, which was adequate. The energy remaining for other CFBs, after including 56 kcal of fortified infant cereal (Table 1), was 0–224 kcal/d at 6 to <9 mo and 124–484 kcal/d at 9 to <12 mo. The lower amounts in these ranges correspond to options with high levels of human milk intake.

Table 1.

Energy (kcal) provided by HM or IF plus 0.5 oz equivalents of fortified infant cereal and remaining energy available for other CFBs for infants, by age and 3 levels of HM or IF intake¹

Energy level, kcal		600			700			800			900			1000
Age, mo	Energy source	H	A	L	H	A	L	H	A	L	H	A	L	H	A	L
6 to <9	HM (or IF)	600	480	390	700	560	455	800	640	520
	Total CFBs	02	120	210	02	140	245	02	160	280
	Infant cereal	0	56	56	0	56	56	0	56	56
	Remaining CFBs	0	64	154	0	84	189	0	104	224
9 to <12	HM (or IF)	420	330	240	490	385	280	560	440	320	630	495	360
	Total CFBs	180	270	360	210	315	420	240	360	480	270	405	540
	Infant cereal	56	56	56	56	56	56	56	56	56	56	56	56
	Remaining CFBs	124	214	304	154	259	364	184	304	424	214	349	484
12 to <24	HM (or IF)				350	245	140	400	280	160	450	315	180	500	350	200
	Total CFBs				350	455	560	400	520	640	450	585	720	500	650	800
	Infant cereal				56	56	56	56	56	56	56	56	56	56	56	56
	Remaining CFBs				294	399	504	344	464	584	394	529	664	444	594	744

¹Energy from HM or IF was modeled at 3 levels (low, average, and high) applied to each age interval (6 to <9 mo, 9 to <12 mo, and 12 to <24 mo). The average level was based on the mean percentage of total energy requirements expected to come from HM at those ages, using data from published studies conducted in high-income countries (12), which indicated average milk volumes of 688, 529, and 448 mL/d at 6 to <9, 9 to <12, and 12 to <24 mo, respectively (28). The low and high levels were set at 15% lower and 15% higher than the mean, respectively. For the modeling exercises for infants fed IF at ages 6 to <9 mo and 9 to <12 mo, the proportion of total energy expected to come from IF was the same as for HM. This table represents a third step to examine how much energy remained available for other CFBs, after including 56 kcal/d (0.5 oz eq/d) of fortified dry infant cereal. A, average; CFB, complementary food and beverage; H, high; HM, human milk; IF, infant formula; L, low.

²With high HM or IF intake modeled as described in footnote 1, no energy remains for CFBs at 6 to <9 mo; however, introduction of CFBs is recommended at ∼6 mo of age.

The final models (Table 2) reflect the allocation of energy to CFBs across food groups and subgroups to fill nutrient gaps as much as possible, particularly for iron, zinc, potassium, and choline. These combinations include a relatively high proportion of protein foods, particularly meat owing to its high content and bioavailability of iron and zinc. Seafood, eggs, and nuts are included to adhere to recommendations to introduce such foods during infancy.

Table 2.

Approximate amounts of food groups and subgroups in example combinations of CFBs for infants and toddlers ages 6 to <12 mo and 12 to <24 mo¹

	For infants and toddlers receiving human milk2						For toddlers 12 to <24 mo old no longer receiving human milk or infant formula
	6 to <9 mo		9 to <12 mo		12 to <24 mo		Healthy US-Style Pattern		Healthy Vegetarian Pattern
Food groups	Daily	Weekly	Daily	Weekly	Daily	Weekly	Daily	Weekly	Daily	Weekly
Total fruits, cup eq	$\frac{1}{8}\text{}-\text{}\frac{1}{4}$	—	$\frac{1}{8}\text{}-\text{}\frac{1}{2}$	—	$\frac{1}{3}\text{}-\text{}\frac{3}{4}$	—	$\frac{1}{2}\text{}-\text{}1$	—	$\frac{1}{2}\text{}-\text{}1$	—
Total vegetables, cup eq	$\frac{1}{8}\text{}-\text{}\frac{1}{4}$	—	$\frac{1}{8}\text{}-\text{}\frac{1}{2}$	—	$\frac{2}{3}$	—	$\frac{2}{3}\text{}-\text{}1$	—	1	—
Red and orange	—	$\frac{1}{4}\text{}-\text{}\frac{2}{3}$	—	$\frac{1}{2}\text{}-\text{}1\frac{1}{2}$		$1\frac{1}{2}$	—	$1-2\frac{1}{2}$	—	$2\frac{1}{2}$
Starchy	—	$\frac{1}{4}\text{}-\text{}\frac{1}{2}$	—	$\frac{1}{3}\text{}-\text{}1$		$\frac{1}{2}\text{}-\text{}\frac{3}{4}$	—	1–2	—	2
Dark green	—	$<\frac{1}{8}$	—	$\frac{1}{4}$		$1-1\frac{1}{2}$	—	$\frac{1}{3}-1$	—	$\frac{1}{2}$
Legumes	—	$<\frac{1}{8}$	—	$\frac{1}{4}$		$\frac{1}{2}$	—	$\frac{1}{3}\text{}-\text{}\frac{3}{4}$	—	$\frac{3}{4}$
Other	—	$\frac{1}{4}\text{}-\text{}\frac{1}{2}$	—	$\frac{1}{4}\text{}-\text{}\frac{3}{4}$		$\frac{3}{4}$	—	$\frac{3}{4}-1\frac{1}{2}$	—	$1\frac{1}{2}$
Total grains, oz eq	$\frac{1}{2}\text{}-\text{}\frac{3}{4}$	—	$\frac{1}{2}-1$	—	$1\frac{1}{4}-2\frac{1}{4}$		$1\frac{3}{4}-3$	—	$1\frac{3}{4}-3$	—
Fortified infant cereals (dry)	$\frac{1}{2}$	—	$\frac{1}{2}$	—	—	—	—	—	—	—
Whole grains	3	—	3	—	3	—	$1\frac{1}{2}-2$	—	$1\frac{1}{4}-2$	—
Total protein foods,4 oz eq	$\frac{3}{4}\text{}-\text{}2\frac{2}{3}$	—	2–3	—	$2\frac{1}{4}-3$	—	2	—	1	—
Meats	—	$4\frac{2}{3}-16$	—	$8\frac{1}{2}-15\frac{1}{2}$	—	$9\frac{1}{4}-15\frac{3}{4}$	—	$7–8\frac{3}{4}$5	—	—
Poultry	—	$\frac{1}{2}-1\frac{1}{4}$	—	1	—	1–3	—	$7–8\frac{3}{4}$5	—	—
Seafood	—	<1	—	≥3	—	≥3	—	3	—	—
Eggs	—	<1	—	≥1	—	≥1	—	$2-2\frac{3}{4}$	—	$3\frac{1}{2}$
Nuts and seeds	—	<1	—	$\ge \frac{1}{2}$	—	$\ge \frac{1}{2}$	—	$1-1\frac{1}{4}$	—	4
Total dairy,6 cup eq	$\frac{1}{4}$	—	$\frac{1}{2}$	—	$\frac{1}{4}-1\frac{3}{4}$	—	$1\frac{2}{3}-2$	—	$1\frac{1}{2}-2$	—
Total added oils/fats, g	0	—	$0–7\frac{3}{4}$	—	2–11	—	8–13	—	$8\frac{1}{2}-15$	—

¹The amounts shown represent the quantities of food items (cup or oz eq) that infants ages 6 to <12 mo could consume as CFBs from different food groups and subgroups to approach nutrient adequacy for iron, zinc, potassium, and choline (the nutrients with the most critical gaps) within the energy allocation for CFBs for their age group (0–224 kcal at 6 to <9 mo and 124–484 kcal at 9 to <12 mo). Energy from human milk or infant formula was modeled at 3 levels (low, average, and high) applied to each age interval (6 to <9 mo, 9 to <12 mo, and 12 to <24 mo). The average level was based on the mean percentage of total energy requirements expected to come from human milk at those ages, using data from published studies conducted in high-income countries (12) and the low and high levels were set at 15% lower and 15% higher than the mean, respectively (see Table 1). CFB, complementary food and beverage.

²Can also be applied to infants fed infant formula ages 6 to <12 mo.

³At least half of grains (from sources other than fortified infant cereals) should be whole grains.

⁴Total protein foods includes a majority from meats rather than poultry because meat has higher iron content than poultry. The weekly amounts of seafood, eggs, and nuts and seeds represent minimum amounts; greater quantities from these subgroups may be accommodated within the quantities allocated to total protein foods and the energy allocation for CFBs for this age group.

⁵Combined value for meats plus poultry.

⁶Not including human milk or infant formula; could include calcium-fortified soy beverages [among plant-based milk alternatives, only fortified soy beverage is currently considered a dairy equivalent (4)].

These combinations come close to meeting almost all nutrient recommendations (Table 3). However, projected potassium intakes fall short of 90% of the AI at several energy levels. In all of these models, no energy remains for added sugars (other than the small amounts of added sugars present in some of the foods in the nutrient profile, such as certain breakfast cereals). The percentage of energy from fat in these models is 41%–44% at 6 to <9 mo and 35%–42% at 9 to <12 mo. The percentage of energy from protein is 11%–16% at 6 to <9 mo and 16%–19% at 9 to <12 mo.

Table 3.

Summary of energy and select nutrient amounts as a percentage of the RDA or AI in example combinations of complementary foods and beverages for 6 to <12 mo and Patterns for 12 to <24 mo¹

						For toddlers 12 to <24 mo no longer receiving HM or IF
	For infants and toddlers receiving HM or IF					Healthy	Healthy Vegetarian
	6 to <9 mo		9 to <12 mo		12 to <24 mo	US-Style Pattern	Pattern
Infant milk source	HM1	IF1	HM	IF	HM	—	—
Energy range modeled, kcal/d	600–800	600–800	600–900	600–900	700–1000	700–1000	700–1000
Protein, % RDA	144–264	158–279	226–334	238–345	237–360	288–346	240–308
Iron, % RDA	77–90	150–186	84–109	134–194	74–109	88–120	89–1262
Zinc, % RDA	103–170	226–302	143–217	232–339	162–254	198–243	163–224
Choline, % AI	96–137	96–137	107–164	107–163	88–132	84–100	88–102
Potassium, % AI	62–105	86–130	79–139	99–157	50–82	65–90	66–87
Calcium3	108–156	186–262	112–190	197–294	46–114	87–112	87–115
Vitamin D3	2–12	64–100	11–25	60–112	9–39	36–43	31–40
Vitamin E3	98–124	127–183	81–133	103–191	68–104	60–81	71–93
α-Linolenic acid,4 % AI	—	—	—	—	81–156	130–178	129–196
Linoleic acid,4 % AI	—	—	—	—	66–121	87–123	90–137

¹AI, Adequate Intake; HM, human milk; IF, infant formula.

²If the RDA for iron is increased by a factor of 1.8 for vegetarian diets (18), the Pattern meets only 50%–71% of the RDA.

³6 to <12 mo, % AI; 12 to <24 mo, % RDA.

⁴At 6 to <12 mo, no AIs are available for the essential fatty acids [α-linolenic acid (18:3n–3) and linoleic acid (18:2n–6)]; at 12 to <24 mo, the AI values were used for these calculations but the intake values are total undifferentiated ω-3 and ω-6 fatty acids.

Infants 6 to <12 mo old fed infant formula

For infants fed infant formula, human milk was replaced with infant formula in the models, and the same combinations of CFBs were included. Because these models include fortified infant cereal as well as infant formula, there are few shortfall nutrients (Table 3), except for vitamin D and omega (ω)-3 fatty acids at some energy levels. However, the potential for excess intakes of certain nutrients exists. In these models, expected iron intake reaches 150%–175% of the RDA at 6 to <9 mo and 134%–194% of the RDA at 9 to <12 mo, although it does not exceed the Tolerable Upper Intake Level (UL) for iron (40 mg/d). Expected zinc intake reaches 226%–302% (7–9 mg/d, respectively) of the RDA at 6 to <9 mo and 232%–339% (7–10 mg/d, respectively) of the RDA at 9 to <12 mo. These estimates all exceed the UL for zinc (5 mg/d), although this UL has been challenged as being too low (29).

Toddlers fed neither human milk nor infant formula

In the first model for toddlers, which included food group amounts in proportion to the amounts in the 1000 kcal Pattern for ages 2 y and older (25), there were shortfalls for several nutrients. To fill nutrient gaps, adjustments were made to quantities for meats (for iron), dairy products (for calcium), whole grains (for potassium), and oils (for ω-3 and ω-6 fatty acids). Seafood was set at 3 oz eq per wk for all energy levels. Table 2 shows the resulting Healthy US Style Pattern for 12 to <24 mo.

The percentage breakdown of energy from macronutrients in this Pattern is ∼44%–50% carbohydrate, 31%–36% fat, and 17%–20% protein (Table 3). Expected intakes achieve ≥90% of the RDA or AI for most nutrients, with a few exceptions particularly at the lower energy levels (e.g., ω-6 fatty acids, iron, potassium, vitamin E, choline, and vitamin D). In FPM, any energy remaining after meeting nutrient goals was allocated to oils (8–13 g/d) to help meet recommended essential fatty acid intakes, leaving no additional energy for added sugars apart from the 2–3 g/d of added sugars in the Pattern as a result of some of the foods in the nutrient profile (mostly refined grains).

Toddlers fed a lacto-ovo vegetarian diet, and fed neither human milk nor infant formula

In the first model for toddlers consuming a vegetarian diet, which was adapted from the Healthy Vegetarian Pattern at the 1000-kcal level for ages 2 y and older (25), nutrient shortfalls included choline, potassium, vitamin E, vitamin D, and ω-3 and ω-6 fatty acids. Adjustments were made to include 3 eggs/wk (for choline) and increase the proportion of whole grains (for several other nutrients). Table 2 shows the resulting Healthy Vegetarian Pattern for 12 to <24 mo.

The percentage breakdown of energy from macronutrients in this Pattern is ∼48%–53% carbohydrate, 32%–36% fat, and 16%–17% protein (Table 3). Expected intakes achieve >90% of the RDA or AI for most nutrients, with some exceptions particularly at the lower energy levels (e.g., iron, potassium, calcium, vitamin E, choline, and vitamin D). However, most of the iron in this Pattern comes from whole grains, soy products, nuts/seeds, and legumes, and the bioavailability of iron (and zinc) from these types of foods is low (18). If the RDA for iron is increased by a factor of 1.8 for vegetarian diets (18), the Pattern meets only 50%–71% of the RDA. As for the Healthy US Style Pattern for toddlers, there is no available energy for added sugars (apart from the 2–3 g/d of added sugars present in some of the foods in the nutrient profile).

Toddlers fed human milk

For toddlers fed human milk, there were numerous nutrient shortfalls in the first model, which was based on food group amounts in proportion to the amounts in the 1000-kcal Pattern for ages 2 y and older (25). Adjustments were made to protein foods (to increase iron and calcium), vegetable subgroups (to emphasize good sources of calcium and/or iron), and grains (to minimize refined grains and allow some energy for oils to increase fatty acid adequacy). Table 2 shows the resulting combinations.

The percentage breakdown of energy from macronutrients in this Pattern is ∼44%–48% carbohydrate, 35%–40% fat, and 15%–20% protein (Table 3). These combinations come close to meeting most nutrient recommendations for a variety of scenarios differing in the proportions of energy coming from human milk and CFBs. However, some nutrients fall below 90% of the RDA or AI at most or all energy levels, including calcium, iron, potassium, vitamin E, and vitamin D. The Committee did not label the combinations for toddlers fed human milk as “recommended food patterns” because of uncertainty about nutrient requirements for this age range and challenges in meeting the DRIs.

Discussion

Overview of approach

Developing recommended food patterns for infants and toddlers ages 6 to <24 mo is challenging because nutrient needs are high relative to energy requirements at this age, and the amounts of CFBs that can be consumed are relatively low, especially at the younger ages. The Committee opted to start with modeling the contributions of food groups in proportion to the amounts in the 1000-kcal Pattern for ages 2 y and older (25), with adaptations as needed to correspond to estimated energy intakes and nutritional goals for infants and toddlers ages 6 to <24 mo. This approach has the advantage of developing patterns that are feasible with respect to the types of foods consumed in the United States, and that become consistent, by age 24 mo, with the patterns recommended for older age groups. However, the results do not necessarily represent the optimal combinations of foods and beverages for meeting nutritional goals, which requires a different modeling approach, such as optimization modeling (5., 6., 7.).

One strength of our approach was the modeling of various scenarios with respect to the potential contribution from human milk or infant formula, as well as several options reflecting total energy needs at ages 6 to <12 mo and 12 to <24 mo. Another strength, which is true of the USDA Food Patterns in general, is that the patterns provide examples of amounts of food groups and subgroups that could be consumed, but do not dictate the specific foods to be consumed, providing a large amount of flexibility for foods to be tailored to an individual's needs and preferences. This flexibility is very important during the CFB period, because it accommodates cultural preferences and cost considerations, and permits multiple approaches for the introduction of a wide variety of foods, flavors, and textures important in shaping healthy eating patterns (1, 30).

One challenge in developing food patterns for infants and toddlers fed human milk is uncertainty regarding the nutrient composition of human milk. The models described herein generally used the mean concentrations of each nutrient in human milk cited in the descriptions of the DRIs for infants as the nutrient profile (26), but in many cases these values are based on relatively few samples and/or outdated methods. Currently, no suitable approach or database is available that represents the variability of human milk composition in the United States. Several nutrients in human milk vary because of maternal nutritional status, diet and/or supplement intake, and other factors (13), including total fat, fatty acids, most vitamins, choline, iodine, and selenium. This has implications for the modeling exercises. For example, because the DRI report for vitamin D (21) states that human milk is not a meaningful source of vitamin D, the nutrient profile for human milk used in the modeling included no vitamin D. However, it is known that milk vitamin D concentrations can increase substantially in response to maternal supplementation with high doses of vitamin D and sun exposure (13, 31); thus, infants are expected to get some vitamin D through human milk.

Another challenge faced by the Committee was that for older infants, RDAs are available only for protein, iron, and zinc, so the nutritional goals for the modeling exercises for ages 6 to <12 mo were based mainly on AI values. The primary approach used by DRI committees for setting the AIs for older infants was to sum the estimated mean content coming from reported CFB intakes and from 600 mL/d of human milk. However, when the value was judged to be unreasonable, the AI was set by extrapolating up from the AI for ages 0 to <6 mo (for vitamins K, E, and B-12, selenium, and iodine), down from estimates of adult requirements (for thiamin and niacin), or a combination of the 2 (for riboflavin, vitamin B-6, folate, and choline) (26). The lack of RDAs made it difficult for the Committee to evaluate risk of inadequacy for potential shortfall nutrients, such as potassium and choline, for which the AI may or may not represent the correct target. In addition, the potential for overconsumption was assessed in the models (particularly those with infant formula) based on ULs for older infants (and toddlers), but some ULs have been criticized as having been established with too little available data and are considered to be too low for certain nutrients (29), specifically zinc and retinol. For both age intervals (6 to <12 mo and 12 to <24 mo), published nutrient reference values vary considerably across authoritative bodies (32), which suggests some uncertainty about nutrient requirements.

Nutrient shortfalls

In the first set of models for infants ages 6 to <12 mo, estimated intake of several nutrients fell short of nutritional goals. Some of these nutrients were also reported as underconsumed in national data (2) (Part D: Chapter 1), including iron and zinc among infants fed human milk, and vitamin D, potassium, and choline among all infants 6 to <12 mo of age. Although the intake data suggested that 27% of infants fed human milk at this age had protein intakes that were low enough to be at risk of inadequacy, protein was not a limiting nutrient in the FPM exercises in any of the scenarios, because a focus on iron naturally led to the inclusion of iron-rich foods, like meat, that also have a high protein content. Indeed, the percentage of energy from protein at ages 9 to <12 mo (16%–19%) was on the high side, and evidence suggests that protein intakes >15% of energy in early life may increase the risk of excess weight gain (33). However, this is an area of active research and it is not clear which types of protein (e.g., dairy compared with meat) may or may not be contributing to this association.

The FPM results confirmed the challenges of meeting iron and zinc needs for infants fed human milk. Fortified infant cereal helped to close some of the gap between the amount provided in the example combinations of foods and the RDA for both iron and zinc. However, there were still some shortfalls for certain nutrients, including iron, potassium, magnesium, and choline. In the final step of the modeling exercises, most of these gaps were filled by prioritizing protein foods, particularly meat. For infants fed infant formula at ages 6 to <12 mo, the combinations included fortified infant cereal as well as infant formula, so there were few shortfall nutrients except for vitamin D and ω-3 fatty acids at some energy levels. These FPM exercises did not attempt to model mixed-feeding scenarios, in which infants receive both human milk and infant formula. However, the example combinations that meet most nutrient recommendations for infants fed human milk also are likely to be nutritionally adequate for mixed-fed infants.

For ages 12 to <24 mo, the shortfall nutrients (for some or all of the energy levels) in the first set of models for toddlers fed neither human milk nor infant formula were calcium, iron, potassium, vitamin E, vitamin D, choline, and ω-3 and ω-6 fatty acids. Some of these, such as potassium and vitamin D, were also reported as underconsumed at this age in national data (2) (Part D: Chapter 1), and choline and linoleic acid were categorized as “special challenges.” Small increases in protein foods and dairy and an emphasis on whole grains rather than refined grains closed some of these gaps, but potassium, vitamin E, and vitamin D were still consistently below the goals. As was the case at ages 9 to <12 mo, the percentage of energy from protein was on the high side (17%–21%), which warrants further consideration.

For toddlers fed lacto-ovo vegetarian diets, most of the iron in the Vegetarian Pattern comes from whole grains, soy products, nuts and seeds, and legumes, for which bioavailability of iron is likely to be low owing to relatively high concentrations of phytate and the absence of heme iron (18). The projected percentage of the iron RDA provided in that Pattern is likely an overestimate of the amount that is physiologically available. If one assumes that iron requirements are 1.8 times higher for vegetarian diets than for nonvegetarian diets (18), the Vegetarian Pattern for toddlers would meet only 50%–71% of the RDA for iron. Further work is needed to evaluate whether the foods and beverages in this Pattern can adequately support iron and zinc status during the second year of life.

For toddlers fed human milk at ages 12 to <24 mo, the FPM exercises revealed challenges in meeting nutrient goals for both calcium and iron simultaneously, given that 1) human milk contains considerably less calcium than cow milk (although calcium absorption from human milk is high, i.e., ∼60%) (21); and 2) inclusion of sufficient amounts of dairy products to meet calcium needs meant that iron became a shortfall nutrient, because dairy products contain very little iron. Of note, the RDA for calcium at ages 1–3 y (700 mg) is much higher than the AI for calcium at ages 7 to <12 mo (270 mg), and the recommended calcium intake for ages 1–3 y published by the European Food Standards Authority is only 450 mg/d (34). As was true for toddlers fed neither human milk nor infant formula, other shortfall nutrients included potassium, vitamin E, and vitamin D. Further modeling work is needed that incorporates estimates of mineral absorption under various circumstances. Using tools such as linear programming would be helpful in addressing multiple nutritional constraints and food sources of nutrients simultaneously, to identify combinations of foods and beverages that meet all nutritional goals.

The percentage of energy from fat in these models was 41%–44% at 6 to <9 mo, 35%–42% at 9 to <12 mo, and 29%–40% at 12 to <24 mo of age, within recommended ranges capable of meeting the AI for infants and the Acceptable Macronutrient Distribution Range for toddlers. The AI is 30 g fat/d at ages 7 to <12 mo, which represents ∼30%–45% of energy for total energy intakes of 600–900 kcal/d. Projected intakes of ω-6 fatty acids in these models were more than adequate at 6 to <12 mo, but were lower than the AI of 7 g/d for linoleic acid at 12 to <24 mo for several scenarios. For ω-3 fatty acids, projected intakes in these models were >90% of the AI for most energy levels (except the lowest, 600-kcal, level) for the infants fed human milk or infant formula at ages 6 to <12 mo. At ages 12 to <24 mo, they also were generally >90% of the AI.

Iron is a key nutrient at 6 to ≤12 mo of age

As expected, the most limiting nutrient for infants fed human milk at ages 6 to <12 mo was iron. It was not possible to meet the RDA without the inclusion of iron-fortified infant foods. Because the iron concentration of human milk is low (∼0.3 mg/L after 5 mo of lactation) (18), the FPM exercises assumed 0 iron coming from that source. Absorption of iron from human milk is variable (35), but even if 100% is absorbed, the amount of iron that an infant would receive from 600 mL/d would be <0.2 mg/d, a trivial amount relative to the RDA of 11 mg/d. This discrepancy may seem counterintuitive, but it is likely that the iron content of CFBs fed to infants during most of human evolution, when humans relied completely on hunting and gathering before the invention of agriculture, was much higher than it is today, and that iron deficiency was rare (36). The estimated iron density of the preagricultural diet was 2.9 mg/100 kcal at age 9 mo, whereas typical modern-day (unfortified) complementary food diets have an iron density of only 0.4–1.3 mg/100 kcal (36).

Fortified infant foods are not necessarily the only way for infants fed human milk to achieve the RDA, however. For example, certain animal-source foods (e.g., red meat) are good sources of iron, particularly when taking into account the fact that heme iron (as found in meat) is much better absorbed than nonheme iron (as found in plant-based foods). Assuming that infants at ages 6 to <12 mo need 1.1 mg/d of absorbed iron (back-calculated from the RDA of 11 mg/d, which assumes 10% absorption), and 25% absorption of heme iron, infants would need 4.4 mg Fe/d from animal-source foods. Obtaining that amount solely from beef, which has ∼1 mg Fe/100 g (81 kcal of “baby food” beef), would require consuming 440 g of beef (356 kcal), which is not feasible. Organ meats such as liver have far more iron. For example, the iron content of chicken liver is ∼11.5 mg/100 g (166 kcal), so infants would need only 38 g (64 kcal) to meet the target of 4.4 mg. However, feeding liver to infants is not common in the United States. Further work is needed to estimate the quantities of iron-rich foods that would be needed by infants fed human milk, in the absence of fortified infant foods, to support adequate iron status at 6 to <12 mo, recognizing that 1) the RDA is set to meet the needs of 97.5% of infants, and many infants require less than the RDA; 2) iron absorption is upregulated when iron stores begin to become depleted; and 3) the Recommended Nutrient Intake for iron at this age set by the WHO/FAO (i.e., 9.3 mg/d) (37) is lower than the RDA (i.e., 11 mg/d), both of which assume 10% absorption. In the meantime, it should be noted that iron-fortified infant foods have been an important strategy for reducing iron deficiency among infants in the United States for several decades (38).

On the other hand, infants fed infant formula have the potential for excess intakes of iron (and other nutrients), because the iron content of the formulas most commonly used in the United States is relatively high (∼1.8 mg/100 kcal), ∼40 times the iron content of human milk. In the FPM exercises for infants fed infant formula, inclusion of iron-fortified infant cereal in addition to infant formula would result in total iron intakes that are 123%–181% of the RDA, although the bioavailability of iron in fortified infant cereals is highly variable, depending on the type of cereal and form of iron that is added (39). Although the estimated iron intakes in these scenarios did not exceed the UL of 40 mg/d, iron-fortified infant foods are not necessary if infant formula intake is >760 mL/d at ages 6 to <9 mo or >690 mL/d at ages 9 to <12 mo.

Potassium, iodine, and sodium

It was challenging to meet the AI for potassium (860 mg/d at 7–12 mo; 2000 mg/d at 1–3 y) in all of the FPM exercises. The AI for 7 to <12 mo is based on 260 mg/d from human milk plus 600 mg/d from CFBs, and it is possible that the latter is an overestimate of actual intakes. After 12 mo, the AI is based on the highest median intakes at ages 1–3 y, and this may overestimate needs at ages 12 to <24 mo. The recommended potassium intakes published by the European Food Standards Authority are lower than the AI values: 750 mg/d at 7 to <12 mo and 800 mg/d at 1–3 y (34). This suggests some uncertainty regarding potassium requirements for infants and toddlers. Nonetheless, choosing potassium-rich foods is important at these ages.

Iodine intakes could not be predicted because food composition data are not available for iodine. The AI for iodine at 6 to <12 mo is 130 μg/d, which was extrapolated up from the AI of 110 μg/d for 0 to <6 mo. Estimated iodine intake at 6 to <12 mo is ∼141 μg/d, based on Total Diet Study estimates of iodine in the US food supply and predicted intakes based on food consumption data reported in What We Eat in America 2007–2008, 2009–2010, and 2011–2012 (40). However, this daily estimate is based on 59% (83 μg) from “baby food,” which includes infant formula, and 33% (46 μg) from dairy products. It is not clear what the estimated intake would be among infants fed human milk. Infant formula generally has 15 μg I/100 kcal, which would provide the AI if energy intake from formula is ≥866 kcal/d, although the minimum content required is only 5 μg I/100 kcal (14). The average estimated iodine concentration of human milk is similar, but maternal diet influences milk iodine concentration (13), and wide variability in iodine intakes of women who are lactating is likely because of differences in intakes of dairy products and iodized salt (41). In situations in which neither the mother nor the infant consumes iodized salt or obtains adequate iodine from other sources (e.g., dairy products), iodine intakes of infants could be deficient. Only 53% of table salt sold at the retail level in the United States in 2009 was iodized (42), and the iodine content of cow milk in the United States is highly variable (41). Underconsumption of iodine during infancy has important potential consequences for brain development, especially if maternal intake was also low during pregnancy (43).

The FPM exercises for infants 6 to <9 mo of age provided relatively little sodium (179–400 mg/d), although by 9 to <12 mo the estimates were adequate (384–566 mg/d in the models for infants fed human milk). The AI for sodium at 7 to <12 mo is based on estimated sodium intake from human milk (70 mg/d, from 600 mL/d of human milk) plus CFBs (300 mg/d), for a total of 370 mg/d. Physiological requirements for sodium during infancy (44, 45) correspond to an intake of ∼300–460 mg/d at 6 to <12 mo. Projected intakes at 6 to <9 mo for “average” human milk intake models were <300 mg/d. Infant feeding guidance usually recommends not to add salt to foods for infants. This has implications not only for adequacy of sodium intake, but also adequacy of iodine intake, because iodized salt is a key contributor to the latter. If infants are fed some prepared foods to which salt has been added, sodium intakes may not be low, but if the recommendation to avoid added salt were fully implemented, underconsumption may be a concern. The estimated sodium provided by the Patterns for 12 to <24 mo (with no human milk or infant formula) was 613–729 mg/d, which is below the AI of 800 mg/d (which is based on extrapolation down from adult values).

Added sugars

The amounts of added sugars in the example combinations of foods developed for infants 6 to <12 mo of age (0.5–1.1 g/d) and in the Patterns for 12 to <24 mo (2–3 g/d) are negligible. This low amount of added sugars is logical given that the foods selected for the Patterns were in the most nutrient-dense forms, and would thus, by definition, be low in added sugars. The observed intake patterns among infants and toddlers suggest that much higher amounts of added sugars are currently being consumed: ∼4 g/d at 6 to <12 mo and ∼26 g/d at 12 to <24 mo (2) (Part D: Chapter 1). The FPM exercises demonstrate that aiming to achieve recommended intakes of key nutrients for ages 6 to <24 mo leaves virtually no remaining energy for added sugars apart from those already present in the foods in the nutrient profile. Shifts in the dietary intakes of infants and toddlers are needed to ensure that nutrient-dense foods are provided and added sugars are decreased.

Conclusions and recommendations

For infants ages 6 to <12 mo, the Committee was not able to establish a recommended food pattern. Further work is needed to explore various options for meeting all nutrient recommendations during that age range, using tools such as linear programming and taking into account differences in iron bioavailability from different sources. In the meantime, the FPM exercises revealed the importance of prioritizing certain food groups and making careful food choices within food groups. For example, certain animal-source foods are important sources of key “shortfall” nutrients at this age, including iron, zinc, choline, and long-chain PUFAs. Fortified infant cereals can contribute a substantial amount of some of these nutrients, particularly iron and zinc, but prioritizing consumption of meat, egg, and seafood is an important strategy for providing all of these crucial nutrients. By contrast, dairy products (such as yogurt and cheese) are less crucial than other types of animal-source foods at ages 6 to <12 mo because infants are still receiving human milk or infant formula, and dairy products tend to have low amounts of iron. Prioritizing fruits and vegetables, particularly those that are rich in potassium, vitamin A, and vitamin C, is another key element of healthy complementary food diets at ages 6 to <12 mo, not only to provide adequate nutrition but also to foster acceptance of these foods in forms that are consistent with a healthy dietary pattern. In addition, introduction of peanut products and egg in the first year of life is advised, to build tolerance to food antigens (i.e., help prevent food allergies) and to provide good sources of fatty acids and choline.

For toddlers ages 12 to <24 mo, the Committee was able to establish a recommended Food Pattern for toddlers fed neither human milk nor infant formula that resembles the Pattern established for ages 2 y and older. The Pattern allows for a variety of nutrient-rich animal-source foods, including meat, poultry, seafood, eggs, and dairy products, as well as nuts and seeds, fruits, vegetables, and grain products. Key aspects to emphasize include choosing potassium-rich fruits and vegetables, prioritizing seafood, making whole grains the predominant type of grains offered, and choosing oils over solid fats. Figure 2 shows that the Healthy US-Style Pattern for toddlers is an achievable pattern, with a few shifts from current consumption patterns needed. Average intakes of fruits, grains, dairy products, and total protein foods meet or exceed the range of recommended intakes in the Pattern. However, average intake of vegetables (particularly dark green vegetables) is below the target range. In addition, within the grains group, consumption of whole grains is well below the target range, whereas intake of refined grains is well above the target range. Within the protein foods group, average intake of seafood is well below the target range. Thus, shifts toward greater consumption of vegetables and seafood, and a higher proportion of total grains as whole grains, are needed.

Figure 2.

Recommended food group and whole grain and refined grain subgroup intake ranges compared with average intakes for ages 12 to <24 mo. Average intake data from Food Group Intake Distributions, What We Eat in America NHANES 2013–2016, ages 12 to <24 mo. Prepared by the National Cancer Institute, 2019.

For toddlers fed lacto-ovo vegetarian diets and fed neither human milk nor infant formula at ages 12 to <24 mo, a Pattern was established that includes regular consumption of eggs, dairy products, soy products, and nuts or seeds, in addition to fruits, vegetables, grains, and oils. Because of concerns about iron bioavailability in the vegetarian pattern, further modeling work that takes this into account is needed. Careful choices of CFBs within vegetarian diets are very important to meet nutrient needs. It should be noted that the Healthy Vegetarian Eating Pattern developed is not a vegan diet, because the former includes substantial amounts of animal-source foods (egg and dairy). Without supplements and/or fortified products, it is not possible to meet all nutrient goals with a vegan diet at this age (46).

For toddlers fed human milk at ages 12 to <24 mo, the Committee was not able to establish a recommended food pattern, but provides examples of potential combinations of CFBs that come close to meeting almost all nutrient recommendations. Further work is needed to examine predicted nutrient intakes of toddlers fed human milk, taking into account mineral bioavailability under various conditions.

Taken together, these findings are not intended to provide a combination of CFBs or food pattern that is right for every infant or toddler, because children develop at different rates, and many different circumstances influence feeding needs and decisions. In the Patterns developed for toddlers ages 12 to <24 mo, the lowest energy level (700 kcal) presented challenges for meeting certain nutritional goals (e.g., iron and fatty acids). Toddlers with relatively low energy intakes may benefit from food combinations that resemble those for infants ages 6 to <12 mo, with a gradual shift to the patterns presented for ages 12 to <24 mo. A general principle is to view the period from ages 6 to <24 mo as a continuous transition from diets appropriate for infants, to diets that resemble family food patterns. Figure 3 depicts this transition. For most of the food groups, amounts to be consumed gradually rise as energy from CFBs increases, which is correlated with age. However, the energy from Protein Foods is relatively constant, and is a substantial proportion of total energy from CFBs, at all energy levels between 700 and 1000 kcal. This is a reflection of the need for nutrient-rich foods for children younger than age 24 mo. Another important feature of Figure 3 is the high proportion of whole grains, relative to total grains, until age 2 y.

Figure 3.

Relative amounts of food groups and subgroups in Healthy US-Style Patterns across energy levels for toddlers and young children. Inclusive of complementary foods and beverages and not human milk or infant formula; modeled complementary food includes fluid milk, calcium-fortified soy beverage, and 100% fruit and vegetable juice. Reproduced from (2) (Part D: Chapter 7).

In all the CFB combinations and Patterns examined, the energy allocated to oils and solid fats is minimal, and no energy remains for added sugars. The Committee recommended that added sugars be avoided for infants and toddlers <2 y of age. The energy contributed by CFBs with added sugars is likely to displace energy from nutrient-dense foods, increasing the risk of nutrient inadequacies. Moreover, consumption of sugar-sweetened beverages is linked with increased risk of overweight or obesity (2) (Part D: Chapter 5). Because food preferences and patterns are beginning to form during this developmental stage, and taste and flavor preferences appear to be more malleable in this life stage than in older children, it is important that caregivers limit consumption of foods and beverages that contain added sugars, while encouraging consumption of nutrient-dense foods. For these reasons, one of the hallmark recommendations of the Dietary Guidelines for Americans, 2020–2025 is to “make every bite count” (4).

For future FBDGs for infants and children <2 y of age, the Committee recommended several important next steps (Box 1) to 1) establish a reference human milk composition profile; 2) update and strengthen the DRI values for this age group; and 3) use optimization modeling, in combination with FPM, to identify combinations of CFBs that meet all nutritional goals. Advancements in these 3 areas are envisioned before the development of the Dietary Guidelines for Americans, 2025–2030.

BOX 1.

Recommended next steps for development of FBDGs for infants and young children

• Establish a reference human milk profile:

  • Continue the ongoing Federal initiative to expand research on human milk composition and how it relates to maternal and infant health (47).
  • Develop an accurate and current database of representative values for the energy and nutrient composition of human milk across the full course of lactation, including beyond 1 y postpartum.

    • Milk samples should be collected from diverse groups of individuals and linked to dietary intake and other metadata (e.g., age, parity).
  • Update USDA databases with a human milk composition profile that incorporates data from diverse populations and across lactation.
• Update the DRI values for infants and children younger than age 24 mo:

  • Incorporate new data on human milk composition.
  • Include estimates of average requirements whenever possible, as well as RDA and UL values.
  • Coordinate efforts with other organizations, such as the WHO and the European Food Standards Authority.
  • If possible, update DRI values for women during pregnancy and lactation at the same time.
• Use optimization modeling in combination with FPM to develop FBDGs:

  • Incorporate multiple nutritional constraints and food sources of nutrients in the models, to identify combinations of foods and beverages that meet all nutritional goals.
  • Take into account nutrient bioavailability (especially iron, zinc, and calcium) from various food sources, including human milk.
  • Consider models that allow for amounts of individual foods to vary within each food group or subgroup.