Maternal Sedentary Behavior and Physical Activity across Pregnancy and Early Childhood Growth

Abstract

Background:

Evidence suggests in utero exposures are related to lifespan health of the offspring. Whether maternal activity profile during pregnancy impacts offspring health remains unknown.

Methods:

This follow-up study recruited mothers with objectively measured sedentary behavior (SED) and moderate-to-vigorous intensity physical activity (MVPA) from a previous cohort study. Maternal activity was analyzed across pregnancy (trajectory groups) and continuously by trimester. Offspring anthropometrics up to 24 months were abstracted from medical records (n = 62). Outcomes included childhood growth rate (incremental rate of BMI z-score change up to 24 months) and rapid growth (increased BMI z-score >0.67 at 12 months). Associations of maternal activity with growth rate were examined using mixed linear models and rapid growth using generalized linear models.

Results:

Forty percent of participants were in the high SED and 20% in the high MVPA trajectories during pregnancy. Higher SED, across pregnancy [slope (95% confidence interval; CI): 0.080 (0.024–0.061) ΔBMI z-score/month] and in the first trimester [standardized beta; std β (95% CI): 0.017 (0.007–0.026)], was related to accelerated growth rate. Higher MVPA, in the second and third trimesters, was associated with accelerated growth rate [std β (95% CI): trimester 2: 0.013 (0.002–0.024) and trimester 3:0.011 (0.003–0.020)] and greater risk of rapid growth [risk ratio (95% CI): trimester 2: 1.25 (1.009–1.555) and trimester 3: 1.25 (1.056–1.475)].

Conclusions:

These findings add to growing evidence on the deleterious effects of high SED during pregnancy. The increased risk for accelerated growth with higher MVPA elicits further investigation. Overall, maternal activity profile shows promise as a modifiable behavior to improve intergenerational health.

Introduction

Noncommunicable diseases, such as cardiovascular disease and type 2 diabetes, are highly prevalent in the United States¹ and prevention efforts are critically important. There is substantial evidence to suggest that the fetal environment impacts long-term noncommunicable disease risk of the offspring. Specifically, the Developmental Origins of Health and Disease and the Freinkel theories hypothesize that undernutrition or overnutrition delivery in utero results in permanent alterations to the offspring's physiology.^2–5 Therefore, prevention of this cascade of cardiovascular risk that begins in utero is of great public health importance.

Size at birth has been extensively studied as an indicator of health and disease risk across the lifespan. Smaller size at birth is consistently associated with an increased risk for cardiometabolic disease development in childhood^6–9 and adulthood.^2,4,5,10,11 Furthermore, the child's rate of growth in early life (<24 months) has also been associated with poorer health outcomes. There is substantial evidence demonstrating that children who grow more rapidly in early life have greater risk for poor cardiometabolic health and overweight and obesity in childhood and adulthood.^8,12–16 Infants with lower birth weight, lower weight-for-length, or those born small for gestational age are at a greater risk for accelerated or rapid growth in early life.¹⁶ Improved understanding of modifiable factors that impact the fetal environment and risk for adverse growth patterns is needed to inform intervention targets and public health recommendations for pregnant women.

The sufficiency of nutrient delivery to the fetus and subsequent fetal growth are the result of the intrauterine environment.¹⁷ This environment is sensitive to maternal nutrition, obesity, inflammation, and other perturbations, likely through alterations in the placental capacity to deliver nutrients to the fetus.¹⁷ As such, maternal health behaviors during pregnancy may play an important role in fetal programming of disease and provide opportunity for future interventions.

One potential behavioral target for improving the fetal environment is maternal activity profile during pregnancy. Activity profile includes time spent in sedentary behavior (SED), defined as low-intensity behaviors in a sitting, reclining, or lying position,¹⁸ and moderate-to-vigorous intensity physical activity (MVPA). This profile may have an effect on early childhood growth due to an effect on placental nutrient transport, energy metabolism,^19,20 and birth size. Our group recently reported that high maternal SED during pregnancy is related to smaller (worse) ponderal index²¹ (g × 100/cm³) at birth.

Other studies have demonstrated that physical activity impacts birth size. Specifically, leisure time MVPA is related to a reduced risk for large birth size without increasing the risk for low birth weight or small weight for gestational age,^22–26 while other domains such as occupational physical activity may negatively impact birth size.²⁷ However, although activity profile appears to be related to size at birth, research investigating whether these effects persist after birth to impact growth rate in early childhood is extremely limited and the few available studies report conflicting results.^28–30 This is an important research gap as improving maternal activity profile is a potential novel approach to lifespan cardiometabolic disease prevention.

The primary objective of this study was to examine associations of objectively measured SED and MVPA, across pregnancy and by trimester, with early childhood growth patterns. A secondary objective was to examine whether these associations are independent of birth size. We hypothesized that higher maternal SED and lower MVPA during pregnancy would be related to more accelerated growth rate or rapid growth. Furthermore, we expected that associations of maternal activity with infant growth would not be independent of birth size.

Methods

This study added follow-up assessments on the children of women enrolled in the MOnitoring Movement and Health (MoM Health) study, which was a prospective cohort study that collected objectively measured SED and MVPA across trimesters of pregnancy.³¹ This follow-up study collected additional health measures on children born to women in the MoM Health study between 12 and 24 months of age. Women were eligible for the parent study (MoM Health) if they were between the ages of 18–45, <14 weeks pregnant, not taking any blood pressure- or glucose lowering-medications, and able to walk half a mile and climb two flights of stairs.

Those with ≥1 trimester of valid objective activity monitoring and a singleton live birth (n = 103) were contacted for follow-up recruitment for this study and were deemed eligible if they had primary custody of their child and the child did not have any congenital or chromosomal abnormalities that would affect growth or development.

The parent study included three study visits, one in each trimester of pregnancy occurring between 8 to <14 weeks, 20 to <23 weeks, and 30 to <35 weeks gestation. To measure SED and MVPA, participants received two activity monitors (described below) at each study visit to be worn for the following 7 days. The follow-up data collection that is the focus of this report included medical record abstraction of childhood growth anthropometrics since birth and an electronic questionnaire battery completed by the mother. All participants completed informed consent and all procedures were approved by the University of Pittsburgh's Institutional Review Board.

Maternal-Child Characteristics

Child age, race/ethnicity, and childhood feeding type (i.e., breastfeeding or formula fed), maternal education, and household income were self-reported by the mother on standard questionnaires using the REDCap (Research Electronic Data Capture) web-based software.

Exposure: Objective Measurement of SED and MVPA

Maternal activity profile data were initially collected during pregnancy in the MoM Health study. Data were collected during each trimester using best practice, objective assessment methodology.^32,33 SED was measured using a thigh-mounted activPAL3 micro accelerometer (PALtechnologies, Glasgow, Scotland) following a 24-hour wear protocol, including sleeping, showering, and bathing, only to be removed for swimming. Using participant diaries, sleep and nonwear times were removed by trained research staff from data exported in the event format.^32,34 MVPA was measured by a waist-worn Actigraph GT3X accelerometer (Actigraph, Pensacola, FL).

Participants were instructed to wear the Actigraph on the right hip, below the gravid abdomen and during all waking hours, only to be removed for sleep or water activities (bathing or swimming). Data were analyzed using Actilife v6 12.2 software using Freedson 2011 VM cut points for classifying MVPA.³⁵ Participants were instructed to wear monitors concurrently for 7 days, and data were considered valid if worn for ≥10 hours on ≥4 days.^32,36

To describe activity across pregnancy, trajectory groups were generated separately for SED and MVPA using group-based trajectory analysis in the parent study cohort.³⁷ Best fit for trajectory groups was selected based on the Bayesian criterion index, greatest percentage of participants placed in groups with posterior probability of ≥70%, and clinical relevance of derived groups. Patterns of both SED and MVPA remained relatively stable across trimesters and therefore, women were assigned to one of three trajectory groups (high, medium, and low) that most closely fit their activity pattern across pregnancy separately for each SED and MVPA. High, medium, and low SED trajectories correspond to an average of 10.9, 9.2, and 7.8 h/day, respectively, while high, medium, and low MVPA trajectories correspond to 52.7, 32.1, and 16 min/day, respectively. These trajectories have been previously reported elsewhere.³¹

Outcome: Early Childhood Growth Patterns

Access to medical records from pediatric visits since birth was obtained for children through the consent process. Medical records not available through the University of Pittsburgh Medical Center electronic health record were requested using clinic-specific medical record request forms signed by the mother.

All available height/length and weight measures were abstracted from well child visits (1 week, 1, 2, 4, 6, 9, 12, 15, 18, and 24 months) in medical records from birth up to 24 months of age. Sick visits or weight check visits were not included to prevent oversampling among children presenting to the medical system more frequently. Anthropometric measures from each visit were converted into age-specific BMI non-scores using the STATA World Health Organization z-score calculator plug-in.³⁸ Growth pattern was assessed using two definitions: growth rate (primary approach) and rapid growth (secondary approach).

Growth rate was examined using incremental rate of BMI z-score change estimated by a line of best fit, using all available BMI z-scores between birth and 24 months from medical records. Rapid growth was operationalized as a dichotomous variable defined as an increase in z-score from birth to 12 months of age ≥0.67.³⁹ Rapid growth (dichotomous outcome) was calculated from birth to 12 months because all participants had reached at least this age at the time of follow-up.

Statistical Analysis

Statistical analyses were performed using Stata 14 software (StataCorp, College Station, TX, USA). Descriptive statistics described the characteristics of the sample, including maternal and child demographics, feeding type, and age at questionnaire completion overall and by trajectory group.

Associations of maternal activity profile with growth rate and rapid growth were analyzed using two methods: (1) categorical maternal SED and MVPA trajectory groups and (2) separate models by trimester with continuous maternal SED and MVPA. In models including continuous SED or MVPA by trimester, all beta coefficients and risk ratios (RRs) were standardized to estimated change in outcome per one standard deviation (SD) of SED (trimester one: 87.2, trimester two: 75.8, and trimester three: 81.2 minutes) or MVPA (trimester one: 16.6, trimester two: 17.3, and trimester three: 17.3 minutes) to facilitate comparison of results.

Mixed linear regression models examined the relationships between maternal SED and MVPA trajectory groups, and trimester-specific SED and MVPA (independent variables) with early childhood growth rate as measured by incremental rate of BMI z-score change (dependent variable; four total models). Models included repeated measures of all child anthropometric data abstracted from pediatric appointment medical records from birth up to 24 months and tested for differences in slope of BMI z-score change with increasing age by maternal activity.

Generalized linear models examined the risk of rapid growth occurring by SED and MVPA trajectory groups and trimester-specific SED and MVPA. BMI z-score at birth was added to rapid growth models to examine associations before and after inclusion and changes in magnitude and significance of associations were assessed qualitatively to explore whether associations were independent of birth size. Similar analyses testing attenuation of the association between maternal SED and MVPA and growth rate were not conducted as BMI z-score at birth was already included in the model.

Due to small sample size and limited ability to adjust for covariates, potential confounders, including race, income, education, child feeding type, maternal prepregnancy BMI, gestational age at delivery, and maternal age, were added one at a time to each model to test for influence on associations. If inclusion of a covariate suggested meaningful confounding, it was to be retained in the final model. However, inclusion of any covariate did not impact statistical significance or magnitude of beta coefficients by >10% in any tested model; thus, unadjusted associations are presented throughout.

Results

Of the n = 103 participants contacted for this follow-up study, n = 74 replied and consented to participate. All participants who responded were deemed eligible (had custody of their child and no chromosomal or congenital abnormality). Medical records were available through University of Pittsburgh Medical Center for 50 participants and received through medical record request for 12 participants. Eight records were not received from clinics after medical record request and four consents to release medical records were not signed due to loss to follow-up, resulting in an analytic sample of 62 participants with available growth data. A flow chart of these recruitment numbers is presented in Supplementary Figure S1. Women who responded and enrolled in this study were significantly younger, more highly educated, and less racially diverse than the parent study sample.

In this study sample, children were 52% male, primarily white (82%), and mostly had mothers who were highly educated (65% had a masters or doctoral degree). Final age of growth data abstracted from medical records was 18.9 ± 4.9 months (range: 12.5–24.9). Patterns of SED and MVPA across pregnancy are presented by high, medium, and low trajectory groups in Figure 1. Demographic and clinical characteristics are summarized in Table 1 overall and by maternal SED and MVPA trajectory. In this follow-up sample, participants had three (n = 96), two (n = 6), or one (n = 1) valid trimesters of objectively measured SED and MVPA.

Figure 1.

Maternal SED and MVPA trajectory groups across pregnancy trimesters. MVPA, moderate-to-vigorous-intensity physical activity; SED, sedentary behavior.

Table 1.

Participant Characteristics Overall and by Sedentary Behavior and Moderate-to-Vigorous-Intensity Physical Activity Trajectory Groups

	SED trajectory (n)					MVPA trajectory (n)
Mean (SD)	Overall (62)	Low (10)	Med (27)	High (25)	p-value	Low (15)	Med (34)	High (13)	p-value
Child age, months	20.9 (5.0)	22.4 (5.5)	20.5 (4.9)	20.7 (5.0)	0.588	20.6 (5.7)	21.0 (4.9)	22.1 (5.0)	0.950
Gestational age at birth, weeks	40.0 (1.6)	39.2 (0.9)	38.9 (1.4)	38.8 (1.9)	0.791	39.4 (0.8)	39.0 (1.8)	38.4 (1.7)	0.253
Maternal prepregnancy BMI, kg/m2	25.9 (6.6)	30.1 (9.6)	24.8 (5.8)	25.3 (5.4)	0.075	26.8 (6.6)	25.9 (5.3)	25.9 (9.4)	0.812
Maternal EPDS scorea	3.3 (3.2)	3.0 (3.5)	2.4 (2.8)	4.1 (3.3)	0.227	3.7 (2.4)	2.9 (3.2)	3.9 (4.5)	0.647
n (%)
Sex					0.110				0.361
Male	32 (52)	4 (40)	11 (41)	17 (68)		10 (67)	15 (44)	7 (54)
Female	30 (48)	6 (60)	16 (59)	8 (32)		5 (33)	19 (56)	6 (46)
Feeding type					1.00				0.812
Exclusively breastfed	31 (50)	5 (50)	13 (48)	13 (52)		7 (47)	17 (50)	7 (53)
Partial breastfeeding	28 (45)	5 (50)	12 (45)	11 (44)		8 (53)	14 (41)	6 (46)
Exclusively formula fed	3 (5)	0	2 (7)	1 (4)		0	3 (9)	0
Maternal education					0.133				0.159
Some college or training	9 (15)	4 (40)	3 (11)	2 (8)		4 (27)	3 (9)	2 (15)
College graduate	13 (21)	2 (20)	4 (15)	7 (28)		5 (33)	7 (21)	1 (8)
Masters/doctoral	40 (65)	4 (40)	20 (74)	16 (64)		6 (40)	24 (71)	10 (77)
Household income					0.084				0.528
<50,000	5 (8)	3 (30)	1 (4)	1 (4)		1 (7)	3 (9)	1 (8)
50,000 to <75,000	7 (11)	1 (10)	4 (15)	2 (8)		3 (20)	3 (9)	1 (8)
>75,000	48 (78)	5 (50)	21 (78)	22 (88)		10 (68)	28 (82)	10 (76)
Do not know/refused to answer	2 (3)	1 (10)	1 (4)	0		2 (7)	0	1 (8)
Race					0.486				0.742
White	51 (82)	7 (70)	24 (89)	20 (80)		11 (73)	28 (82)	12 (92)
Black	6 (10)	2 (20)	2 (7)	2 (8)		2 (13)	3 (9)	1 (8)
Other	5 (8)	1 (10)	1 (4)	3 (12)		2 (13)	3 (9)	0

p-Value represents tests for differences (one-way ANOVA or Fisher's exact test) in participant characteristics across SED or MVPA trajectory groups.

^a Fewer observations available for EPDS score (n = 46).

EPDS, Edinburgh Postpartum Depression Score; MVPA, moderate-to-vigorous-intensity physical activity; SED, sedentary behavior.

While no characteristic significantly differed across activity trajectory groups, mean maternal prepregnancy BMI was meaningfully lower in the medium and high SED trajectory groups compared to low SED, which was in the obese category. Mean (SD) BMI z-score was −0.69 (1.17) at birth and 0.03 (0.97) at 12 months, with an average change of 0.70 (1.41) between birth and 12 months.

Growth Rate

Predicted slopes of childhood BMI z-score growth rate by maternal SED and MVPA trajectories are displayed in Figure 2. These unadjusted analyses included n = 62 children with an average of 9.2 (range: 4–11) observations per child. Growth rate was significantly different by maternal SED, but not MVPA, trajectory. Mothers in both medium and high maternal SED trajectories had infants with significantly steeper growth rates compared to infants from mothers in the low SED trajectory. Participants in the high SED group also had the smallest BMI z-score at birth (BMI z-score low SED: 0.02, medium SED: −0.57, and high SED: −1.10) and the steepest slope of growth [slope (95% confidence interval; CI): 0.080 (0.02–0.06) ΔBMI z-score/month] with increasing age.

Figure 2.

Incremental rate of BMI z-score change by sedentary and MVPA trajectory.

Unadjusted mixed linear models examining growth rate and trimester-specific maternal SED and MVPA are displayed in Table 2. Higher SED in the first trimester [standardized beta; std β (95% CI): 0.017 (0.007–0.026)] as well as higher MVPA in the second [std β (95% CI): 0.013 (0.002–0.024)] and third [std β (95% CI): 0.011 (0.003–0.020)] trimesters were associated with steeper slope of growth (i.e., greater ΔBMI z-score/month per 1 SD increase in SED or MVPA).

Table 2.

Unadjusted Incremental Rate of BMI z-score Change with Age by Trimester Specific Maternal Sedentary and Moderate-to-Vigorous-Intensity Physical Activity

	Trimester 1	Trimester 2	Trimester 3
	Standard coefficient (95% CI)	Standard coefficient (95% CI)	Standard coefficient (95% CI)
SED	0.017 (0.007 to 0.026)	−0.002 (−0.013 to 0.009)	0.004 (−0.005 to 0.014)
MVPA	0.003 (−0.008 to 0.014)	0.013 (0.002 to 0.024)	0.011 (0.003 to 0.020)

Standardized beta coefficient represents change in BMI z-score/month slope per one SD increase in SED or MVPA.

SD SED: Trimester 1: 87.2, Trimester 2: 75.8, Trimester 3: 81.2 minutes.

SD MVPA: Trimester 1: 16.6, Trimester 2: 17.3, Trimester 3: 17.3 minutes.

Rapid Growth

Overall, there were 28 cases of rapid growth (increase >0.67 in BMI z-score from birth to 1 year). Rates of rapid growth occurrence were 4/9 (11%) in low, 10/26 (19%) in medium, and 14/25 (32%) in high SED trajectory groups and 6/15 (7%) in low, 16/32 (25%) in medium, and 6/13 (38%) in high MVPA trajectory groups. Unadjusted generalized linear models examining RRs of rapid growth by maternal SED or MVPA trajectory groups are presented in Table 3. Risk of rapid growth did not significantly differ by SED or MVPA trajectory groups and inclusion of BMI z-score at birth in these models further attenuated the risk for experiencing rapid growth.

Table 3.

Risk of BMI z-score Rapid Growth between Birth and 12 Months of Age by Maternal Activity Trajectory with and without Adjustment for BMI z-score at Birth

	Low	Medium	High
	RR (95% CI)	RR (95% CI)	RR (95% CI)
SED
Unadjusted	1.00 (Reference)	0.95 (0.395–2.291)	1.26 (0.551–2.883)
Adjusted	1.00 (Reference)	0.73 (0.367–1.450)	0.74 (0.360–1.504)
MVPA
Unadjusted	1.00 (Reference)	1.33 (0.643–2.743)	1.15 (0.478–2.784)
Adjusted	1.00 (Reference)	1.04 (0.552–1.973)	0.90 (0.440–1.857)

Adjusted models include BMI z-score at birth.

RR, risk ratio.

Generalized linear models for risk of rapid growth by trimester-specific maternal SED and MVPA are presented in Table 4. Maternal SED in any trimester and first or second trimester MVPA were not significantly associated with risk of rapid growth. Higher maternal MVPA in the third trimester (RR 1.95, 95% CI 1.04–3.63) was significantly associated with higher risk of early childhood rapid growth. This association was partially attenuated and no longer significant with the addition of BMI z-score at birth.

Table 4.

Risk of BMI z-score Rapid Growth between Birth and 12 Months of Age by Trimester-Specific Activity with and without Adjustment for BMI z-score at Birth

	Trimester 1	Trimester 2	Trimester 3
	RR (95% CI)	RR (95% CI)	RR (95% CI)
SED
Unadjusted	1.05 (0.819–1.353)	0.92 (0.685–1.232)	0.95 (0.691–1.292)
Adjusted	0.71 (0.721–1.144)	0.90 (0.718–1.138)	0.89 (0.686–1.144)
MVPA
Unadjusted	1.11 (0.876–1.399)	1.25 (1.009–1.555)	1.25 (1.056–1.475)
Adjusted	1.08 (0.885–1.318)	0.99 (0.810–1.212)	1.10 (0.935–1.297)

Risk ratios represent risk of rapid growth associated with a one standard deviation increase in SED or MVPA. SD SED: Trimester 1: 87.2, Trimester 2: 75.8, Trimester 3: 81.2 minutes. SD MVPA: Trimester 1: 16.6, Trimester 2: 17.3, Trimester 3: 17.3 minutes. Adjusted models include BMI z-score at birth.

Discussion

This study was conducted to better understand how maternal activity profile across pregnancy may relate to early childhood growth patterns. The main findings are that maternal SED and MVPA during pregnancy are related to early childhood growth patterns. Higher maternal SED, across pregnancy by trajectory and in the first trimester, was associated with a more accelerated slope of BMI z-score increase up to 24 months of age, but was not associated with the risk of rapid growth at 12 months. Higher maternal MVPA, specifically in late pregnancy, was associated with a more accelerated slope of BMI z-score increase from birth up to 24 months of age and increased risk of rapid growth in the child at 12 months of age.

Aligned with our hypothesis, offspring of women with higher amounts of SED were born smaller²¹ and, particularly for those women with high SED in the first trimester, were more likely to have rapid growth up to 24 months of age. Rapid growth, when accompanied by small size at birth, is associated with a higher long-term risk for cardiovascular disease, overweight, and obesity.^16,40 These findings provide novel evidence to suggest that high maternal SED during pregnancy may affect offspring health into early childhood. The potential negative impact of high SED during early pregnancy on long-term health of the offspring requires further investigation.

Contrary to our hypothesis, we found that being in the higher maternal MVPA was associated with more adverse childhood growth pattern. Considering the evidence that this accelerated or rapid growth is associated with cardiovascular disease risk^12,16,40,41 and overweight and obesity risk,^8,9,14 our findings suggest that higher levels of maternal MVPA may have a deleterious effect on long-term health of the offspring. However, further consideration of growth patterns may be necessary to interpret these findings. BMI z-score at birth was not associated with maternal MVPA trajectory in this cohort, and some studies have suggested that rapid growth, in the absence of small birth size, is not as strongly associated with long-term poorer health.^16,40

It is important to note that physical activity during pregnancy is considered safe and encouraged in healthy, uncomplicated pregnancies.⁴² Therefore, our findings should elicit further investigation into the long-term impact that higher levels of maternal MVPA may have on offspring and, specifically, whether associated rapid growth in these offspring engenders deleterious long-term health outcomes in broader population studies.

Our study contributes novel findings to an important research gap by examining associations of objectively measured maternal activity across pregnancy with within-subject changes in offspring BMI z-score over time, which may better capture growth patterns in early childhood. While no previous study has used within-subject measures, our results can be compared to previous studies that have reported on time point-specific differences in offspring body mass between experimental and control groups following antenatal exercise interventions.

One randomized control trial, including a 12-week maternal exercise intervention in the second and third trimester, found lower birthweight, no difference in weight at 1-year follow-up, and higher body fat percentage at the 7-year follow-up in children born to mother in the experimental (n = 33) versus control groups (n = 24).²⁸ While within-subject changes in infant weight were not directly reported, birthweight was lower in the intervention offspring, but did not differ from controls at 1 year. This may support our findings in which higher MVPA was related to more accelerated growth, although the ability to compare findings is limited by the time point-specific analysis as opposed to patterns of growth examined in this study.

Contrasting our findings, one study found that at 5 years of age, children born to mothers who maintained at least 30 minutes of aerobic activity (n = 20) three times per week during pregnancy had lower body fat percentage than those whose mothers discontinued exercise during their pregnancy (n = 20).²⁹ Finally, a randomized controlled trial, which included dietary counseling and daily moderate aerobic activity in overweight and obese pregnant women (intervention: n = 77 and control: n = 73) found no difference in offspring BMI z-score at 2.5–3 years of age.³⁰

The differences in populations and methodology between the existing studies make it difficult to compare findings. Yet, taken together, maternal physical activity may have an effect on early childhood growth. However, this effect may vary by time of exercise introduction, prepregnancy activity, age of follow-up on the children, and maternal BMI, suggesting a need for more research.

Due to the nature of this study, the mechanisms by which maternal SED or MVPA may result in accelerated or rapid growth cannot be ascertained. One proposed mechanism by which this may occur is the mismatch theory that suggests differences between in utero and postnatal nutritional availability result in a metabolic mismatch,^43,44 which may lead to quicker rate of growth postnatally. The placenta's vascular capacity to transport nutrition to the fetus may be reduced with high amounts of SED.¹⁷ This proposed mechanism is consistent with previous findings from our group and others in which high maternal SED was related to smaller ponderal index at birth²¹ and lower ponderal index are related to insufficiency of nutrient delivery in the placenta.⁴⁵

Insufficient early pregnancy nutrient delivery may inhibit the structure and function of developing organ systems, preparing the fetal physiology for less nutrition than what is then available postnatally, resulting in the metabolic mismatch. Furthermore, MVPA in late pregnancy is related to intermittent reductions in maternal glucose after exercise, which may result in placental adaptations and reduced nutrient delivery to the fetus over time.⁴⁶ Therefore, higher MVPA in late pregnancy may be protective of a nutritional excess to the fetus and, when mismatched with abundant nutrition postnatally, may increase the offspring's risk for more accelerated growth. Further research should seek to ascertain the impact of accelerated growth, within the context of high MVPA, on future health risks.

Our findings could also be related to postnatal exposures rather than in utero programming. While factors such as feeding type, primary caregiver interactions, and maternal diet have not been directly related to adverse growth patterns, they are related to childhood health.⁴⁷ Furthermore, birth size and early childhood weight gain differ by maternal education, race, and socioeconomic status.^48,49 Yet, the addition of these covariates where available to our models did not change the significance or magnitude of associations. Altogether, our findings in which BMI z-score at birth attenuated the associations between maternal MVPA and rapid growth would support in utero development as the leading hypothesis for increased risk of rapid growth as opposed to postnatal exposures. Further research on postnatal correlates of early childhood growth rates is needed to disentangle fetal programming or postnatal environmental factors as mechanisms for accelerated or rapid growth.

Strength and Limitations

The primary strength of this study was the objective measurement of SED and MVPA across all trimesters of pregnancy. Having multiple measures during pregnancy allowed for analysis by pattern across pregnancy and within each trimester. Furthermore, the prospective, observational design allowed for assessment of habitual SED and MVPA and the ability to establish temporality of the prenatal exposures with postnatal outcomes. Other strengths include data collection from medical records, which included anthropometrics objectively measured by clinicians. This also allowed for analyses using frequent, longitudinal within-subject measurements for each subject to get a more accurate measure of growth rate during early childhood.

This study also had limitations. The small sample size affected our ability to adjust for possible confounders in our analyses. While adjustment for confounders did not change our results, this could be attributed to a lack of power to detect differences. Due to our homogenous sample, the results found in our study may differ in other samples of mothers and children (i.e., racial/ethnic minority groups). In addition, our limited racial/ethnic, socioeconomic, and demographic diversity limit the generalizability and our ability to evaluate subgroup differences.

As we did not experimentally manipulate maternal activity patterns during pregnancy, we cannot determine causality such as whether changing SED or MVPA would elicit the same effects. Finally, this study only examined waking activity as no assessment of sleep was included. Future research should consider compositional analysis of 24-hour activity. Despite these limitations, our findings inspire future study of the association between maternal activity profiles and early childhood growth.

Conclusions

There is substantial evidence to suggest that early life exposures and health are strongly related to health outcomes across the lifespan. Our findings indicate that maternal activity behaviors during pregnancy may have implications on early childhood health. High amounts of SED were related to more accelerated early childhood growth, adding to a growing body of evidence on the adverse effects of high amounts of SED during pregnancy. This reinforces the need for SED research and consideration of SED recommendations for pregnant women. The potentially deleterious effects of higher levels of MVPA on early childhood growth patterns also warrant further investigation. Within the context of the Developmental Origins of Health and Disease theory, our findings identify maternal activity profile as a possible modifiable prenatal exposure by which risk for noncommunicable diseases of the offspring may be impacted.

Authors' Contributions

M.A.J., S.E.T.R., K.D., K.L., and B.B.G. were involved with conceptualization and methodology of this project, M.A.J., K.W., and B.B.G. were involved with writing of the original draft, and M.A.J., S.E.T.R., K.D., and K.L were involved with writing—review and editing of subsequent drafts involved. All authors have read and agreed to the published version of the article.

Funding Information

This work was supported by the University of Pittsburgh's School of Education Student Research Grant awarded to M.J.; by the American Heart Association (17GRNT3340016) awarded to B.B.G.; and by the National Heart, Lung, and Blood Institute (R01HL153095) awarded to K.W.

Author Disclosure Statement

No competing financial interests exist.

Supplementary Material

Supplementary Figure S1