The interplay between birth weight and obesity in determining childhood and adolescent cardiometabolic risk

Abstract

Background: Birth weight (BW) is associated with risk of cardiometabolic disease (CMD) in adulthood, which may depend on the state of obesity, in particular if developed at a young age. We hypothesised that BW and a polygenic score (PGS) for BW were associated with cardiometabolic risk and related plasma protein levels in children and adolescents. We aimed to determine the modifying effect of childhood obesity on these associations.

Methods: We used data from The cross-sectional HOLBAEK Study with 4263 participants (median [IQR] age, 11.7 [9.2, 14.3] years; 57.1% girls and 42.9% boys; 48.6% from an obesity clinic and 51.4% from a population-based group). We gathered information on BW and gestational age, anthropometrics, cardiometabolic risk factors, calculated a PGS for BW, and measured plasma proteins using Olink Inflammation and Cardiovascular II panels. We employed multiple linear regression to examine the associations with BW as a continuous variable and performed interaction analyses to assess the effect of childhood obesity on cardiometabolic risk and plasma protein levels.

Findings: BW and a PGS for BW associated with cardiometabolic risk and plasma protein levels in childhood and adolescence. Childhood obesity modified the associations between BW and measures of insulin resistance, including HOMA-IR (βadj [95% CI per SD] for obesity: -0.12 [-0.15, -0.08]; normal weight: -0.04 [-0.08, 0.00]; Pinteraction = 0.004), c-peptide (obesity: -0.11 [-0.14, -0.08]; normal weight: -0.02 [-0.06, 0.02]; Pinteraction = 5.05E-04), and SBP SDS (obesity: -0.12 [-0.16, -0.08]; normal weight: -0.06 [-0.11, -0.01]; Pinteraction = 0.0479). Childhood obesity also modified the associations between BW and plasma levels of 14 proteins (e.g., IL15RA, MCP1, and XCL1; Pinteraction < 0.05).

Interpretation: We identified associations between lower BW and adverse metabolic phenotypes, particularly insulin resistance, blood pressure, and altered plasma protein levels, which were more pronounced in children with obesity. Developing effective prevention and treatment strategies for this group is needed to reduce the risk of future CMD.

Funding: Novo Nordisk Foundation (NNF15OC0016544, NNF0064142 to T.H., NNF15OC0016692 to T.H. and A.K., NNF18CC0033668 to S.E.S, NNF18SA0034956 to C.E.F., NNF20SA0067242 to DCA, NNF18CC0034900 to NNF CBMR), The Innovation Fund Denmark (0603-00484B to T.H.), The Danish Cardiovascular Academy (DCA) and the Danish Heart Foundation (HF) (PhD2021007-DCA to P.K.R, 18-R125-A8447-22088 (HF) and 21-R149-A10071-22193 (HF) to M.A.V.L., PhD2023009-HF to L.A.H), EU Horizon (668031, 847989, 825694, 964590 to A.K.), Innovative Health Initiative (101132901 for A.K.), A.P. Møller Foundation (19-L-0366 to T.H.), The Danish National Research Foundation, Steno Diabetes Center Sjælland, and The Region Zealand and Southern Denmark Health Scientific Research Foundation.

Keywords: Adolescent; Birth weight; Child; Inflammation; Obesity; Polygenic score.

Fig. 1

(a) Estimated regression β-effects (95% CI:s) for associations of birth weight (SD units) on CMR factors (SD units), pooling the obesity clinic and population-based groups, adjusted for gestational age (weeks), sex, age (for anthropometric traits), and childhood BMI SDS (for remaining CMR factors). Right-skewed risk factors were log-transformed, affecting all but BMI SDS, waist circumference SDS, waist-hip SDS, SBP SDS, and DBP SDS. (b–d) Top interactions between birth weight and obesity (pink) or normal weight (blue) on HOMA-IR, c-peptide, and TG/HDL-C. Sample sizes are specified in Supplementary Tables S1 and S3. Abbreviations: ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, CMR = cardiometabolic risk, DBP = diastolic blood pressure, GGT = gamma-glutamyltransferase, GLP-1 = glucagon-like peptide-1, HbA1c = haemoglobin A1c, HOMA-IR = homeostasis model assessment of insulin resistance, hs-CRP = high-sensitivity C-reactive protein, LDL-C = low-density lipoprotein cholesterol, SBP = systolic blood pressure, SDS = standard deviation score; TC = total cholesterol, TG/HDL-C = triglyceride to high-density lipoprotein cholesterol ratio, WBC = white blood cell count.

Fig. 2

(a) Estimated regression β-effects for associations of birth weight (SD units) on plasma protein levels (SD units), pooling the obesity clinic and population-based groups, adjusted for gestational age, sex, age, childhood BMI SDS, and blood sample storage time. Plasma proteins (non-normal distribution) were inverse-normal transformed. FDR-corrected P < 0.05 (blue = negative, red = positive), FDR-corrected P ≥ 0.05 (grey). Labelled proteins are FDR-corrected. (b–d) Top interactions between birth weight and obesity (pink) or normal weight (blue) on IL15RA, XCL1, and MCP1. Sample sizes are specified in Supplementary Tables S5 and S7. Abbreviations: IL15RA = interleukin-15 receptor subunit alpha, MCP1 = monocyte chemoattractant protein 1 or C–C motif chemokine 2; XCL1 = lymphotactin.

Fig. 3

(a) Estimated regression β-effects (95% CI:s) for associations of birth weight PGS (SD units) on CMR factors (SD units), pooling the obesity clinic and population-based groups, adjusted for gestational age, sex, age (for anthropometric traits), childhood BMI SDS (for remaining CMR factors), genotype batch, and the first five genetic PCs. Right-skewed CMR factors were log-transformed, affecting all but BMI SDS, waist SDS, waist-hip SDS, SBP SDS, and DBP SDS. (b) Estimated regression β-effects for associations of birth weight PGS (SD units) on plasma protein levels (SD units), pooling obesity clinic and population, adjusted for gestational age, sex, age, childhood BMI SDS, blood sample storage time, genotype batch, and the first five genetic PCs. Plasma proteins (non-normal distribution) were inverse-normal transformed. FDR-corrected P ≥ 0.05 (grey). Sample sizes are specified in Supplementary Tables S8 and S10. Abbreviations: ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, CMR = cardiometabolic risk, DBP = diastolic blood pressure, FABP2 = fatty acid binding protein 2, GGT = gamma-glutamyltransferase, GLP-1 = glucagon-like peptide-1, HbA1c = haemoglobin A1c, HDL-C = high-density lipoprotein cholesterol, HOMA-IR = homeostasis model assessment of insulin resistance, hs-CRP = high-sensitivity C-reactive protein, LDL-C = low-density lipoprotein cholesterol, LPL = lipoprotein lipase, PC = principal component, PGS = polygenic score, PRSS8 = prostasin, SBP = systolic blood pressure, SDS = SD score, TC = total cholesterol, TG = triglycerides, WBC = white blood cell count.