The Independent and Cumulative Effect of Early Life Risk Factors on Child Growth: A Preliminary Report

Abstract

Background:

Early life risk factors may promote faster infant growth leading to childhood obesity. We examined growth patterns (birth to 12 months) and their association with early life risk factors.

Methods:

Participants were drawn from an ongoing birth cohort of 351 mothers. Child weight and length were obtained at birth, 6 weeks, and 3, 6, 9, and 12 months. Independent variables and demographics were tested as risk factors for inclusion in the cumulative risk score if they were significantly associated with change in weight-for-length z-scores (WFLZ; month 12 minus birth) or based on established evidence for an association with child growth. Multiple regression was used to determine the association of change in WFLZ with low maternal education, low month 3 postpartum maternal weight loss, and nonexclusive breastfeeding by month 3 (use of formula) or their cumulative risk. Trajectory groups were identified using semiparametric mixture models and their association with the risk factors and cumulative risk score was tested using logistic regression.

Results:

Nonexclusive breastfeeding by month 3 was associated with greater WFLZ increase. We identified three trajectory groups: low-rising (14.5%), mid-stable (59.4%), and high-rising (26.1%). Low-rising versus mid-stable group membership doubled with nonexclusive breastfeeding by month 3 (odds ratio [OR] = 2.24; 95% confidence interval [CI] = 1.05–4.80), but increased the most among children who had three risk factors compared to none (OR = 4.82; 95% CI = 1.14–20.32). No relationships were observed with high-rising growth.

Conclusions:

Early life risk factors influence the development of growth trajectories during the first year in a cumulative fashion.

Introduction

Growth patterns in the first year of life, particularly rapid growth, have been associated with increased risk for obesity into childhood and adolescence.^1,2 Factors associated with rapid growth in early childhood include maternal education,³ maternal weight during the child's first year of life,⁴ and breastfeeding duration.^5,6 Typically, these risk factors are considered as individual determinants of weight gain. An alternative approach is to consider the cumulative effects of individual factors on child growth patterns in the first year of life. Understanding the combined effect of early risk factors on child growth patterns is paramount to developing strategies for improving child health later in life.⁷

Cumulative risk approaches have been applied to the study of childhood obesity in several studies.^8–11 Overall, results from these studies suggest that early life experiences, including sociodemographic risk^10,11 (e.g., income), family relationship factors,^8,9 and maternal health (including obesity)^2,4,11 accumulate over time to predict child and adolescent obesity risk. The vast majority of these studies focus on cumulative risk during early childhood and trajectories to adolescence. To date, little attention has been paid to the cumulative effects of known risk factors on growth patterns during the first year of life.

One way to assess growth patterns is to measure rate of growth, which may lead to early detection of growth problems that precede obesity.^12,13 Another approach, group-based trajectory modeling,¹⁴ identifies clusters of children who follow a similar progression of weight. This approach is suitable, because fluctuations in early child growth, including rapid growth, may be markers of risk for obesity in childhood and adulthood.¹ Previous research has identified different growth trajectories that may be associated with obesity risk. Carling and colleagues⁶ identified four distinct growth trajectories according to rate of increase in weight-for-length (WFL) by 24 months (high-rising, low-rising, mid-stable, and low-stable). In comparing the rising and stable growth groups, there were significant differences for maternal BMI and education level. Infants who evidenced the rising growth pattern had mothers who were more overweight or obese and of lower educational level.⁶ The Carling study examined how breastfeeding duration modified the association of obesity risk factors with a rising weight trajectory. Our study focused on weight gain by 12 months, and breastfeeding duration was included in a cumulative risk score.

The purpose of this report is to examine, in an ongoing study, the cumulative effects of maternal education level, early maternal postpartum weight loss, and breastfeeding duration in predicting rate of growth during the first year of life. Further, we propose to compare different growth trajectories (low-rising, mid-stable, and high-rising) in relation to cumulative risk categories (low to high). We hypothesized that infants who experience high levels of cumulative risk during the first year of life will evidence greater change in growth and rising growth trajectories.

Methods

Participants

Mothers and infants were recruited as part of an ongoing cohort, STRONG Kids2: A Cells-to-Society Approach to Nutrition in Early Childhood, which will consist of 440 families. Women are recruited in their third trimester of pregnancy from healthcare facilities (e.g., obstetrics and gynecology) and birthing classes in East-Central Illinois. Exclusion criteria include premature birth (<37 weeks), birth conditions precluding normal feeding (e.g., phenylketonuria), and low birthweight (<2.5 kg). Parents complete surveys online or on paper at 1 and 6 weeks and 3 and 12 months. Written informed consent was obtained at study entry. This research was approved by the institutional review board at University of Illinois at Urbana–Champaign.

Demographics

Mother's age, education (1 = less than college graduate, 0 = college graduate/post graduate), monthly household income (0 = $3000 and under, 1 = $3001-$5000, 2 = ≥$5001), marital status (0 = not single, 1 = single), employment status (1 = unemployed/other, 0 = employed), participation in assistance programs (0 = none, 1 = yes), child race (0 = Hispanic/Latino, 1 = non-Hispanic/Latino White, 2 = non-Hispanic/Latino other) and gender.

Child Anthropometric Variables

Child birth weight (kg) and length (cm) were obtained by parental report at 1 week. Values at 6 weeks and 3 and 12 months were measured at the child's home by trained research assistants. Length was obtained by measuring the distance from head to foot while the child lay flat on a scale (model 728; seca GmbH & Co. KG, Hamburg, Germany). Weight was measured using a digital scale (model 349KLX; HealthOmeter, McCook, IL). Parents were asked to provide, from the well-child visit records, the weight and length when the child was 6 and 9 months. Age- and sex-specific WFL z-scores (WFLZ) at birth, week 6, and months 3, 6, 9, and 12 were calculated using the 2006 World Health Organization growth charts.^15,16 Rate of growth (change) over the first year was calculated as WFLZ at month 12 minus WFLZ at birth (month 12 minus birth). Higher values indicate increased weight gain.

Independent Variables

Mother's weight and height at prepregnancy and study entry was collected by parental report. Values at 6 weeks and 3 and 12 months were objectively measured at the child's home. Maternal BMI (kg/m²) was converted into a binary variable of nonoverweight (0) consisting of underweight (BMI <18.5) and normal weight (18.5 ≤ BMI <25)¹⁷, and overweight (1) consisting of overweight (25 ≤ BMI <30) and obese (BMI ≥30).¹⁷ Combining the underweight with normal weight and the overweight with obese ensured sufficient analytical sample size within groups and did not impact study findings related to maternal weight. Early maternal weight loss since pregnancy (EWLP) was used to measure postpartum weight change and calculated as month 3 postpartum weight minus third trimester weight (study entry) and dichotomized according to the 75th percentile (−3.9 kg). Compared with 75% of the sample, maternal weight change <−3.9 kg indicated that mothers lost more weight at month 3 since the third trimester and classified as high (0); values ≥−3.9 kg indicated that mothers lost less weight, remained stable, or gained weight (low [1]). Other variables include delivery method (1 = vaginal, 2 = C-section) and exclusive breastfeeding (0 = yes, 1 = no). Exclusive breastfeeding refers to children who were not fed any formula.

Statistical Analysis

All analyses were conducted using the Statistical Analysis Software (version 9.3; SAS Institute Inc., Cary, NC). The association of each independent and demographic variable with change in WFLZ was tested using the t-test, analysis of variance (ANOVA), or correlation analyses. Change in WFLZ was similar across children who were exclusively breastfed at week 6 (0.8 ± 1.6) or not (1.6 ± 1.2) and month 3 (0.8 ± 1.6) or not (1.5 ± 1.3). Thus, a binary variable was created by combining both time points (exclusive breastfeeding by month 3). Risk factors were selected for inclusion in the cumulative risk score if they were significantly associated with change in WFLZ (p < 0.05; maternal education, exclusive breastfeeding by month 3) or based on established evidence for an association with child growth (EWLP^4,18). Selected risk factors were summed to calculate the cumulative risk score. Low risk is indicated as 0 (mother with high EWLP, college graduate/postgraduate education, and exclusively breastfed by month 3). Higher scores indicate greater risk (≥1 risk factor of low EWLP, less than college degree, and nonexclusive breastfeeding by month 3).

Mean change in WFLZ across categories of the cumulative risk score was tested using ANOVA with Tukey's multiple comparisons test. Separate linear regression models tested the association of change in WFLZ with the selected risk factors and cumulative risk score after controlling for binary birth weight (1 = high birth weight [≥3.9 kg], 0 = not [<3.9 kg]),^19,20 mother's age at study entry, child sex, and race. The condition index and variance proportion were used to detect multicollinearity.²¹ None of the condition indices for binary birth weight or risk factors indicated multicollinearity²¹ in the adjusted risk factors only and cumulative risk score only models. In both models, none of the independent variables contributed strongly to the variance (proportion >0.5) of two or more other independent variables.²¹ Hence, multicollinearity did not pose a problem in either model.

Trajectory Group Modeling

A normal trajectory model was fit to the repeated WFLZ over age using semiparametric mixture models.^5,10 The trajectory group probabilities were adjusted for mother's age at study entry, child sex, and race.

The final trajectory model was selected after considering several factors. The optimal order of polynomials was determined based on significance of the beta coefficients (β).The Bayesian Information Criterion²³ was used to assess model fit; lower values indicated improved model fit. Group membership probability was examined to ensure that it was ≥0.70 for each group.³ Finally, we preferred to have ≥10% of the sample within each group. Significance of a trajectory's β was determined using Wald's tests. Trajectories were considered to be distinct (not parallel with minimal overlap) because linear and second-order terms of the polynomials were unequal.¹⁴

Once the trajectory groups were identified, the proportion of risk factors across the groups was examined using chi-square tests. Pair-wise comparisons of trajectory group membership were conducted using logistic regression. The models were adjusted for binary birth weight. All p values are two-tailed.

Results

General characteristics of the study sample are presented in Table 1. The mothers were 29.9 ± 4.9 years of age at study entry. Just under half were overweight at prepregnancy and the majority (72%) delivered vaginally. Over half were employed (66%) and completed college or a postgraduate degree (67%). Over half of the children were female; the majority were non-Hispanic/Latino White (75%).

Table 1.

Characteristics of Study Sample (n = 351)^a

Characteristic	No. (%) or mean ± SD
Time point
Week 1	351 (100.0)
Week 6	340 (96.9)
Month 3	328 (93.5)
Month 6	93 (26.5)
Month 9	93 (26.5)
Month 12	223 (63.5)
Mother's age at study entry (years)	29.9 ± 4.9
Mother's BMI
Prepregnancy
Nonoverweight	160 (45.6)
Overweight	175 (49.9)
Unknown/missing	16 (4.6)
Study entry
Nonoverweight	58 (16.5)
Overweight	279 (79.5)
Unknown/missing	14 (3.9)
Week 6
Nonoverweight	115 (32.8)
Overweight	235 (66.9)
Unknown/missing	1 (0.3)
Month 3
Nonoverweight	120 (34.2)
Overweight	204 (58.1)
Unknown/missing	27 (7.7)
Month 12
Nonoverweight	92 (26.2)
Overweight	117 (33.3)
Unknown/missing	142 (40.5)
Early maternal weight loss since pregnancy
High	233 (66.4)
Low	78 (22.2)
Unknown/missing	40 (11.4)
Delivery method
Vaginal	252 (71.8)
C-section	94 (26.8)
Unknown/missing	5 (1.4)
Mother's education
Less than college graduate	101 (28.8)
College graduate/postgraduate degree	234 (66.7)
Unknown/missing	16 (4.6)
Monthly household income
$3000 and under	106 (30.2)
$3001–$5000	86 (24.5)
≥$5001	105 (29.9)
Unknown/missing	54 (15.4)
Marital status
Not single	292 (83.2)
Single	41 (11.7)
Unknown/missing	18 (5.1)
Employment status
Unemployed/other	100 (28.5)
Employed	232 (66.1)
Unknown/missing	19 (5.4)
Healthcare coverage	327 (93.2)
Participation in WIC	78 (22.2)
Participation in childcare assistance program	6 (1.7)
Receives welfare/cash assistance	5 (1.4)
Child sex
Female	182 (51.9)
Race/ethnicity
Hispanic/Latino	13 (3.7)
Non-Hispanic/Latino White	263 (74.9)
Non-Hispanic/Latino other	33 (9.4)
Unknown/missing	42 (11.9)
Birth weight
Not high (<3.9 kg)	311 (88.6)
High (≥3.9 kg)	37 (10.5)
Unknown/missing	3 (0.9)
Exclusive breastfeeding
Week 1	249 (70.9)
Week 6	225 (64.1)
Month 3	210 (59.8)
Month 12	93 (26.5)
Child WFLZ
Birth	−0.5 ± 1.3
Week 6	−0.1 ± 1.1
Month 3	−0. 2 ± 1.2
Month 6	0.1 1.1
Month 9	0.4 ± 1.2
Month 12	0.6 ± 0.9
Change in child WFLZ (month 12 minus birth) (n = 206)	1.1 ± 1.5

Data that are unknown or missing were not provided by the mother. Data collected at months 6, 9, and 12 are less than that at earlier time points because the study is still ongoing and recruitment is not completed. The only data collected at months 6 and 9 are parental report of child height and weight measurements and were added after the study was initiated. Maternal education, household income, marital status, employment status, healthcare coverage, participation in WIC, or other assistance programs were collected at week 6.

Mother's BMI was classified as nonoverweight (underweight [BMI <18.5] and normal weight [18.5 ≤ BMI <25]) and overweight (overweight [25 ≤ BMI <30] and obese [BMI ≥30]). Early maternal weight loss since pregnancy was used as a measure of postdelivery maternal weight change. It was calculated as month 3 postpartum weight minus third trimester weight (study entry) and dichotomized according to the 75th percentile as high, indicating losing more than 3.9 kg, and low, indicating weight loss less than 3.9 kg, no weight loss, or gaining weight at month 3 postpartum. Exclusive breastfeeding was determined based on mothers who reported feeding their child breast milk in response to the question “Did you feed your baby formula, breast milk, or both in the past 7 days?”

Percentages may not add up to 100 because of rounding.

WIC, Special Supplemental Nutrition Program for Women, Infants, and Children; WFLZ, child weight-for-length z-score; SD, standard deviation.

Mean change in WFLZ was lowest among children with a cumulative risk score of 0 (0.7 ± 1.6) compared to 1 (1.2 ± 1.6), 2 (1.3 ± 1.1), and 3 (1.9 ± 1.5; p = 0.014; Table 2).

Table 2.

Mean Change in WFLZ (Month 12 Minus Birth) Across Cumulative Risk Score Categories

Cumulative risk score	No. of risk factors (category)	No. (%) of observations within each score category (n = 344)	No. (%) with available data on change in WFLZ (month 12 minus birth) (n = 206)	Change in WFLZ (month 12 minus birth) Mean ± SD
0	0 (low)	134 (38.9)	87 (42.2)	0.7 ± 1.6
1	1 (low-medium)	130 (37.8)	72 (34.9)	1.2 ± 1.6
2	2 (medium-high)	63 (18.3)	36 (17.5)	1.3 ± 1.1
3	3 (high)	17 (4.9)	11 (5.3)	1.9 ± 1.5^*

The cumulative risk score was calculated by summing the number of risk factors (mother's education, exclusive breastfeeding by month 3, and early maternal weight loss since pregnancy [EWLP]) that were classified as high risk. Low cumulative risk is indicated as 0 (mother with high EWLP, college graduate/postgraduate education, and exclusively breastfed by month 3). Higher scores (1–3) indicate greater risk (i.e., addition of one or more of the following risk factors: low EWLP, less than college degree, and nonexclusive breastfeeding by month 3). The cumulative risk score could not be calculated for seven participants because of missing data.

p < 0.05 when compared to cumulative risk score of 0 using Tukey's Studentized range test. p = 0.014 using analysis of variance.

WFLZ, child weight-for-length z-score; SD, standard deviation.

Regression findings for the association of change in WFLZ as the outcome with selected risk factors and cumulative risk score are summarized in Table 3. In the risk factors only model, nonexclusive breastfeeding by month 3, lesser education, and low EWLP were associated with greater change in WFLZ. Standardized βs are 0.21 (p < 0.001), 0.13 (p > 0.05), and 0.003 (p > 0.05), respectively, interpreted as magnitude of change in standard deviation (SD) units in the predicted outcome. High birthweight was associated with less change in WFLZ (−0.16; p < 0.05). In the cumulative risk score only model, compared to zero risk factors, having three risk factors yielded the greatest increase in the predicted outcome (0.21; p < 0.01) than one or two risk factors (0.17 and 0.16, respectively; p < 0.05).

Table 3.

Adjusted Multiple Regression of Selected Risk Factors and Cumulative Risk Score with the Outcome of Change in WFLZ (Month 12 Minus Birth)

	Risk factors only Change in WFLZ (month 12 minus birth) Standardized β (95% CI)	Cumulative risk score only Change in WFLZ (month 12 minus birth) Standardized β (95% CI)
Cumulative risk score	NA
1 vs. 0		0.17 (0.02–0.33)^*
2 vs. 0		0.16 (0.005–0.32)^*
3 vs. 0		0.21 (0.06–0.34)^**
EWLP (low vs. high)	0.003 (−0.14–0.15)	NA
Education (less than college graduate vs. college graduate/postgraduate degree)	0.13 (−0.02–0.29)	NA
Exclusive breastfeeding by month 3 (no vs. yes)	0.21 (0.06–0.38)^**	NA
Birth weight (high vs. not)	−0.16 (−0.29 to −0.02)^*	−0.19 (−0.30 to −0.05)^**
Mother's age at study entry	−0.004 (−0.17–0.16)	−0.07 (−0.22–0.08)
Child sex (male vs. female)	0.15 (0.002–0.29)^*	0.12 (−0.02–0.26)
Child race
Non-Hispanic/Latino other vs. Non-Hispanic/Latino White	−0.07 (−0.24–0.08)	−0.08 (−0.22–0.07)
Hispanic/Latino vs. non-Hispanic/Latino White	0.008 (−0.15–0.17)	0.06 (−0.09–0.21)

Exclusive breastfeeding by month 3 was determined based on mothers who reported feeding their child breast milk in response to the question “Did you feed your baby formula, breast milk, or both in the past 7 days?” at week 6 and month 3. Both models were adjusted for birth weight (0 = not high, 1 = high), mother's age at study entry, child sex, and race.

The cumulative risk score was calculated by summing scores for EWLP (0 = high, 1 = low), education (1 = less than college graduate, 0 = college graduate/postgraduate degree), and exclusive breastfeeding by month 3 (0 = yes, 1 = no). Risk factors only model: F_{(8, 163)} = 3.16, p = 0.002; adjusted R² = 0.092. Cumulative risk score only model: F_{(8, 179)} = 3.36, p = 0.001; adjusted R² = 0.092.

p < 0.05; ^**p < 0.01; ^{^***}p < 0.001.

WFLZ, child weight-for-length z-score; β, beta coefficient; CI, confidence interval of the standardized β; EWLP, early maternal weight loss since pregnancy; NA, not applicable.

Three trajectory groups were identified. Figure 1 plots the trajectory groups' point estimates and 95% confidence intervals (CIs). Group 1 consists of children who were born with the least WFLZ (−1.58; 95% CI = −2.08 to −1.07), which gradually increased over time. The proportion of children in this low-rising group was 14.5%. Group 2 comprised 59.4% of the children, born with an intermediate WFLZ (–0.21; 95% CI = −0.58–0.16) that remained constant over time and were characterized as mid-stable. Group 3 consisted of 26.1% of the children, born with the highest WFLZ (−0.08; −0.39–0.24) that sharply increased, peaking around 9 months. This group was characterized as high-rising. None of the variables used to adjust group membership were significant.

Figure 1.

Trajectory group point estimates and 95% confidence intervals obtained by fitting weight-for-length z-scores (WFLZ) with child age using semiparametric mixture models. WFLZ from birth, week 6, and months 3, 6, 9, and 12 were used as outcomes for developing the trajectories. Group 1 = low-rising (n = 44); group 2 = mid-stable (n = 180); group 3 = high-rising (n = 79). The trajectory group probabilities were allowed to vary (adjusted) as a function of mother's age at study entry, child sex, and race. None of the variables used to adjust group membership were significant.

The trajectory group probabilities, which determine probability of belonging to a specific group, were optimal (≥0.70; Table 4). Table 5 provides the profile of the risk factors within trajectory groups. The low-rising group had the highest proportion of mothers who exhibited low EWLP (39.0%; nonsignificant) and who practiced nonexclusive breastfeeding by month 3 (55.8%; p < 0.05).

Table 4.

Trajectory Group Probabilities

		Mean probability
Trajectory group	No. (%) of observations within each trajectory group (n = 303)^a	Group 1 Low-rising	Group 2 Mid-stable	Group 3 High-rising
1	44 (14.5)	0.78	0.21	0.01
2	180 (59.4)	0.09	0.76	0.15
3	79 (26.1)	0.01	0.23	0.76

The mean probability for each trajectory group is bolded.

Of 351 observations, 303 were used in the trajectory model.

Table 5.

Profile of Selected Risk Factors within Trajectory Groups

	Trajectory group
	Group 1 Low-rising (n = 44)	Group 2 Mid-stable (n = 180)	Group 3 High-rising (n = 79)
EWLP (low) (%)	39.0	25.6	18.7
Education (less than college graduate) (%)	23.8	26.8	29.1
Exclusive breastfeeding by month 3 (no) (%)^*	55.8	33.3	37.9

EWLP was used as a measure of postdelivery maternal weight change. It was calculated as month 3 postpartum weight minus third trimester weight (study entry) and dichotomized according to the 75th percentile as high, indicating losing more than 3.9 kg, and low, indicating weight loss less than 3.9 kg, no weight loss, or gaining weight at month 3 postpartum. Exclusive breastfeeding by month 3 was determined based on mothers who reported feeding their child breast milk in response to the question “Did you feed your baby formula, breast milk, or both in the past 7 days?” at week 6 and month 3. Percentages were calculated within each trajectory group.

p < 0.05

EWLP, early maternal weight loss since pregnancy.

Table 6 shows the logistic regression findings predicting the association of the selected risk factors (risk factors only model) and cumulative risk score (cumulative risk score only model) with inclusion in a trajectory group. Membership in the low-rising versus mid-stable group doubled among children who were nonexclusively breastfed by month 3 (odds ratio [OR] = 2.24; 95% CI = 1.05–4.80). A smaller positive effect was observed for low versus high EWLP, but was not significant. Compared to low risk, a score of 1 or 2 tended to increase membership as evidenced by positive, but nonsignificant, ORs (2.23; 95% CI = 0.98–5.08 and 1.79; 95% CI = 0.68–4.75). Presence of three risk factors significantly increased membership in the low-rising versus mid-stable group by 4.8 times (4.82; 1.14–20.32). A cumulative risk score of 3 versus 0 tended to increase membership in the low-rising versus high-rising group (2.26; 95% CI = 0.47–10.76) and decrease likelihood of the mid-stable versus high-rising group (0.58; 95% CI = 0.15–2.21); these trends were not significant.

Table 6.

Logistic Regression of the Association of Risk Factors and Cumulative Risk Score with Pairs of Trajectory Groups

	Low-rising vs. mid-stable OR (95% CI)	Low-rising vs. high-rising OR (95% CI)	Mid-stable vs. high-rising OR (95% CI)
Risk factors only
EWLP (low vs. high)	1.89 (0.87–4.13)	2.36 (0.95–5.89)	1.49 (0.75–2.98)
Education (less than college graduate vs. college graduate/postgraduate degree)	0.67 (0.28–1.61)	0.72 (0.28–1.85)	1.02 (0.54–1.92)
Exclusive breastfeeding by month 3 (no vs. yes)	2.24 (1.05–4.80)^*	1.35 (0.58–3.14)	0.76 (0.42–1.39)
Birth weight (high vs. not)	0.22 (0.03–1.69)	0.24 (0.03–2.05)	0.88 (0.37–2.08)
Cumulative risk score only
1 vs. 0	2.23 (0.98–5.08)	1.38 (0.56–3.41)	0.69 (0.39–1.27)
2 vs. 0	1.79 (0.68–4.75)	2.03 (0.64–6.43)	1.36 (0.59–3.09)
3 vs. 0	4.82 (1.14–20.32)^*	2.26 (0.47–10.76)	0.58 (0.15–2.21)
Birth weight (high vs. not)	0.16 (0.02–1.26)	0.19 (0.02–1.65)	0.95 (0.43–2.13)

EWLP was used as a measure of postpartum maternal weight change. It was calculated as month 3 postpartum weight minus third trimester weight (study entry) and dichotomized according to the 75th percentile as high, indicating losing more than 3.9 kg, and low, indicating weight loss less than 3.9 kg, no weight loss, or gaining weight at month 3 postpartum. Exclusive breastfeeding by month 3 was determined based on mothers who reported feeding their child breast milk in response to the question “Did you feed your baby formula, breast milk, or both in the past 7 days?” at week 6 and month 3.

Membership in 3 trajectory groups (low-rising [n = 44], mid-stable [n = 180], and high-rising [n = 79] was determined using semiparametric mixture models and adjusted for mother's age at study entry, child sex, and race. All logistic regression models were adjusted for birth weight. The cumulative risk score was calculated by summing scores EWLP (0 = high, 1 = low), education (1 = less than college graduate, 0 = college graduate/postgraduate degree), and exclusive breastfeeding by month 3 (0 = yes, 1 = no).

p < 0.05.

EWLP, early maternal weight loss since pregnancy; OR, odds ratio; CI, confidence interval.

Discussion

Using an ongoing longitudinal birth cohort, we investigated the association of a cumulative risk score and its components with pattern of child growth in the first year of life. The pattern of child growth was measured using two methods: change in WFLZ (month 12 minus birth) and trajectory groups. Among the three risk factors (low EWLP, lesser education, and nonexclusive breastfeeding by month 3), nonexclusive breastfeeding was associated with the greatest increase in WFLZ. This finding is consistent with other reports.²⁴ We also observed that children with one or more risk factors evidenced greater WFLZ increases than low-risk children. High cumulative risk and nonexclusive breastfeeding yielded a similar increase in WFLZ, suggesting that breastfeeding status is a main driver of WFLZ change. Although the effect of the cumulative risk factors on growth, including factors identified in this study, was previously observed in preschool and school-age children,^25,26 it is less commonly studied in the first year of life. Using an alternative cutoff for rapid weight gain since birth, we observed that 19% of the children had >0.67 SD increase in WFLZs by 12 months, compared to 14% by 4–7 months.²⁷

Three distinct growth trajectories were identified. Their mean predicted WFLZ estimates mostly lay within normal growth cutoffs (between 2 and −2).¹⁶ This is expected given that the cohort is relatively healthy. However, the trajectories may signal susceptibility for excess weight gain later in life. Although the majority of the children were mid-stable growers (59%), two other groups emerged at the extremes of the WFLZ distribution—low-rising (15%) and high-rising (26%). Low-rising children were smallest at birth (although full term and of average weight) with gradually accelerating growth. High-rising children were born with the largest WFLZ that steeply increased beyond the other groups for most of the first year. Similar to our study, Carling and colleagues found that children who were breastfed for shorter durations were more likely to belong to a rising than stable weight gain trajectory.⁶ Giles and colleagues identified four trajectory patterns during the first 3.5 years of life²⁸; the low group in their study showed similar growth patterns as the low-rising group in this cohort. Their high or accelerating groups, comparable to the high-rising group in this cohort, were associated with weight gain at 9 years.²⁸ In another cohort of 5- to 15-year-old girls, four growth trajectories were identified; the trajectory with the highest weight gain had the highest metabolic risk.²⁶ Future follow-up of our cohort will determine whether similar trajectories are associated with unhealthy weigh gain.

Consistent with previous reports focusing on older children,²⁹ we found that growth trajectories are multidetermined. Even focusing on a limited number of risk factors, we find that with increasing exposure of risk, infants are more likely to evidence faster growth. Mothers of children in the low-rising group tended to lose the least amount of weight at month 3 postpartum and practiced more formula feeding than exclusive breastfeeding by month 3. Likelihood of inclusion in a low-rising versus mid-stable trajectory more than doubled among children who were nonexclusively breastfed by month 3. Thus, exclusive intake of human milk by month 3 inhibits faster growth and promotes relatively consistent growth during the first year of life. Notably, the proportion of breastfed infants was highest among mid-stable compared to low-rising and high-rising groups at week 6 (74%, 52%, and 66%, respectively) and month 3 (69%, 51%, and 66%). Moreover, compared to low-risk children, children with all three risk factors were almost 5 times more likely to be classified in the low-rising than mid-stable group. This effect was double that of nonexclusive breastfeeding by month 3, suggesting a cumulative effect. Future research is warranted to consider additional risk factors, such as genetic contributions, biological variations, and more complete assessments of poverty and economic hardship.

A limitation of this study is that fewer children were sampled at 12 months than earlier time points. This is because the cohort is ongoing. Despite this fact, study findings were suggestive of a cumulative effect of maternal risk factors during the first year of life. Findings will need to be confirmed upon completion of study recruitment. Values that were missing for the outcome or independent variables were excluded from the analyses, which may have impacted findings. The children were predominantly non-Hispanic/Latino White and the majority was more educated, thus the cohort is not representative of the general population. However, around 30% had an annual household income of $36,000 or less, which is less than the US median in 2014 ($53,657).³⁰ Child growth assessments were obtained at 3-month intervals, causing paucity of data for the intervening time periods. Although EWLP was used to assess maternal postpartum weight change, it potentially presents a mixture of prepregnancy weight, gestational weight gain, and postpartum weight retention. Finally, unmeasured risk factors (e.g., smoking during pregnancy, age at introduction of solids, and paternal BMI) could have influenced the growth trajectories.

Conclusions

Nonexclusive breastfeeding by month 3 was related to the largest increase in WFLZ during the first year of life. We observed that the combined effect of three risk factors (low maternal weight loss at 3 months postpartum, lesser maternal education, and nonexclusive breastfeeding by month 3) was greater than nonexclusive breastfeeding by month 3 alone on inclusion in a low-rising versus mid-stable growth trajectory. These findings suggest that early life risk factors act in concert to influence growth trajectories over the course of the first year.

Footnotes

Acknowledgments

The authors thank the participating families as well as the following funding sources: the Dairy Research Institute to Barbara Fiese and Sharon Donovan (Principal Investigators; PIs); USDA (Hatch 793-328) to Barbara Fiese (PI); the Gerber Foundation to Sharon Donovan (PI); and the Christopher Family Foundation Food and Family Program.

Author Disclosure Statement

No competing financial interests exist.

References

Baird

, Fisher

, Lucas

, et al. Being big or growing fast: Systematic review of size and growth in infancy and later obesity. BMJ, 2005; 331:929.

Ziyab

, Karmaus

, Kurukulaaratchy

, et al. Developmental trajectories of body mass index from infancy to 18 years of age: Prenatal determinants and health consequences. J Epidemiol Community Health, 2014; 68:934–941.

Balistreri

, Van Hook

. Trajectories of overweight among US school children: A focus on social and economic characteristics. Matern Child Health J, 2011; 15:610–619.

Robinson

, Cohen

, Rehkopf

, et al. Pregnancy and post-delivery maternal weight changes and overweight in preschool children. Prev Med, 2014; 60:77–82.

Oddy

, Mori

, Huang

, et al. Early infant feeding and adiposity risk: From infancy to adulthood. Ann Nutr Metab, 2014; 64:262–270.

Carling

, Demment

, Kjolhede

, et al. Breastfeeding duration and weight gain trajectory in infancy. Pediatrics, 2015; 135:111–119.

Macaulay

, Donovan

, Leask

, et al. The importance of early life in childhood obesity and related diseases: A report from the 2014 Gravida Strategic Summit. J Dev Orig Health Dis, 2014; 5:398–407.

Jones

, Fiese

. Parent routines, child routines, and family demographics associated with obesity in parents and preschool-aged children. Front Psychol, 2014; 5:374.

Lumeng

, Wendorf

, Pesch

, et al. Overweight adolescents and life events in childhood. Pediatrics, 2013; 132:e1506–e1512.

10.

Margerison-Zilko

, Cubbin

. Dynamic poverty experiences and development of overweight in a prospective cohort of US children aged 4–14 years. Obesity (Silver Spring), 2013; 21:1438–1445.

11.

Wells

, Evans

, Beavis

, et al. Early childhood poverty, cumulative risk exposure, and body mass index trajectories through young adulthood. Am J Public Health, 2010; 100:2507–2512.

12.

Tanner

. The assessment of growth and development in children. Arch Dis Child, 1952; 27:10–33.

13.

Van Den Berg

, Van Eijsden

, Galindo-Garre

, et al. Low maternal education is associated with increased growth velocity in the first year of life and in early childhood: the ABCD study. Eur J Pediatr, 2013; 172:1451–1457.

14.

Jones

, Nagin

. Advances in group-based trajectory modeling and an SAS procedure for estimating them. Sociol Methods Res, 2007; 35:542–571.

15.

World Health Organization (WHO).. WHO Multicenter Growth Reference Study Group. 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. World Health Organization: Geneva, Switzerland,. 2006. Available at www.who.int/childgrowth/standards/technical_report/en Last accessed January 7, 2016.

16.

Grummer-Strawn

, Reinold

, Krebs

. Use of World Health Organization and CDC growth charts for children aged 0–59 months in the United States. MMWR Recomm Rep, 2010; 59:1–15.

17.

National Institutes of Health (NIH).. NIH publication no. 98-4083. National Heart Lung and Blood Institutes. National Institutes of Health. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults. The evidence report. 1998. Available at www.nhlbi.nih.gov/files/docs/guidelines/ob_gdlns.pdf Last accessed January 13, 2016.

18.

Olson

, Demment

, Carling

, et al. Associations between mothers' and their children's weights at 4 years of age. Child Obes, 2010; 6:201–207.

19.

Dalenius

, Borland

, Smith

, et al. Pediatric Nutrition Surveillance 2010 Report. US Department of Health and Human Services, Centers for Disease Control and Prevention: Atlanta, GA,. 2012. Available at www.cdc.gov/pednss/pdfs/PedNSS_2010_Summary.pdf Last accessed January 4, 2016.

20.

Weng

, Redsell

, Swift

, et al. Systematic review and meta-analyses of risk factors for childhood overweight identifiable during infancy. Arch Dis Child, 2012; 97:1019–1026.

21.

Velleman

, Welsch

. Efficient computing of regression diagnostics. Am Stat, 1981; 35:234–242.

22.

Jones

, Nagin

, Roeder

. A SAS procedure based on mixture models for estimating developmental trajectories. Sociol Methods Res, 2001; 29:374–393.

23.

Schwarz

. Estimating the dimension of a model. Ann Stat, 1978; 6:461–464.

24.

Ong

, Preece

, Emmett

, et al. Size at birth and early childhood growth in relation to maternal smoking, parity and infant breast-feeding: Longitudinal birth cohort study and analysis. Pediatr Res, 2002; 52:863–867.

25.

Robinson

, Crozier

, Harvey

, et al. Modifiable early-life risk factors for childhood adiposity and overweight: An analysis of their combined impact and potential for prevention. Am J Clin Nutr, 2015; 101:368–375.

26.

Ventura

, Loken

, Birch

. Developmental trajectories of girls' BMI across childhood and adolescence. Obesity (Silver Spring), 2009; 17:2067–2074.

27.

Mihrshahi

, Battistutta

, Magarey

, et al. Determinants of rapid weight gain during infancy: Baseline results from the NOURISH randomised controlled trial. BMC Pediatr, 2011; 11:99.

28.

Giles

, Whitrow

, Davies

, et al. Growth trajectories in early childhood, their relationship with antenatal and postnatal factors, and development of obesity by age 9 years: results from an Australian birth cohort study. Int J Obes (Lond), 2015; 39:1049–1056.

29.

, Goran

, Kaur

, et al. Developmental trajectories of overweight during childhood: role of early life factors. Obesity (Silver Spring), 2007; 15:760–771.

30.

DeNavas-Walt

, Proctor

. US Census Bureau, Current Population Reports, P60-252, Income and Poverty in the United States. 2014. US Government Printing Office: Washington, DC, 2015. Available at www.census.gov/library/publications/2015/demo/p60-252.html Last accessed January 7, 2016.