Antenatal Steroid Exposure, Aerobic Fitness, and Physical Activity in Adolescents Born Preterm With Very Low Birth Weight

Abstract

Objective:

To determine if antenatal corticosteroid exposure is associated with aerobic fitness or physical activity participation in adolescents born preterm with very low birth weight (VLBW).

Study design:

Observational cohort study of 173 14-year-olds born with VLBW between 1992 and 1996 at a regional perinatal center with 91 exposed to ANCS. Aerobic fitness was determined from peak oxygen uptake ($\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$) obtained via maximal exercise testing on a cycle ergometer. PA levels for the past year and past two months were estimated from questionnaire. Between-group) comparisons for continuous variables were evaluated using independent t-tests or Mann-Whitney U tests. Generalized linear models were used to compare differences in fitness and PA between ANCS+ and ANCS− groups, with race and sex in models.

Results:

Regression analysis revealed an ANCS x sex x race interaction for $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ (p ≤ .001). Nonblack ANCS+ males had significantly higher $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ than nonblack ANCS− males expressed relative to body mass [mean difference (95%CI); 8.5 (2.1, 15.0) mL•kg⁻¹•min⁻¹] and lean body mass [LBM; 9.0 (1.1, 16.9) mL•kg_LBM⁻¹•min⁻¹]. No ANCS group differences in $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ were evident in black males, or black and nonblack females. ANCS+ males reported participating in significantly more total PA (medians: 14.6 vs. 8.5) and vigorous PA (3.0 vs. 0.95) per week for the past two months than ANCS− males.

Conclusions:

Exposure to ANCS was associated with higher PA participation and aerobic fitness in VLBW adolescents, particularly in nonblack males, which may confer health benefits in this at risk population.

Approximately 1.4% of infants are born in the US with very low birth weight (VLBW; <1500 g) and up to 87% of them are exposed to antenatal corticosteroid therapy to promote fetal lung maturation and survival when preterm delivery appears imminent.(1–3) Although beneficial in the short-term, the long-term effects of ANCS on health and functional outcomes are less clear. Some evidence suggests that ANCS exposure is associated with an increased risk for cardiometabolic disease.(4,5) We have previously shown that ANCS is associated with greater prevalence of larger airway obstruction in adolescence, and preclinical studies suggest that ANCS may disrupt normal skeletal and cardiac muscle development.(6–9) These systems play a major role in aerobic fitness and ability to participate in physical activity (PA), which in turn are associated with risk for cardiometabolic disease.(10) Several studies report that persons born preterm and/or with VLBW have lower levels of fitness and PA participation.(11–15) With nearly 48,000 infants born each year in the US exposed to ANCS (ANCS+), it is important to determine the long-term effects of ANCS exposure on these functional outcomes.(2)

In this study, we examined aerobic fitness and habitual PA in a cohort of 14-year-old individuals born with VLBW, of whom approximately one-half were exposed to ANCS. We hypothesized that aerobic fitness and PA levels would be lower in ANCS+ adolescents compared with their unexposed peers (ANCS−). Based on our previous work showing sex- and race-specific effects of ANCS, we examined the interactions of ANCS with sex and race in relation to fitness and PA.(16–18)

METHODS

The study was approved by the Institutional Review Boards of Wake Forest Baptist and Forsyth Medical Centers (Protocol No. BG05-047). Written informed consent was from a parent/legal guardian and assent was from the adolescent.

Participants were recruited from a cohort of 479 VLBW infants born between 1992 and 1996 at a regional perinatal center (Forsyth Medical Center) meeting the inclusion criteria of 14 years of age, singleton birth with no major congenital anomaly, clinical evaluation at one-year adjusted age, and physical ability to undergo testing. Participants were excluded if the mother took oral corticosteroids during pregnancy. The study protocol included three study visits.

Aerobic Fitness

Aerobic fitness was assessed at the 2^nd study visit and was determined from expired gases collected during progressive exercise testing on a cycle ergometer (CPX Metabolic Cart and Corival cycle ergometer, Medical Graphics, St. Paul, MN) following the Godfrey Protocol (work increments of 10, 15, or 20 W each minute based on child’s height of <125, 125-150, or >150 cm, respectively).(19) The participant was verbally encouraged to give a maximal effort, defined as peak heart rate > 195 bpm, peak respiratory exchange ratio (RER) > 1.05, and/or two experienced testers agreeing that a maximal effort was given. The test was terminated if significant abnormalities were detected on the electrocardiogram (eg, ventricular arrhythmia) or if systolic blood pressure exceeded 220 mmHG or diastolic blood pressure exceeded 100 mmHg. Participants were asked to refrain from eating, drinking caffeinated beverages, or exercising the morning of the study visit. Aerobic fitness was determined from the highest 20-second average of oxygen uptake attained ($\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$) and expressed relative to body weight (mL•kg⁻¹•min⁻¹), lean body mass (LBM; mL•kg_LBM⁻¹•min⁻¹), and as a percent of predicted age- and sex-specific reference values.(20) Fitness was considered below normal if $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ was less than 80% of predicted. Lean body mass was determined from a total body scan using dual energy x-ray absorptiometry (Hologic Delphi with pediatric software).

Habitual Physical Activity

Physical activity was assessed using the Kriska Modifiable Activity Questionnaire (MAQ) for which validity and reliability have been demonstrated in other pediatric populations.(21,22) The MAQ was administered to the participant by a trained interviewer with a parent present to assist if necessary. The participant was read a list of common leisure activities and asked to indicate those in which he/she had engaged at least five times in the past year. The participant could also add leisure activities not included on the list. For indicated activities, the participant was then asked to indicate the specific months in the past year, the average number of times per month or week, and the average number of minutes per session. The total minutes of activity were summed and expressed as an average total hours of activity (TOT-hrs) per week for the past year. Activities were assigned estimated metabolic equivalents (METs), and those with an estimated intensity >6 METs were summed and expressed as average hours per week spent in vigorous PA (VIG-hrs) for the past year.(23) In view of seasonal variation in PA and its potential effect on aerobic fitness, we also calculated average TOT-hrs and VIG-hrs per week for the previous two months.(24) We further examined the proportions meeting the national recommendations for PA.(25) Physical activity reports were considered unreliable by experienced interviewers when the reporter did not appear to be fully engaged.

Neonatal Characteristics

Neonatal characteristics including birth weight, gestational age, and ANCS exposure (including date and dose of ANCS, when available) were obtained from a research database as well as maternal and participant medical records reviewed by a research nurse. Gestational age was determined in order of availability from either first trimester ultrasound, maternal report of last menstrual period, or clinical assessment of the newborn infant. Chronic lung disease was defined as supplemental oxygen requirement at 36 weeks post-menstrual age.(26) Birth weight z-values were determined from gestational age- and sex-specific reference data,(27) and a z-value < 2 SD was considered small-for-gestational age. Additional information regarding neonatal morbidities were derived from diagnoses recorded by the attending neonatologist at the time when the infant was discharged from neonatal intensive care.

Adolescent Characteristics

Height and weight were measured in light clothing without shoes in triplicate using a wall-mounted stadiometer and digital platform scale, respectively. Body mass index (BMI) was calculated from the ratio of weight in kg to height in cm², and age- and sex-specific percentiles were determined.(28) A BMI ≥ 85^th percentile was considered overweight or obese. The participant’s race was categorized as black or nonblack based on the report by parent/guardian via questionnaire. Puberty stage was self-reported by the participant in private using the Tanner sexual maturation scale comprised of drawings of the five progressive stages of secondary sexual characteristics (breast development in females, genital development in males, and pubic hair for both sexes).(29)

Statistical analyses

Data were analyzed using SPSS version 25. Measures of central tendency and dispersion were examined, and transformations were applied to variables with Shapiro-Wilk test of normality results having a p <0.05 in attempt to improve distributional characteristics prior to analyses. Data are presented as means ± SD, median (25^th, 75^th percentiles), or n (%). Between-group (ANCS+ v. ANCS−) differences were examined using independent samples t-tests for continuous variables, or Mann-Whitney U tests, when transformation (e.g. natural log, square root) did not improve a variable’s distributional characteristics. Proportional differences in categorical variables were examined via Pearson χ² analysis. Generalized linear models were used to compare differences in fitness and PA between ANCS+ and ANCS− groups, and based on our previous work, we decided a priori to include sex, race, and their interaction product terms with ANCS in the model. For interaction terms with P ≤ .10, stratified regression analyses were run. Logistic regression analysis was used for dichotomous outcome variables. Post hoc pairwise comparisons of estimated marginal means (95%CI) were made with Bonferroni correction.

RESULTS

As shown in Figure 1 (available at www.jpeds.com), 193 of 479 eligible VLBW adolescents were enrolled in the study. This sample size met our recruitment goal for the overall study. After enrollment, six participants were found to be ineligible, and one was unable to participate because of severe cerebral palsy. After further exclusions for missing data or unreliable reporting of PA, the final analysis for participant characteristics and PA included 91 ANCS+ (43 male, 64 nonblack) and 82 ANCS− (33 male, 37 nonblack) adolescents. Final analysis of aerobic fitness included 85 ANCS+ (40 male, 61 nonblack) and 75 ANCS− (28 male, 31 nonblack) who gave a maximal effort and had valid exercise test data. Two adolescents (both ANCS−) were not able to do the exercise test due to neuromotor limitations associated with cerebral palsy. One adolescent (ANCS−) was unable to perform the test due to cognitive impairment. Three adolescents (2 ANCS−) had evidence of significant airway obstruction and were not tested. In addition, the exercise test was terminated prematurely in 3 adolescents (all ANCS−) including 2 who developed ventricular arrhythmia and one whose systolic blood pressure exceeded 220 mm Hg.

Figure 1.

Online. Eligibility, enrollment, exclusions, and inclusion for final analysis at adolescence. (VLBW: very low birth weight; ANCS: antenatal corticosteroid; V2: visit 2; CP: cerebral palsy; CV: cardiovascular)

Participants’ characteristics as neonates and adolescents are presented in Table 1. Additional information regarding prevalence of neonatal morbidities can be found in Table 2 (available at www.jpeds.com). In the ANCS+ group, 89 of the 91 mothers received betamethasone, one received dexamethasone (black male offspring), and one received both betamethasone and dexamethasone (nonblack female offspring). More detailed information about the number of doses and the time of dose relative to birth was available in 60 (26 male, 13 black) of 91 ANCS−exposed women. The majority of women (75%) received 2 doses of betamethasone (one full course), and the majority of births (62%) occurred at least 24 hours after, but within 7 days, of administration of treatment. No differences in neonatal or current characteristics were found between offspring of ANCS−treated mothers with and without detailed data on exposure. Nonblack participants were more likely to have been exposed to ANCS than black participants. Gestational age, birth weight, and birth weight z-value did not differ between the two ANCS groups. Only five neonates (1 ANCS+) had birth weight z-values < 2 SD. Twenty-nine adolescents (15 ANCS+) participated postnatally in a randomized controlled trial of dexamethasone to decrease dependency on mechanical ventilation, with 7 ANCS+ and 8 ANCS− receiving dexamethasone. Demographic and neonatal characteristics were similar among eligible VLBW survivors who did or did not participate (data not shown).

Table 1.

Participants’ characteristics as neonates and at 14 years of age by ANCS exposure. Values are medians (25^th, 75^th percentiles) or n (%).

	ANCS+ n=91	ANCS− n=82
Neonatal
Male, n (%)	43 (47)	33 (40)
Non-black, n (%)	64 (70)a	37 (45)
Gestational age, wk	28 (26.0, 30.0)	27 (25.8, 30.0)
Birth weight, g	1010 (785, 1250)	1058 (899, 1308)
Birth weight Z-value	−0.240 (−0.659, 0.266)	−0.026 (−0.722, 0.482)
Birth weight Z-values < 2 SD	1 (1)	4 (5)
Chronic Lung Disease, n (%)	23 (25)	19 (23)
14 years of age
Age, yr	14.5 (14.3, 14.8)	14.5 (14.3, 14.8)
Weight, kg	55.1 (48.2, 68.4)	56.4 (45.4, 70.4)
Weight Z-value	0.306 (−0.405, 1.266)	0.540 (−0.788, 1.429)
Lean body mass, kgb	42.5 (36.1, 51.8)	40.6 (36.5, 49.7)
Height, cm	162.8c (156.2, 168.7)	158.9 (151.8, 167.0)
Height Z-value	−0.146d (−0.845, 1.266)	−0.461 (−1.514, 0.387)
BMI Z-value	0.268 (−0.635, 1.403)	0.777 (−0.247, 1.569)
BMI ≥ 85th percentile, n (%)	27 (30)	31 (38)
Puberty Stage n(%)e
Breast/Genital Development
Stage 1 (pre-pubertal)	1 (1)	0 (0)
Stage 2 (early pubertal)	0 (0)	1 (1)
Stage 3 (mid-pubertal)	14 (16)	11 (13)
Stage 4 (late pubertal)	48 (53)	33 (40)
Stage 5 (post-pubertal)	27 (30)	37 (45)
Pubic Hair
Stage 1 (pre-pubertal)	1 (1)	0 (0)
Stage 2 (early pubertal)	0 (0)	0 (0)
Stage 3 (mid-pubertal)	3 (3)	7 (9)
Stage 4 (late pubertal)	39 (43)	26 (32)
Stage 5 (post-pubertal)	47 (52)	49 (60)

^ap<.05 from χ² test.

^bn = 85 ANCS+, n= 69 ANCS−

^cp<.05 from independent t-tests

^dp<.05 from Mann-Whitney U tests for variables with non-normal distributions

^en = 90, 1 ANCS+ male refused.

Table 2.

Prevalence of neonatal morbidities

	ANCS+ n=91	ANCS− n=82
Necrotizing Enterocolitis, n (%)	10 (11)	9 (11)
Intraventricular Hemorrhage (Grade 3 or 4), n (%)	1 (1)	2 (1)
Retinopathy of Prematurity (Grade 3, 4, 5 or surgery), n (%)	13 (14)	16 (20)
Sepsis, n (%)	1 (1)*	8 (10)
Patent Ductus Arteriosus
Treated with Indocin, n (%)	3 (3)	8 (10)
Treated with surgery, n (%)	42 (59)	34 (47)

^*p<0.05 from χ² test.

At follow-up, ANCS+ adolescents were taller than their unexposed peers. When stratified by sex, height of ANCS+ males was 4.84 cm (mean difference; 95% CI:0.63, 9.06; p=0.03) higher than ANCS− males. ANCS+ females were 1.51 cm taller (95% CI:−1.29, 4.30), but the difference was not statistically significant (p=0.29). Overweight/obesity was prevalent in 30% or more of both ANCS groups. For pubertal stages, the largest proportion of ANCS+ males reported stage four (late pubertal) for both pubic hair (47%) and genital development (57%), whereas the largest proportion of ANCS− males reported stage five (post-pubertal) for both pubic hair (55%) and genital development (46%). In females, the majority of ANCS+ (60%) and ANCS− (63%) reported stage five for pubic hair. For breast development, 50% of ANCS+ reported stage 4, and an equal percentage of ANCS− females reported stage four (45%) and stage five (45%). For neonatal and other adolescent characteristics, stratification by sex did not reveal any significant ANCS group differences.

Aerobic Fitness

Aerobic fitness results for the 85 ANCS+ and 75 ANCS− participants who gave a maximal effort and had valid exercise test data are presented in Table 3. Peak oxygen uptake was higher in the ANCS− exposed participants compared with those not exposed when expressed relative to body mass as well as LBM. When stratified by sex, the significant difference was evident in males only.

Table 3.

Comparison of aerobic fitness by ANCS exposure for the total group and stratified by sex^†.

	Total group		Male		Female
	ANCS+ (n = 85)	ANCS− (n = 75)	ANCS+ (n = 40)	ANCS− (n = 28)	ANCS+ (n = 45)	ANCS− (n = 47)
$\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$, mL•kg−1•min−1	39.4 ± 10.1*	35.7 ± 9.4	45.8 ± 9.3*	40.9 ± 9.6	33.6 ± 6.8	32.7 ± 7.9
$\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$, mL•kgLBM−1•min−1	51.2 ± 9.2*	47.9 ± 9.4	55.7 ± 9.1*	50.2 ± 8.9	47.2 ± 7.2	46.5 ± 9.5
$\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$, % of predicted	85 ± 19*	78 ± 19	92 ± 19*	82 ± 19	86 ± 17	83 ± 20

^†Values presented are unadjusted means ± SD. Between-group (ANCS+ v. ANCS−) comparisons were made using independent t-tests.

^*p≤.05 ANCS+ > ANCS−

Despite giving maximal efforts, some participants were unable to reach the maximal criterion of RER ≥ 1.05. Three (1 ANCS−) of the 8 participants with cerebral palsy (4 ANCS−) exhibited a neuromotor limitation as indicated by difficulty pedaling at higher workloads as well as a peak RER < 1.05. In addition, four participants (3 ANCS−) with a history of tracheomalacia had a peak RER < 1.05 and reported breathing as their reason for stopping. Exclusion of these participants with limitations resulted in slightly higher $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ for both ANCS+ and ANCS− males (mean ± sd: 47.0 ± 8.1 and 41.9 ± 9.5 mL•kg⁻¹•min, respectively) and a greater difference between groups. In females, $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ increased only in the ANCS+ group (to 34.0 ± 6.3 mL•kg⁻¹•min⁻¹), but the comparison with ANCS− females remained nonsignificant. Similar changes were evident in $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ expressed per Kg_LBM.

Multivariable analysis, with ANCS exposure, sex, race, and interaction terms in the model, revealed a 3-way interaction of ANCS with sex and race for $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ expressed relative to both body mass (p=0.001) as well as LBM (p<0.001). ANCS differences in $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ were evident in males and were significant in nonblack but not in black males (Figure 2, A). The mean adjusted difference in $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ was 8.5 (95% CI:2.1, 15.0) mL•kg⁻¹•min⁻¹ with higher values in nonblack ANCS+ males compared with their unexposed nonblack male peers. Similar differences between nonblack ANCS+ and ANCS” males were observed in $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ when expressed relative to LBM (adjusted mean difference: 9.0, (95% CI:1.1, 16.9) mL•kg_LBM⁻¹•min⁻¹. No significant ANCS group differences in fitness were observed in females, irrespective of race (Figure 2, B). The analysis was repeated excluding participants with apparent neuromotor or pulmonary limitations. The nonblack ANCS+ males continued to have higher $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ than nonblack ANCS” males, with the mean adjusted difference in $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ reduced slightly to 7.7 (95% CI: −0.5, 15.9; p=0.08) mL•kg_LBM⁻¹•min⁻¹. Exclusion of the data of 15 subjects (7 ANCS+, 8 ANCS−) who received dexamethasone postnatally (as part of a randomized controlled trial to reduce mechanical ventilator dependence) reduced the mean adjusted difference in $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ between nonblack ANCS+ and ANCS− males as well (7.2 mL•kg_LBM⁻¹•min⁻¹) and increased the p value to 0.1. No significant ANCS group differences in $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ were evident for black males, or females, irrespective of race.

Figure 2.

Adjusted Mean $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ in mL•kg⁻¹•min⁻¹ by ANCS exposure, race, and sex. Bars represent estimated marginal means and error bars represent the 95% confidence interval. Post hoc analysis (with Bonferroni adjustment) indicated that ANCS group differences were significant in nonblack males but not black males, black females, or nonblack females.

When $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ was expressed as a percent of age- and sex-specific reference values, 31% of ANCS+ and 48% of ANCS− had a $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ < 80% of predicted (p=0.024). Logistic regression analysis, with ANCS exposure, race, sex, and interaction terms in the model, revealed an interaction between ANCS and race (p=0.10). Stratification by race further revealed that ANCS differences were significant in nonblack but not black participants. Nonblack ANCS+ participants were 2.7 times (adjusted OR; 95% CI: 1.04, 6.83) more likely to have a $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ ≥ 80% of predicted compared with their nonblack unexposed peers.

Physical Activity

Physical activity data are presented in Table 4. Physical activity variables were not normally distributed, and distributions were not improved by various transformations. Consequently, Mann-Whitney U tests were used to examine between-group (ANCS+ vs. ANCS−) differences for those variables. For males and females combined, TOT-hrs and VIG-hrs per week for the past year did not differ significantly by ANCS exposure. For the past two months, TOT-hrs per week was not different, but VIG-hrs per week was higher in the ANCS+ group compared with the ANCS− group. When stratified by sex, VIG-hrs per week for the past year as well as TOT-hrs and VIG-hrs per week for the past two months were significantly higher in the ANCS+ males compared with ANCS− males. No ANCS group differences in PA were observed in females. Exclusion of data of participants who were unable to do the exercise test due to neuromotor or pulmonary limitations increased the TOT-hrs and VIG-hrs for the past year and VIG-hrs for the past two months in the ANCS− males, but the TOT-hrs and VIG-hrs for the past two months remained significantly lower than their ANCS+ male peers. In females, PA levels did not change negligibly and no differences by ANCS group remained.

Table 4.

Comparison of PA Levels by ANCS exposure and Sex.^†‡

	Total group		Male		Female
	ANCS+ (n = 91)	ANCS− (n = 72)	ANCS+ (n = 43)	ANCS− (n = 33)	ANCS+ (n = 48)	ANCS− (n = 49)
Physical Activity (past year)
TOT-hrs per wk	9.34 (4.46, 14.87)	8.05 (3.81, 14.64)	12.91 (7.67, 20.95)	9.84 (6.09, 19.05)	5.66 (2.66, 12.08)	6.21 (3.00, 13.69)
VIG-hrs per wk	1.16 (0.05, 4.59)	0.75 (0, 1.99)	4.59* (0.99, 8.54)	1.65 (0.13, 6.60)	0.36 (0, 1.72)	0.25 (0, 1.59)
Physical Activity (past 2 months)
TOT-hrs per wk	8.24 (4.50, 18.23)	7.03 (2.76, 15.48)	14.60** (6.54, 24.88)	8.5 (3.97, 17.15)	6.36 (2.74, 12.61)	5.47 (1.98, 15.10)
VIG-hrs per wk	0.93** (0, 3.73)	0.15 (0, 1.96)	3.00** (0.50, 9.00)	0.95 (0, 4.13)	0 (0, 1.62)	0 (0, 1.00)

^†Values are median (25^th, 75^th percentiles).

^‡Mann-Whitney U tests used to compare rank differences.

^**p<0.05

^*p<0.10

We also examined the proportion of participants meeting the national PA recommendations of at least seven hours per week.(25) For the past year, 59% of ANCS+ and 57% of ANCS− adolescents reported meeting this recommendation. When stratified by sex, more males (77 % of ANCS+ and 73% of ANCS−) than females (44% of ANCS− and 47% of ANCS+) met the recommendation, with no ANCS group differences. Similar results were observed for total PA in the previous two months. The national recommendation also specifies that most of the 60 minutes or more per day should be vigorous PA on at least 3 days per week. For the past year, 34% of ANCS+ and 20% of ANCS− groups reported engaging in vigorous PA for at least 3 hours per week (p=0.03). When stratified by sex, a greater proportion of ANCS+ males (57%) met this recommendation compared with ANCS− males (43%) (p =0.04) The proportions were slightly different for vigorous PA participation in the previous two months (54% of ANCS+ males and 36% of ANCS− males). In females, only 5 ANCS+ (10%) and 4 ANCS− (8%) females met the recommendation for vigorous PA for the past year, with slightly greater proportions observed in the previous two months (21% and 10% in ANCS+ and ANCS−, respectively). Logistic regression analysis, with ANCS exposure, sex, and race in the model, revealed a significant main effect of ANCS exposure on meeting the vigorous PA recommendation for the past 2 months. Interactions terms were not significant and were subsequently excluded from the analysis. ANCS+ adolescents were twice as likely to meet the recommendation (adjusted OR = 2.14, 95% CI 1.02, 4.50; p=.045) than their unexposed peers. Notably, 42% of ANCS− males compared with 14% of ANCS+ males and more than half of females in both groups (52% ANCS+ and 53% ANCS−) reported that they did not participate in any vigorous PA in the previous two months.

DISCUSSION

The results of our study suggest that ANCS exposure is associated with greater aerobic fitness and PA participation in males born preterm with VLBW, but only among nonblack ANCS+ males. The sex- and race-specific associations are consistent with our previous work showing sex and race differences of ANCS exposure on other outcomes such as heart rate variability, cholesterol levels, and renin-angiotensin system peptide levels.(16–18) The findings, however, are inconsistent with our hypotheses and are somewhat surprising in view of our previous work showing adverse effects of ANCS exposure on pulmonary function in this cohort.(6) Despite a greater prevalence of larger airway obstruction in ANCS−exposed adolescents compared with their unexposed peers (35% vs. 21%, respectively),(6) ANCS exposure was not associated with lower aerobic fitness or PA participation. Other studies have shown that exercise capacity and PA levels are more likely to be associated with impaired pulmonary diffusion, ventilatory inefficiency, and ventilatory flow limitations during exercise rather than resting measures of pulmonary function.(30,31) Future research should examine these parameters and their role in determining aerobic fitness in the population.

ANCS exposure during critical periods of fetal growth may have affected cardiac development and function. Preclinical studies have shown ANCS promotes myocyte maturation, and increases atrial mass, ventricular filling, and aortic flow in preterm-born piglets compared with their unexposed littermates.(9) In humans, Kelly et al reported no differences in left ventricular mass or ejection fraction between 16 ANCS+ and 16 ANCS− young adults, but the long-term effects of ANCS on cardiac structure and function, particularly in response to exercise, remain unknown.[4]

In contrast to cardiac muscle development, ANCS exposure has been shown to be detrimental to skeletal muscle growth and development. In animal models, fetal exposure to glucocorticoids has been associated with inhibited myogenesis, reduced myonuclei per myofiber, smaller fiber cross-sectional area and muscle mass, and fewer satellite cells.(7,8) However, in vitro studies suggest that the effects are dependent on the dose and timing of exposure.(32,33) Muscle satellite cells are muscle-specific stem cells that play an important role in fiber plasticity, repair, and regenerative capacity.(34) Aerobic exercise training has been shown to activate satellite cells and promote proliferation and differentiation toward a more oxidative phenotype with greater mitochondrial protein content, oxidative capacity, and capillary density.(35) In our cohort, ANCS exposure was associated with greater participation in vigorous PA, particularly in males, which may have contributed to their higher levels of aerobic fitness. Given the cross-sectional examination of functional parameters, it is also possible that greater fitness may have enabled more participation in vigorous PA.

The ANCS group differences in PA participation were evident in estimates for the past two months but not the past year. Although the past year is more likely to reflect habitual activity levels, it is subject to greater recall error than estimates for the previous two months. Estimates for the previous two months are more likely to reflect current season-specific PA levels and correspond to current aerobic fitness level. The low levels of participation in vigorous PA particularly in unexposed males and both groups of females is concerning and may contribute to their risk for developing cardiometabolic disease. The large sex differences were not totally unexpected and are consistent with data from the Youth Risk Behavior Survey showing that the percentage of adolescent males meeting the national guidelines for vigorous activity is more than double that of adolescent females in North Carolina.(36) The relatively high percentage of adolescents reporting that they met the national recommendation for total PA in the previous two months was somewhat surprising, and may reflect recall error as well as possible inaccuracy associated with self-report measures. Objective assessment with motion sensors would help to reduce these problems, and provide a more accurate measurement of short-term PA levels.

It is possible that fitness and PA in our cohort might be attributed to differences in body size. The ANCS+ males were significantly taller than their unexposed peers, and it may be speculated that the shorter stature of the ANCS− males might diminish their ability to compete with their peers in sports and ultimately lead to less participation in vigorous PA and lower aerobic fitness. Furthermore, the ANCS+ and ANCS− groups had similar LBM, and the group differences in fitness persisted when $\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$ was expressed relative to LBM. The differences are also not explained by sexual maturation as the majority of ANCS+ males reported a lower stage of development (Stage 4 late pubertal) than their ANCS− peers (Stage 5 post-pubertal).

Finally, ANCS exposure has been associated with reduced risk for cerebral palsy. Although four participants in both ANCS groups had a history of cerebral palsy, the group differences persisted when excluding data of participants with neuromotor limitations. ANCS exposure has also been associated with better fine and gross motor control in infants and young children,(37,38) and earlier infant motor development has been associated with higher levels of PA including sports participation at 14 years of age.(39) Consequently, the ANCS−exposed children may have developed better motor proficiency that facilitated motor learning, and instilled better self-efficacy or confidence in their ability to engage in PA. Future research should examine long-term effects of ANCS on motor proficiency and PA self-efficacy and their potential roles as mediators of PA participation in this population.

In 2 studies of young adult males with birth weights < 10^th percentile, aerobic training led to improvements in aerobic fitness, body composition, glucose regulation, and markers of insulin resistance.(40,41) Growing evidence suggests that persons with VLBW have increased risk for developing Type 2 diabetes.(42) Consequently, aerobic training may prove beneficial for improving cardiometabolic outcomes and reducing disease risk in the VLBW population.

A limitation is that this study is observational study and not a randomized controlled trial. However, treatment with ANCS for preterm labor is now standard practice with up to 87% of VLBW infants currently exposed,(2) and a randomized controlled trial would not be ethical. A major strength is the 50% exposure rate (due to birth years surrounding the 1994 NIH Consensus Panel promoting ANCS use) provided us with a unique opportunity to examine the long-term effects of ANCS exposure. Our cohort is also racially diverse and was born when medical advances such as surfactant were available, making our results more generalizable to current VLBW cohorts. However, given that most of the cohort was exposed to only one course of ANCS and birth occurred within one week, the results of repeated courses of ANCS and timing relative to birth warrant further examination.

In summary, our study results suggest that ANCS exposure is not detrimental to aerobic fitness and PA participation in VLBW adolescents, and may have sex- and race-specific benefits. The low levels of participation in vigorous PA particularly in unexposed males and both groups of females is concerning and may contribute to their risk for developing cardiometabolic disease. Further research is warranted to identify modifiable determinants of PA participation, such as motor proficiency and PA self-efficacy, which may aid in the development of PA interventions. In view of the current ANCS exposure rates and associated risk for cardiometabolic disease, manipulation of PA in randomized control trials is needed to determine if aerobic fitness and habitual activity can be increased and ultimately reduce the risk for developing cardiometabolic disease in the VLBW population. The timing of intervention relative to growth and maturation should be considered and may prove to be critical.

ABBREVIATIONS LIST

BMI

Body mass index

RER

Respiratory exchange ratio

TOT-hrs

Total hours of PA per week

VIG-hrs

Hours spent in vigorous PA per week

VLBW

Very low birth weight

$\stackrel{.}{\text{V}}{\text{O}}_2\text{peak}$

Peak oxygen uptake