You are here

Article Text

Article menu

Download PDFPDF

Original article
What reduction in BMI SDS is required in obese adolescents to improve body composition and cardiometabolic health?
  1. Anna L Ford1,2,
  2. Linda P Hunt1,2,
  3. Ashley Cooper3,
  4. Julian P H Shield1,2
  1. 1Institute of Child Life and Health, Clinical Sciences South Bristol, University of Bristol, Bristol, UK
  2. 2Bristol Royal Hospital for Children, University Hospitals Bristol NHS Foundation Trust, Bristol, UK
  3. 3Department of Exercise, Nutrition and Health Sciences, University of Bristol, Bristol, UK
  1. Correspondence to Professor Julian P H Shield, University of Bristol and Bristol Royal Hospital for Children, Upper Maudlin Street, Bristol BS2 8AE, UK; j.p.h.shield{at}bristol.ac.uk

Abstract

Objective To study the impact of body mass index (BMI) SD score (SDS) improvement through lifestyle modification on metabolic risk and body composition over 12 months.

Design Prospective cohort study.

Setting Hospital outpatient weight management clinic in the UK.

Patients 88 adolescents (40 males, 86% Caucasian) of median age 12.4 years (range 9.1–17.4) and mean (SD) BMI SDS 3.23 (0.49).

Main outcome measures BMI at baseline and 12 months was adjusted for age and gender providing BMI SDS using British 1990 growth reference data. Body composition was measured by bioimpedance. A standard oral glucose tolerance test (OGTT) examined glucose metabolism. Fasting lipid profiles, high sensitivity C-reactive protein (HsCRP) and blood pressure (BP) were measured.

Results Reducing BMI SDS by ≥0.5 achieved significant improvements in important measures of body composition with mean waist circumference SDS reducing by 0.74 units and body fat SDS by 0.60 units, while also leading to significant reductions in key metabolic risk factors (triglycerides (−30%), low-density lipoprotein-cholesterol (−15%), HsCRP (−45%)). A lesser reduction of ≥0.25 improved insulin sensitivity, total cholesterol/high-density lipoprotein ratio and BP. The greater the BMI SDS reduction, the better the improvement seen in insulin sensitivity. The most insulinsensitive individuals at baseline were most likely to achieve BMI SDS changes of ≥0.5 regardless of baseline BMI SDS.

Conclusions Improvement in body composition and cardiometabolic risk can be seen with BMI SDS reductions of ≥0.25 in obese adolescents, while greater benefits accrue from losing at least 0.5 BMI SDS. The most insulin-sensitive individuals seem best able to effect these changes.

https://doi.org/10.1136/adc.2009.165340

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    While the apparent worldwide obesity epidemic continues unabated among children1,2 and clinicians try and identify successful methods of addressing this problem, there remains a paucity of data on the effectiveness of lifestyle interventions in terms of metabolic health. In adult practice, a loss in weight of at least 5% has been shown to positively impact on long-term health.3 Childhood obesity is also associated with a significantly increased risk of cardiovascular and endocrine morbidity, which in most cases only manifests as disease in adult life.4,5 While these increased risks in terms of hypertension, dyslipidaemia and glucose metabolism are documented in many obese cohorts,6 few studies have quantified change in these parameters secondary to weight loss. With a few notable exceptions,7 those studies that have been performed are on very small numbers of children or have only studied very short-term effects. In this study we took a cohort of obese children participating in a weight management programme of behavioural modification of diet and exercise, and followed them prospectively for 12 months in order to investigate the body composition and metabolic changes associated with change in body mass index (BMI) SD score (SDS).

    What is already known on this topic

    • Obesity is associated with reduced insulin sensitivity and increased markers of cardiovascular risk.

    • The minimal changes in body mass index necessary to improve adiposity and metabolic health have yet to be fully established.

    What this study adds

    • A minimum reduction in body mass index SD score (BMI SDS) of at least 0.25 is required to improve adiposity and metabolic health, with greater reductions eliciting greater benefits.

    • Interventions for childhood obesity from 9 years of age onwards should aim to achieve a minimum 0.25 BMI SDS reduction for clinical efficacy.

    Methods

    Patients

    Newly referred patients aged 9–17 years were recruited from the care of a childhood obesity clinic in Bristol (UK) between September 2004 and March 2007. The vast majority of referrals were from primary care with a minority from hospital subspecialities. The study was approved by United Bristol Hospitals Health Care Trust Ethics Committee (E5472). Parents provided written informed consent. A total of 106 children entered into the trial and 88 (83%) completed the 12-month evaluation and are included in this analysis. The median age at baseline was 12.4 years (range 9.1–17.4 years) and 40 (46%) were male. Seventy six (86%) were Caucasian, five Afro-Caribbean and six South-Asian, while one was of Hispanic origin.

    The patients in this study were also taking part in a randomised controlled trial in which they received either a 1-year programme based on behavioural modification of dietary and exercise habits delivered within a multidisciplinary clinic8 or a similar framework with the addition of a novel, computerised device to gradually slow speed of eating and reduce portion size.9 As both trial arms essentially received similar lifestyle modification therapy aimed solely at improving eating and physical activity behaviours, for the purpose of this study both trial arms were joined together and we explored prospectively across the whole cohort the impact of BMI SDS change on body composition, metabolic and cardiovascular risk factors.

    Measures

    Measurements were taken in a clinical investigations unit at the hospital with a parent in attendance, at baseline and on study completion (12 months).

    Body weight (kg) was measured using Seca (Birmingham, UK) digital scales to one decimal point in light indoor clothing and bare feet. Height was measured to the nearest 0.1 cm using a stadiometer. BMI was calculated as weight/height squared (kg/m2). Waist circumference was measured in centimetres to one decimal point using a standard anthropometric tape at the maximal abdominal girth. BMI SDS and waist circumference SDS were calculated using the British 1990 growth reference data (Child Growth Foundation10). While the best measure reflecting change in adiposity is debatable,11 we chose to examine change in BMI SDS as this has been cited in the National Obesity Observatory's Standard evaluation framework for weight management interventions as desirable for statistical evaluation.12

    Blood pressure (BP) was measured in the right arm while seated, after a 10 min rest using a random zero sphygmomanometer. The measurement protocol followed the British Hypertension Society guidelines,13 with three measurements taken, each separated by a 5 min rest. The first BP measurement was discarded and an average of the final two BP measurements taken.

    Fasting blood was taken for glucose, insulin, lipid profile and high sensitivity C-reactive protein (HsCRP). Samples for glucose and insulin were taken during an oral glucose tolerance test (OGTT) at 30, 60, 90 and 120 min after glucose ingestion (1.75 g/kg to a maximum of 75 g).

    Fasting insulin levels were determined using either an ELISA method (Dako K6219; Dako, Glostrup, Denmark) or a two-site immunoradiometric assay.14 Insulin resistance was measured by the homeostasis model assessment (HOMA-R) using the following equation15: Embedded Image

    Glucose area under the curve (AUC) during the OGTT was calculated for each child using the trapezium rule: (glucose AUC = (½×(G0+2×G30+2×G60+2×G90+G120)×0.5) mmol/l×h, where G0, G30, G60, G90 and G120 are the concentrations of glucose at time 0, 0.5, 1, 1.5 and 2 h, respectively). This measure provides an integrated assessment of postprandial glucose handling, thus supplying more information on metabolic status than isolated glucose measures during an OGTT.16

    Total body and truncal fat percentage were estimated using a Tanita bioimpedance segmental body composition analyser (model BC-418MA; Tanita, Arlington Heights, Illinois) and were measured to the nearest 0.1%. The BC-418MA has previously been validated against dual energy x-ray absorptiometry (DEXA) in children and adults of mixed race17 and against DEXA and air-displacement plethysmography (Bod Pod) in the paediatric population.18,19 Individual measures of percentage body fat were also converted to a SDS using the Child Growth Foundation's anthropometry software.20

    Patients were pubertally staged using the method of Tanner and Whitehouse by an experienced endocrinologist at baseline and at 12 months.21 It is well documented that puberty is associated with a transient state of insulin resistance with this reduction in insulin sensitivity associated with a compensatory increase in insulin secretion.17 Cross-sectional studies have demonstrated that insulin resistance increases at the beginning of puberty and peaks at Tanner stage 3, with reports of up to a 30% reduction in insulin sensitivity, while returning to prepubertal levels at the end of puberty.17,18

    In order to capture associated physiological changes in body composition and insulin sensitivity,18 pubertal status was categorised by puberty groups 1–3 as described below:

    1. Boys: pre to early puberty (G1–2, testes <6 ml); girls: prepubertal (<breast stage 2).

    2. Boys: mid-puberty and in puberty (G3, testes ≥6 to 25 ml, with height velocity >2 cm/year); girls: pubertal (≥breast stage 2±menarche in the last 2 years).

    3. Boys: postpubertal (testes >20 ml and height velocity <2 cm/year); girls: postpubertal (postmenarche of >2 years and height velocity <2 cm/year).

    Pubertal status was also categorised by the change in puberty status over the 12-month study period:

    1. Children who remained in puberty group 1 or 3.

    2. Children who progressed from puberty group 1 to 2 (or remained in group 2).

    3. Children who advanced from puberty group 2 to 3.

    Statistical analysis

    While weight loss may be looked at in several different ways, here we have chosen to divide the children into four subgroups on the basis of their change in BMI SDS:

    Group (i): participants with increased BMI SDS

    Group (ii): participants with BMI SDS decreased by >0 to <0.25

    Group (iii): participants with BMI SDS decreased by ≥0.25 to <0.5

    Group (iv): participants with BMI SDS decreased by ≥0.5.

    These groups were chosen on the basis of previous data from our clinic relating to the BMI SDS change needed to reduce adiposity22 and to facilitate comparisons with studies from Reinehr et al in Germany.23,–,25 We also felt that the relationships between the changes in BMI SDS and other factors may not be linear, making regression-type models unnecessarily complex.

    Statistical analyses were carried out using the software package SPSS (SPSS V.14; SPSS, Chicago, Illinois, USA). Positively skewed variables (ie, HOMA-R, fasting insulin, total cholesterol/high-density lipoprotein-cholesterol (HDL-C) ratio, triglycerides and HsCRP) were logarithmically transformed (base 10) prior to analysis with geometric means and ranges used for their data summary.

    One-way analyses of variance were used to compare the four BMI SDS change subgroups with respect to mean measures of the body composition, metabolic and cardiovascular risk factor variables. The subgroups were first compared at baseline, to see if any variable was indicative of a propensity to improve BMI SDS and then compared again at 12 months, with and without covariate adjustment for baseline differences,26 finally followed by a test of linear trend. Adjustment was made for gender difference and puberty as appropriate (gender and puberty for body composition and HOMA-R, gender for triglyceride).

    For completeness, mean changes over the 12-month period (or geometric means of the ratios for the variables needing transformation) were calculated for the four subgroups and regarded as significant if their 95% CIs excluded 0 (or 1 for the ratios). Changes are ‘final minus initial’ unless the direction is stated explicitly. A 5% level of significance was used throughout.

    Results

    Baseline characteristics

    Table 1 shows the baseline measurements together with their relationships with BMI SDS at baseline. Mean BMI SDS, truncal fat, body fat SDS, HOMA-R, lipid parameters, glucose levels during the OGTT, BP and HsCRP for males and females were similar, although females had a significantly greater mean waist circumference SDS and geometric mean for triglyceride levels. HsCRP was significantly positively correlated with BMI SDS (r=0.40, p<0.001), % body fat (r=0.33, p=0.002), waist circumference SDS (r=0.29, p=0.006) and % truncal fat (r=0.30, p=0.005).

    View this table:
    Table 1

    Baseline measurements and their relationship to BMI SDS at baseline

    Over the 12-month period, 19 individuals (22%) reduced their BMI SDS by ≥0.5, 20 (23%) reduced by ≥0.25 to <0.5, 29 (33%) reduced by >0 to <0.25, and 20 (23%) increased their BMI SDS. The mean initial and final values for these subgroups are shown in table 2. Initial BMI SDS did not influence ability to improve BMI SDS over the 12-month period. Children who showed greatest falls in BMI SDS tended to have lower BMI SDSs initially, although the correlation was not statistically significant (r=−0.11, p=0.3; table 2). Given the strong correlation between the baseline and final BMI SDS (r=0.84), for normal children a small positive correlation might have been expected.

    View this table:
    Table 2

    Subgroups formed according to change in BMI SDS over 12 months

    Although weight is not a good indicator of adiposity and subsequent health risks in children and adolescents, there has been much interest in the clinical setting as to the degree of weight loss required to achieve a reduction of ≥0.5 BMI SDS. In the present study, mean weight changes of +9.1 kg (range +0.3 to +20.5), +4.9 kg (range −3.4 to +11.5), −1.7 kg (range −11.7 to 4.0) and −10.3 kg (range −27.2 to +1.5) were associated with an increase in BMI SDS and reductions in BMI SDS of <0.25, ≥0.25 to <0.5, and ≥0.5, respectively, for this age group.

    Change in body composition and metabolic parameters at 12 months, after adjustment for baseline and covariates, with change in BMI SDS

    Body composition

    At 12 months, there were significant differences between the four subgroups with respect to mean waist circumference SDS, body fat SDS and % truncal fat (table 3). There was a trend of increasing reduction in all three variables with increasing BMI SDS loss (p<0.001), which might have been predicted from the relationship seen between these variables and BMI SDS at baseline (table 1).

    View this table:
    Table 3

    Change in body composition, insulin sensitivity and glucose metabolism with change in BMI SDS

    Insulin sensitivity

    HOMA-R at 12 months was similarly related to the degree of change in BMI SDS (p<0.001; table 3). Greatest reduction in HOMA-R was achieved with greatest reduction in BMI SDS loss.

    Glucose AUC

    There was a trend towards lower mean glucose AUC with greater reduction in BMI SDS over the 12-month period (p=0.009; table 3). Given that the glucose AUC was not related to BMI SDS at baseline (table 1), this result might not have been predicted.

    Dyslipidaemia

    Mean HDL-C at 12 months was not significantly associated with the range of changes in BMI SDS over the year (table 4). Mean low-density lipoprotein-cholesterol (LDL-C), total cholesterol/HDL-C ratio and triglycerides at 12 months were significantly related to change in BMI SDS. Significant changes were observed within subgroups (iii) and (iv).

    View this table:
    Table 4

    Change in lipids, blood pressure and HsCRP with change in BMI SDS

    Blood pressure

    Mean 12-month systolic and diastolic BP were significantly associated with change in BMI SDS (p=0.045 and p=0.007 for systolic and diastolic, respectively; table 4). Reductions were observed for subgroups (iii) and (iv).

    Low-grade inflammation (HsCRP)

    HsCRP at 12 months was significantly associated with change in BMI SDS (p=0.036; table 4). The change in HsCRP was only significant for subgroup (iv).

    Discussion

    This study sought to explore the changes in key metabolic risk parameters and body composition that accompany various degrees of BMI SDS improvement through lifestyle interventions alone. Our hypothesis was that a minimum reduction in BMI SDS of 0.5 or more would be needed to improve measures of body fatness and metabolic risk based on previous data from our clinic relating to reducing adiposity22 and that of Reinehr et al.24

    In this sample we observed no improvement in mean waist circumference SDS or body fat SDS with a BMI SDS reduction of >0 to <0.25. Improvement by ≥0.25 was associated with small reductions in mean waist circumference SDS and body fat SDS. Concomitant with these findings, we observed significant improvements in the key metabolic risk factors triglyceride, LDL-C and HsCRP levels (30%, 15% and 45%, respectively) with a minimum loss of ≥0.5 BMI SDS. Improvements were also seen (13%, 12% and 11%) for losses of between ≥0.25 and <0.5. A similar picture emerged with BP with losses of ≥0.25 BMI SDS associated with significant reduction.

    Previously, using a broader age range (2–18 years) we documented a trend for those with better insulin sensitivity measured by HOMA-R to achieve greater BMI SDS improvement over 1 year, although the results were not statistically significant.8 In the present study a similar picture developed, with those achieving the best improvements being more insulin sensitive compared to all three other groups regardless of similar mean baseline BMI SDS.

    Our findings are complementary to those of Reinehr et al,24 who reported significant improvements in triglycerides, insulin sensitivity (HOMA-R) and systolic and diastolic BP with a minimum reduction of ≥0.5 BMI SDS, although our study suggests that a value of ≥0.25 BMI SDS may be the minimum requirement for improvement in BP, some lipid parameters and insulin sensitivity.

    Cross-sectional studies show that obesity in childhood is associated with chronic inflammation.8 HsCRP, a marker of low-grade inflammation, is produced in the liver in reaction to interleukin 6, a proinflammatory cytokine produced in adipocytes especially from visceral fat. Adiposity is the main determinant of low-grade inflammation with elevated HsCRP linked to increased risk of cardiovascular disease27 and type 2 diabetes.28 We have shown that an improvement in BMI SDS of ≥0.5 is associated with a clinically significant reduction in low-grade, obesity-driven inflammation.

    Substantial evidence now exists which demonstrates the importance of obesity in childhood in creating the metabolic milieu for cardiovascular disease and type 2 diabetes in young adulthood. This study has demonstrated significant improvements in body composition, insulin sensitivity, lipids, diastolic BP and low-grade inflammation with reduction in BMI SDS and adds to our understanding of the minimum requirements for successful weight management in childhood obesity.

    The fact that body fat SDS was significantly improved with a BMI SDS reduction of ≥0.25 and that at this level indices of insulin sensitivity also improved, suggests that improving the level of adiposity is key to improved metabolic health. Moreover, we suggest that any intervention should aim to improve BMI SDS in obese children by at least 0.25 for clinical effectiveness, while greater benefits may be accrued with improvements of 0.5 or more. We suggest that those evaluating trials in children use a reduction in BMI SDS of ≥0.25 as a minimum for clinical effectiveness and that future studies are judged on ability to attain this level as a minimum criterion for success.

    Acknowledgments

    We thank the patients and their families for participating in this study.

    References

    1. ↵
      1. Jackson-Leach R,
      2. Lobstein T
      . Estimated burden of paediatric obesity and co-morbidities in Europe. Part 1. The increase in the prevalence of child obesity in Europe is itself increasing. Int J Pediatr Obes 2006;1:26–32.
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Lobstein T,
      2. Jackson-Leach R
      . Child overweight and obesity in the USA: prevalence rates according to IOTF definitions. Int J Pediatr Obes 2007;2:62–4.
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Klein S,
      2. Burke LE,
      3. Bray GA,
      4. et al
      . Clinical implications of obesity with specific focus on cardiovascular disease: a statement for professionals from the American Heart Association Council on Nutrition, Physical Activity and Metabolism: endorsed by the American College of Cardiology Foundation. Circulation 2004;110:2952–67.
      OpenUrlAbstract/FREE Full Text
    4. ↵
      1. Freedman DS,
      2. Dietz WH,
      3. Tang R,
      4. et al
      . The relation of obesity throughout life to carotid intima-media thickness in adulthood: the Bogalusa Heart Study. Int J Obes Relat Metab Disord 2004;28:159–66.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Srinivasan SR,
      2. Myers L,
      3. Berenson GS
      . Predictability of childhood adiposity and insulin for developing insulin resistance syndrome (syndrome X) in young adulthood: the Bogalusa Heart Study. Diabetes 2002;51:204–9.
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. Lobstein T,
      2. Jackson-Leach R
      . Estimated burden of paediatric obesity and co-morbidities in Europe. Part 2. Numbers of children with indicators of obesityrelated disease. Int J Pediatr Obes 2006;1:33–41.
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Reinehr T,
      2. de Sousa G,
      3. Toschke AM,
      4. et al
      . Long-term follow-up of cardiovascular disease risk factors in children after an obesity intervention. Am J Clin Nutr 2006;84:490–6.
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Sabin MA,
      2. Ford A,
      3. Hunt L,
      4. et al
      . Which factors are associated with a successful outcome in a weight management programme for obese children? J Eval Clin Pract 2007;13:364–8.
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Blass E
      1. Bergh CSM,
      2. Shield J,
      3. Hellers G,
      4. et al
      . A framework for the treatment of obesity: early support. In: Blass E ed. Obesity: causes, mechanisms, and prevention. Sunderland, Massachusetts, USA: Sinauer Associates Inc, 2008:399–425.
    10. ↵
      1. Cole TJ,
      2. Freeman JV,
      3. Preece MA
      . Body mass index reference curves for the UK, 1990. Arch Dis Child 1995;73:25–9.
      OpenUrlAbstract/FREE Full Text
    11. ↵
      1. Cole TJ,
      2. Faith MS,
      3. Pietrobelli A,
      4. et al
      . What is the best measure of adiposity change in growing children: BMI, BMI %, BMI z-score or BMI centile? Eur J Clin Nutr 2005;59:419–25.
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Roberts K,
      2. Cavill N,
      3. Rutter H
      . Standard Evaluation Framework for weight management interventions: National Obesity Observatory, 2009.
    13. ↵
      1. O'Brien E,
      2. Petrie J,
      3. Littler W,
      4. et al
      . An outline of the revised British Hypertension Society protocol for the evaluation of blood pressure measuring devices. J Hypertens 1993;11:677–9.
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Sobey WJ,
      2. Beer SF,
      3. Carrington CA,
      4. et al
      . Sensitive and specific two-site immunoradiometric assays for human insulin, proinsulin, 65-66 split and 32-33 split proinsulins. Biochem J 1989;260:535–41.
      OpenUrlAbstract/FREE Full Text
    15. ↵
      1. Matthews DR,
      2. Hosker JP,
      3. Rudenski AS,
      4. et al
      . Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985;28:412–19.
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Sabin MA,
      2. Hunt LP,
      3. Ford AL,
      4. et al
      . Elevated glucose concentrations during an oral glucose tolerance test are associated with the presence of metabolic syndrome in childhood obesity. Diabet Med 2008;25:289–95.
      OpenUrlCrossRefPubMed
    17. ↵
      1. Pietrobelli A,
      2. Rubiano F,
      3. St-Onge MP,
      4. et al
      . New bioimpedance analysis system: improved phenotyping with whole-body analysis. Eur J Clin Nutr 2004;58:1479–84.
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Pietrobelli A,
      2. Malavolti M,
      3. Fuiano N,
      4. et al
      . The invisible fat. Acta Paediatr Suppl 2007;96:35–8.
      OpenUrlPubMed
    19. ↵
      1. Pietrobelli A,
      2. Tatò L
      . Body composition measurements: from the past to the future. Acta Paediatr Suppl 2005;94:8–13.
      OpenUrlCrossRefPubMed
    20. ↵
      1. McCarthy HD,
      2. Cole TJ,
      3. Fry T,
      4. et al
      . Body fat reference curves for children. London: Int J Obes 2006;30:598–602.
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. Tanner JM
      . Normal growth and techniques of growth assessment. Clin Endocrinol Metab 1986;15:411–51.
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Hunt LP,
      2. Ford A,
      3. Sabin MA,
      4. et al
      . Clinical measures of adiposity and percentage fat loss: which measure most accurately reflects fat loss and what should we aim for? Arch Dis Child 2007;92:399–403.
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Reinehr T,
      2. Kiess W,
      3. Kapellen T,
      4. et al
      . Insulin sensitivity among obese children and adolescents, according to degree of weight loss. Pediatrics 2004;114:1569–73.
      OpenUrlAbstract/FREE Full Text
    24. ↵
      1. Reinehr T,
      2. Andler W
      . Changes in the atherogenic risk factor profile according to degree of weight loss. Arch Dis Child 2004;89:419–22.
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. Reinehr T,
      2. Stoffel-Wagner B,
      3. Roth CL,
      4. et al
      . High-sensitive C-reactive protein, tumor necrosis factor α, and cardiovascular risk factors before and after weight loss in obese children. Metabolism 2005;54:1155–1161.
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Vickers AJ,
      2. Altman DG
      . Statistics notes: analysing controlled trials with baseline and follow up measurements. BMJ 2001;323:1123–4.
      OpenUrlFREE Full Text
    27. ↵
      1. Bassuk SS,
      2. Rifai N,
      3. Ridker PM
      . High-sensitivity C-reactive protein: clinical importance. Curr Probl Cardiol 2004;29:439–93.
      OpenUrlPubMedWeb of Science
    28. ↵
      1. Pradhan AD,
      2. Manson JE,
      3. Rifai N,
      4. et al
      . C-reactive protein, interleukin 6, and risk of developing type 2 diabetes mellitus. JAMA 2001;286:327–34.
      OpenUrlCrossRefPubMedWeb of Science

    Footnotes

    • Funding The researchers are grateful to the BUPA Foundation for supporting this study.

    • Competing interests None.

    • Patient consent Parental consent obtained.

    • Ethics approval This study was conducted with the approval of the United Bristol Hospitals Health Care Trust Ethics Committee (E5472).

    • Provenance and peer review Not commissioned; externally peer reviewed.