Physical Activity and Trajectories of Cardiovascular Health Indicators During Early Childhood

Details of the HOPP study design and methodology have been published.¹⁴ Briefly, each year for 3 years, children completed a series of assessments distributed over 2 visits separated by 19 ± 14 days. At the first visit, anthropometry and fitness assessments were conducted. At the end of the visit, the child was given a physical activity monitor to wear for 1 week. After the activity monitoring period, the child attended a second visit, at which resting arterial stiffness and blood pressure were assessed. Data were collected from August 2010 to September 2014. Written informed consent from the child’s legally authorized representative (eg, parent) was obtained at the initial visit. The Hamilton Integrated Research Ethics Board approved the study.

The HOPP study was conducted on a community-based sample through enrollment of children when they were 3, 4, or 5 years old. Recruitment took place from August 2010 to August 2012 in south-central Ontario, Canada, through early-years centers, preschools, day cares, local schoolboards, community events, and word of mouth. Children with diagnosed medical conditions or known developmental or cognitive delays were excluded. There were 418 children who completed the initial visit; 42 participants withdrew or were lost to follow-up by the final visit (5% attrition per year; Fig 1).

Free-living physical activity was assessed by using accelerometry. Children were instructed to wear an accelerometer (ActiGraph GT3X line) over their right hip for 7 days during all waking hours except prolonged water activities. Data were downloaded in 3-second epochs and analyzed with ActiLife software (version 6.6.3; ActiGraph). Nonwear time, identified as periods of at least 60 minutes of continuous 0 counts or as indicated by the parent in a logbook, was removed. Data were analyzed in the vertical plane for average daily TPA and MVPA by using the Pate cut-points.^18,19 The 15-second epoch cut-points were divided by 5 to set thresholds at ≥8 counts per 3 seconds for TPA and ≥84 counts per 3 seconds for MVPA. At least 3 days of valid wear (defined as at least 10 hours of wear) were required for inclusion in subsequent statistical analyses.²⁰ 

To provide a comprehensive perspective of cardiovascular health, we assessed 3 related indicators: cardiovascular fitness (exercise time on a treadmill test and heart rate recovery [HRR]), arterial stiffness (pulse wave velocity [PWV] and carotid artery β stiffness index), and seated brachial artery systolic blood pressure (SBP).

Exercise time on a graded maximal treadmill test (hereafter, treadmill time), the Bruce Protocol,²¹ and HRR 1 minute after test termination were used as indicators of cardiovascular fitness.²² To accommodate the children’s abilities and small size, participants held handrails, and a researcher was positioned behind the child for safety.²³ A heart rate (HR) monitor (Polar Electro) was worn on the chest throughout. The treadmill (General Electric Marquette Series 2000) increased in speed and incline every 3 minutes until the child could no longer maintain running cadence to position themselves at the front of the treadmill or refused to continue. The time of cessation was recorded as the treadmill time. The child was seated immediately after, and HRR was calculated as the difference in peak HR achieved during the test and HR at 1 minute of recovery. Fitness data were excluded from analyses if the participant did not reach an HR of at least 180 beats per minute.²⁴ HRR data were also excluded if the child did not sit calmly during recovery (eg, coughing, being upset, or moving). A longer treadmill time and faster HRR indicate more favorable cardiovascular fitness.

Children lay supine, and a movie was displayed overhead to encourage them to remain still and quiet during the assessment. After at least 10 minutes of rest, arterial stiffness was measured regionally in a large segment of the arterial tree (whole-body PWV) and locally in the common carotid artery (β stiffness index). PWV was determined as the speed of the pulse wave between the carotid and dorsalis pedis arteries. A handheld pressure transducer (tonometer model SPT-301; Millar Instruments) was placed on the neck over the right carotid artery, and a photoplethysmograph probe (MLT1020PPG; ADInstruments) was placed on the right foot over the dorsalis pedis artery. The time delay between the arrival of the pulse wave at these 2 sites was determined by applying a 5- to 30-Hz bandpass filter,²⁵ as previously described²⁶ (LabChart 7 Pro version 7.3.4; ADInstruments). The time delay for 20 consecutive beats was averaged and divided into the measured distance between the 2 sites ([suprasternal notch to sensor on foot] – [suprasternal notch to location of tonometer on carotid]).^27,28 A faster PWV is indicative of greater arterial stiffness.

Local stiffness in the common carotid artery was calculated by using β stiffness index²⁷:

where P_s and P_d are carotid artery systolic and diastolic blood pressure, respectively, and AD_max and AD_min are the maximum and minimum artery diameters over a heart cycle. We chose to use the β stiffness index because this takes into account the child’s artery size and blood pressure, which increase with age,^29,30 and it is more reliable than compliance and distensibility in young children.²⁶ Carotid artery systolic and diastolic blood pressure were determined by calibrating 10 heart cycles of carotid artery pressure waveforms, obtained from the tonometer, to discrete measures of supine brachial artery blood pressure, as previously described.²⁶ Simultaneous to tonometry, a 12-MHz linear ultrasound probe (Vivid q; General Electric Medical Systems) was placed over the left common carotid artery with an orientation to view the long axis. Video clips of brightness mode ultrasound were captured for 10 consecutive heart cycles (23 frames per second). Frames corresponding to the maximum and minimum artery diameters over a heart cycle were extracted (Sante DICOM Editor, Santesoft version 3.3.1), and arterial wall boundaries were analyzed by using semiautomated edge tracking software (Artery Measurement System version 2.0).

Automated measures of seated blood pressure (Dinamap Pro 100; Critikon Inc) were obtained in the right arm in triplicate with a 1-minute delay between each measure. The second and third measures were averaged³¹; additional measures were taken if SBP differed by more than 5 mm Hg. Seated SBP was used as a cardiovascular health indicator.³² 

Linear mixed-effects modeling with the restricted maximum likelihood method was used to determine the effect of physical activity on trajectories of cardiovascular health indicators (SAS University Edition release 3.6; SAS Institute, Inc, Cary, NC). This method accommodates missing data by creating estimates using all data available for each participant. The same approach was taken for modeling each cardiovascular health indicator (treadmill time, HRR, PWV, β stiffness index, and SBP). TPA and MVPA were always in separate models. We determined the effect of time (ie, study years 1, 2, and 3) and physical activity (TPA or MVPA) on each cardiovascular health indicator (model 1) and subsequently assessed for a TPA × time or MVPA × time interaction (model 2). All models were adjusted for the child’s sex and/or gender (parent reported), age at enrollment, and height z score³³ as a surrogate for growth at each visit. To determine if the impact of physical activity on cardiovascular health differed by sex and/or gender, TPA × sex and/or gender and MVPA × sex and/or gender interactions and 3-way interactions with time (TPA × time × sex and/or gender and MVPA × time × sex and/or gender) were added. To test if the effects of physical activity were confounded by body size, we further adjusted models by BMI. A random intercept at the participant level was included in all models.

Participant characteristics at baseline are shown in Table 1. Physical activity (exposure) and indicators of cardiovascular health (outcome) variables are presented in Tables 2 and 3, respectively. Children lost to follow-up (25–42 depending on outcome measure) had shorter treadmill time and lower β stiffness index at baseline but were otherwise not different compared with those retained in the cohort (Supplemental Table 5). Results are summarized below with linear mixed-effects models 1 and 2, presented in Table 4. Estimates are unstandardized, and all effects were independent of the child’s age at enrollment and height z score. Estimates (Est) were similar in all models after adjustment for BMI (results not shown).

Treadmill time increased over the study. TPA and MVPA had positive main effects on treadmill time (Table 4), showing higher levels of physical activity are associated with greater endurance. A significant TPA × time × sex and/or gender interaction (Est = −0.005; P = .049) revealed that TPA had a positive effect on the rate of change of treadmill time for girls but not boys. HRR did not change over the study period; there were positive main effects of TPA and MVPA on HRR, and this association was not moderated by sex and/or gender.

PWV and β stiffness index increased over the 3 years, indicating arteries became stiffer over time. There was an inverse (favorable) main effect of TPA on PWV, and this relationship was consistent over time (Fig 2A). The significant MVPA × time interaction indicates that the rate of change of PWV differs on the basis of levels of MVPA, with higher amounts of MVPA being associated with a slower increase in PWV (Fig 2B). Similarly, there was a significant MVPA × time interaction on the β stiffness index. The TPA × time interaction on the β stiffness index was not significant (P = .051). There were no moderating effects of sex and/or gender on either stiffness outcome.

SBP rose over the study period. Although neither TPA nor MVPA were associated with SBP on average or over time (Table 4), a significant MVPA × time × sex and/or gender interaction (Est = 0.06; P = .009) revealed that over time, MVPA had a favorable effect on the rate of change of SBP for girls. The TPA × time × sex and/or gender interaction was not significant (Est = 0.02; P = .06).

Our findings demonstrate that in early childhood, trajectories of cardiovascular fitness and arterial stiffness are favorably impacted by higher levels of physical activity independent of age, height, and BMI. The positive effects of physical activity on cardiovascular health were found for boys and girls, suggesting that both benefit from engagement in physical activity and particularly MVPA; however, MVPA appears to be associated with SBP in girls only.

Cardiovascular fitness is an important indicator of cardiovascular risk, and it has been suggested that it be included as a vital sign when assessing patient health and risk in adults.³⁴ We found that higher objectively measured physical activity levels during early childhood have a positive effect on cardiovascular fitness, as measured by a laboratory-based assessment of exercise time on a graded maximal treadmill test and HRR independent of age and growth. Our findings are consistent with the Ballabeina³⁵ and Mobile-Based Intervention Intended to Stop Obesity in Preschoolers³⁶ studies that reported positive associations between baseline MVPA and field-based measures of fitness 9 to 12 months later in 4- to 6-year-olds. However, these studies involved data from the control arm of an intervention trial and did not control for physical activity at follow-up. By including repeated assessments of physical activity and fitness, we determined that the positive effect of MVPA on fitness is consistent over a 3-year span in early childhood for boys and girls, but TPA appears to improve endurance in girls only. This may reflect boys’ average higher physical activity and fitness levels (∼25 minutes more TPA and lasting 30 seconds longer on the treadmill than girls), thereby requiring more intense physical activity to improve fitness.

Our study advances our understanding of the role of physical activity on cardiovascular health with novel determinations of the relationship between physical activity and HRR during early childhood. Although we included HRR as a proxy of fitness, it is more commonly used as an indicator of autonomic function because the drop in HR after exercise is primarily due to parasympathetic reactivation.³⁷ Therefore, our results suggest that young children who engaged in greater amounts of physical activity also have better autonomic function, as previously identified in studies of older children³⁸ and adults.^39,40 

Both PWV and carotid artery stiffness are strong predictors of cardiovascular events, even after accounting for traditional risk factors.^41,42 In a cross-sectional study involving 5- and 8-year-olds, Idris et al⁴³ found no relationship between parent-reported physical activity and sport participation and carotid artery stiffness. In contrast, significant inverse relationships between objectively measured physical activity and arterial stiffness were reported in cross-sectional examinations of 6- to 8-year-olds¹⁰ and 10-year-olds.¹¹ Using a longitudinal design and objective physical activity assessment, we found that age-related increases in arterial stiffness were attenuated in children who engaged in greater amounts of MVPA but not TPA (Fig 2). This suggests that more intense physical activity is required to slow the progressive stiffening of arteries and results in better vascular health trajectories. Indeed, a recent systematic review concluded that positive associations between physical activity and most indices of cardiovascular health in school-aged children are more robust at higher intensities of activity.⁴⁴ Future work should further delineate the role of intensity and determine if certain bout lengths of MVPA are optimal for promoting cardiovascular health in young children.

Although not a specific objective of our article, we note that PWV and β stiffness index increased with age, in contrast to Hidvégi et al,⁴⁵ who did not find differences across 3- to 8-year-olds in PWV measured by a pulse wave analysis technique. Our results align with the literature in school-aged children and adults, which consistently show central arteries stiffen with age.^46,–48 

Our findings show that SBP increases more slowly for girls engaging in greater amounts of MVPA but not boys. The association between MVPA and SBP seems to emerge over time, which is consistent with cross-sectional studies, which more often find relationships between physical activity and SBP in school-aged⁶ but not preschool-aged^13,49,–51 children. It is unclear why we found this relationship only in girls; however, given that physical activity was associated with arterial stiffness in both boys and girls, blood pressure, which has low day-to-day reliability in young children,²⁶ may not have been a sensitive enough indicator to detect changes in vascular health in this cohort of boys.

This study is strengthened by repeated measures over time, a high retention of participants over the study (90%), objective measurement of physical activity, a laboratory-based maximal fitness test, and sensitive indicators of arterial health. Notwithstanding these strengths, we acknowledge the following limitations: Only 22% of our families were below the median income for the geographic area, and 13% of participants were non-white, potentially limiting the generalizability of our sample. Accelerometry data lack information during water-based activities, such as swimming, and are sensitive to epoch and cut-point selection; however, it remains the best available tool for assessment of free-living physical activity in children. We were unable to determine peak oxygen consumption, the gold standard measure of fitness, during the maximal exercise test because collecting breath samples in young children was not feasible.⁵² Nevertheless, exercise time on the Bruce Protocol is a valid indicator of fitness in children,⁵³ and HRR was used to complement our results. Finally, our measure of PWV incorporated central and peripheral arteries, whereas the gold standard measure includes only central arteries (carotid to femoral). As a result, we may have underestimated the impact of physical activity on PWV because studies in adults report a relationship between physical activity and central, but not peripheral, PWV.^54,55 

Engagement in physical activity results in greater cardiovascular fitness, better autonomic function, and lower arterial stiffness during early childhood. More intense physical activity (ie, MVPA) provides additional benefits because it is associated with slowing the progressive stiffening of arteries, which is a marker of atherosclerosis. MVPA also slows the increase in SBP in girls. This study adds to the evidence that physical activity is beneficial to cardiovascular health^4,5,44 and fills an important gap in the literature demonstrating that the protective effects of physical activity on cardiovascular health begin early in childhood. Future research should evaluate if the effects of physical activity on cardiovascular health indicators during early childhood carry over into later childhood and adulthood.

We are grateful to the families who participated in the HOPP Study for their time, energy, and commitment. We also thank all of the trainees and staff who contributed to the study. In particular, we thank Hilary Caldwell and Leigh Gabel (who assisted with data collection for fitness outcomes), Ninette Shenouda (who assisted with data collection for vascular outcomes), and Natascja Di Cristofaro (who provided accelerometry analysis and administrative support).

HOPP

Health Outcomes and Physical Activity in Preschoolers

HR

heart rate

HRR

heart rate recovery

MVPA

moderate-to-vigorous physical activity

PWV

pulse wave velocity

SBP

systolic blood pressure

TPA

total physical activity