ARTICLE IN PRESS

THE JOURNAL OF PEDIATRICS • www.jpeds.com ORIGINAL

ARTICLES
Early Life Adversity with Height Stunting Is Associated with
Cardiometabolic Risk in Adolescents Independent of Body Mass Index
Brie M. Reid, MA1, Michelle M. Harbin, MS2, Jessica L. Arend, BA1,3, Aaron S. Kelly, PhD4, Donald R. Dengel, PhD2,4, and
Megan R. Gunnar, PhD1

Objective To evaluate cardiovascular and metabolic function in youths adopted internationally from orphanages/
institutions (postinstitutionalized) who were height-stunted at adoption.
Study design A total of 30 postinstitutionalized youths (age, 9-18 years; body mass index [BMI] percentile, 7.2-
90.4) who were height-stunted at adoption were compared with age- and BMI percentile-matched youths (n = 90).
Measurements included total body fat and visceral adipose tissue (dual radiograph absorptiometry), arterial stiff-
ness (augmentation index and pulse wave velocity), cardiac autonomic function (heart rate variability), blood pres-
sure, and fasting lipid, glucose, and insulin levels. Linear regression analyses were computed controlling for parent
education, age, trunk tissue fat, height-for-age, sex, and race.
Results Compared with controls of the same age, sex, and BMI, the postinstitutionalized children had higher sys-
tolic blood pressure (P = .018), augmentation index (P = .033), total cholesterol (P = .047), low-density lipoprotein
cholesterol (P = .03), triglycerides (P = .048), insulin (P = .005), and HOMA-IR (P = .01) values. The postinstitutionalized
children had a lower low-frequency to high-frequency ratio (P = .008), indicating lower sympathetic tone, as well
as a lower total lean mass (P = .016), a lower gynoid lean mass (P = .039), and a higher proportion of trunk tissue
fat (P = .017). The postinstitutionalized and control children did not differ in any other body composition measures.
Conclusions Early life stress, as represented by height-stunted growth in institutional care, may be associated
with early pathways to cardiovascular and metabolic risk in youths even after moving into well-resourced homes
early in life and in the absence of increased adiposity. These findings suggest that postinstitutionalized youths with
a history of height stunting may need to be closely monitored for emergent cardiometabolic risk factors. (J Pediatr
2018;■■:■■-■■).

O
ver the last 25 years, more than 300 000 children have been adopted internationally into the US, with the majority
coming from institutional (ie, orphanage) care.1,2 Many arrive to their new families growth-delayed, with a signifi-
cant percentage meeting the criteria for mild (−1.5 z-score height-for-age) or moderate height stunting (−2 z-score
height-for-age). Families who adopt internationally tend to be of higher income and education,3 and within a short period fol-
lowing adoption, most children fall within age norms for height and weight.4 The early life histories of these children, com-
bined with the evidence of poor growth before adoption and with rapid catch-up
thereafter, theoretically should increase their risk for early-onset puberty and
obesity,5,6 ultimately leading to increased risk of developing cardiometabolic
problems.7 From the 1Institute of Child Development, University of
Our research group found that youths adopted internationally from orphan- Minnesota; 2School of Kinesiology, University of
Minnesota; 3Department of Psychiatry, University of
ages (postinstitutionalized) at age 7-14 years tended to be lean, and females did Minnesota Medical School; and 4Department of
Pediatrics, Center for Pediatric Obesity Medicine,
not exhibit early pubertal onset.8 This was true also of the 25% of the sample who University of Minnesota Medical School, Minneapolis, MN
were severely height-stunted at adoption.8 However, because body mass index (BMI) Funded by the National Heart, Lung, and Blood Institute
(R01 HL110957, to A.K.), the National Center for
is only one measure of metabolic risk, it is possible that postinstitutionalized youths Advancing Translational Sciences (UL1TR000114), the
may have higher levels of total body fat and/or a greater degree of fat mass dis- National Institute of Diabetes and Digestive and Kidney
Diseases (NORC Grant P30 DK050456), and the
9
tributed in the visceral adipose tissue (VAT) depots, which may increase cardio- Canadian Institute for Advanced Research Child and
Brain Development program (to M.G.). The content is
vascular risk.10,11 It is also possible that postinstitutionalized youths have additional solely the responsibility of the authors and does not
necessarily represent the official views of the National
metabolic and cardiovascular risks that are not captured by a measure of BMI. Institutes of Health. This material is based on work
supported by the National Science Foundation Graduate
Research Fellowship (Grant 00039202, to B.R.). Any
opinion, findings, and conclusions or recommendations
expressed in this material are those of the authors(s) and
AIx Augmentation index do not necessarily reflect the views of the National
BMI Body mass index Science Foundation. A.K. serves as a consultant for Novo
Nordisk, Orexigen, and Vivus Pharmaceuticals and
ECG Electrocardiography receives research support from Astra Zeneca
HF High frequency Pharmaceuticals in the form of drug/placebo. D.D. serves
HOMA-IR Homeostasis model assessment–estimated insulin resistance as a paid consultant for Hologic, Inc. The other authors
declare no conflicts of interest.
LF Low frequency
PWV Pulse wave velocity 0022-3476/$ - see front matter. © 2018 Elsevier Inc. All rights
VAT Visceral adipose tissue reserved.
https://doi.org10.1016/j.jpeds.2018.06.047

FLA 5.5.0 DTD ■ YMPD10083_proof ■ August 23, 2018

THE JOURNAL OF PEDIATRICS • www.jpeds.com Volume ■■ • ■■ 2018

Early life stress has been found to contribute to metabolic and analyzed using enCore version 16.2 (GE Healthcare). VAT
dysregulation in cholesterol, insulin, glucose, triglycerides, and was estimated using CoreScan (GE Healthcare) as described
adipose tissue.12,13 Cardiac autonomic function, character- previously.27-29 Seated blood pressure in the right arm was mea-
ized by activation of the sympathetic nervous system, and mea- sured after the participant rested for 10 minutes with legs un-
sures of vascular health, such as arterial stiffness, may contribute crossed using an automated sphygmomanometer (BP-8800;
independently to the development of cardiovascular disease Colin Medical Instruments, San Antonio, Texas). Blood pres-
later in life.14-21 This study follows our previous work and in- sure percentiles were determined from age, sex, and height using
cludes predictors of metabolic and cardiovascular functioning. the Fourth Report on the Diagnosis, Evaluation, and Treat-
ment of High Blood Pressure in Children and Adolescents.30
Methods A blood lipid panel, fasting glucose, and fasting insulin were
measured (Fairview Diagnostic Laboratories, Fairview Uni-
One-hundred and twenty youths (60 females), aged 8-18 years, versity Medical Center, Minneapolis, Minnesota). Plasma
were included in this study. Data were obtained from 2 cohorts: glucose levels were measured by the bichromatic endpoint (Fair-
a cohort of postinstitutionalized youths and a control cohort view Diagnostic Laboratories). Insulin levels were measured
of nonadopted youths living in the Minneapolis-Saint Paul met- with a chemiluminescent immunoassay (ADVIA Centaur
ropolitan area. Both study protocols were approved by the Uni- System; Siemens Healthcare Diagnostics, Tarrytown, New York).
versity of Minnesota’s Institutional Review Board, and consent/ The homeostasis model assessment–estimated insulin resis-
assent was obtained from parents/participants. tance (HOMA-IR) score was calculated from a standard equa-
The postinstitutionalized cohort (n = 30; 19 females) con- tion from fasting plasma glucose and insulin levels.31
sisted of youths recruited from a registry of internationally Right radial, and carotid artery waveforms, as well as carotid-
adopting families. Participants were included if their adop- radial pulse wave velocity (PWV), were recorded by applana-
tion medical records indicated a height-for-age z-score ≤-1.5, tion tonometry using SphygmoCor MM3 version 8.0 software
defined as mild height stunting. 22 Mild height stunting (AtCor Medical, Sydney, Australia). Radial and carotid artery
served as an inclusion criterion both to increase the augmentation index (AIx), both corrected to a heart rate of
postinstitutionalized recruitment pool and to include a gra- 75 bpm, were derived from a validated integral transfer func-
dation of growth faltering that often continues to decline until tion applied by SphygmoCor MM3. PWV was measured by
around age 2 years.23,24 Exclusion criteria included current the sequential acquisition of pressure waveforms from the
steroid medication use, ongoing or chronic and infections, acute carotid and radial artery using the same tonometer. Carotid-
congenital and endocrine disorders, fetal alcohol syndrome, radial PWV was calculated from the transit time between the
and pregnancy. Exclusion criteria also included adoption after 2 arteries relative to the R-wave within the electrocardiogra-
age 3 years, to investigate early insults to growth. All phy (ECG) complex, using the foot-to-foot method and the
postinstitutionalized cohort testing was conducted in the Dela- intersecting tangent algorithm.32-34
ware Clinical Research Unit at the University of Minnesota, Autonomic nervous system activity was assessed by heart
Twin Cities. rate variability (HRV) using the same SphygmoCor MM3
In the control cohort (n = 90; 57 females), data were col- system. Using a 3-lead ECG in a modified lead II configura-
lected from youths who participated in a cross-sectional study tion, Heart rate was recorded continuously for 5 minutes. All
examining cardiovascular risk factors in youths ranging in ECG recordings were reviewed and analyzed for time and fre-
weight from normal to severe obesity. Participants were re- quency domains, with segments demonstrating arrhythmias
cruited from local medical clinics and advertisements. Par- excluded from the analysis. Time domain variables, such as the
ticipants were excluded if they were taking medications known mean R-R interval length, standard deviation of R-R inter-
to influence cardiovascular function or had known/diagnosed vals, number of adjacent intervals >50 ms, and percentage of
cardiovascular disease. This cohort has been described in more adjacent intervals >50 ms, were obtained from automated al-
detail elsewhere.25 Controls were identified from this sample gorithms provided by the SphygmoCor software.25,35 Spectral
of 300 youths by matching 3:1 with cases by age and weight analysis calculated frequency-domains, including low fre-
status. Testing in both cohorts was performed in the morning, quency (LF), high frequency (HF), LF:HF ratio, and total
with the participants fasted and also fully abstinent from power.25 Frequencies of 0.04-0.15 Hz and 0.15-0.40 Hz were
tobacco, alcohol, and caffeine, for a minimum of 8 hours before defined as LF and HF, respectively.25
their study visit. Self-report questionnaires determined parent education and
Measurements for height and body mass were obtained using race. Parent education was recorded as either 1, obtained a
a wall-mounted stadiometer and electronic scale, respec- college degree (associate’s degree or above), or 0, no college
tively. BMI was calculated as body mass in kilograms divided degree. Race was defined as 0, white, or 1, nonwhite.
by height in meters squared. BMI percentile was calculated using
Centers for Disease Control normality scales and stratified into Statistical Analyses
3 categories: normal weight (>5th to <85th percentile), over- Descriptive statistics were tabulated separately for
weight (≥85th to <95th percentile), and obese (≥95th postinstitutionalized and control subjects. The P values pre-
percentile).26 Body composition was measured by dual X-ray sented in Table I were based on t tests for continuous covariates
absorptiometry (iDXA; GE Healthcare, Madison, Wisconsin) and c2 tests for categorical covariates. Outliers were winsorized,
2 Reid et al

FLA 5.5.0 DTD ■ YMPD10083_proof ■ August 23, 2018

■■ 2018 ORIGINAL ARTICLES

Table I. Participant demographic, anthropometric, and Results

covariate characteristics by group status
Postinstitutionalized Control
The groups did not differ in age or sex distribution, al-
Measures (n = 30) (n = 90) P value though the proportion of white patients was higher in the
Demographic data control group compared with the postinstitutionalized group
Age, y, mean (SD) 13.5 (3.26) 13.0 (2.94) .476 (Table I). The height-for-age z-score showed that the
Age at adoption, mo, 18.6 (9.98) NA — postinstitutionalized youths were smaller for age than the com-
mean (SD)
Female sex, n (%) 19 (63.3) 41 (45.6) .092
parison group, though the averages for both groups were in
White race, n (%) 13 (43.3) 82 (91.1) < .001 the normal range. The postinstitutionalized youths were more
Anthropometric data likely to have lower overall lean mass and lower gynoid lean
Height-for-age z-score −2.2 (.48) — —
for at time of adoption,
mass (Table II). They were also were more likely to have a
mean (SD) higher percentage of trunk tissue fat, but did not differ from
Height-for-age z-score, −0.8 (0.93) 0.2 (0.93) <.001 controls in the amount of VAT. The postinstitutionalized youths
mean (SD)
BMI, kg/m2, mean (SD) 19.8 (4.5) 19.8 (4.8) .989 and control youths did not differ in any other measures of body
BMI percentile, mean (SD) 51.5 (31.5) 53.9 (26.8) .681 fat mass.
Weight category, n (%) .953 The postinstitutionalized youths exhibited higher systolic
Normal 25 (83.3) 75 (83.3)
Overweight 4 (13.3) 12 (13.3) blood pressure than control youths, though there was no dif-
Obese 1 (3.3) 3 (3.3) ference in either systolic blood pressure percentile or in either
Waist circumference, cm, 67.9 (12.2) 66.2 (12.5) .505 measure of diastolic blood pressure (Table III). 10 The
mean (SD)
Hip circumference, cm, 79.3 (15.3) 82.1 (13.8) .355
postinstitutionalized youths had a lower LF:HF compared with
mean (SD) controls. They also had higher carotid Aix values. Carotid AIx
and systolic blood pressure percentile were both indepen-
NA, not applicable.
BMI percentile reported as age- and sex-adjusted percentile (SD) based on Centers for Disease dently predicted by group membership (Table IV).
Control growth curves. Significant P values are in bold type. Postinstitutionalized status did not predict carotid-radial PWV.
Blood test results are displayed in Table IV. The
postinstitutionalized group had higher total cholesterol, low-
and skewed distributions were log-transformed for analyses. density lipoprotein cholesterol, triglycerides, and insulin levels.
Multiple imputation by chained equations was used with 50 There were no between-group differences in high-density li-
repetitions to impute missing data.36 Results between the poprotein and glucose levels. HOMA-IR was also higher in the
complete-case analysis and nonimputed dataset were similar; postinstitutionalized youths.
thus, imputed results are presented. Regression analyses were
conducted on all outcomes to assess differences between groups Discussion
and adjusted for sex, race, age, BMI percentile, height-for-
age, and parent education. Vascular, hemodynamic, and bio- These findings demonstrate a potential association between
chemical variables were also adjusted for trunk tissue fat early life adversity and later childhood development of
percentage. Covariates were selected a priori for their role as cardiometabolic risk factors and arterial stiffness. These risks
potential confounders or precision variables. All statistical analy- appear to be independent of BMI and trunk tissue fat in these
ses were performed with R version 3.4.3 (R Foundation for Sta- postinstitutionalized youths who were mildly to severely height-
tistical Computing, Vienna, Austria) and the “mice” library stunted at adoption. Postinstitutionalized youths, exposed to
version 2.46.0. psychosocial adversity and growth adversity before age 3 years,

Table II. Anthropometric dual X-ray absorptiometry characteristics by group status

Measures Postinstitutionalized (n = 30) Control (n = 90) Adjusted difference (95% CI) P value
Lean mass, g 30 657.6 (11 636.7) 34 074.0 (12180.3) −3039.6 (−5506.90 to −572.38) .016
Fat mass, g 13 330.5 (7624.9) 13 163.7 (8759.3) 482.914 (−1593.46 to 2559.29) .646
Total tissue fat, % 29.1 (8.2) 26.8 (7.4) 2.5 (−0.42 to 5.39) .093
Trunk lean mass, g 14 399.1 (5623.9) 15 637.4 (5831.8) −1200.9 (−2787.77 to 385.97) .136
Trunk fat mass, g 5918.4 (4561.5) 5307.4 (4782.0) 740.5 (−594.96 to 2075.90) .274
Trunk tissue fat, % 26.6 (10.6) 23.0 (9.1) 4.11 (0.74 to 7.49) .017
Android lean mass, g 2050.8 (833.1) 2258.9 (834.6) −183.1 (−366.67 to 0.51) .051
Android fat mass, g 818.9 (847.4) 734.3 (872.7) 137.3 (−118.24 to 392.85) .289
Gynoid lean mass, g 4586.8 (2092.1) 5075.6 (2189.8) −487.2 (−949.97 to −24.42) .039
Gynoid fat mass, g 2136.1 (1289.9) 2288.8 (1574.3) −66.5 (−465.73 to 332.69) .742
VAT mass, g 164.6 (208.2) 117.1 (151.4) 38.3 (−52.8 to 129.45) .406

Values presented are mean (SD) and mean difference (95% CI), adjusted for age, sex, BMI percentile, current height-for-age, parent education (college vs others), and race (white vs others).
Significant P values are in bold type.

Early Life Adversity with Height Stunting Is Associated with Cardiometabolic Risk in Adolescents Independent of 3
Body Mass Index
FLA 5.5.0 DTD ■ YMPD10083_proof ■ August 23, 2018
THE JOURNAL OF PEDIATRICS • www.jpeds.com Volume ■■

Table III. Autonomic nervous system activity and arterial stiffness by group status
Postinstitutionalized Control (n = 90), Adjusted difference
Measure (n = 30), mean (SD) mean (SD) (95% CI) P value
Autonomic nervous system activity
Heart rate, bpm 65.2 (7.71) 67.5 (10.12) −3.4 (−8.76 to 2.01) .216
SBP, mmHg 116.9 (10.5) 106.6 (10.46) 5.7 (1.02-10.4) .018
SBP percentile 68.6 (25.06) 47.9 (27.45) 13.6 (−1.95 to 29.25) .086
DBP, mmHg 60.7 (10.7) 58.7 (8.65) 1.4 (−5.60 to 8.36) .693
DBP percentile 35.0 (24.0) 41.0 (29.32) 3.3 (−17.41 to 24.08) .749
NN50 160.2 (60.77) 141.3 (65.82) 20.5 (−17.48 to 58.4) .287
PNN50 50.0 (19.76) 45.6 (22.83) 6.3 (−6.66 to 19.30) .336
Mean R-R interval, ms 933.1 (115.0) 912.9 (142.16) 42.1 (−30.88 to 115.15) .255
SD R-R interval, ms 86.1 (40.8) 100.7 (58.97) 1.0 (0.72-1.44) .921
LF:HF .6 (.41) 1.1 (1.4) −0.68 (−1.18 to −0.18) .008
Arterial stiffness
Carotid AIx 7.2 (10.97) −6.4 (16.92) 10.0 (0.81-19.21) .033
Radial AIx 4.6 (11.28) 1.2 (15.29) 0.4 (−9.31 to 10.12) .934
Carotid-radial PWV 7.1 (1.35) 6.5 (1.27) 0.4 (−0.35 to 1.11) .303

DBP, diastolic blood pressure; NN50, number of adjacent intervals over 50 ms; PNN50, percentage of adjacent intervals over 50 ms; SBP, systolic blood pressure; SD R-R, standard deviation of
R-R intervals.
Values presented are mean (SD) and mean difference (95% CI), adjusted for age, sex, BMI percentile, current height-for-age, trunk tissue fat (%), parent education (college vs others), and race
(white vs others).

demonstrated higher systolic blood pressure, evidence of ar- physiological mechanisms of catch-up growth promote the ac-
terial stiffening, and evidence of higher total cholesterol, low- cumulation of VAT after nutritional deprivation,5 this was not
density lipoprotein cholesterol, triglycerides, and insulin levels observed in these postinstitutionalized youths. Instead, the
and HOMA-IR scores. Importantly, these associations were postinstitutionalized youths were also more likely to have a
found to be independent of body composition as the major- higher percentage of trunk tissue fat, but not of VAT specifi-
ity of youths in the postinstitutionalized group were well within cally. Indeed, results on early growth stunting and later vis-
healthy norms for body composition and, because the com- ceral adiposity have been mixed.37 An obesogenic environment
parison group was selected to be consistent with the may be the critical component in the association of early life
postinstitutionalized group, this was true of the comparison catch-up growth with later obesity and/or increased VAT.38
group as well. In addition, the postinstitutionalized youths dem- Arterial stiffness is indicated by both AIx and PWV. Aug-
onstrated a less sympathetically driven cardiovascular system. mentation index measures the relative magnitude of the re-
These results suggest that children with a history of early life flected (ie, retrograde) pulse wave early in the cardiac cycle,
stress and malnutrition might be at higher risk for with greater values indicating increased arterial stiffening.14 Pulse
cardiometabolic health problems even compared with youths wave transit time increases within stiffer arteries and is mani-
with a similar BMI. fested by higher PWV values.14 The postinstitutionalized youths
Although previous work by our group had shown normal had higher carotid Aix values after controlling for parent edu-
BMI in postinstitutionalized youths,8 we questioned whether cation, race, age, sex, height for age, trunk tissue fat, and BMI
more rigorous tests might reveal greater accumulation of fat percentile. In contrast, there were no differences in PWV
in the viscera. Nonetheless, although evidence suggests that the between the postinstitutionalized and control groups.

Table IV. Participant blood lipid, glucose, insulin, and HOMA-IR results by group status
Postinstitutionalized Control (n = 90), Adjusted difference
Measure (n = 30), mean (SD) mean (SD) (95% CI) P value
Total cholesterol, mg/dL 169.8 (29.3) 155.4 (27.9) 15.2 (0.19-30.11) .047
Total cholesterol, mmol/L 4.4 (0.8) 4.0 (0.7)
LDL, mg/dL 92.0 (25.4) 82.3 (24.0) 14.4 (1.40-27.45) .03
LDL, mmol/L 2.4 (0.7) 2.1 (0.6)
HDL, mg/dL 59.7 (16.6) 57.3 (15.4) −1.4 (−9.66 to -6.90) .742
HDL, mmol/L 1.6 (0.4) 1.5 (0.4)
Triglycerides, mg/dL 88.9 (39.3) 79.4 (34.7) 1.2 (1.01-1.54) .048
Triglycerides, mmol/L 1 (0.4) 0.9 (0.4)
Glucose, mg/dL 79.0 (8.7) 77.2 (8.7) −0.7 (−5.76 to 4.36) .785
Glucose, mmol/L 4.4 (0.5) 4.3 (0.5)
Insulin, mU/L 7.8 (4.3) 5.0 (3.9) 1.6 (0.06-0.34) .005
Insulin, mmol/L 54.2 (29.9) 34.7 (27.1)
HOMA-IR 1.5 (0.9) 1.0 (0.8) 1.5 (1.10-2.12) .01

HDL, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol.

Values presented are mean (SD) and mean difference (95% CI), adjusted for age, sex, race (white vs others), BMI percentile, current height-for-age, parent education (college education or above),
and trunk tissue fat (%).

4 Reid et al

FLA 5.5.0 DTD ■ YMPD10083_proof ■ August 23, 2018

■■ 2018 ORIGINAL ARTICLES

Arterial stiffness is an independent risk factor for cardiovas- comment on how gestational age and birth weight play a
cular disease in adults16 and has been shown to be associated role in metabolic or cardiovascular differences between the
with cardiovascular risk factors and excess adiposity in groups. Catch-up growth from preterm or low birth weight
children.39-41 might influence the results and represent a risk to the
The HRV results were mixed. The postinstitutionalized postinstitutionalized youths distinct from their postnatal
youths exhibited a lower LF:HF than control youths, yet dif- adversity. Low birth weight is overrepresented in
fered in no other measures of HRV. The LF:HF is considered postinstitutionalized youths compared with other groups,
a measure of the autonomic nervous system’s overall and thus the effects seen in this study might come not from
sympathovagal balance, with lower values indicating a less sym- early life stress and malnutrition, but rather from low birth
pathetically driven cardiovascular system.42,43 This less sym- weight.52 In a similar vein, we are unable to make claims
pathetically driven system is consistent with a Romanian study about the adversity that the control cohort may or may not
of postinstitutionalized youths who showed lower sympa- have experienced early in life, in either a psychosocial or a
thetic nervous system reactivity to a psychosocial challenge com- physical domain. We do not have data on the height-for-age
pared with children reared in their families.44 In both this study of the control cohort at an age comparable with the age of
and the study of postinstitutionalized youths in Romania, adoption in the postinstitutionalized cohort, and thus we are
youths with a history of early life adversity were prone to ex- unable to comment on the linear growth of the control cohort
periencing a “down-regulated” sympathetic nervous system early in life.
response.45,46 The postinstitutionalized mean LF:HF of 0.6 is A third limitation is the absence of the measurement of
also lower than the expected range of 1.1-11.6 reported in other health behaviors in this study. We are unable to make claims
studies with adults.47 Because sympathetic tone is associated about the participants’ current health behaviors. Diet may play
with increasing age, postinstitutionalized youths might a large role in mediating the relationship between early life ad-
experience increased LF:HF as they age. 47 Alternatively, versity, growth restriction, and later risks to cardiovascular and
LF:HF may be an early indicator of less optimal HRV for metabolic health. It is not yet clear how these youths’ diets may
postinstitutionalized youths later in life, despite other mea- impact their current or future health. As continued research
sures of HRV that do not exhibit increased risk. suggests that psychosocial stress may induce emotional
Postinstitutionalized children exhibited higher systolic, but overeating,53-56 emotional eating in individuals with a history
not diastolic, blood pressure. Previous studies of the associa- of adversity could lay the groundwork for later obesity and even
tion between adverse childhood experiences and blood pres- higher risks of cardiometabolic disease.57,58 Thus, future studies
sure have shown mixed results. There have been few studies would benefit from rigorously measuring and assessing diet
of adolescents with a history of adverse childhood experi- as a covariate to account for differences that may influence the
ences that report blood pressure metrics, and the results of these metabolic outcomes.
studies have shown either no relationship between adverse early Finally, many outcomes were studied in this pilot study, and
experiences and blood pressure or a relationship between early thus, with multiple testing, there could be increased risk for
adversity and diastolic, but not systolic, blood pressure.48,49 chance findings. Results should be interpreted cautiously. Future
Increased plasma cholesterol has been identified as playing studies of cardiometabolic risk in postinstitutionalized youths
a key role in arterial impairment and a higher augmentation would benefit from additional statistical power.
index in a pediatric population. 50 The results from this This study provides preliminary evidence that youths exposed
postinstitutionalized cohort suggest that the combination of to early life stress and growth stunting may be at greater risk
higher plasma cholesterol levels and increased carotid aug- of developing heart disease and metabolic disorders indepen-
mentation index could put children with early adversity at risk dent of BMI. Postinstitutionalized youths with a history of
for cardiometabolic dysfunction without the presence of obesity. growth stunting may require close monitoring for emergent
Postinstitutionalized youths in this study exhibited higher cardiometabolic risk. ■
insulin levels than controls. The higher HOMA-IR scores seen
in the postinstitutionalized group, although consistent with We thank the families for their participation in the International Adop-
the group’s higher levels of insulin, exist without a signifi- tion Project, and thank Bonny Donzella, Bao Moua, Cameron Naughton,
cantly higher glucose level. Although the mean HOMA-IR and Heather Taylor for their assistance with the study. We also thank
the Center for Neurobehavioral Development and Clinical Transla-
scores are not above the HOMA-IR cutoff scores for insulin tional Science Institute at the University of Minnesota.
resistance for either prepubertal or pubertal youths, this in-
crease without the presence of obesity could signal Submitted for publication Mar 12, 2018; last revision received May 22, 2018;
cardiometabolic risk for youths with a history of early life accepted Jun 14, 2018
adversity and height stunting.51 Reprint requests: Brie M. Reid, MA, Institute of Child Development, University
of .Minnesota, Minneapolis, MN 55455. E-mail: reidx189@umn.edu.
This study has a number of important limitations. First,
the study consists of a small, cross-sectional sample, and so
we are unable to make causal claims. Second, we do not have References
information on gestational age or birth weight for either 1. Jacobs E, Miller LC, Tirella LG. Developmental and behavioral perfor-
cohort. This information is not available at all for most mance of internationally adopted preschoolers: a pilot study. Child Psy-
postinstitutionalized youths, and thus we are unable to chiatry Hum Dev 2010;41:15-29.

Early Life Adversity with Height Stunting Is Associated with Cardiometabolic Risk in Adolescents Independent of 5
Body Mass Index
FLA 5.5.0 DTD ■ YMPD10083_proof ■ August 23, 2018
THE JOURNAL OF PEDIATRICS • www.jpeds.com Volume ■■

2. U.S. Department of State. Immigrant Visas Issued to Orphans Coming underweight, and progress towards MDG 1 in 141 developing coun-
into the U.S. https://travel.state.gov/content/travel/en/Intercountry tries: a systematic analysis of population representative data. Lancet
-Adoption/adopt_ref/adoption-statistics.html. Accessed February 8, 2012;380:824-34.
2018. 23. de Onis M, Branca F. Childhood stunting: a global perspective. Matern
3. Hellerstedt WL, Madsen NJ, Gunnar MR, Grotevant HD, Lee RM, Johnson Child Nutr 2016;12(Suppl 1):12-26.
DE. The International Adoption Project: population-based surveillance 24. Victora CG, de Onis M, Hallal PC, Blössner M, Shrimpton R. World-
of Minnesota parents who adopted children internationally. Matern Child wide timing of growth faltering: revisiting implications for interven-
Health J 2008;12:162-71. tions. Pediatrics 2010;125:e473-80.
4. Johnson DE, Gunnar MR IV. Growth failure in institutionalized chil- 25. Ryder JR, O’Connell M, Bosch TA, Chow L, Rudser KD, Dengel DR, et al.
dren. Monogr Soc Res Child Dev 2011;76:92-126. Impaired cardiac autonomic nervous system function is associated with
5. Sawaya AL, Roberts S. Stunting and future risk of obesity: principal physi- pediatric hypertension independent of adiposity. Pediatric Res 2016;79:49-
ological mechanisms. Cad Saude Publica 2003;19(Suppl 1):S21-8. 54.
6. Monteiro PO, Victora CG. Rapid growth in infancy and childhood and 26. Kumar S, Kelly AS. Review of childhood obesity: from epidemiology, eti-
obesity in later life–a systematic review. Obes Rev 2005;6:143-54. ology, and comorbidities to clinical assessment and treatment. Mayo Clin
7. Fernandez-Twinn DS, Ozanne SE. Mechanisms by which poor early growth Proc 2017;92:251-65.
programs type-2 diabetes, obesity and the metabolic syndrome. Physiol 27. Bosch TA, Steinberger J, Sinaiko AR, Moran A, Jacobs DR Jr, Kelly AS,
Behav 2006;88:234-43. et al. Identification of sex-specific thresholds for accumulation of vis-
8. Reid BM, Miller BS, Dorn LD, Desjardins C, Donzella B, Gunnar M. Early ceral adipose tissue in adults. Obesity (Silver Spring) 2015;23:375-
growth faltering in post-institutionalized youth and later anthropomet- 82.
ric and pubertal development. Pediatr Res 2017;82:278-84. 28. Bosch TA, Burruss TP, Weir NL, Fielding KA, Engel BE, Weston TD, et al.
9. Hoffman DJ, Martins PA, Roberts SB, Sawaya AL. Body fat distribution Abdominal body composition differences in NFL football players. J Strength
in stunted compared with normal-height children from the shanty- Cond Res 2014;28:3313-9.
towns of São Paulo, Brazil. Nutrition 2007;23:640-6. 29. Kaul S, Rothney MP, Peters DM, Wacker WK, Davis CE, Shapiro MD,
10. National Heart, Lung, and Blood Institute. Expert panel on integrated et al. Dual-energy X-ray absorptiometry for quantification of visceral fat.
guidelines for cardiovascular health and risk reduction in children and Obesity (Silver Spring) 2012;20:1313-8.
adolescents: summary report. Pediatrics 2011;128(Suppl 5):S213- 30. National High Blood Pressure Education Program Working Group on
56. High Blood Pressure in Children and Adolescents. The fourth report on
11. Demerath EW, Schubert CM, Maynard LM, Sun SS, Chumlea WC, Pickoff the diagnosis, evaluation, and treatment of high blood pressure in
A, etc. Do changes in body mass index percentile reflect changes in body children and adolescents. Pediatrics 2004;114(2 Suppl 4th Report):555-
composition in children? Data from the Fels Longitudinal Study. Pedi- 76.
atrics 2006;117:e487-95. 31. Levy J, Morris R, Hammersley M, Turner R. Discrimination, adjusted cor-
12. Lehman BJ, Taylor SE, Kiefe CI, Seeman TE. Relation of childhood so- relation, and equivalence of imprecise tests: application to glucose tol-
cioeconomic status and family environment to adult metabolic function- erance. Am J Physiol 1999;276(2 Pt 1):E365-75.
ing in the CARDIA study. Psychosom Med 2005;67:846-54. 32. Kelly RP, Millasseau SC, Ritter JM, Chowienczyk PJ. Vasoactive drugs in-
13. Taylor SE. Mechanisms linking early life stress to adult health outcomes. fluence aortic augmentation index independently of pulse-wave veloc-
Proc Natl Acad Sci U S A 2010;107:8507-12. ity in healthy men. Hypertension 2001;37:1429-33.
14. Kelly AS, Rudser KD, Dengel DR, Kaufman CL, Reiff MI, Norris AL, et al. 33. Millasseau SC, Stewart AD, Patel SJ, Redwood SR, Chowienczyk PJ. Evalu-
Cardiac autonomic dysfunction and arterial stiffness among children and ation of carotid-femoral pulse wave velocity: influence of timing algo-
adolescents with attention deficit hyperactivity disorder treated with stimu- rithm and heart rate. Hypertension 2005;45:222-6.
lants. J Pediatr 2014;165:755-9. 34. Rajzer MW, Wojciechowska W, Klocek M, Palka I, Brzozowska-Kiszka M,
15. Sutton-Tyrrell K, Najjar SS, Boudreau RM, Venkitachalam L, Kupelian V, Kawecka-Jaszcz K. Comparison of aortic pulse wave velocity measured
Simonsick EM, et al. Elevated aortic pulse wave velocity, a marker of ar- by three techniques: Complior, SphygmoCor and Arteriograph. J Hypertens
terial stiffness, predicts cardiovascular events in well-functioning older 2008;26:2001-7.
adults. Circulation 2005;111:3384-90. 35. Farah BQ, Barros MV, Balagopal B, Ritti-Dias RM. Heart rate variabil-
16. Weber T, Auer J, O’Rourke MF, Kvas E, Lassnig E, Berent R, et al. Arte- ity and cardiovascular risk factors in adolescent boys. J Pediatr
rial stiffness, wave reflections, and the risk of coronary artery disease. Cir- 2014;165:945-50.
culation 2004;109:184-9. 36. White IR, Royston P, Wood AM. Multiple imputation using chained
17. Schindler TH, Hornig B, Buser PT, Olschewski M, Magosaki N, Pfisterer equations: issues and guidance for practice. Stat Med 2011;30:377-
M, et al. Prognostic value of abnormal vasoreactivity of epicardial coro- 99.
nary arteries to sympathetic stimulation in patients with normal 37. Walker SP, Chang SM, Powell CA. The association between early child-
coronary angiograms. Arterioscler Thromb Vasc Biol 2003;23:495- hood stunting and weight status in late adolescence. Int J Obes (Lond)
501. 2007;31:347-52.
18. Halcox JP, Donald AE, Ellins E, Witte DR, Shipley MJ, Brunner EJ, et al. 38. Skidmore PM, Cassidy A, Swaminathan R, Richards JB, Mangino M,
Endothelial function predicts progression of carotid intima-media thick- Spector TD, et al. An obesogenic postnatal environment is more impor-
ness. Circulation 2009;119:1005-12. tant than the fetal environment for the development of adult adiposity:
19. Halcox JP, Schenke WH, Zalos G, Mincemoyer R, Prasad A, Waclawiw a study of female twins. Am J Clin Nutr 2009;90:401-6.
MA, et al. Prognostic value of coronary vascular endothelial dysfunc- 39. Urbina EM, Kimball TR, Khoury PR, Daniels SR, Dolan LM. Increased
tion. Circulation 2002;106:653-8. arterial stiffness is found in adolescents with obesity or obesity-related
20. Rubinshtein R, Kuvin JT, Soffler M, Lennon RJ, Lavi S, Nelson RE, et al. type 2 diabetes mellitus. J Hypertens 2010;28:1692-8.
Assessment of endothelial function by non-invasive peripheral arterial to- 40. Urbina EM, Gao Z, Khoury PR, Martin LJ, Dolan LM. Insulin resistance
nometry predicts late cardiovascular adverse events. Eur Heart J and arterial stiffness in healthy adolescents and young adults. Diabetologia
2010;31:1142-8. 2012;55:625-31.
21. Yoshida T, Kawano H, Miyamoto S, Motoyama T, Fukushima H, Hirai 41. Kaufman CL, Kaiser DR, Steinberger J, Dengel DR. Relationships between
N, et al. Prognostic value of flow-mediated dilation of the brachial artery heart rate variability, vascular function, and adiposity in children. Clin
in patients with cardiovascular disease. Intern Med 2006;45:575-9. Auton Res 2007;17:165-71.
22. Stevens GA, Finucane MM, Paciorek CJ, Flaxman SR, White RA, 42. Heart rate variability. Standards of measurement, physiological interpre-
Donner AJ, et al. Trends in mild, moderate, and severe stunting and tation, and clinical use. Task Force of the European Society of

6 Reid et al

FLA 5.5.0 DTD ■ YMPD10083_proof ■ August 23, 2018

■■ 2018 ORIGINAL ARTICLES

Cardiology and the North American Society of Pacing and Electrophysi- 51. Kurtoğlu S, Hatipoğlu N, Mazıcıoğlu M, Kendirici M, Keskin M, Kondolot
ology. Eur Heart J 1996;17:354-81. M. Insulin resistance in obese children and adolescents: HOMA-IR cut-
43. Malliani A, Pagani M, Furlan R, Guzzetti S, Lucini D, Montano N, off levels in the prepubertal and pubertal periods. J Clin Res Pediatr
et al. Individual recognition by heart rate variability of two different Endocrinol 2010;2:100-6.
autonomic profiles related to posture. Circulation 1997;96:4143- 52. Johnson DE. Medical and developmental sequelae of early childhood in-
5. stitutionalization in Eastern European adoptees. In: Nelson CA, ed. The
44. McLaughlin KA, Sheridan MA, Tibu F, Fox NA, Zeanah CH, Nelson CA Minnesota Symposia on Child Psychology. Vol 31: The effects of early ad-
3rd. Causal effects of the early caregiving environment on development versity on neurobehavioral development. Mahwah (NJ): Lawrence Erlbaum;
of stress response systems in children. Proc Natl Acad Sci USA 2001, p. 113-62.
2015;112:5637-42. 53. Arce M, Michopoulos V, Shepard KN, Ha QC, Wilson ME. Diet choice,
45. McEwen BS, Seeman T. Protective and damaging effects of mediators of cortisol reactivity, and emotional feeding in socially housed rhesus monkeys.
stress. Elaborating and testing the concepts of allostasis and allostatic load. Physiol Behav 2010;101:446-55.
Ann N Y Acad Sci 1999;896:30-47. 54. Michopoulos V, Toufexis D, Wilson ME. Social stress interacts
46. Danese A, McEwen BS. Adverse childhood experiences, allostasis, allostatic with diet history to promote emotional feeding in females.
load, and age-related disease. Physiol Behav 2012;106:29-39. Psychoneuroendocrinology 2012;37:1479-90.
47. Abhishekh HA, Nisarga P, Kisan R, Meghana A, Chandran S, Raju TR, 55. Machado TD, Dalle Molle R, Reis RS, Rodrigues DM, Mucellini AB,
et al. Influence of age and gender on autonomic regulation of heart. J Clin Minuzzi L, et al. Interaction between perceived maternal care, anxiety symp-
Monit Comput 2013;27:259-64. toms, and the neurobehavioral response to palatable foods in adoles-
48. Pretty C, O’Leary DD, Cairney J, Wade TJ. Adverse childhood experi- cents. Stress 2016;19:287-94.
ences and the cardiovascular health of children: a cross-sectional study. 56. Machado TD, Dalle Molle R, Laureano DP, Portella AK, Werlang IC, Benetti
BMC Pediatr 2013;13:208. Cda S, et al. Early life stress is associated with anxiety, increased stress
49. Gooding HC, Milliren CE, Austin SB, Sheridan MA, McLaughlin KA. Child responsivity and preference for “comfort foods” in adult female rats. Stress
abuse, resting blood pressure, and blood pressure reactivity to psycho- 2013;16:549-56.
social stress. J Pediatr Psychol 2016;41:5-14. 57. Dallman MF, Pecoraro NC, la Fleur SE. Chronic stress and comfort foods:
50. Riggio S, Mandraffino G, Sardo MA, Iudicello R, Camarda N, Imbalzano self-medication and abdominal obesity. Brain Behav Immun 2005;19:275-
E, et al. Pulse wave velocity and augmentation index, but not intima- 80.
media thickness, are early indicators of vascular damage in hypercholes- 58. Dallman MF. Stress-induced obesity and the emotional nervous system.
terolemic children. Eur J Clin Invest 2010;40:250-7. Trends Endocrinol Metab 2010;21:159-65.

Early Life Adversity with Height Stunting Is Associated with Cardiometabolic Risk in Adolescents Independent of 7
Body Mass Index
FLA 5.5.0 DTD ■ YMPD10083_proof ■ August 23, 2018