Overweight, Obesity, and Body Composition in 3.5- and 7-Year-Old Swedish Children Born with Marginally Low Birth Weight

Overweight, Obesity, and Body Composition in 3.5- and 7-Year-Old Swedish Children Born with Marginally Low Birth Weight 4 € Josefine Lindberg1, Mikael Norman, MD, PhD2, Bj€orn Westrup, MD, PhD3, Tove Ohrman , Magnus Domell€of, MD, PhD1, 1 and Staffan K. Berglund, MD, PhD

Objectives To assess the prevalence of overweight/obese children and to explore body composition in a Swedish cohort of preschool children born with marginally low birth weight (MLBW, ie, 2000-2500 g).

Study design We included 285 Swedish children with MLBW (44% small for gestational age), and 95 control children with normal birth weights. At 3.5 years and 7 years of age, we assessed anthropometrics, including the prevalence of overweight/obese children. At 7 years, dual-energy X-ray was used for body composition. Results There were no significant differences between groups in the prevalence of overweight/obesity or in skinfold thickness; however, at 3.5 years, mean height, weight, and BMI in children with MLBW were 2.1 cm (95% CI 1.2-3.1), 1.2 kg (95% CI 0.7-1.6), and 0.47 kg/m2 (95% CI 0.17-0.76) lower compared with controls. The corresponding mean differences also were lower in children with MLBW compared with control children at 7 years; 2.5 cm (95% CI 0.9-4.1), 1.6 kg (95% CI 0.6-2.8), and 0.48 kg/m2 (95% CI 0.01-0.94). The differences were greater in those born small for gestational age. Dual-energy X-ray analyses showed lower fat-free mass index in MLBW infants and a similar trend in fat mass index. Within children with MLBW, BMI at 7 years correlated positively to growth velocity in infancy. Conclusion Children with MLBW had lower BMI and did not show increased risk of overweight or obesity up to 7 years. Nevertheless, the BMI in MLBW children was positively correlated to growth-velocity in infancy. (J Pediatr 2015;-:---). Trial registration Clinicaltrials.gov: NCT00558454.

T

he prevalence of overweight and obese adults as well as children and adolescents is increasing worldwide. This is a major public health problem, considering the well-documented adverse outcomes that accompany these conditions, such as sedentariness, cardiovascular diseases, noninsulin-dependent diabetes mellitus, sleep apnea, and other noncommunicable diseases. Of further concern is that these correlations to later morbidity also are observed in overweight and obese children. Thus, more knowledge of when and how to identify and prevent young children at risk of overweight and obesity should be prioritized.1,2 In 1989, Barker et al3 correlated low birth weight (LBW) with coronary heart disease later in life. They suggested that the origins of metabolic syndrome (MS), which includes obesity, dyslipidemia, insulin resistance, glucose intolerance, and elevated blood pressure, could be found already in the intrauterine and immediate postnatal stages of life as a consequence of early malnutrition.3-5 This concept of “early metabolic programming” has since generated substantial interest in epidemiologic and experimental research.6-8 Particularly, the roles of LBW, preterm birth, and intrauterine growth restriction (IUGR) have been explored and discussed.8-11 In common for many infants born too early or too small is an increased growth velocity at some point after birth. Mismatch in nutrition, resulting in poor prenatal growth followed by accelerated postnatal growth triggered by overnutrition, may be a particularly important link in the causal chain of metabolic programming.8,12,13 Overweight and obesity are frequently studied outcomes with regard to early metabolic programming after LBW. Several studies have confirmed an association between LBW and later increased body mass index (BMI)14,15; however, From the Department of Clinical Sciences, Pediatrics, Ume a University, Ume a; Division of Pediatrics, the mechanisms of this association are unclear, and there are several studies, Department of Clinical Science, Intervention and 1

2

Technology, 3Division of Neonatology, Department of Women’s and Children’s Health, Karolinska Institutet; and 4Department of Medical Physics, Danderyd Hospital, Stockholm, Sweden

AGA BMI FFMI FMI IOTF IUGR LBW MLBW MS SGA

Appropriate for gestational age Body mass index Fat-free mass index Fat mass index International Obesity Task Force Intrauterine growth restriction Low birth weight Marginally low birth weight Metabolic syndrome Small for gestational age

Supported by the Swedish Research Council (Formas222-2005-1894), Swedish Research Council for Health, Working Life and Welfare (FORTE-2012-0708), €sterbotten County Council (ALF), the Jerring FoundaVa tion, the Oskar Foundation, the Swedish Society of Medicine (SLS-331751), the Childhood Foundation of the Swedish Order of Freemasons, and by a regional agreement on medical training and clinical research (ALF) between Stockholm County Council and Karolinska Institutet. The authors declare no conflicts of interest. 0022-3476/$ - see front matter. Copyright ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2015.08.045

1

THE JOURNAL OF PEDIATRICS



www.jpeds.com

including a recent meta-analysis, which failed to show any correlation between LBW and BMI/obesity.16,17 The reason for these previously diverging results is most likely because LBW occurs at different degrees and subgroups and because of different settings with regard to nutrition, time periods, and sociodemographic factors.8,16 The majority of children with LBW are born only with marginally LBW (MLBW; 2000-2500 g), with or without IUGR, and the magnitude and emergence of overweight and obesity in these subjects are less known, especially in healthy and well-nourished populations. Furthermore, the proportions of lean and fat mass may be altered in children born with LBW. More knowledge regarding the association between birth weight and body composition is needed.18,19 The aim of this cohort study was to investigate the possible associations between MLBW and early childhood risk of overweight and obesity as well as other anthropometric signs of MS in a Swedish setting. We hypothesized that preschool children born MLBW would show early anthropometric signs of MS as a consequence of early metabolic programming.

Methods This prospective cohort study included 285 children with MLBW and 95 control children born between March 2004 and November 2007. The MLBW participants originally took part in an iron supplementation trial and were identified by the use of delivery records for inclusion at 6 weeks of age.20 Inclusion criteria were birth weight 2000-2500 g, no signs of disease at inclusion, no chronic disease, no previous blood transfusion, and never having received iron supplements. The participants were collected at 2 tertiary hospitals in Sweden—Ume a University Hospital, Ume a, and Karolinska University Hospital, Stockholm. This trial was approved by the Ethical Review Boards at Umea University and the Karolinska Institute and registered with Clinicaltrials.gov, number NCT00558454. The MLBW infants were randomized into 3 intervention groups receiving different doses of iron supplementation from 6 weeks to 6 months of age (0 mg/kg/day [placebo], 1 mg/kg/day, or 2 mg/kg/day). The children were assessed for iron status and growth as described previously.20 Because iron supplements did not have any effect on growth in our previous analyses, we analyzed all 3 intervention groups as 1 cohort in the present analyses. Before the 3.5-year control, every third MLBW child was chosen as an index case for recruitment of controls (matched for age, sex, and study center). For each index child, a list of 10 possible control children was made. These were the children born closest in time at the same study center and with the same sex as the index child as well as fitting into the following criteria: born in gestational weeks 37-42, born with birth weight between 2501 and 4500 g, and not being admitted to neonatal unit. The parents of the child born closest in time were contacted and offered to participate with their child as a control and if the parents declined, the next one was con2

Vol. -, No. tacted until each index child had a corresponding control child, or the list of eligible controls was exhausted. At inclusion of MLBW and control children, background data were collected from parents and from delivery records, including gestational age at birth, sex, anthropometric data, neonatal diagnoses, maternal birth country, smoking habits, income, and family situation. Using a Swedish gestational age–corrected growth standard,21 we calculated the weight for age SDS (SDS for weight) and defined small for gestational age (SGA) at birth as a birth weight SDS for gestational age less than 2 and appropriate for gestational age (AGA) in all other cases. As a part of the original iron supplementation trial, the children with MLBW were examined at 6, 12, and 19 weeks and at 6 and 12 months regarding anthropometric data such as weight, height, and waist and head circumference. At each visit, we calculated the SDS for weight and height and its change (DSDS) since last visit. The present follow-up trial included visits at 3.5 and 7 years of age. At each visit, skinfold thickness (triceps and subscapular) was assessed with the use of a skinfold measuring instrument (Harpenden Skinfold Caliper; Baty International, West Sussex, United Kingdom), waist and head circumference with a measuring tape (Seca 212; seca, Hamburg, Germany), weight with an electronic scale (Seca 701; seca), and height with a wall stadiometer (Hyssna Measuring Equipment AB, Hyssna, Sweden). On one occasion, we also measured the weight and length of the parents and, if any parent was not present at the 3.5- or 7-year follow-up, we used the reported values (18.5% of mothers and 58.2% of fathers). BMI values of children and mothers were calculated according to the International Obesity Task Force (IOTF), and obesity and overweight were determined using the ageand sex-specific SD by Cole at al.22 At the 7-year visit, the children were examined with dualenergy X-ray absorptiometry. These dual-energy X-ray absorptiometry measurements were performed at both study centers by use of the Lunar Prodigy Advance (GE Medical Systems Lunar, Madison, Wisconsin) with enCORE software version 13.31016 (GE Medical Systems Lunar). The X-ray spectrum was produced with a cerium filter resulting in 2 energy peaks at approximately 38 and 70 keV, and the 2 machines were cross-calibrated by the use of an aluminum and acrylic lumbar spine phantom showing a less than 1% difference in measured bone mineral density values. Fatfree mass was calculated as bone mineral content + lean mass, and fat mass index (FMI) and fat-free mass index (FFMI) were calculated as fat mass/height2 (kg/m2) and fat free mass/height2 (kg/m2), respectively.23 Statistical Analyses Statistical analyses were performed using SPSS 22.0 for Windows (SPSS Inc, Chicago, Illinois). Differences between the groups were analyzed using the independent t test for continuous variables and the c2 test for categorical variables. The main outcomes of the present article were compared between all MLBW and control subjects. To further explore the Lindberg et al

- 2015

ORIGINAL ARTICLES

contribution of being born at different degree of growth restriction, we also stratified the analyses in the MLBW cohort according to SGA or AGA at birth. Finally, we used univariate and multivariate linear regression models to explore birth weight, infancy growth data, and other possible predictors of BMI, FMI, and FFMI in MLBW children.

Results A flow chart for included infants is presented in Figure 1 (available at www.jpeds.com). Four children were diagnosed with congenital disorders after inclusion but before the age of 7, and they were excluded in all analyses. A total of 327 children (87%) were examined at 3.5 years and 275 (72%) children at 7 years. There were no significant differences in the prevalence of overweight (P = .753), BMI, or any other anthropometric characteristics at 3.5 years between the cases who later dropped out and those who remained at 7 years, suggesting a nonselective drop out. The perinatal and sociodemographic characteristics of those examined at 3.5 or 7 years are presented in Table I. By definition, the MLBW children were smaller in size at birth and, as expected, there was a lower prevalence of vaginal deliveries compared with control children, especially in those born SGA. The SGA children had mothers who were significantly shorter compared with mothers of control children. Maternal age, parity, weight, BMI, and education did not differ between the 2 groups. Paternal education also was similar.

Anthropometric Differences between MLBW and Control Children Only one child at 3.5 years of age (MLBW group) and one at 7 years of age (control group) were obese according to the IOTF definition. These 2 children were analyzed together with the overweight children. The prevalence of overweight/obese children at 3.5 and 7 years is presented in Figure 2, with no significant differences between the groups shown. We observed a nonsignificant trend of increasing prevalence of overweight from 3.5 to 7 years of age in the MLBW children, particularly in the SGA group, suggesting a possible ongoing change. We found at 3.5 years that mean height, weight, and BMI were 2.1 cm (95% CI 1.2-3.1, P < .001), 1.2 kg (95% CI 0.71.6, P < .001), and 0.47 kg/m2 (95% CI 0.17-0.76, P = .002) lower in MLBW children compared with controls (Table II). The MLBW children remained lighter and shorter at 7 years of age with corresponding mean differences of 2.5 cm (95% CI 0.9-4.1, P = .001), 1.6 kg (95% CI 0.6-2.8, P = .001), and 0.48 kg/m2 (95% CI 0.01-0.94, P = .046). Furthermore, the head circumferences were significantly lower at both ages and the waist circumferences were lower at 3.5 years in MLBW children. There were no differences in skinfold thickness between the groups. The body composition analyses at 7 years showed significantly lower total body fat, total lean mass, and bone mineral content in MLBW than in controls (Table II). These differences did not remain significant when we adjusted for height (data not shown); however, FFMI was significantly lower in MLBW born children and FMI

Table I. Cohort characteristics of MLBW children, including subgroups of SGA and AGA, and control children with normal birth weights Controls, n = 95 Pregnancy Maternal smoking during pregnancy Delivery Vaginal delivery Single gestation Admission for neonatal care Hypoglycemia at birth Infant Female sex Gestational age at birth, wk Apgar score at 5 min, median Born preterm, <37 wk Born SGA (< 2 SD) Birth weight, kg Birth length, cm Head circumference at birth, cm Parental characteristics Mother born in Scandinavia Maternal weight, kg Maternal height, cm Maternal BMI, kg/m2 Mother’s age at birth Parity Mother educated at university Father educated at university

MLBW, n = 236

P1

SGA, n = 105

P2

7 (3.0%)

.941

5 (4.8%)

68 (73.9%) 94 (98.9%) N/A N/A

133 (56.6%) 161 (68.2%) 114 (48.5%) 48 (20.3%)

.004 <.001

52 (49.5%) 69 (65.7%) 28 (26.9%) 24 (22.9%)

<.001 <.001

81 (62.3%) 92 (70.2%) 86 (65.6%) 24 (18.3%)

.070 <.001

48 (50.5%) 39.99 (1.18) 10 (10; 10) 0 (0.0%) 0 (0.0%) 3.56 (0.43) 50.56 (2.04) 35.07 (1.43)

121 (51.3%) 36.47 (1.88) 10 (9; 10) 132 (55.9%) 105 (44.5%) 2.29 (0.14) 45.30 (1.44) 32.23 (1.22)

.902

48 (45.7%) 38.02 (1.09) 10 (9; 10) 18 (17.1%)

.496

73 (55.7%) 35.23 (1.40) 10 (9; 10) 114 (87.0%)

.439

69 (93.2%) 65.19 (9.59) 167.24 (6.12) 23.43 (3.77) 35.19 (4.53) 1.6 (0.72) 58 (61.1%) 45 (47.4%)

198 (84.3%) 65.92 (12.87) 165.85 (6.81) 24.00 (4.92) 35.55 (4.77) 1.6 (0.84) 129 (54.7%) 140 (59.8%)

.140

2.28 (0.14) 45.70 (1.34) 32.49 (1.06) .049 .625 .090 .322 .535 .972 .289 .039

83 (79.0%) 66.29 (14.35) 164.88 (6.98) 24.39 (5.39) 36.30 (4.94) 1.75 (0.93) 61 (58.1%) 61 (58.7%)

2 (1.6%)

P3

3 (3.2%)

.046

.554

AGA, n = 131

.421

.038

2.30 (0.15) 44.98 (1.44) 32.02 (1.31) .009 .535 .014 .161 .104 .215 .670 .111

115 (88.5%) 65.62 (11.62) 166.61 (6.60) 23.70 (4.50) 34.94 (4.57) 1.49 (0.76) 68 (51.9%) 79 (60.8%)

.269 .773 .474 .653 .692 .269 .172 .046

N/A, not applicable. Data are mean (SD) or no (%). Apgar score at 5 minutes is presented as median (IQR). P value for differences between groups, P1: controls vs MLBW, P2: controls vs SGA, P3: controls vs AGA. c2 test for categorical variables, independent t test for continuous variables and MannWhitney U test for Apgar score.

Overweight, Obesity, and Body Composition in 3.5- and 7-Year-Old Swedish Children Born with Marginally Low Birth Weight

3

THE JOURNAL OF PEDIATRICS 3.5 years 9% 8% 7% 6% 5% 4% 3% 2% 1% 0%



circumference, total lean mass, and bone mineral content; total body fat was similar.

P =.089 P >.999 P=.439

Controls

Overweight Obese

MLBW (all)

MLBW (AGA)

P=.768 P =.743

7 years 9% 8% 7% 6% 5% 4% 3% 2% 1% 0%

MLBW (SGA)

Overweight Obese

P =.573

Controls

Vol. -, No. -

www.jpeds.com

MLBW (all)

MLBW (AGA)

MLBW (SGA)

Figure 2. Prevalence of overweight or obesity at 3.5 years and 7 years of age in children born with MLBW (2000-2500 g) compared with control children born with normal birth weight.

Predictors of BMI, FMI, and FFMI In univariate (Table III; available at www.jpeds.com) linear regression analyses, we explored possible early predictors of later BMI, FMI, and FFMI. Maternal weight and maternal BMI turned out to be the strongest univariate predictor of BMI and FMI in the MLBW offspring; however, these predictors did not correlate with FFMI, which instead correlated positively to preterm birth and weakly and negatively to female sex and maternal length. The degree of growth restriction at birth (SDS for weight) did not correlate with BMI and FMI but weakly and positively with FFMI. When exploring early weight and height gain rate, we found several positive correlations, suggesting that the magnitude of catch-up growth in infancy predicted greater BMI, FMI, and FFMI at 7 years. In multivariate regression models, including all significant predictors, some of these associations to early growth remained significant (Table IV; available at www.jpeds.com). In fact, when we adjusted for maternal BMI and maternal weight, respectively, in the multivariate models predicting BMI and FMI, the models suggested that weight gain in infancy was the only additional predictor. Similarly, the multivariate model of FFMI suggested a weak positive association to rate of height gain in early infancy when we adjusted for the other predictors.

showed a similar trend, although the difference did not reach statistical significance.

Discussion

Stratified Analyses Based on Degree of Growth Restriction at Birth Further stratified analyses were made to investigate the impact of being born SGA or AGA respectively (Figure 2 and Table II). We observed a nonsignificant trend of lower overweight/obesity prevalence in the SGA subgroup at 3.5 years (P = .089) compared with controls but not at 7 years, where the prevalence was greater in SGA (7.5 vs 4.3%), a difference that did not reach significance. With regard to other anthropometric measures, height, weight, and BMI were lower in both subgroups compared with controls, but the differences were more pronounced in SGA children compared with controls. A corresponding observation was found for FFMI in these subgroups; both SGA and AGA had lower values, but they did only reach significance in the subgroup of SGA children. In fact, the SGA infants had also a significantly lower FMI compared with controls. When we compared the anthropometric outcomes and body composition between the subgroups born SGA and AGA, we found at 3.5 years of age that height, weight, BMI, and head and waist circumference were significantly greater in the AGA children. At 7 years of age, the difference remained significant with regard to height, weight, head

We explored early anthropometric indicators of MS in a cohort of otherwise-healthy Swedish children with MLBW until 7 years of age. In contrast to our hypothesis, the prevalence of overweight and obesity was not greater than in a control group of children born with normal birth weight. Indeed, the weight, height, head circumference, BMI, and FFMI levels were lower in those born with MLBW compared with controls, a difference that was most pronounced in those born SGA. The concept of early metabolic programming is well established. According the hypothesis originally suggested by Barker et al,4,9 different cells and organs have different critical periods during fetal growth under which they are particularly sensitive to stress. If oxygen and/or nutrients are altered during this period, the development of these particular organs is affected, leading to consequences later in life.4,9,16,24 LBW is one of the most explored early risk factors of metabolic programming24; however, because both the settings of children with LBW as well as the outcomes are widely diverging, previous research still leaves unanswered questions. With regard to later risk of overweight and obesity, several previous papers have suggested a strong association with LBW19,25-27; however, none of these studies included children with MLBW alone and furthermore, others have suggested that there is no or even a negative association

4

Lindberg et al

- 2015

ORIGINAL ARTICLES

Table II. Anthropometric measures at 3.5 years and 7 years of age in children born with MLBW compared with healthy control children born term with normal birth weight Controls Year 3.5 y Anthropometrics Height, cm Weight, kg BMI, kg/m2 Head circumference, cm Waist circumference, cm Skinfold thickness triceps, mm Skinfold thickness subscapular, mm 7y Anthropometrics Height, cm Weight, kg BMI, kg/m2 Head circumference, cm Waist circumference, cm Skinfold thickness triceps, mm Skinfold thickness subscapular, mm Body composition Total body fat, kg Total lean mass, kg Bone mineral content, kg FMI, kg/m2 FFMI, kg/m2

MLBW (all)

MLBW (AGA)

MLBW (SGA)

n

Mean

SD

n

Mean

SD

P1

n

Mean

SD

P2

n

Mean

SD

P3

P4

94 94 94 94 94 93

100.1 16.1 16.0 50.7 50.9 10.1

4.1 2.0 1.2 1.4 3.2 1.9

232 232 232 232 231 225

98.0 14.9 15.5 50.0 49.6 10.2

3.9 1.9 1.2 1.4 3.1 2.1

<.001 <.001 .002 <.001 .001 .535

129 129 129 129 128 125

98.6 15.3 15.7 50.2 50.4 10.5

3.8 1.9 1.2 1.4 3.0 2.0

.004 .006 .135 .017 .209 .132

103 103 103 103 103 100

97.2 14.4 15.2 49.7 48.7 9.9

4.0 1.8 1.2 1.4 3.0 2.2

<.001 <.001 <.001 <.001 <.001 .613

<.001 .008 .002 .014 <.001 .050

93

5.6

1.5

223

5.4

1.3

.380

124

5.6

1.3

.979

99

5.3

1.3

.106

.065

70 70 70 70 70 68

124.9 24.3 15.5 52.7 55.2 9.2

5.4 3.8 1.5 1.8 4.7 2.8

204 204 204 204 204 198

122.4 22.7 15.1 51.6 54.3 8.8

5.4 3.6 1.6 1.4 4.9 2.5

.001 .001 .046 <.001 .192 .215

111 111 111 111 111 107

123.4 23.2 15.2 51.9 54.8 9.0

5.4 3.5 1.5 1.4 4.5 2.5

.073 .046 .146 .001 .615 .530

93 93 93 93 93 91

121.2 22.1 15.0 51.3 53.7 8.6

5.2 3.6 1.6 1.4 5.4 2.5

<.001 <.001 .030 <.001 .062 .110

.003 .025 .349 .003 .092 .229

68

5.3

2.3

199

5.2

2.2

.674

108

5.3

2.3

.895

91

5.1

2.1

.508

.557

68 68 68 68 68

4.47 18.91 0.88 2.81 12.65

2.34 1.93 0.13 1.25 0.82

198 198 198 198 198

3.80 17.80 0.82 2.50 12.40

2.16 2.06 0.12 1.30 0.82

.031 <.001 .001 .087 .033

111 111 111 111 111

3.97 18.20 0.84 2.58 12.47

2.14 2.09 0.13 1.29 0.79

.151 .024 .060 .248 .155

87 87 87 87 87

3.57 17.30 0.80 2.39 12.31

2.17 1.91 0.12 1.31 0.84

.014 <.001 <.001 .047 .014

.187 .002 .009 .311 .175

Data are mean, SD, and P values are unadjusted t test. P1: MLBW vs controls, P2: AGA vs controls, P3: SGA vs controls, P4: AGA vs SGA.

between LBW and overweight and obesity.16,17,28 We used the IOTF definition of overweight and obesity and found no correlation to being born with MLBW; however, of note is that the prevalence found were generally low, lower than previously shown in Swedish children,29 and our study might have been underpowered to show small but clinically relevant differences. The reason for our low prevalence may be a biased selection of families of greater socioeconomic status, but they might also reflect a suggested recent decline in the prevalence of overweight among Swedish children.30 In contrast to the hypothesis, our results showed that children with MLBW had lower weight and BMI at 3.5 and 7 years of age. Previous research has suggested that the association between LBW and high BMI becomes more prominent with increasing age, and it is possible that we would have seen a negative impact of MLBW at a greater age10; however, the results also may suggest that an association between LBW and greater BMI is low or absent in otherwise-healthy MLBW children from a well-nourished setting. In fact, that conclusion is in agreement with several previous studies, including a recent meta-analysis in which the authors found that high birth weight but not LBW correlated to later obesity.16,17 We found no previous paper in which the authors studied children with MLBW separately. LBW may correlate differently to lean and fat mass later in life. Some studies have shown that LBW is most strongly associated with increased fat mass.15,31,32 Singhal et al19 sug-

gested in a work from 2003 that birth weight is closely associated with lean mass but not with fat mass, and Claris et al33 suggested that increases of both lean and fat mass are associated with postnatal growth after IUGR. In the present study, we found, in concordance with the effect on BMI and body weight, that children with MLBW had lower levels of both FMI and FFMI. Interestingly, however, a significant association to LBW only was seen for FFMI and total lean mass, similarly to the results from Singhal et al. Again, a possible explanation is that the increased FFMI and FMI associated to LBW in other studies may be an effect emerging only after 7 years of age, or that this programming effect is not present in children with MLBW. Nevertheless, it underscores the need of body composition analyses when exploring early programming of future BMI. Other important measures of body composition are skinfold thickness and waist circumference. These outcomes have been associated previously with metabolic programming but showed no such trends in the present study.28,34 The LBW definition includes a continuum of more or less preterm and more or less growth restricted newborns. This also was the case in the present cohort. Because the current opinions regarding mechanisms for metabolic programming are related to fetal undernutrition and/or rapid catch-up growth with overnutrition, it is relevant to explore the associations stratified according to the degree of prematurity or the degree of growth restriction.8,35,36 In the present paper,

Overweight, Obesity, and Body Composition in 3.5- and 7-Year-Old Swedish Children Born with Marginally Low Birth Weight

5

THE JOURNAL OF PEDIATRICS



we chose the latter cut off for stratified analyses because of the fact that other research on metabolic programming focused on SGA infants and not preterm infants. The SGA group included mostly term infants and the AGA mostly preterm infants (Table I). Even though children born SGA consequently had a greater degree of catch-up growth (unpublished data), they still did not show any anthropometric signs of MS. Instead, the lower BMI, FFMI, and body weight was more pronounced in the subgroup of SGA, suggesting a lower instead of greater risk. Of note, however, is the observed trend between 3.5 years and 7 years of age (Figure 2), suggesting decreasing prevalence of overweight in controls, but it increased in the MLBW group. This ongoing change seems to be most rapid in the SGA subgroup and may be an early indicator of changes that appear after the age of 7 years. Several researchers have suggested that it is the postnatal catch-up growth that follows LBW and not the LBW itself that constitutes the risk factor for children with LBW. The speed and duration of the catch-up growth has proved to be essential in the development of adverse health outcomes.8,12,13,35,37 This opens for a possibility to modulate later risks by dietary interventions. However, exactly who is at risk and whether dietary interventions are applicable is still not entirely known.6,25 Approximately 90% of children born with LBW have a postnatal catch-up growth, which causes problems in analyzing the contributing effect from the actual birth weight vs the catch-up growth. It is also the case in this study where approximately 70% of MLBW infants increased their SDS for weight >1.0 during their first year of life. SGA children had a greater weight growth velocity during infancy but as discussed previously, they showed no increased risk of MS. When we analyzed the predictors of BMI, FMI, and FFMI, however, we found a significant association with the rate of weight gain and the BMI and FMI at 7 years. These data support that an early and rapid catch-up growth is positively associated with later body composition. Whether this also is a risk factor for later obesity requires further trials including follow-up beyond 7 years of age. Nevertheless, they support that childhood body fat can be sensitive to early programming, opening the way for future interventions. There was also an interesting association between velocity of height growth at 6-12 weeks and FFMI, suggesting that longitudinal growth and weight gain may have different but still significant predicative effect in early of programming. In summary, the latter results support the hypothesis that the growth velocity and particularly weight change rather than LBW itself correlates to future body composition. Children with MLBW in a well-nourished population with low prevalence of overweight and obesity had no increased signs of obesity or other anthropometric measures of MS to 7 years of age, which may be attributable to a lack of an association or that the effect of metabolic programming appears later in life. Our results suggest that children born MLBW in general and SGA MLBW children in particular have lower BMI, FMI, and FFMI at 7 years. Nevertheless, 6

Vol. -, No. -

www.jpeds.com

these measures were positively correlated to the weight gain in infancy, supporting the hypothesis that body composition in school aged children can be partly predicted from the early growth patterns and possibly modulated by dietary interventions in infancy. n

We want to thank all participating families and our dedicated research  Sundstr€om (Ume nurses, Kerstin Andersson (Stockholm) and Asa a). Submitted for publication Apr 30, 2015; last revision received Jul 28, 2015; accepted Aug 21, 2015. Reprint requests: Staffan K. Berglund, MD, PhD, Department of Clinical Sciences, Pediatrics, Ume a University, SE-901 85 Ume a, Sweden. E-mail: [email protected]

References 1. World Health Organization. Obesity: preventing and managing the global epidemic. Report of a WHO consultation. World Health Organ Tech Rep Ser 2000;894:i-xii. 1-253. 2. Wang Y, Lobstein T. Worldwide trends in childhood overweight and obesity. Int J Pediatr Obes 2006;1:11-25. 3. Barker DJ, Winter PD, Osmond C, Margetts B, Simmonds SJ. Weight in infancy and death from ischaemic heart disease. Lancet 1989;2:577-80. 4. Barker DJ. Fetal origins of coronary heart disease. BMJ 1995;311:171-4. 5. Ozanne SE, Const^ancia M. Mechanisms of disease: the developmental origins of disease and the role of the epigenotype. Nat Clin Pract Endocrinol Metab 2007;3:539-46. 6. Valerio Nobili M, Alisi A, Panera N, Agostoni C. Low birth weight and catch-up-growth associated with metabolic syndrome: a ten year systematic review. Pediatr Endocrinol Rev 2008;6:241-7. 7. Warner MJ, Ozanne SE. Mechanisms involved in the developmental programming of adulthood disease. Biochem J 2010;427:333-47. 8. Kelishadi R, Haghdoost AA, Jamshidi F, Aliramezany M, Moosazadeh M. Low birthweight or rapid catch-up growth: which is more associated with cardiovascular disease and its risk factors in later life? A systematic review and cryptanalysis. Paediatr Int Child Health 2015;35:110-23. 9. de Boo HA, Harding JE. The developmental origins of adult disease (Barker) hypothesis. Aust N Z J Obstet Gynaecol 2006;46:4-14. 10. Stroescu R, Micle I, Bizerea T, Puiu M, Marginean O, Doros¸ G. Metabolic monitoring of obese children born small for gestational age. Obes Res Clin Pract 2014;8:e592-8. 11. Parkinson JR, Hyde MJ, Gale C, Santhakumaran S, Modi N. Preterm birth and the metabolic syndrome in adult life: a systematic review and meta-analysis. Pediatrics 2013;131:e1240-63. 12. Lei X, Chen Y, Ye J, Ouyang F, Jiang F, Zhang J. The optimal postnatal growth trajectory for term small for gestational age babies: a prospective cohort study. J Pediatr 2015;166:54-8. 13. Singhal A, Cole TJ, Fewtrell M, Deanfield J, Lucas A. Is slower early growth beneficial for long-term cardiovascular health? Circulation 2004;109:1108-13. 14. Ong KK, Dunger DB. Perinatal growth failure: the road to obesity, insulin resistance and cardiovascular disease in adults. Best Pract Res Clin Endocrinol Metab 2002;16:191-207. 15. Murtaugh MA, Jacobs DR, Moran A, Steinberger J, Sinaiko AR. Relation of birth weight to fasting insulin, insulin resistance, and body size in adolescence. Diabetes Care 2003;26:187-92. 16. Yu ZB, Han SP, Zhu GZ, Zhu C, Wang XJ, Cao XG, et al. Birth weight and subsequent risk of obesity: a systematic review and meta-analysis. Obes Rev 2011;12:525-42. 17. Gomes FM, Subramanian SV, Escobar AM, Valente MH, Grisi SJ, Brentani A, et al. No association between low birth weight and

Lindberg et al

- 2015

18. 19.

20.

21. 22.

23. 24.

25. 26.

27. 28.

cardiovascular risk factors in early adulthood: evidence from S~ao Paulo, Brazil. PLoS One 2013;8:e66554. Battista M, Murray RD, Daniels SR. Use of the metabolic syndrome in pediatrics: a blessing and a curse. Semin Pediatr Surg 2009;18:136-43. Singhal A, Wells J, Cole TJ, Fewtrell M, Lucas A. Programming of lean body mass: a link between birth weight, obesity, and cardiovascular disease? Am J Clin Nutr 2003;77:726-30. Berglund S, Westrup B, Domell€ of M. Iron supplements reduce the risk of iron deficiency anemia in marginally low birth weight infants. Pediatrics 2010;126:e874-83. Niklasson A, Albertsson-Wikland K. Continuous growth reference from 24th week of gestation to 24 months by gender. BMC Pediatr 2008;8:8. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ 2000;320:1240-3. Wells JC. A critique of the expression of paediatric body composition data. Arch Dis Child 2001;85:67-72. Hanley B, Dijane J, Fewtrell M, Grynberg A, Hummel S, Junien C, et al. Metabolic imprinting, programming and epigenetics—a review of present priorities and future opportunities. Br J Nutr 2010;104(Suppl 1): S1-25. Weaver LT. Rapid growth in infancy: balancing the interests of the child. J Pediatr Gastroenterol Nutr 2006;43:428-32. ~ez L, Ong K, Dunger DB, de Zegher F. Early development of Iban adiposity and insulin resistance after catch-up weight gain in smallfor-gestational-age children. J Clin Endocrinol Metab 2006;91:2153-8. Benjamin C. Obesity as a consequence of undernutrition. J Pediatr 2006; 149:S97-9. Roswall J, Karlsson AK, Allvin K, Tangen GA, Bergman S, Niklasson A, et al. Preschool children born moderately preterm have increased waist

ORIGINAL ARTICLES

29.

30.

31. 32.

33.

34. 35.

36.

37.

circumference at two years of age despite low body mass index. Acta Paediatr 2012;101:1175-81. Sj€ oberg A, Moraeus L, Yngve A, Poortvliet E, Al-Ansari U, Lissner L. Overweight and obesity in a representative sample of schoolchildren - exploring the urban-rural gradient in Sweden. Obes Rev 2011;12: 305-14. Bergstr€ om E, Blomquist HK. Is the prevalence of overweight and obesity declining among 4-year-old Swedish children? Acta Paediatr 2009;98: 1956-8. Wells JC, Chomtho S, Fewtrell MS. Programming of body composition by early growth and nutrition. Proc Nutr Soc 2007;66:423-34. Carlsen EM, Renault KM, Nørgaard K, Nilas L, Jensen JE, Hyldstrup L, et al. Newborn regional body composition is influenced by maternal obesity, gestational weight gain and the birthweight standard score. Acta Paediatr 2014;103:939-45. Claris O, Beltrand J, Levy-Marchal C. Consequences of intrauterine growth and early neonatal catch-up growth. Semin Perinatol 2010;34: 207-10. Wells JC, Fewtrell MS. Measuring body composition. Arch Dis Child 2006;91:612-7. Leunissen RW, Kerkhof GF, Stijnen T, Hokken-Koelega AC. Effect of birth size and catch-up growth on adult blood pressure and carotid intima-media thickness. Horm Res Paediatr 2012;77:394-401. Stettler N, Zemel BS, Kumanyika S, Stallings VA. Infant weight gain and childhood overweight status in a multicenter, cohort study. Pediatrics 2002;109:194-9. Escribano J, Luque V, Ferre N, Mendez-Riera G, Koletzko B, Grote V, et al. Effect of protein intake and weight gain velocity on body fat mass at 6 months of age: the EU Childhood Obesity Programme. Int J Obes (Lond) 2012;36:548-53.

Overweight, Obesity, and Body Composition in 3.5- and 7-Year-Old Swedish Children Born with Marginally Low Birth Weight

7

THE JOURNAL OF PEDIATRICS



www.jpeds.com

Vol. -, No. -

Included MLBW children at 6 wk N = 285

Excluded = 4

Drop out before 3.5 years = 45

Abstained from 3.5 year visit = 4

Analyzed at 3.5 years = 232

Drop out = 31

Included control children at 3 years N = 95

Analyzed at 3.5 years = 95

Drop out = 25

Analyzed at 7 years = 205

Analyzed at 7 years = 70

Figure 1. Flow chart.

7.e1

Lindberg et al

BMI at 7 Univariate prediction Birth weight, kg SDS for weight Birth length, cm BMI at birth, kg/m2 Born SGA Born preterm Vaginal delivery Single gestation Female sex Exclusive breastfeeding 6 wk Any breastfeeding at 6 wk Any formula feeding at 6 wk Exclusive breastfeeding 6 mo Any breastfeeding at 6 mo Any formula feeding at 6 mo Mean daily formula intake 6 wk-6 mo Iron supplementation 6 wk-6 mo Weight change 0-6 wk (DSDS) Weight change 6-12 wk (DSDS) Weight change 12-19 wk (DSDS) Weight change 19 wk-6 mo (DSDS) Weight change 6-12 mo (DSDS) Height change 0-6 wk (DSDS) Height change 6-12 wk (DSDS) Height change 12-19 wk (DSDS) Height change 19 wk-6 mo (DSDS) Height change 6-12 mo (DSDS) Mother’s length, cm Mother’s weight, kg Mother’s BMI, kg/m2 Father’s weight, kg Father’s BMI, kg/m2 Mother born in Scandinavia Mothers age at birth of child Mother university education

B (95% CI) 1.45 ( 2.95; 0.059) 0.12 ( 0.077; 0.31) 0.23 ( 0.38; 0.07) 0.13 ( 0.16; 0.42) 0.21 ( 0.65; 0.23) 0.49 (0.05; 0.92) 0.012 ( 0.43; 0.46) 0.10 ( 0.38; 0.59) 0.047 ( 0.48; 0.39) 0.078 ( 0.37; 0.52) 0.14 ( 0.90; 0.62) 0.078 ( 0.52; 0.37) 0.43 ( 0.54; 1.4) 0.22 ( 0.25; 0.69) 0.21 ( 0.65; 0.24) 0.00 ( 0.002; 0.001) 0.36 ( 0.10; 0.81) 0.13 ( 0.37; 0.10) 0.62 (0.25; 1.0) 0.31 ( 0.17; 0.78) 0.72 (0.16; 1.3) 0.68 (0.34; 1.0) 0.016 ( 0.10; 0.07) 0.051 ( 0.16; 0.12) 0.012 ( 0.057; 0.08) 0.015 ( 0.036; 0.067) 0.054 ( 0.002; 0.11) 0.011 ( 0.045; 0.023) 0.041 (0.024; 0.059) 0.12 (0.068; 0.16) 0.012 ( 0.008; 0.033) 0.059 ( 0.02; 0.13) 0.054 ( 0.52; 0.62) 0.015 ( 0.034; 0.064) 0.44 ( 0.88; 0.010)

FMI at 7 2

r

P value

0.017 0.007 0.041 0.004 0.004 0.023 0.000 0.001 0.000 0.001 0.001 0.001 0.004 0.004 0.004 0.003 0.012 0.006 0.051 0.008 0.031 0.071 0.001 0.011 0.001 0.002 0.018 0.002 0.098 0.107 0.007 0.013 0.000 0.002 0.019

.060 .238 .004 .375 .349 .029 .958 .674 .833 .731 .715 .731 .384 .354 .366 .458 .122 .266 .001 .206 .012 <.001 .702 .142 .731 .555 .059 .518 <.001 <.001 .236 .122 .853 .544 .047

- 2015

Overweight, Obesity, and Body Composition in 3.5- and 7-Year-Old Swedish Children Born with Marginally Low Birth Weight

Table III. Univariate linear regression models assessing sociodemographic and perinatal predictors of BMI, FMI, and FFMI at 7 years of age in children born with MLBW FFMI at 7

B (95% CI)

2

r

P value

1.30 ( 2.5;-0.05) 0.08 ( 0.08; 0.24) 0.14 ( 0.27; 0.01) 0.020 ( 0.22; 0.26) 0.19 ( 0.56; 0.18) 0.29 ( 0.08; 0.65) 0.099 ( 0.47; 0.27) 0.086 ( 0.31; 0.49) 0.58 (0.22; 0.94) 0.042 ( 0.33; 0.41) 0.26 ( 0.89; 0.38) 0.042 ( 0.41; 0.33) 0.15 ( 0.65; 0.96) 0.020 ( 0.41; 0.37) 0.038 ( 0.41; 0.35) 0.00 ( 0.001; 0.000) 0.21 ( 0.17; 0.59) 0.17 ( 0.37; 0.031) 0.30 ( 0.02; 0.62) 0.33 ( 0.06; 0.72) 0.81 (0.36; 1.26) 0.54 (0.26; 0.83) 0.035 ( 0.14; 0.067) 0.036 ( 0.048; 0.12) <0.001 ( 0.085; 0.086) 0.031 ( 0.033; 0.096) 0.083 (0.014; 0.15) 0.010 ( 0.02; 0.04) 0.039 (0.02; 0.05) 0.095 (0.05; 0.1) 0.006 ( 0.01; 0.02) 0.040 ( 0.02; 0.10) 0.24 ( 0.24; 0.72) 0.007 ( 0.04; 0.05) 0.30 ( 0.67; 0.067)

0.021 0.005 0.024 0.000 0.005 0.012 0.001 0.001 0.049 0.000 0.003 0.000 0.001 0.000 0.000 0.002 0.006 0.014 0.017 0.014 0.061 0.068 0.002 0.004 <0.001 0.005 0.028 0.003 0.129 0.108 0.003 0.009 0.005 0.001 0.013

.042 .329 .029 .872 .311 .126 .599 .671 .002 .825 .432 .825 .705 .918 .842 .561 .278 .097 .067 .100 .001 <.001 .497 .399 .992 .335 .019 .478 <.001 <.001 .483 .203 .331 .745 .107

B (95% CI) 0.17 ( 0.98; 0.65) 0.11 (0.002; 0.21) 0.10 ( 0.19; 0.023) 0.15 ( 0.005; 0.30) 0.17 ( 0.40; 0.07) 0.38 (0.15; 0.62) 0.13 ( 0.10; 0.37) 0.003 ( 0.26; 0.25) 0.74 ( 0.95; 0.52) 0.084 ( 0.16; 0.32) 0.12 ( 0.29; 0.54) 0.084 ( 0.32; 0.16) 0.36 ( 0.16; 0.87) 0.22 ( 0.03; 0.46) 0.16 ( 0.40; 0.07) 0.00 ( 0.001; 0.00) 0.10 ( 0.15; 0.34) 0.026 ( 0.16; 0.10) 0.40 (0.20; 0.60) 0.045 ( 0.30; 0.21) 0.16 ( 0.47; 0.14) 0.093 ( 0.095; 0.28) 0.020 ( 0.14; 0.18) 0.112 ( 0.021; 0.24) 0.026 ( 0.11; 0.16) 0.058 ( 0.16; 0.039) 0.003 ( 0.11; 0.11) 0.024 ( 0.042; 0.006) 0.002 ( 0.008; 0.012) 0.020 ( 0.006; 0.047) 0.006 ( 0.005; 0.016) 0.017 ( 0.02; 0.06) 0.14 ( 0.45; 0.16) 0.007 ( 0.020; 0.033) 0.18 ( 0.4; 0.06)

r2

P value

0.001 0.020 0.032 0.018 0.010 0.051 0.006 0.000 0.194 0.002 0.002 0.002 0.010 0.016 0.009 0.001 0.003 0.001 0.074 0.001 0.006 0.005 <0.001 0.009 0.001 0.242 0.936 0.038 0.001 0.012 0.006 0.004 0.004 0.001 0.011

.687 .045 .012 .059 .163 .001 .264 .982 <.001 .488 .553 .488 .172 .083 .177 .709 .443 .693 <.001 .728 .290 .330 .811 .098 .695 .242 .963 .008 .726 .136 .283 .368 .357 .627 .142

B is the unstandardized regression coefficient with 95% CI using linear regression, and r is standardized coefficient.

ORIGINAL ARTICLES

7.e2

THE JOURNAL OF PEDIATRICS

Vol. -, No. -

www.jpeds.com



Table IV. Multivariate linear regression models assessing sociodemographic and perinatal predictors of BMI, FMI, and FFMI at 7 years of age in children born with MLBW BMI at 7 Multivariate prediction

B (95% CI)

2

BMI at birth, kg/m Born preterm Female sex Weight change 6-12 wk (DSDS) Weight change 19 wk-6 mo (DSDS) Weight change 6-12 mo (DSDS) Height change 6-12 wk (DSDS) Mother’s length, cm Mother’s weight, kg Mother’s BMI, kg/m2 Model summary R square:

2

r

FMI at 7 P value

NS NS NS 0.65 (0.27; 1.0)

0.055

.001

0.043

.003

NS 0.54 (0.18; 0.89) NS NS NS 0.10 (0.05; 0.15)

0.077 0.194

<.001

B (95% CI) NS NS 0.52 (0.16; 0.89) 0.37 (0.06; 0.68) 0.45 (0.00; 0.91) 0.32 (0.04; 0.61) NS NS 0.029 (0.01; 0.04) NS

r

FFMI at 7 2

0.040 0.026 0.018 0.024 0.071

P value

.005 .021 .050 .026 <.001

B (95% CI) 0.22 (0.08; 0.35) 0.52 (0.31; 0.72) 0.71 ( 0.91; 0.50) NS NS NS 0.14 (0.010; 0.27) 0.021 ( 0.04; 0.01) NS NS

0.233

r2

P value

0.038 0.091 0.179

.002 <.001 <.001

0.017 0.028

.035 .08

0.363

NS, not significant. Best predicting model achieved by stepwise multivariate linear regression including predictors from univariate regression with P < .20 (Table III). B is unstandardized regression coefficient with 95% CI using linear regression, and r is standardized coefficient.

7.e3

Lindberg et al