Early Life Protein Intake: Food Sources, Correlates, and Tracking across the First 5 Years of Life

RESEARCH

Original Research

Early Life Protein Intake: Food Sources, Correlates, and Tracking across the First 5 Years of Life Karen J. Campbell, PhD, MPH; Gavin Abbott, PhD; Miaobing Zheng, PhD; Sarah A. McNaughton, PhD, MHSc, AdvAPD* ARTICLE INFORMATION Article history: Submitted 1 September 2016 Accepted 21 March 2017 Available online 17 May 2017

Keywords: Infant Child Protein intakes Food sources Tracking

Supplementary materials: Table 1 is available at www.jandonline.org 2212-2672/Copyright ª 2017 by the Academy of Nutrition and Dietetics. http://dx.doi.org/10.1016/j.jand.2017.03.016 *

AdvAPD¼advanced accredited practising dietitian (certified in Australia).

ABSTRACT Background High consumption of protein has been associated with accelerated growth and adiposity in early childhood. Objective To describe intake, food sources, correlates, and tracking of protein in young children. Design Secondary analysis of Melbourne Infant Feeding Activity and Nutrition Trial (InFANT). Dietary data were collected using three 24-hour dietary recalls at ages 9 and 18 months as well as 3.5 and 5 years. Participants/setting First-time mothers and their child (n¼542) participated in an 18month intervention to prevent childhood obesity and the cohort was followed-up with no intervention when children were aged 3.5 and 5 years. Main outcome measures Protein intake, food sources, correlates, and tracking of protein. Statistical analyses performed Child and maternal correlates of protein intake were identified using linear regression and tracking of protein intake was examined using Pearson correlations of residualized protein scores between time points. Results Mean protein (grams per day) intake was 29.711.0, 46.311.5, 54.213.8, and 60.014.8 at 9 and 18 months and 3.5 and 5 years, respectively. Protein intakes at all ages were two to three times greater than age-appropriate Australian recommendations. The primary source of protein at 9 months was breast/formula milk. At later ages, the principal sources were milk/milk products, breads/cereals, and meat/meat products. Earlier breastfeeding cessation, earlier introduction of solids, high dairy milk consumption (500 mL), and high maternal education were significant predictors of high protein intake at various times (P<0.05). Slight tracking was found for protein intakes at 9 months, 18 months, and 5 years (r¼0.16 to 0.21; P<0.01). Conclusions This study provides unique insights into food sources and correlates of young children’s high protein intakes, and confirms that early protein intakes track slightly up to age 5 years. These finding have potential to inform nutrition interventions and strategies to address high protein intakes and protein-related obesity risk. J Acad Nutr Diet. 2017;117:1188-1197.

A

YOUNG CHILD’S CONSUMPTION OF A PROTEIN-RICH diet has been long considered the fulcrum for healthy growth and development. Historically, the focus has been on ensuring that rapidly growing infants consume enough protein to meet increased needs for growth, with predominant interest focused on the

The Continuing Professional Education (CPE) quiz for this article is available for free to Academy members through the MyCDRGo app (available for iOS and Android devices) and via www.eatrightPRO.org. Simply log in with your Academy of Nutrition and Dietetics or Commission on Dietetic Registration username and password, go to the My Account section of My Academy Toolbar, click the “Access Quiz” link, click “Journal Article Quiz” on the next page, then click the “Additional Journal CPE quizzes” button to view a list of available quizzes. Non-members may take CPE quizzes by sending a request to [email protected]. There is a fee of $45 per quiz (includes quiz and copy of article) for non-member Journal CPE. CPE quizzes are valid for 1 year after the issue date in which the articles are published.

1188

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

association between inadequate protein and constrained growth or malnutrition (eg, the description of kwashiorkor in the 1930s).1 Although a focus on protein malnutrition remains vitally important in many countries with developing economies, there has been a shift in countries with developed economies, to concerns regarding the association of high protein consumption and childhood health. A recent systematic review of protein intake and health across the first 18 years of life reported that there was convincing evidence (grade 1) that higher protein intakes were associated with increased growth and high body mass index (BMI) in childhood.2 This review also reported limitedsuggestive evidence (grade 3) that intake of animal protein, especially from diary, has a stronger positive association with growth than vegetable protein, and that bone density in childhood is positively associated with increased protein intake. ª 2017 by the Academy of Nutrition and Dietetics.

RESEARCH The timing of the introduction of higher protein intakes also appears likely to be an important determinant of growth. The review by Hornell and colleagues2 concluded that the available data suggested it was probable that children will be most sensitive to high protein intakes during the first 2 years of life. Since that review, Mihrshahi and colleagues3 identified infant formula consumption as one of only two modifiable predictors (the other being parent feeding infants to a schedule) of rapid weight gain (birth to age 4 to 7 months) in a sample of Australian infants and hypothesized that this growth was likely to have been stimulated by the increased protein intakes associated with formula feeding. In an older sample of children, Günther and colleagues4 suggest that the transition from breastmilk or infant formula to “family foods” at around age 12 months is a critical phase for increased protein intake and subsequent childhood obesity. Despite the potential negative health consequences of high protein intakes, many protein-dense foods are also nutrient-dense, and in some instances may be a child’s principal source of key nutrients such as calcium and iron. It is important, therefore, to ensure that any proposed reduction in protein intake will not compromise other essential nutrients. Identifying food sources of protein (and sources of other nutrients) is necessary to inform these considerations. In addition, describing modifiable correlates of high protein intakes in early life will identify appropriate targets for interventions seeking to reduce early protein intake. Because there is evidence that children’s diets track over time, it is important also to understand whether early protein intakes set a trajectory for higher protein intake across childhood. Therefore, the aims of this study were to describe dietary protein intakes, food sources, and modifiable child and maternal correlates of protein intake in a cohort of young Victorian (Australian) children over the first 5 years of life (9 and 18 months and 3.5 and 5 years). In addition, this study examined tracking of protein intake across these early years.

MATERIALS AND METHODS Study Design and Participants The Melbourne Infant Feeding Activity and Nutrition Trial (InFANT) Program was a cluster randomized controlled trial (Clinical Trial registration: Current controlled Trials ISRCTN81847050) involving first-time mothers attending parents’ groups, from when their infants were aged 3 to 20 months. This lifestyle intervention was conducted during 2008 to 2010 within Melbourne, Australia (population w4 million), and the cohort was followed-up (no further intervention) when children were approximately age 3.5 and 5 years. Primary aims of the Melbourne InFANT Program focused on reducing a range of children’s obesity-risk behaviors. There was no focus on protein consumption at that time and no difference in protein consumption was observed between trial arms (data not shown). The study design has been previously reported.5,6 Eighty-six percent of eligible parents consented to participate (n¼542). The Melbourne InFANT Program was approved by the Deakin University Human Research Ethics Committee (ID no. EC 175-2007) and the Victorian Government Department of Human Services, August 2017 Volume 117 Number 8

Office for Children, Research Coordinating Committee (Ref no. CDF/07/1138). This study was deemed exempt under federal regulation 45 46.101 (b) CFR. Because there were no differences at any time in protein intakes between intervention and control group children, data for this article have been pooled. We present data from children at approximately ages 9 and 18 months and 3.5 and 5 years, herein referred to as time two (T2), three (T3), four (T4), and five (T5) for consistency with other publications arising from these data.7 As outlined in the Figure, data were excluded for children from nonefirst-time mothers (n¼14) and those missing key baseline maternal variables (mother’s marital status, country of birth, education, employment, age, and BMI) or missing data on age of breastfeeding cessation and starting solids (n¼76). A total of 86 participants were excluded (some participants were excluded on more than one criteria), leaving an initial sample of 456 children. From this sample, further time point-specific exclusions were made to be consistent with our previously published data from this cohort.8 Children aged younger than age 7 months or older than age 11 months were excluded from the T2 sample (n¼14), whereas children younger than age 16 months or older than age 20 months were excluded from the T3 sample (n¼32). In addition, children lost to follow-up since baseline (T2 n¼2, T3 n¼14, T4 n¼122, and T5 n¼123), children with fewer than 3 days of dietary recalls (T2 n¼61, T3 n¼89, T4 n¼94, and T5 n¼97), children with outlier (3 standard deviations) total energy intakes (T2 n¼1, T3 n¼4, T4 n¼3, and T5 n¼1), and children with missing BMI z score data (T2 n¼0, T3 n¼15, T4 n¼4, and T5 n¼0) were excluded from their respective time-specific sample. After exclusion of children who met one or more of these criteria, the final analysis samples were n¼381 for T2, n¼321 for T3, n¼237 for T4, and n¼235 for T5. Analyses were undertaken to examine the difference between participants included in the study and those excluded at each time point. Children included in each of the time point-specific samples were significantly younger and had higher maternal education than children who were excluded. In addition, children included in the T2 analysis sample (mean¼0.11.0) had lower T2 BMI z scores than those excluded (mean¼0.31.0), whereas children included in the T3 analysis sample (mean¼46.30.5 g) had lower T3 protein intake than children excluded from T3 analyses (mean¼49.814.9 g), and finally children included in the T5 analysis sample (mean¼0.61.0) had higher T5 BMI z scores than those excluded (mean¼0.40.9). No differences were found with regard to children’s sex or maternal age (data not shown).

Maternal Factors Self-administered paper-based questionnaires were provided to parents at the recruitment meeting and returned to program staff at the first InFANT Program session. These provided demographic and socioeconomic data at baseline (T1¼children aged 3 months), including maternal age, maternal employment, education level, and country of birth. Maternal employment was dichotomized as not currently employed in paid work or currently employed full- or parttime. Maternal education level was dichotomized as university educated or noneuniversity educated. Country of birth was classified as born in Australia or overseas. Mothers JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

1189

RESEARCH

542 Parcipang children Non-first me mothers: 14 a Other exclusions: 72 Baseline: 456 inial sample

T2: 454 parcipang children Outside age range (7-11months): 14 b Other exclusions: 59 381 included in analysis

T3: 442 parcipang children Outside age range (16-20months): 32 b Other exclusions: 89 321 included in analysis

T4: 334 parcipang children b Other exclusions: 97 237 included in analysis T5: 333 parcipang children b Other exclusions: 98 235 included in analysis Figure. Flow chart showing the number of participants included in analyses at each time point in the Melbourne Infant Feeding Activity and Nutrition Trial Program. aParticipants excluded due to missing data on baseline maternal variables and/or missing data on child age of breastfeeding cessation or starting solids. bParticipants excluded due to incomplete 3-day dietary recall data, extreme outlier total energy intake, and/or missing body mass index. self-reported their current height and their prepregnancy weight, and from these measures BMI were calculated.

Child Dietary Intake and Feeding Behaviors Children’s diet was assessed by nutritionists trained by the researchers (K. J. C., S. A. M.) using standardized telephoneadministered, 5-pass 24-hour recalls.9 Study-specific food measurement books aided parents’ estimation of infants’ food consumption. Three nonconsecutive days of dietary data (including 1 weekend day) were collected via a purpose designed computer program facilitating reporting fidelity and consistency between interviewers. Overall, 96% of telephone calls were unscheduled. Nutrient intakes were evaluated using the 2007 Australian Food, Supplement, and Nutrient Database food composition database (Food Standard 1190

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Australia New Zealand. AUSNUT 2007 food measures database file. Available from: http://www.foodstandards.gov.au/ science/monitoringnutrients/ausnut/Pages/ausnut2007.aspx). Consistent with the Avon Longitudinal Study of Parents and Children10 and the United Kingdom’s 2011 Diet and Nutrition Survey of Infants and young Children,11 breastfeeding was recorded as minutes spent breastfeeding, and breastmilk quantity was estimated using a conversion factor of 10 mL/min up to a maximum of 100 mL at any one feed. The nutrient contribution after 10 minutes of breastfeeding is considered minimal.12 In cases where breastmilk was expressed, volumes estimated by caregiver report were used. Single questions assessed the age in weeks at which solids were introduced, and at which breastfeeding was ceased. To determine the contributions of protein from food groups, the 1,688 individual food items consumed by children in the August 2017 Volume 117 Number 8

RESEARCH InFANT Program were grouped before analysis using the standard food groupings in the 2007 Australian Food, Supplement and Nutrient Database food coding system developed by Food Standards Australia New Zealand (Food Standard Australia New Zealand. AUSNUT 2007 food measures database file. Available from: http://www.foodstandards.gov. au/science/monitoringnutrients/ausnut/Pages/ausnut2007.aspx). The mean values and standard deviations of protein (grams) consumed and percentage contribution to total protein intake were then calculated for each food group to determine the food sources of protein. Protein density was reported as protein in grams per 1,000 kcal (grams per 1,000 kcal) consumed. Intakes of calcium and iron (milligrams per day) at each time point were calculated and along with protein intake were compared with Australian Estimated Average Requirement (EAR) and Recommended Dietary Intake (RDI).13 Milk intake was further classified to examine the source and quantity of milk consumed in early childhood. Milk intakes at T2 (child aged 9 months) according to primary milk source are described in Table 1 (available online at www. jandonline.org). Children’s primary source of milk at T2 was classified according to whether the major source was breast, formula, or dairy milk. Where children derived more than two-thirds (66.7%) of their milk from a given source, this was determined to be their primary source of milk. Children with less than two-thirds of their milk from a single source had their primary source classified as mixed. Only four children had dairy milk as their primary milk source at this age and thus these children were combined with the formula group because these milk sources are more similar in protein than they are to breastmilk. Therefore, three categories of primary milk source were created: breastmilk, formula/dairy, and mixed. Finally, milk intake at T3 (child age 18 months) was categorized according to recommendations in the Australian Infant Feeding Guidelines,14 which specify that children at aged 1 to 2 years should not consume more than 500 mL milk per day, based on concerns that milk will displace other foods and in turn limit food variety and nutrient adequacy.

Child Anthropometric Characteristics Children’s height/length and weight (light clothing only) were measured by trained staff at each time point. Height/ length was measured to 0.1 cm using a calibrated measuring mat (Seca 210; Seca Deutschland) or portable stadiometer (Invicta IPO955; Oadby). Weight was measured to 10 g using calibrated infant digital scales (Tanita 1582; Tanita Corp). The average of two measures was used in analyses. BMI and BMI z scores were calculated using World Health Organization sexspecific BMI-for-age growth charts.15

Statistical Analyses Descriptive statistics were used to summarize protein intake, food group contributions to total energy, and protein intakes at all time points. For each of the time points between T2 and T5, a separate random-intercepts linear regression model, with participants nested within mothers’ group, was conducted to examine correlates of protein intake. The following predictor variables: child age; child sex; age of breastfeeding cessation; age when child began eating solids; and mothers’ prepregnancy BMI, age at baseline, whether working at baseline, education, and whether they were born in Australia were August 2017 Volume 117 Number 8

examined and were included simultaneously in the same model at each time point. All models also included child BMI z scores and intervention status as a covariate. Therefore, the associations of predictor variables with protein intake at each time point are independent of each other and of child BMI z score. In the T2 model (children aged 9 months), children’s primary milk source (as described above) was also included as a predictor. In the T3 model, children’s consumption of more than 500 mL milk per day was also included. To assess tracking of protein intakes across the time points, residualized protein intake scores were created by regressing children’s protein intake on their age, sex and total energy intake.16 Tracking of protein intake was assessed by examining Pearson product-moment correlations of these residualized protein scores between the different time points. This method has been widely used to assess the tracking of nutrient intakes.17 Protein intake at each time point was near normally distributed, and statistical assumptions were met for all models tested. Interpretation of these correlation coefficients was based on the following guidelines: <0.3 for slight tracking, 0.3 to 0.6 for moderate tracking and >0.6 for high tracking, consistent with our previous approach.8 Statistical analyses were conducted using Stata version 14.018 with statistical significance set at P<0.05.

RESULTS Sample Characteristics Among the initial sample (n¼456), 53.3% of children were girls, 55.3% of mothers had a university education, 79.6% of mothers were born in Australia, and 8.8% of mothers were working. Mothers’ mean prepregnancy BMI and age was 24.35.1 and 32.24.1 years, respectively.

Protein Intakes Protein intakes in grams per day and grams per kilogram body weight at each time point are presented in Table 2. Energy intakes (in kilocalories), protein density (grams per 1,000 kcal), and the percentage of total energy from protein are also presented. At T2, 95% of children exceeded the Australian RDI of 14 g/day protein, and 99.4% of children exceeded the Australian RDI of 1.08 g/kg. At all other time points, these age-specific Australian RDIs were exceeded by all children.

Iron and Calcium Intakes Calcium intakes in this sample substantially exceeded the Australian Nutrient Reference Values (Table 2). At age 18 months and 3.5 years, mean calcium intakes were 759220 mg and 732239 mg, respectively, thus more than double the Australian EAR of 360 mg and approximately 30% greater than the Australian RDI (500 mg) for children aged 1 to 3 years. At age 5 years, calcium intake was 767271 mg, which was much closer to both the Australian EAR and RDI for children aged 4 to 8 years (520 and 700 mg, respectively). While iron intakes were 20% to 50% greater than the Australian EARs for equivalent ages (7 mg at age 7 to 12 months, 4 mg at age 1 to 3 years, and 4 mg at age 4 to 8 years), these intakes did not meet the Australian RDIs at any age (11 mg at age 7 to 12 months, 9 mg at age 1 to 3 years, and 10 mg at 4 to 8 years). JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

1191

RESEARCH Table 2. Protein intakes (grams per day and grams per kilogram body weight), protein density (grams per 1,000 kcal) and the proportion (%) of energy from protein, energy intake (kcal per day), calcium, and iron intakes (milligrams per day) at T2 (age 9 months), T3 (age 18 months), T4 (age 3.5 years), and T5 (age 5 years) in the Melbourne Infant Feeding Activity and Nutrition Trial Programa Nutrient

T2 (n[381)

T3 (n[321)

T4 (n[237)

T5 (n[235)

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒmeanstandard deviationƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ! Protein intake (g/d)

29.711.0

46.311.5

54.213.8

60.014.8

Protein per bodyweight (g/kg)

3.41.2

4.11.0

3.30.8

3.00.8

Protein density (g/1,000 kcal)

34.77.2

43.96.7

42.66.7

42.66.7

% total energy from protein

14.12.9

17.82.7

17.32.7

17.42.8

Energy intake (kcal/d)

840200

1,057199

1,274261

1,404292

Calcium (mg/d)

643251

759220

732239

767271

9.34.8

6.52.4

7.22.4

8.12.6

Iron (mg/d) a

Nutrient reference values for protein, calcium, and iron (Adequate Intake [AI], Estimated Average Requirement [EAR], Recommended Dietary Intake [RDI]). Protein: age 7 to 12 months (AI: 14 g/day, 1.6 g/kg), age 1 to 3 years (EAR: 12 g/day, 0.92 g/kg; RDI: 14 g/day, 1.08 g/kg); age 4-8 years (EAR: 16 g/day, 0.73 g/kg; RDI: 20 g/day, 0.91 g/kg); Calcium: age 7-12 months (AI: 270 mg/day), 1-3 years (EAR: 360 mg/day, RDI: 500 mg/day); age 4 to 8 years (EAR: 520 mg/day, RDI: 700 mg/day); Iron: age 7 to 12 months (EAR: 7 mg/day, RDI: 11 mg/day); age 1 to 3 years (EAR: 4 mg/day, RDI: 9 mg/day); age 4 to 8 years (EAR: 4 mg/day, RDI: 10 mg/day).

Food Sources of Protein The primary food sources of protein changed across the first 5 years of life (Table 3). At T2 (age 9 months), infant formula was the predominant source of protein followed by milk and milk products, breads/cereals, meat and poultry, and breast milk. Further analyses of the sources of milk (breastmilk, infant formula, and dairy milk) at T2 showed that infant formula, and to a lesser degree dairy milk were the predominant milk sources (providing more than 66.7% of all milk consumed) for nearly two-thirds of infants at age 9 months (data not shown). At all other ages (age 18 months, 3.5 years, and 5 years), milk and milk products were the primary sources of protein, followed by breads/cereals, meat and poultry and together these three groups of food contributed 66% to 71% of total protein intake. Discretionary foods, such as cakes/cookies and processed meats were also major protein sources at ages 3.5 and 5 years, each accounting for approximately 5% of total protein intake by 5 years. Although protein intakes from breads/cereals at T3, T4, and T5 were more than double the consumption at T2, the percentage of total protein from all milk decreased across ages from 47% at T2 to 26% at T5, and the percentage of total protein from animal sources (meat, poultry, egg, fish, and processed meat all combined) increased from T2 and T3 (21%) to T4 and T5 (27% to 28%). Protein intake from fruit and vegetables was consistent among all ages, providing 6% to 8% of total protein. Similarly, the percentage of total protein from fish (3% to 4%) was consistent across all ages.

Correlates of Protein Density Feeding choice (ie, the choice of breastfeeding or formula feeding) and the timing of the introduction of solids were all significantly correlated with protein density (Table 4). At T2, (child age 9 months) both the earlier introduction of solids, and the primary milk source being formula/dairy or mixed (as opposed to the primary milk source being breastmilk) were associated with consuming significantly more protein 1192

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

per 1,000 kcal. At T3 (child age 18 months), earlier cessation of breastfeeding, and the consumption of more than 500 mL/ day dairy milk was significantly associated with higher protein density, whereas at T4 (child age 3.5 years), later introduction of solids was associated with higher protein density. At T4 (child age 3.5 years), a number of baseline maternal characteristics were also significantly associated with protein density such that children of more highly educated mothers consumed diets with higher protein density, and higher maternal prepregnancy BMI and maternal employment at baseline were associated with lower protein density. No child or maternal characteristics were found to be associated with protein density at T5 (child age 5 years).

Tracking of Protein Intakes Residualized protein intake at T2 was significantly associated with T3 (r¼0.16; P¼0.007) and T5 (r¼0.18; P¼0.006) scores, but not T4 scores (r¼e0.02; P¼0.669). Residualized protein intake at T3 was associated with both T4 (r¼0.24; P¼0.001) and T5 scores (r¼0.21; P¼0.003), whereas T4 and T5 scores were also significantly correlated (r¼0.26; P<0.0005).

DISCUSSION High protein intakes in young children are concerning given evidence of associations with early rapid growth, and in turn adiposity and its associated sequelae across the life course.19 Describing protein intakes across childhood, food sources of protein, and the correlates of protein intakes provides insights that may enable public health action to address this important, and likely modifiable, predictor of child adiposity. In this cohort of children, protein intakes were consistently and substantially in excess of existing recommendations. Although mean protein intakes were double the Australian RDI at age 9 months, at all other ages intakes were around three times the recommended intakes for similar ages, using both total grams per day and grams per kilogram body August 2017 Volume 117 Number 8

August 2017 Volume 117 Number 8

Table 3. Main protein food sources at T2 (age 9 months), T3 (age 18 months), T4 (age 3.5 years), and T5 (age 5 years) in Melbourne Infant Feeding Activity and Nutrition Trial Program (food sources that provided at least 2% of total protein at one or more time points were presented) T2 (n[381) Food group

% of total %a energy

% Total protein

T3 (n[321) Protein provided % Total (g) %a energy

ƒmeanstandard deviationƒ!

% Total protein

T4 (n[237) Protein provided % Total (g) %a energy

ƒmeanstandard deviationƒ!

Breads/ cereals

94

9.97.7

10.78.9

3.22.9

100 21.89.4

19.69.3

Pasta

% Total protein

T5 (n[235) Protein provided % Total (g) %a energy

ƒmeanstandard deviationƒ!

8.94.4 100 24.29.4

22.19.3

% Total protein

Protein provided (g)

ƒmeanstandard deviationƒ!

11.85.8 100 25.29.0 22.79.5

13.56.3

43

1.32.3

1.12.1

0.30.6

57 2.43.5

1.82.8

0.91.4

57 2.43.3

1.82.5

11.3

55 2.53.9

23.4

1.11.8

Infant 75 cereals/ products

3.84.9

3.75.0

1.01.3

21 0.41.0

0.31.0

0.10.4

1 0.00.4

0.00.3

0.00.3

0 0.00.0

0.00.0

0.00.0

Cakes/ cookies

48

1.32.6

0.81.6

0.20.5

88 6.05.6

3.03.4

1.31.4

95 9.36.6

4.53.9

2.42.1

96 10.36.6

5.14.1

32.4

Fruit

96

6.25.1

2.62.5

0.70.7

99 9.95.2

3.52.1

1.50.9

99 9.45.3

3.32.3

1.71.1

99 9.14.8

3.11.9

1.81.0

Vegetables 93

5.65.5

1.61.6

89 3.13.9

2.72.8

1.21.2

93 2.73.1

2.52.0

1.31.1

95 3.13.4

3.13.0

1.81.7

1.62.0

10 1.45.7

0.63.4

0.20.8

1 0.00.0

0.00.0

0.00.0

0 0.00.0

0.00.0

0.00.0

70 32.226.4 23.921.4 7.26.0 Infant/ toddler formula

13 2.68.4

2.27.6

0.92.9

1 0.22.9

0.33.1

0.11.5

1 0.22.7

0.22.8

0.11.9

Milk and 91 10.28.1 milk products

98 32.414.4 38.116.3 17.78.6

99 23.910.1 30.211.9 16.47.7

98 19.98.9 26.011.2 15.98.3

15.411.5 4.74.0

Beef, veal, 62 lamb

3.04.2

9.011.4 3.14.4

69 3.13.9

7.68.7

3.94.9

66 3.34.2

8.410

5.06.8

63 3.23.9

8.39.4

5.26.4

Poultry

55

2.33.9

8.111.3 2.84.5

58 2.23.3

5.98.1

2.94.6

62 2.43.2

6.48.0

3.44.4

59 3.54.6

9.010.7

5.46.7

Egg and egg dishes

11

0.31.0

0.51.7

0.20.6

28 1.12.6

1.73.7

0.81.9

37 1.22.1

2.03.6

11.8

39 1.22.3

2.03.5

1.22.1

Fish

33

0.81.8

3.26.3

1.02.1

37 1.32.6

3.05.7

1.53.1

39 1.62.9

3.96.7

2.24.1

43 1.93.2

4.47.0

2.95.0

Processed 13 meats

0.20.8

0.62.2

0.21.0

47 1.32.4

2.44.0

1.11.9

72 3.44.3

6.46.8

3.53.9

73 2.53.1

5.05.5

3.03.3

Infant foods

3.74.7

2.74.4

0.71.2

33 1.22.5

0.71.6

0.30.7

5 0.31.2

0.10.3

0.00.2

3 0.10.3

0.00.1

0.00.1

1193

a

58

Percentage of children who consumed the food.

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

5.85.3 7.812

Breastmilk 44 15.721.3

RESEARCH

1194 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Table 4. Child and maternal correlates of protein density (g/1,000 kcal) at T2 (age 9 months), T3 (age 18 months), T4 (age 3.5 years), and T5 (age 5 years) in the Melbourne Infant Feeding Activity and Nutrition Trial programa T2 (n[381)

b

T3 (n[321)

T4 (n[237)

P value b

a

0.002

.41 e0.32 to 1.14

0.270

.44 e0.88 to 1.76

0.512

.53 e0.85 to 1.92

0.449

e.12 e1.11 to 0.27

0.267

a

95% CI

95% CI

P value b

a

95% CI

T5 (n[235) P value b

a

(95% CI)

P value

Child factors Child age Child sex female (reference category¼male) Child age at breastfeeding cessation Child age when started solids

1.21 0.44 to 1.99

e1.27 e2.06 to e0.48 0.002

.28 e0.09 to 0.65

0.140

.20 e0.43 to 0.83 .530

1.13 0.50 to 02.76 0.174

.19 e1.53 to 1.92 .827

e.21 e0.34 to e0.07 0.003

e.10 e0.25 to 0.05

0.196

e.01 e0.03 to 0.08 .673

e.03 e0.98 to 0.92

0.951

1.51 0.40 to 2.61

0.008

e.02 e1.20 to 1.16 .978

1.68 0.05 to 3.31

0.044

e.19 e0.37 to 0.00

0.045

.17 e0.00 to 0.35 .057

e.04 e0.24 to 0.17

0.729

.06 e0.17 to 0.29 .601

e3.34 e6.34 to e0.35 0.029

e2.30 e5.86 to 1.27 .207

Primary milk source (reference category¼breastmilk) Formula/dairy

2.58 0.14 to 5.02

0.038

b

4.82 2.01 to 7.64

0.001

Mixed

Child consumed >500 g/d dairy milk at T2 Maternal factors Prepregnancy body mass index Age at baseline Working at baseline Tertiary educated at baseline August 2017 Volume 117 Number 8

Born overseas

e.02 e0.15 to 0.11

0.773

.01 e0.13 to 0.15

0.913

e.04 e0.20 to 0.13

0.644

.13 e0.04 to 0.30

0.142

e1.21 e3.60 to 1.18

0.321

.59 e2.07 to 3.26

0.664

.40 e1.01 to 1.80

0.581

1.23 e0.26 to 2.72

0.105

e.53 e2.19 to 1.12

0.530

e.15 e1.87 to 1.57

0.864

2.37 0.62 to 4.13 .12 e2.10 to 2.34

0.008

.89 e0.97 to 2.74 .350

0.918

.82 e1.42 to 3.05 .474

a b is the regression coefficient of association between child/maternal factor and protein density. For categorical child/maternal factors, b is interpreted as the difference in protein density compared with the reference category. For continuous child/ maternal variable, b is interpreted as change in protein density associated with per unit change in child/maternal factors. b Mixed¼breastmilk and formula/dairy milk.

RESEARCH weight. These findings are consistent with Australian national dietary data20 for children aged 2 to 17 years, and with data internationally where reported intakes in infants and toddlers are 250% to 300% greater than country-specific recommendations in the United Kingdom, Europe, and the United States.21-23 The current study elucidates the food sources of protein at several ages across early life, thus enabling identification of foods that may be targeted to reduce protein intakes in this life phase. Milk and milk products were the principle sources of protein across the first 5 years of life, peaking at age 18 months when this food group provided nearly 40% of all protein. This high consumption of milk and milk products was accompanied by the consumption of a range of other animal products: meats, poultry, fish, eggs, and processed meats. Although the relative contribution of each of these groups to total protein intakes varied across each time point, in total, meats, poultry, fish, eggs, and processed meats were supplying around one-quarter (23%) of total protein consumed. Consistent with milk and milk products, these animal protein foods also contribute important nutrients essential for healthy growth in childhood. In seeking to reduce protein intakes it is important that other key nutrients (eg, calcium and iron) are not compromised. At all ages, calcium intakes substantially exceeded the Australian RDIs in this population and suggest a reduction in milk intake in the first 3 years of life is not likely to compromise children’s calcium intakes. Although it was not surprising to find that milk was the principal source of protein at T2 (9 months) a time when children are moving from a predominantly milk-based diet to family foods, it is important to note that infant formula was consumed by most infants at this time. Around half of women (46.7%) reported they were breastfeeding when infants were 9 months old; however, many of these women frequently used a mixture of breastfeeding and formula feeding, thus increasing total protein intakes (compared with breastfeeding only).24 At age 18 months, breastmilk, infant formula, and toddler milk each provided around 5% of total milk consumed with dairy milk becoming the predominant milk source (average volume 335 mL, 85% of all milk [data not shown]). As previously reported,25 dairy milk was consumed by 90% of children aged 18 months and a quarter of these children consumed more than 500 mL/day. This high level of milk consumption is concerning as unmodified cows’ milk is a poor source of iron and is likely to displace potentially ironrich foods such as red meat. Indeed, this level of cow’s milk consumption in the first 2 years of life has been associated with substantially lower ferritin concentrations.26,27 Furthermore, evidence suggests high intake of animal, especially dairy protein intake at 12 months is predictive of an unfavorable body composition in childhood.28 Meat and meat products, particularly red meats are important sources of many nutrients and of specific interest in early life are their contribution to heme iron intake. These nutrients are known to frequently be at risk in this population.25 Previous analyses from the InFANT cohort identified that 33% of infants and 19% of toddlers were at risk of inadequate iron intakes,25 consistent with other research demonstrating that children aged 4 to 24 months did not consume enough iron-rich meat sources with processed meat August 2017 Volume 117 Number 8

intake displacing other sources.29 Given the importance of iron for young children’s growth and development,30,31 it is important to acknowledge that a reduction in iron-rich sources of red meat as a means by which to reduce total protein is not recommended. There is good rationale to encourage parents to replace processed meats, which in this sample provided around 5% of protein at ages 3.5 and 5 years, with fresh meats. Processed meats are relatively poor sources of iron and have also been identified in this cohort as important contributors of sodium.7 The findings of modifiable correlates of protein intake in this study provide useful leads for how to appropriately focus public health messaging regarding young children’s protein intakes. The finding that at age 9 months protein density was correlated both to earlier introduction of solids and the primary milk source being formula/dairy or mixed (as opposed to the primary milk source being breastmilk), reinforces the importance of health messaging promoting increased duration of exclusive breastfeeding; and relatedly, delaying the introduction of both infant formula and solid foods (to around age 6 months). These messages are broadly consistent with dietary guidelines in Australia,32 the United Kingdom,33 and the United States.34 The Australian Infant Feeding Guidelines recommend limiting consumption of dairy milk to 500 mL/ day.15 As outlined by Cameron and colleagues35 early cessation of breastfeeding (and in turn introduction of infant formula) and early introduction of solids are socioeconomically patterned, with mothers of lower socioeconomic position more likely to engage in these behaviors. Given the significantly increased risk of adiposity in children from lower socioeconomic position environments,35 these findings provide further support for the need to address the barriers to the adoption of these key health behaviors in at-risk groups. A number of baseline maternal characteristics emerged as predictors of protein density in later childhood. Maternal university education (a proxy for increased socioeconomic advantage) was directly associated with a child’s increased protein density at age 3.5 years, whereas lower protein density at this age was associated with higher maternal prepregnancy BMI and maternal employment at baseline, both potential proxies for lower maternal socioeconomic position. These associations may reflect financial capacity to purchase foods rich in protein but relatively lower in energy (eg, lean chicken rather than chicken nuggets, fresh/tinned fish rather than fish fingers, and lean meat vs sausages). It is also possible that the societal trend positing protein as a socalled good food to be consumed in preference to carbohydrates36 may be more strongly adopted by women with higher levels of education. These findings are important and may indeed represent barriers to reducing children’s protein intakes in higher socioeconomic position families. No baseline child and maternal correlates of protein density were observed at T5 (child age 5 years). This may reflect the dilution of association over time and/or unmeasured confounders and predictors likely to change substantially over a child’s early life.

Tracking Slight tracking of protein intake was observed between most age groups. Intake at age 9 and 18 months predicted intake at age 5 years, although the tracking coefficients were low. JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

1195

RESEARCH Similarly, previous studies also found slight tracking of nutrient intakes starts early in preschool age.37,38 Tracking of protein intake reinforces the premise that dietary behaviors expressed in early life may herald the maintenance of these behaviors across life.

with body mass index and percentage body fat at 7 y of age. Am J Clin Nutr. 2007;85(6):1626-1633. 5.

Campbell K, Hesketh K, Crawford D, Salmon J, Ball K, McCallum Z. The Infant Feeding Activity and Nutrition Trial (INFANT) an early intervention to prevent childhood obesity: Cluster-randomised controlled trial. BMC Public Health. 2008;8:103.

6.

Campbell KJ, Lioret S, McNaughton SA, et al. A parent-focused intervention to reduce infant obesity risk behaviors: A randomized trial. Pediatrics. 2013;131(4):652-660.

7.

Campbell KJ, Hendrie G, Nowson C, et al. Sources and correlates of sodium consumption in the first 2 years of life. J Acad Nutr Diet. 2014;114(10):1525-1532 e1522.

8.

Lioret S, McNaughton SA, Spence AC, Crawford D, Campbell KJ. Tracking of dietary intakes in early childhood: The Melbourne InFANT Program. Eur J Clin Nutr. 2013;67(3):275-281.

9.

Blanton CA, Moshfegh AJ, Baer DJ, Kretsch MJ. The USDA Automated Multiple-Pass Method accurately estimates group total energy and nutrient intake. J Nutr. 2006;136(10):2594-2599.

10.

Emmett P, North K, Noble S. Types of drinks consumed by infants at 4 and 8 months of age: A descriptive study. The ALSPAC Study Team. Public Health Nutr. 2000;3(2):211-217.

11.

Lennox A, Sommerville J, Ong K, Henderson H, Allen R. Diet and Nutrition Survey of Infants and Young Children, 2011. https://www. gov.uk/government/uploads/system/uploads/attachment_data/file/ 139571/DNSIYC_Executive_Summary_UK__updated_.pdf. Accessed March 15, 2017.

12.

Kent JC, Mitoulas L, Cox DB, Owens RA, Hartmann PE. Breast volume and milk production during extended lactation in women. Exp Physiol. 1999;84(2):435-447.

13.

Nutrient Reference Values for Australia and New Zealand, Including Recommended Dietary Intakes. Canberra, Australia: National Health and Medical Research Council; 2006.

14.

Infant Feeding Guidelines. Canberra, Australia: National Health and Medical Research Council; 2012.

15.

WHO Child Growth Standards: Length/Height-for-Age, Weight-for-Age, Weight-for-Length, Weight-for-Height and Body Mass Index-for-Age: Methods and Development. Geneva, Switzerland: World Health Organization; 2006.

16.

Willett WC. Nutritional Epidemiology. New York, NY: Oxford University Press; 2013.

17.

Madruga SW, Araujo CL, Bertoldi AD, Neutzling MB. Tracking of dietary patterns from childhood to adolescence. Revista de saude publica. 2012;46(2):376-386.

18.

Stata Statistical Software Release 14. StataCorp LP, College Station, TX; 2015.

19.

Michaelsen KF, Greer FR. Protein needs early in life and long-term health. Am J Clin Nutr. 2014;99(3 suppl):718S-722S.

20.

Australian Health Survey: Nutrition First Results—Foods and Nutrients, 2011-12. Canberra, Australia: Australian Bureau of Statistics; 2014. Catalog No. 4364.0.55.007.

21.

Hardwick J, Sidnell A. Infant nutrition—diet between 6 and 24 months, implications for paediatric growth, overweight and obesity. Nutr Bull. 2014;39(4):354-363.

22.

Saavedra JM, Deming D, Dattilo A, Reidy K. Lessons from the feeding infants and toddlers study in North America: What children eat, and implications for obesity prevention. Ann Nutr Metab. 2013;62(suppl 3):27-36.

23.

Dalmau J, Moráis A, Martínez V, et al. Evaluation of diet and nutrient intake in children under three years old. ALSALMA pilot study/Evaluación de la alimentación y consumo de nutrientes en menores de 3 años. Estudio piloto ALSALMA. Anales de Pediatría. 2014;81(1):22-31.

24.

Magarey A, Kavian F, Scott JA, Markow K, Daniels L. Feeding mode of Australian infants in the first 12 months of life: An assessment against national breastfeeding indicators. J Hum Lact. 2016;32(4):NP95-NP104.

25.

Atkins LA, McNaughton SA, Campbell KJ, Szymlek-Gay EA. Iron intakes of Australian infants and toddlers: Findings from the Melbourne Infant Feeding, Activity and Nutrition Trial (InFANT) Program. Br J Nutr. 2016;115(2):285-293.

26.

Thorsdottir I, Gunnarsson BS, Atladottir H, Michaelsen KF, Palsson G. Iron status at 12 months of age—Effects of body size, growth and diet in a population with high birth weight. Eur J Clin Nutr. 2003;57(4):505-513.

Strength and Limitations To our knowledge, this is the first study to describe longitudinally the protein intakes, food sources, modifiable correlates, and protein tracking during infancy and early childhood and thus provides comprehensive data that will inform opportunities for safe protein reductions in early childhood. A large and well-retained sample and the assessment of dietary intake via three 24-hour recalls are important strengths of this study design. However, the present study included only healthy full-term infants and is thus not representative of nutritionally at-risk populations, such as preterm infants and infants with low birth weight for gestational age. It is also important to acknowledge potential sources of bias, including an overrepresentation of university-educated women and the potential for reporting bias when collecting self-report data. Misreporting of dietary intake may be due to a number of factors, such as a reliance on memory and inadequate recall and social desirability. However, misreporting was not assessed in this study as there is currently a lack of consensus in the literature in terms of the appropriate methods for this age group, and is complicated by larger day-to-day variation in children’s dietary intake.39 However, the energy intakes of children aged 3.5 and 5 years in the present study were lower than their national representative counterparts.20 Although misreporting may be present, it may have little influence on the findings pertaining to protein sources and examining tracking as misreporting may influence the same individuals over time.40

CONCLUSIONS Protein intake consistently exceeded national recommendations across the first 5 years of life in this cohort. The early cessation of breastfeeding along with the early introduction of infant formula and solid foods were all correlated with children’s higher protein intakes and, therefore, are important potential targets in any efforts to reduce children’s protein intakes. The present study supports a reduction in milk and milk products, but not meats (with the exception of processed meats). These findings need to be considered in both nutrition counseling and nutrition policy formulation.

References 1.

Williams C, Oxon B, Lond H. Kwashiorkor: A nutritional disease of children associated with a maize diet. Bull World Health Organ. 2003;81:912-913.

2.

Hornell A, Lagstrom H, Lande B, Thorsdottir I. Protein intake from 0 to 18 years of age and its relation to health: A systematic literature review for the 5th Nordic Nutrition Recommendations. Food Nutr Res; 2013:57.

3.

Mihrshahi S, Battistutta D, Magarey A, Daniels LA. Determinants of rapid weight gain during infancy: Baseline results from the NOURISH randomised controlled trial. BMC Pediatr. 2011;11:99.

4.

Gunther AL, Buyken AE, Kroke A. Protein intake during the period of complementary feeding and early childhood and the association

1196

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

August 2017 Volume 117 Number 8

RESEARCH 27.

Thane CW, Walmsley CM, Bates CJ, Prentice A, Cole TJ. Risk factors for poor iron status in British toddlers: Further analysis of data from the National Diet and Nutrition Survey of children aged 1.5-4.5 years. Public Health Nutr. 2000;3(4):433-440.

28.

Günther ALB, Remer T, Kroke A, Buyken AE. Early protein intake and later obesity risk: Which protein sources at which time points throughout infancy and childhood are important for body mass index and body fat percentage at 7 y of age? Am J Clin Nutr. 2007;86(6):1765-1772.

29.

Mauch CE, Perry RA, Magarey AM, Daniels LA. Dietary intake in Australian children aged 4-24 months: Consumption of meat and meat alternatives. Br J Nutr. 2015;113(11):1761-1772.

30.

Beard JL. Why iron deficiency is important in infant development. J Nutr. 2008;138(12):2534-2536.

31.

Georgieff MK. Long-term brain and behavioral consequences of early iron deficiency. Nutr Rev. 2011;69(suppl 1):S43-S48.

32.

National Health and Medical Research Council. Australian Dietary Guidelines. Canberra, Australia: National Health and Medical Research Council; 2013.

33.

Public Health England, Welsh Government, Scottish Government, Food Standards Agency in Northern Ireland. Your Guide to the Eatwell Plate Helping you eat a Healthier Diet. London, UK: Public Health England; 2013.

34.

Nutrition and Your Health. 2015-2020 Dietary Guidelines for Americans. 8th ed. Washington, DC: US Government Printing Office; 2015.

35.

Cameron AJ, Spence AC, Laws R, Hesketh KD, Lioret S, Campbell KJ. A Review of the relationship between socioeconomic position and the early-life predictors of obesity. Curr Obes Rep. 2015;4(3):350362.

36.

Talati Z, Pettigrew S, Moore S, Pratt S. Discrepancies between Consumers’ Nutrition Beliefs and Current Nutrition Guides. Melbourne, Australia: Australian & New Zealand Obesity Society 2015 Annual Scientific Meeting; October 15-17, 2015.

37.

Singer MR, Moore LL, Garrahie EJ, Ellison RC. The tracking of nutrient intake in young children: The Framingham Children’s Study. Am J Public Health. 1995;85(12):1673-1677.

38.

Stein AD, Shea S, Basch CE, Contento IR, Zybert P. Variability and tracking of nutrient intakes of preschool children based on multiple administrations of the 24-hour dietary recall. Am J Epidemol. 1991;134(12):1427-1437.

39.

Livingstone MB, Robson PJ. Measurement of dietary intake in children. Proc Nutr Soc. 2000;59(2):279-293.

40.

Black AE, Cole TJ. Biased over- or under-reporting is characteristic of individuals whether over time or by different assessment methods. J Am Diet Assoc. 2001;101(1):70-80.

AUTHOR INFORMATION K. J. Campbell is a professor, Population Nutrition, G. Abbott is a research fellow level 2 (data manager), M. Zheng is a National Health and Medical Research Council Early Career research fellow, and S. A. McNaughton is an associate professor, National Health and Medical Research Council Career Development fellow (level 2), Institute for Physical Activity and Nutrition, School of Exercise and Nutrition Sciences, Deakin University, Geelong, Victoria, Australia. Address correspondence to: Karen J. Campbell, PhD, MPH, Institute for Physical Activity and Nutrition, Deakin University, Locked Bag 20000, Geelong, Victoria, Australia 3220. E-mail: [email protected]

STATEMENT OF POTENTIAL CONFLICT OF INTEREST No potential conflict of interest was reported by the authors.

FUNDING/SUPPORT S. A. McNaughton is supported by a National Health and Medical Research Council Career Development Fellowship Level 2, ID1104636, and was previously supported by an Australian Research Council Future Fellowship (2011-2015, FT100100581). M. Zheng is supported by a National Health and Medical Research Council Early Career Fellowship (ID1124283).

August 2017 Volume 117 Number 8

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

1197

RESEARCH Table 1. Milk intakes (in milliliters) at T2 (age 9 months) according to primary milk source in the Melbourne Infant Feeding Activity and Nutrition Trial program 66.7%-100% breastmilk (n[128)

Milk

66.7%-100% formula/dairy milk (n[225)

Mixeda (n[28)

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ meanstandard deviationƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ! Total milk

440.8

133.6

720.9

181.7

528.0

174.3

Dairy milk

19.3

35.0

34.1

80.6

92.6

101.5

14.8

41.3

680.0

196.2

212.6

155.7

406.8

126.8

6.7

28.5

222.9

124.2

Infant formula Breastmilk a

Mixed¼breastmilk and formula/dairy milk.

1197.e1

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

August 2017 Volume 117 Number 8