Intake, sources, and determinants of free sugars intake in Australian children aged 12–14 months

Abstract

The consumption of free sugars is directly associated with adiposity and dental caries in early childhood; however, intake data in the first 2 years of life are limited. This cross‐sectional analysis aims to identify major food sources of free sugars for Australian children aged 12–14 months and investigate factors associated with meeting the World Health Organisation (WHO) Guideline for sugars intake. Three days of nonconsecutive dietary data were collected via a 24‐hr recall and 2‐day food record for 828 participants. Usual intake of energy, total sugars, and free sugars were estimated, along with food group contributions to free sugars. Multiple logistic regression analysis was used to investigate factors associated with exceeding the WHO conservative recommendation that <5% of energy should come from free sugars. Mean free sugars intake was 8.8 (SD 7.7, IQR 3.7–11.6) g/day, contributing 3.6% (SD 2.8, IQR 1.6–4.8) of energy. Only 2.4% of participants exceeded the WHO recommendation that <10% of energy should come from free sugars, with 22.8% of participants exceeding the <5% recommendation. Children from households with greater socio‐economic disadvantage (IRSAD <5, OR = 1.94) and in the lowest income bracket (OR = 2.10) were more likely to have intakes ≥5% of energy. Major food sources of free sugars were commercial infant foods (26.6%), cereal‐based products (19.7%), namely, sweet biscuits (8.3%) and cakes (7.6%), followed by yoghurt (9.6%), and fruit and vegetable beverages (7.4%). These findings highlight the substantial contribution of infant foods to free sugars intakes and provide further evidence that dietary intakes are influenced by social determinants.

Key messages.

• Only 2.4% of 12‐ to 14‐month olds consumed ≥10% of energy from free sugars, although 22.8% consumed ≥5% energy from free sugars.

• Children from households with greater socio‐economic disadvantage and in the lowest income bracket were more likely to exceed the WHO <5% recommendation.

• With the exception of commercial infant foods, the major food sources of free sugars as children transitioned to the family diet were comparable with those seen later in life.

• These findings highlight the substantial contribution to free sugars made by infant and toddler foods, along with the early emergence of dietary behaviours and influences of social determinants.

1. INTRODUCTION

Food intakes in the first 2 years of life shape taste preferences and healthy eating habits, which can have lasting health implications throughout life (Birch & Doub, 2014; Lillycrop & Burdge, 2011). In particular, intake of foods high in free sugars is associated with increased body weight and dental caries in early childhood (World Health Organisation [WHO], 2015). Dental decay is the leading cause of preventable hospitalisations among Australian children, and prevalence is increasing (Australian Institute of Health and Welfare, 2012; Australian Research Centre for Population Oral Health, 2011). This disease burden is consistently higher among those with lower socio‐economic status, and income‐related inequality in caries experience of Australian children appears to be widening (Armfield & Spencer, 2008; Do et al., 2010).

Diet quality also follows a social gradient (Darmon & Drewnowski, 2008). Socio‐economic disparity has been observed in Australian children's consumption of energy‐dense, nutrient‐poor foods and beverages (Australian Bureau of Statistics [ABS], 2015; Bell et al., 2016; Cameron et al., 2012), many of which are high in added or free sugars. The WHO (2015) recommends reducing intakes of free sugars in adults and children to less than 10% of total energy intake (EI) and conditionally recommends a further reduction to less than 5% of energy. The WHO (2015) definition of free sugars includes “monosaccharides and disaccharides added to foods and beverages by the manufacturer, cook or consumer, and sugars naturally present in honey, syrups, fruit juices and fruit juice concentrates” (p. 4).

An analysis of the National Nutrition and Physical Activity Survey (NNPAS) revealed that just over 50% of children aged 2–3 years exceeded the WHO recommendation that less than 10% of total energy should come from free sugars, and 93% exceeded the less than 5% recommendation (ABS, 2017). Secondary analysis of the Australian National Children's Nutrition and Physical Activity Survey found even higher rates, with 69% of 2‐ to 3‐year‐olds exceeding the 10% recommendation and 98% exceeding the 5% recommendation (Louie, Moshtaghian, Rangan, Flood, & Gill, 2016). Both studies reported increasing intakes of free sugars with age through childhood and adolescence, along with greater percentage of dietary energy from free sugars, and an increasing number of children and adolescents exceeding the WHO recommendations (ABS, 2017; Louie et al., 2016).

Most children commence life on a diet devoid of free sugars, as breastmilk and the majority of infant formula contain no added or free sugars. However, in the absence of national nutrition monitoring in the first 2 years of life in Australia, little is known of the contribution of free sugars to the diets of young children as they transition from a milk‐based diet to one that reflects family intakes, and at what point free sugars enter the diet. The aims of this study are to investigate major food sources of free sugars for Australian children aged 12–14 months; to compare the percentage of energy from free sugars to the WHO guidelines; and to explore sociodemographic factors associated with exceeding the WHO guidelines in this population.

2. METHODS

This study is a cross‐sectional analysis of dietary data collected by the Study of Mothers and Infants Life Events Affecting Oral Health (SMILE), a longitudinal birth cohort study based in Adelaide, Australia. Details of the SMILE protocol are reported elsewhere (Do et al., 2014). Briefly, 2,147 mothers and 2,181 newborns, including 34 pairs of twins, were recruited from the three major maternity hospitals in Adelaide from July 2013 until August 2014. All new mothers with sufficient English competency and not intending to move out of the greater Adelaide area within a year were invited to participate. Mothers in hospitals that service lower socio‐economic areas were oversampled to compensate for anticipated higher attrition rates (Do et al., 2014).

Participants were invited to complete questionnaires at recruitment, and when their child reached 3, 6, 12, and 24 months of age. These questionnaires collected information on dental and dietary behaviours and were available in paper, electronic, and telephone interview forms.

The Southern Adelaide Clinical Human Research Ethics Committee approved the study (HREC/50.13, approval date: 28 Feb 2013) as did the South Australian Women and Children Health Network (HREC/13/WCHN/69, approval date: Aug 7, 2013).

2.1. Dietary data collection at 12 months of age

When the child reached 12 months of age, mothers were invited to complete a 24‐hr recall via telephone interview and 2‐day food record, giving a total of two weekdays and one weekend day of dietary data.

A range of strategies were employed to maximise response rate and quality of dietary data. Participants were posted a food record booklet containing detailed instructions, a 1‐day example food record, and images of food portion sizes and household measures. Approximately 1 week after the materials were sent, one of two research dietitians contacted participants via telephone to conduct the 24‐hr recall. The five‐step multipass method (ABS, 2014) was used, with a modified forgotten foods list that replaced the question about alcohol with “any rusks or teething foods?”, and a uniquely designed probing protocol with specific questions for toddler foods in the detail cycle phase. During the recall, images in the food record booklet were used as a reference for obtaining quantity information over the phone.

At the end of the interview, mothers were allocated two nonconsecutive days over the next 10‐day period, on which to record their child's dietary intake using the food record booklet. Participants who completed the 24‐hr recall received SMS and email reminders to return their food record.

The resulting 3 days of dietary data were entered into FoodWorks version 8 (Xyris Software, 2012–2017, Brisbane, Australia) using the current Australian food composition database, AUSNUT 2011–13 (Food Standards Australia and New Zealand, 2014). The data were double‐entered by trained nutritionists/dietitians, using data entry protocols and calibration procedures for standardisation. When detail was lacking in the food record, the 24‐hr recall was used for clarification purposes. For example, where a participant wrote only “milk” in the food record, the type (full‐cream, skim, etc.) specified in the 24‐hr recall was selected.

Due to the limited number of infant foods in the AUSNUT 2011–13 database and the rapidly changing infant and toddler food market, 187 new foods were added to the database using information from product nutrition panels and manufacturer websites, mapped to a similar product in AUSNUT 2011–13 for missing micronutrient values. These foods included fruit and vegetable pouches, ready to eat meal pouches and jars, dried and ready to eat flavoured infant cereals, infant and toddler yoghurts, snack bars, and other infant and toddler snack foods. Free sugars values for these products were determined by adapting the methodology of Louie et al. (2015) using the WHO definition of free sugars (WHO, 2015). This is the same method that was used to obtain free sugars values for the AUSNUT 2011–13 database (Food Standards Australia and New Zealand, 2018). Each new food was assigned an 8‐digit food code, following the AUSNUT naming conventions. This resulted in the generation of two new 5‐digit food subgroups that were not present in the current database: infant vegetable foods and infant snack foods.

After initial entry, descriptive statistics were used to identify outliers for the weight of food and energy consumed. The dietary data of these participants were rechecked for entry errors in FoodWorks and resolved as appropriate.

To determine food group contribution to nutrient intake, the AUSNUT 2011–13 food group codes were used to group free sugars intakes by major and submajor food groups. The infant foods submajor group was further broken down into minor food groups, to accommodate the relative crudeness of the infant and toddler food group codes. Mean free sugars (g) were calculated for each food group among all participants and consumers only. For the consumer only analysis, consumers were defined as any participant who reported any food item from within the group, regardless of whether it contributed to free sugars.

2.2. Explanatory variables

A range of sociodemographic data were collected at recruitment, from which the following explanatory variables were generated for this analysis: mother's country of birth (“Australia and New Zealand,” “India,” “China,” “Asia—Other,” “UK,” and “Other”); mother's age at baseline; mother's education (“high school/vocational,” and “some university and above”); mother's prepregnancy body mass index (“<18.49 kg/m²,” “18.5–24.99 kg/m²,” “25–29.99 kg/m²,” and “≥30 kg/m²”); total household income (“≤$40,000,” “$40,001–80,000,” and “≥$80,001”); parity (“1,” “2,” and “≥3”); and infant sex (“male” and “female”). Postcode was used to derive a measure of socio‐economic status using the Index of Relative Socio‐economic Advantage and Disadvantage (IRSAD; ABS, 2013). Maternal smoking status at 12 months of age (“yes” or “no”) was obtained in the 12‐month data collection phase. Milk‐feeding method at 12 months of age (“Any breastfeeding,” “formula only,” and “neither”) was derived from the 12‐month dietary data and is reported in detail elsewhere (Scott et al., 2016).

2.3. Outcome variable

The Multiple Source Method (Harttig, Haubrock, Knuppel, & Boeing, 2011) was applied to the 3 days of dietary data to estimate usual daily intake of total sugars, free sugars, and energy for each participant. The nutrient values were then used to calculate percentage of energy from free sugars, and determine the number of children meeting the WHO Guideline (WHO, 2015). As very few children failed to meet the guideline of less than 10% energy from free sugars, failing to meet the more conservative guideline of less than 5% energy from free sugars was chosen as the outcome variable.

2.4. Sensitivity analysis

Sensitivity analysis was undertaken to account for extreme over and under reporting (Magarey et al., 2011). Estimated energy requirement (EER) was determined using a reference value (National Health and Medical Research Council [NHMRC], 2006), based on participant sex and age at 24‐hr recall, as participant weight was unknown. This was used to determine the ratio of reported EI to EER for each child. Children with EI:EER below 0.54 or above 1.46 were deemed to have implausible EIs (Conn, Davies, Walker, & Moore, 2009). Primary analyses were performed on the whole cohort and then repeated on those with plausible EIs after reapplying the Multiple Source Method (Tables S2 and S3).

2.5. Statistical methods

Descriptive statistics, including means, standard deviations (SD), and interquartile ranges (IQR) were performed on the usual intake data. The percentage of children with free sugars intakes ≥5% and ≥10% of energy were also determined. For the food group contributions, the mean grams of free sugars were calculated for each food group among all participants and consumers only, and percentage contributions determined among all participants. Descriptive analysis was performed using SPSS version 24 (IBM SPSS Statistics for Windows, New York, NY, USA).

Multiple logistic regression was applied on the explanatory variables, for the dichotomous outcome <5%/≥5% of energy from free sugars. In the course of building a best model, the first analyses were performed within strata of the sample. Categorisations of continuous and categorical variables, as well as the interaction terms in the final model, are a result of these preliminary analyses. Model selection was also based on the Akaike information criterion (Akaike, 1974). The final model contains interaction between the child's sex and mother's country of birth (MCOB). The interaction coefficient is presented in combination with the sex and MCOB coefficients, as stratified by MCOB and stratified by child's sex. Multiple logistic regression was also used to compare the participant characteristics of responders to the nonresponders. Due to multiple significance tests, a P value below 0.01 was considered significant. Logistic regression analysis was performed using the statistical computing language R, version 3.4.2 (R Core Team, 2017).

3. RESULTS

3.1. Participants

At least 1 day of dietary data was received for 1,175 participants; however, only 828 of these provided 3 days of usable dietary data and were included for analysis. Figure A1 depicts participant flow.

Table 1 describes participant characteristics for the 828 respondents, and those with ≥5% of energy from free sugars. Mean age of the children at time of the 24‐hr recall was 13.1 (±0.8) months with most children (92%) aged between 12 and 14 months. The majority of mothers (74.1%) were born in Australia or New Zealand (ANZ).

Table 1.

Participant characteristics

	All respondents		Respondents with free sugars ≥5% of energy
	n	Mean ± SD	n	Mean ± SD
Mother's age at birth (years)	826	30.8 ± 4.9	189	30 ± 5.3
Child's birthweight (g)	818	3397.9 ± 558.1	186	3402.6 ± 513.0
Child's age at 24‐hour recall (months)	807	13.1 ± 0.8	182	13.2 ± 0.9

	n	%	n	% (row%)
Child's sex
Male	452	54.6	117	61.9 (25.9)
Female	376	45.4	72	38.1 (19.1)
Milk feeding method
Any breastfeeding	288	34.8	50	26.5 (17.4)
Formula only	310	37.4	75	39.7 (24.2)
Neither	230	27.8	64	33.9 (27.8)
Mother's prepregnancy BMIa (kg/m2)
<18.5	41	5.2	5	2.9 (12.2)
18.5–24.99	436	55.6	93	54.1 (21.3)
25–29.99	167	21.3	41	23.8 (24.6)
≥30	140	17.9	33	19.2 (23.6)
Mother's smoking status at 12 months
No	746	92.0	170	91.4 (22.8)
Yes	65	8.0	16	8.6 (24.6)
Total number of children
1	389	48.6	90	48.4 (23.1)
2	291	36.3	62	33.3 (21.3)
≥3	121	15.1	34	18.3 (28.1)
Mother's education
High school/vocational	356	43.2	102	54.5 (28.7)
Some University and above	468	56.8	85	45.5 (18.2)
Household income ($)
≤40,000	92	11.5	35	19.0 (38.0)
40,001–80,000	254	31.8	62	33.7 (24.4)
≥80,001	452	56.6	87	47.3 (19.2)
IRSADb
Deciles 1–4	293	35.6	92	48.7 (31.4)
Deciles 5–10	529	64.4	97	51.3 (18.3)
Mother's country of birth
Australia or New Zealand	610	74.1	151	79.9 (24.8)
Other	213	25.9	38	20.1 (17.8)

^aBody mass index (kg/m², where <18.5 = underweight; 18.5–24.99 = healthy weight range; 25–29.99 = overweight; ≥30 = obese).

^bIndex of Relative Socio‐economic Advantage and Disadvantage (where 1 = most disadvantaged and 10 = least disadvantaged).

Response rates did not differ by prepregnancy body mass index category, education level, IRSAD score, smoking status, parity, household income group, infant's sex, or birthweight. Mothers that were 25–35 and those over 35 years of age were more likely to respond than those under 25 (odds ratio [OR] = 2.24, P < 0.001 and 2.37, P < 0.01, respectively). Mothers who were born in the United Kingdom or China were equally likely to respond as those born in Australia and New Zealand, whereas those from India (OR = 0.35, P < 0.001), other parts of Asia (OR = 0.37, P < 0.001), and other countries combined (OR = 0.43, P < 0.01) were less likely to respond compared to Australia and New Zealand born mothers. Women who provided complete dietary data at 12 months were generally representative of the socio‐economic profile reported by the Pregnancy Outcome unit for South Australian births in 2013 (Scheil et al., 2015). Characteristics of study nonresponders (n = 1353), and those above and below the WHO recommendations for free sugars intakes are provided in Table S1.

3.2. Free sugars intakes at 12 months of age

Mean usual intake of free sugars was 8.8 (SD 7.7, IQR 3.7–11.6) grams per day, contributing an average of 3.6 (SD 2.8, IQR 1.6–4.8) per cent of daily energy (Table 2). In comparison with the WHO guidelines, 22.8% of participants consumed ≥5% of energy from free sugars, and only 2.4% of these consumed ≥10%.

Table 2.

Free sugars intakes at 12–14 months

	Mean ± SD	Range	Percentile 25th, 75th
Free sugars (g/day)	8.8 ± 7.7	0.22–115.3	3.7, 11.6
Total sugars (g/day)	65.3 ± 17.4	18.4–181.05	53.1, 75.1
Energy (kJ/day)	4040 ± 836	1870–7948	3446, 4574
% energy from free sugars	3.6 ± 2.8	0.12–34.3	1.6, 4.8

	n	%
<5% energy from free sugars	639	77.2
≥5% energy from free sugars	189	22.8
≥10% energy from free sugars	20	2.4

3.3. Food group contribution to free sugars

The most common sources of free sugars were commercial infant foods (27%), specifically infant custards and yoghurts (16%) and infant snacks (11%; Table 3, Figure 1a). This was followed by cereal based products and dishes (20%), namely, sweet biscuits (8%) and cakes, muffins, scones, and cake‐type desserts (8%). Other substantial contributions came from yoghurt (10%) that was not specifically identified as infant yoghurt, and fruit and vegetable juices and drinks (7%).

Table 3.

Food group contributions to free sugars intakes

Food groupa	All participants (n = 828)		Consumers
	Mean ± SD free sugars (g)	% contribution to free sugars	n (%)	Mean ± SD free sugars (g)
Nonalcoholic beverages	2.7 ± 14.9	10.3	818 (98.8)	2.7 ± 15.0
Fruit and vegetable juices, and drinks	2.0 ± 9.5	7.4	101 (12.1)	16.0 ± 22.7
Cordials	0.6 ± 10.8	2.1	10 (1.2)	46.6 ± 91.8
Cereals and cereal products	1.3 ± 3.7	4.8	810 (97.8)	1.3 ± 3.7
Breakfast cereals, ready to eat	0.7 ± 2.1	2.6	495 (59.8)	1.1 ± 2.7
Breakfast cereals, hot porridge style	0.4 ± 3.0	1.5	130 (15.7)	2.5 ± 7.2
Cereal based products and dishes	5.2 ± 9.0	19.7	674 (81.4)	6.4 ± 9.6
Sweet biscuits	2.2 ± 4.4	8.3	303 (36.6)	5.9 ± 5.6
Cakes, muffins, scones, cake‐type desserts	2.0 ± 7.0	7.6	132 (15.9)	12.6 ± 13.3
Mixed dishes where cereal is the major ingredient	0.3 ± 1.2	1.2	251 (30.3)	1.0 ± 2.0
Batter‐based products	0.3 ± 1.8	1.3	107 (12.9)	2.6 ± 4.4
Fruit products and dishesb	1.2 ± 4.1	4.7	788 (95.2)	1.3 ± 4.2
Pome fruit	0.3 ± 1.7	1.1	398 (48.0)	0.6 ± 2.5
Mixtures of two or more groups of fruit	0.6 ± 3.5	2.4	69 (8.3)	7.6 ± 9.6
Milk products and dishes	4.0 ± 8.1	15.1	760 (91.8)	4.3 ± 8.4
Yoghurt	2.5 ± 6.3	9.6	373 (45.0)	5.6 ± 8.4
Frozen milk products	0.5 ± 2.4	1.9	70 (8.5)	6.0 ± 5.9
Custards	0.5 ± 3.0	2.0	61 (7.4)	7.2 ± 8.4
Other dishes where milk or a milk product is the major component	0.3 ± 2.3	1.1	21 (2.5)	11.8 ± 8.8
Savoury sauces and condiments	0.4 ± 1.2	1.5	297 (35.9)	1.1 ± 1.9
Gravies and savoury sauces	0.4 ± 1.2	1.3	255 (30.8)	1.1 ± 1.8
Sugar products and dishes	2.3 ± 7.5	8.5	225 (27.1)	8.3 ± 12.4
Sugar, honey and syrups	1.2 ± 5.7	4.6	130 (15.7)	7.7 ± 12.7
Jam and lemon spreads, chocolate spreads, sauces	0.6 ± 2.3	2.4	101 (12.2)	5.2 ± 4.6
Dishes and products other than confectionery where sugar is the major component	0.4 ± 3.7	1.5	24 (2.9)	13.9 ± 17.5
Confectionery and cereal/nut/fruit/seed bars	0.9 ± 3.5	3.2	112 (13.5)	6.3 ± 7.5
Chocolate and chocolate‐based confectionery	0.5 ± 2.3	1.8	57 (6.9)	6.9 ± 5.7
Infant formulae and foods	7.5 ± 10.1	28.4	769 (92.9)	8.1 ± 10.2
Infant foods	7.0 ± 9.6	26.6	583 (70.4)	9.9 ± 10.0
Infant foods: Infant custards or yoghurts	4.1 ± 7.0	15.6	331 (40.0)	10.3 ± 7.7
Infant foods: Infant snack foodsc	2.8 ± 5.7	10.6	382 (46.1)	6.0 ± 7.1
Infant drinksd	0.3 ± 1.9	1.0	21 (2.5)	9.9 ± 6.7

^aFood groups providing <1% contribution to free sugars not listed.

^bFree sugars in fruit products and dishes come only from canned fruit in syrup and stewed fruit with added sugar.

^cNew category code created for infant and toddler specific snack foods not part of AUSNUT11–13.

^dInfant drinks category includes infant fruit juices, but does not include infant formula or human breast milk.

Figure 1.

Contribution to free sugars by food group. (a) All participants. (b) By participant free sugars intake. ^aFood groups providing <1% contribution to free sugars not listed. ^bFree sugars in fruit products and dishes come only from canned fruit and stewed fruit with added sugar. ^cNew category code created for infant‐ and toddler‐specific snack foods not part of AUSNUT11‐13. ^dInfant drinks category includes infant fruit juices but does not include infant formula or human breast milk

Figure 1b shows food group contribution to free sugars by participant adherence to the WHO 5% recommendation. The most common sources of free sugars for participants below 5% were similar to those of the whole group. A notable exception was sugar sweetened beverages. Both cordials and fruit and vegetable juices and drinks contributed a much lower proportion of free sugars to those with intakes below 5% of energy than to those above.

Among consumers only, food groups that supplied the greatest mean intakes of free sugars were cordials (46.6 g per consumer, n = 10) and fruit and vegetable juices and drinks (16.0 g, n = 101). These were followed by dishes and products other than confectionery where sugar is the major component (13.9 g, n = 24), which includes sugar‐based desserts, water ice confection, gelato, and sorbet; then by cakes, muffins, scones and cake‐type desserts (12.6 g, n = 132), and other dishes where milk or a milk product is the major component (11.8 g, n = 21), which includes dairy desserts (Table 3).

3.4. Factors associated with exceeding the WHO guidelines

Table 4 reports factors associated with exceeding the <5% WHO recommendation for free sugars. Children from families in the lowest household income bracket (≤40,000) had higher odds of exceeding the recommendation compared with families with higher incomes—(OR = 2.10; 99% confidence interval [99% CI] [1.05, 4.20]). Similarly, children in the lowest four IRSAD deciles were more likely to exceed the recommendation, when compared with IRSAD ≥5 (OR = 1.94; 99% CI [1.18, 3.18]).

Table 4.

Factors associated with free sugars intakes ≥5% of energy

Variable	Strata	OR	99% CI
Male (vs. female)	MCOBa: Australia or New Zealand	1.96**	[1.13, 3.42]
Male (vs. female)	MCOB: elsewhere	0.66	[0.23, 1.94]
MCOBa: Australia or New Zealand (vs. elsewhere)	Female	0.80	[0.34, 1.89]
MCOB: Australia or New Zealand (vs elsewhere)	Male	2.38*	[0.95, 5.96]
MFMb: Formula only (vs. any breastfeeding)		1.30	[0.70, 2.40]
MFM: Neither formula nor breastfeeding (vs. any breastfeeding)		1.65*	[0.88, 3.10]
Pre‐pregnancy BMI: ≥30 (vs. <30)		0.84	[0.44, 1.59]
Education: High school/vocational (vs. university)		1.33	[0.78, 2.27]
Household income: ≤$40,000 (vs. >$40,000)		2.10**	[1.05, 4.20]
IRSADc score: 1–4 (vs. 5–10)		1.94***	[1.18, 3.18]
Total number of children: ≥2 (vs. 1)		0.96	[0.59, 1.57]
Mother's smoking status: Yes (vs. no)		0.57	[0.21, 1.52]

^aMother's country of birth.

^bMilk feeding method.

^cIndex of Relative Socio‐economic Advantage and Disadvantage (where 1 = most disadvantaged and 10 = least disadvantaged).

^* P < 0.05.

^** P < 0.01.

^*** P < 0.001.

Among mothers born in ANZ, sons had twice the odds of exceeding the recommendation compared with daughters (OR = 1.96; 99% CI [1.13, 3.42]). There was no such significant difference among mothers not born in ANZ. Stratifying on the child's sex, we found a tendency (P < 0.05) of increased odds of exceeding the 5% recommendation among boys with ANZ‐born mothers compared with boys with non‐ANZ mothers. This tendency is strengthened and significant when we remove participants with implausible intakes (OR = 4.31, 99% CI [1.28, 14.48]; Table S3). There was no such difference observed among girls.

3.5. Sensitivity analysis

Removal of 129 participants with implausible intakes resulted in similar findings to the whole cohort (Tables S2 and S3). In this subset, the mean percentage of energy from free sugars was 3.4% (SD 2.5), with 80% of participants consuming less than 5%. The statistically significant factors associated with exceeding the WHO guidelines did not change.

4. DISCUSSION

Although Australian Infant Feeding Guidelines recommend that by 12 months of age infants should be consuming a wide variety of family foods (NHMRC, 2012), little is known about the transitional diet of children, as Australian national dietary surveys do not measure food intakes in children under 2 years of age. To our knowledge, this is the first study to investigate the free sugars intake of Australian children at 12–14 months of age. Although few children in this study exceeded recommended levels of intake of free sugars, findings from cohort studies indicate that diet quality varies greatly at this age (Byrne, Magarey, & Daniels, 2014) and that substantial change occurs between 9 and 18 months of age (Lioret, McNaughton, Spence, Crawford, & Campbell, 2013).

Although some children are beginning to consume family foods as they move into the second year of life, many children are already consuming energy‐dense, nutrient‐poor discretionary foods, which may displace core foods (Byrne et al., 2014; Koh, Scott, Oddy, Graham, & Binns, 2010; Lioret et al., 2013). Unsurprisingly, the major food‐group contributors to free sugars in our study were similar to the major sources of discretionary foods reported elsewhere (Koh et al., 2010; Lioret et al., 2013).

These findings suggest that although free sugars contribute a lower proportion of energy at 12 months of age than at 2–3 years, behaviours are beginning to emerge that reflect the intake patterns of older children. The major food sources of free sugars among 2‐ to 3‐year‐olds in the 2011–13 Australian NNPAS were similar to those reported here, including sweet biscuits (8.1% NNPAS, 8.3% here), cakes, muffins, scones and cake‐type desserts (9.7% NNPAS, 7.6% here), and yoghurt (6.2% NNPAS, 9.6% here; ABS, 2017).

A notable difference to the 12‐ to 14‐month‐old children in this study is the apparent nonconsumption of commercial infant foods among 2‐ to 3‐year‐olds in the NNPAS. It may be that at 2–3 years of age these products are less influential and more family foods are consumed. Alternatively, the infant and toddler food market has grown considerably since the development of AUSNUT 2011–13, in terms of both revenue and new product development (IBISWorld Australia, 2017). Toddler foods have recently differentiated from infant foods to form a distinct product line. For example, the Annabel Karmel range that launched in Australia in 2015, in addition to frozen baby purees, includes toddler and children's meals intended for ages up to 5 years (Medianet, 2015). The addition of 187 new infant and toddler foods reported in this study to the limited number of these products in the AUSNUT 2011–13 database likely reflects the growing availability and popularity of commercial toddler foods and the possibility of underestimation of these foods in the NNPAS.

A further divergence from the food sources of free sugars among 2‐ to 3‐year‐olds in the NNPAS is observed in fruit and vegetable juices and drinks (24.8% NNPAS, 7.4% here). The Australian Infant Feeding Guidelines (NHMRC, 2012) advise that fruit juice drinks are not necessary or recommended for infants under 12 months, and the modest number of consumers (n = 101, 12%) in this study suggest that parents are complying with this recommendation. Among consumers, this food group was the second highest contributor of free sugars. It appears that these drinks are introduced in the second and third years of life, as children continue to transition to family foods and milk consumption declines. A recent Australian study observed a decline in milk‐based drinks accompanied by an increase in sugar‐sweetened beverages, particularly fruit juice‐based drinks, between the ages of 2–5 years (Byrne, Zhou, Perry, Mauch, & Magarey, 2018).

We found that sons of Australian born mothers were twice as likely to exceed the WHO <5% recommendation than daughters. Dietary studies frequently adjust for sex due to the higher total energy intakes of boys than girls, and there is limited reporting on gender discrepancies. Among older children, gender differences have been reported in energy‐adjusted consumption of core food groups; however, studies are limited, and findings are inconsistent (Dubois, Farmer, Girard, Burnier, & Porcherie, 2011; Glynn, Emmett, & Rogers, 2005; Jones, Steer, Rogers, & Emmett, 2010). Further research is needed to investigate whether ours is a replicable finding, which may be attributed to sociocultural differences in perceptions of boys and girls. There is evidence that gendered food habits persist among adults (McPhail, Beagan, & Chapman, 2012), that gendered body ideals are internalised from a young age (Damiano, Paxton, Wertheim, McLean, & Gregg, 2015), and that parental eating psychopathology can result in gendered feeding behaviours, in particular by mothers towards their daughters (McPhie, Skouteris, Daniels, & Jansen, 2014).

Food preferences and eating habits established early in life can track into childhood and beyond (Nicklaus, 2009; Scott, Chih, & Oddy, 2012,) and early nutrition can influence susceptibility to obesity and other metabolic outcomes in later life (Lillycrop & Burdge, 2011; Strazzullo, Campanozzi, & Avallone, 2012). There is evidence that infants are born with a predisposition for sweet and probably salty tastes (Mennella, 2014), both of which are frequently associated with energy‐dense, nutrient‐poor foods. This biological drive is thought to have helped infants achieve adequate growth in times of energy scarcity, and may still be an important factor for early growth (Drewnowski, Mennella, Johnson, & Bellisle, 2012). However, this innate sweet preference appears to decline throughout childhood, whereas dietary experiences through early childhood shape later preferences and taste expectations around the sweetness of the food supply (Drewnowski et al., 2012). Lioret et al. (2013) found that for many core and noncore foods, being a consumer or consuming larger amounts at 9 months of age were both predictive of a greater level of consumption of the same foods at 18 months of age. Byrne et al. (2018) reported strong correlations in tracking of sweet drinks from 2 to 5 years of age. The effects of exposure to sugar on the development of food and flavour preferences may be greatest when complementary foods are being introduced. Hence, the weaning period has been identified “as an opportune time to promote the acceptance of foods that are characteristic of healthy diets” (Birch & Doub, 2014, p. 723s). It is particularly important that young children do not exceed the WHO recommendations for free sugars intakes when these early taste preferences are being formed.

High levels of total and added sugars have been reported in commercial ready‐to‐feed infant and toddler foods in the United States (Maalouf et al., 2017), Canada (Elliott, 2011), and the United Kingdom (García, Raza, Parrett, & Wright, 2013). A review of the Australian commercial infant food market found lower levels of added sugars compared with the United Kingdom and Canadian studies, but comparably high total sugars values, noting the presence of fruit as sweeteners in Australian products (Dunford, Louie, Byrne, Walker, & Flood, 2015).

The Australia New Zealand Food Standards code provides limited regulation regarding the composition and labelling of commercial infant foods in Standard 2.9.2 (Australian Government, 2017). Free sugars are not mentioned at all, and the code defines an infant as a person under 12 months of age, so toddler products labelled for 12 months and older are not covered by this standard. Infant foods containing more than 4 g/100 g added sugars must be labelled as “sweetened”; however use of the term “added sugars” allows manufacturers to subvert the standard by sweetening products with fruit juice or fruit juice concentrate. These products are often labelled “no added sugar” to appeal to parents' desire to provide healthy foods. Given the proliferation in the number and sales of toddler foods in recent years, the current standard should be extended, or a new standard developed, to cover toddler foods and incorporate the definition of free sugars.

These findings support the existing evidence that dietary behaviours are influenced by social determinants, with poorer adherence to dietary recommendations reported among those with greater socio‐economic disadvantage (Cameron et al., 2012; Darmon & Drewnowski, 2008). This disadvantage is also seen in access to dental services, with more people in lower household income groups avoiding or delaying a visit to the dentist due to cost compared with those in higher income groups (Australian institue of Health and Welfare, 2015). The possible compounding effect of higher free sugars intakes and lower access to dental services may explain some of the disparity in caries experience, and the interaction pathways between socio‐economic disadvantage and oral health in the SMILE cohort are an ongoing area of investigation (Do et al., 2014).

Determination of free sugars from Australian dietary data has only recently been possible, as AUSNUT 2011–13 is the first national nutrient database to include values for added and free sugars. There are no analytical measures for determining free sugars separately to total sugars, as the definition relates to food source rather than any molecular difference between the types of sugars. A modified 10‐step methodology is used to determine the contribution of free sugars (Louie et al., 2015). The earliest methodology steps are associated with greatest confidence and move from objective to subjective at step 7. For AUSNUT 2011–13, only Steps 1–4 were necessary as there is a corresponding recipe database to accompany AUSNUT that was used to handle all foods that contained a combination of free and intrinsic sugars (Food Standards Australia and New Zealand, 2018). Within our study, where an infant food product was added to the database, we used up to the full 10‐step methodology to determine free sugars as product recipes were not available. Free sugars values for these additional foods may therefore be less accurate than those provided in the AUSNUT 2011–13 database.

The AUSNUT 2011–13 database is limited by the food products available during the update period. The Australian food supply is ever‐changing and nutrient database updates are infrequent. In particular, the infant formula and foods category contains only a few crude groups that do not reflect the current infant and toddler market. In addition, AUSNUT 2011–13 classifies all infant and toddler formula as lacking free sugars despite the presence of sucrose in the ingredients list of soy‐based formula.

In our study, the WHO conditional recommendation of less than 5% energy from free sugars was used as the outcome variable due to relatively low numbers of participants exceeding the 10% recommendation. This conservative recommendation may be the more appropriate target for 12–14 month olds, given the early establishment of taste preferences, the increasing intakes of energy dense, nutrient poor foods through childhood, and the cumulative effect of dental caries, which tracks from childhood to adulthood (WHO, 2015). This also reflects the recommendations of the Australian Infant Feeding Guidelines (NHMRC, 2012) to limit or avoid foods with added sugars in the first 12 months of life, and that no sugar should be added to foods prepared for infants. The current evidence for the 5% recommendation was rated as very low quality, based on older children only, and is an area of ongoing investigation at all ages (Moynihan & Kelly, 2014).

A strength of this study is our dietary data collection methods. The use of a 24‐hr recall and 2‐day food record resulted in 3 nonconsecutive days of data, which included two weekdays and one weekend day. The phone interview 24‐hr recall allowed researchers to issue instructions for the food record and provide basic training to participants in the use of the reference images and level of detail required. Participants could ask questions to ensure they understood what was required in the food record. The effectiveness of this method is demonstrated by the fact that out of the 847 food records that were returned, only 12 were excluded due to poor quality. We coupled this with standardisation protocols at data entry, and every participant's data were double‐entered to reduce error. In addition, the Multiple Source Method (Harttig et al., 2011) allowed us to adjust for day‐to‐day variability in order to determine usual intake of free sugars, rather than averaging the 3 days. Nevertheless, dietary assessment is by nature limited in terms of precision and accuracy. Due to the age of participants, proxy reporting by the mothers was necessary, which may not describe the complete food landscape experienced by the child. Social desirability bias may have influenced reports given by mothers, and other forms of misreporting are likely to have been present (National Institutes of Health & National Cancer Institute, 2018; Thompson & Subar, 2013; Willett, 2013). Study nonresponders were not significantly different from responders across most sociodemographic variables, with the age of the mothers and some countries of birth the only notable differences. As such, the findings here are considered generalizable to the sampling population.

5. CONCLUSION

Our analysis found that the diets of more than four out of every five children aged 12 to 14 months comply with the WHO 5% recommendation, and almost all comply with the 10% recommendation. Children who experienced greater sociodemographic disadvantage were more likely to exceed the recommendations, indicating that efforts to increase compliance should be targeted towards these groups. The major food sources of free sugars at this age were similar to the major sources later in life, suggesting that the establishment of dietary patterns commences from the beginning of solid food introduction. An additional finding was the substantial contribution to free sugars by products promoted as infant and toddler foods, many of which contain fruit juice to subvert food labelling requirements for added sugars. These findings emphasise the importance of dietary intervention and education from the commencement of weaning, including specific advice around selecting appropriate foods within the infant and toddler food market. Future policy decisions around food labelling should consider the different definitions of added and free sugars in providing information that best serves the consumer. In the absence of action by food manufacturers, policymakers should consider regulation to limit the amount of free sugars in commercial infant and toddler foods. At the very least, food standards related to added sugars in infant foods should be extended to include toddler foods.

CONFLICTS OF INTEREST

The authors declare that they have no conflicts of interest.

CONTRIBUTIONS

LD and JS conceived and designed the study, secured funding, and coordinated the research team; GD helped design and administer the dietary assessment components, assisted with statistical analysis, and prepared the manuscript in consultation with all authors; AB assisted with study design; EY performed the statistical analysis. All authors were involved in interpreting the results and preparation of the manuscript.

Supporting information

Table S1: Participant characteristics

Table S2: Free sugars intakes at 12 Months – Participants with plausible intakes (n = 699)

Table S3: Factors associated with Free Sugars intakes ≥5% of Energy – Participants with plausible intakes (n = 699)

APPENDIX A.

Figure A1.

Participant flow

Devenish G, Ytterstad E, Begley A, Do L, Scott J. Intake, sources, and determinants of free sugars intake in Australian children aged 12–14 months. Matern Child Nutr. 2019;15:e12692 10.1111/mcn.12692