Source: Archives of Disease in Childhood Preprint Date: September 5, 2007 URL: http://adc.bmj.com/cgi/rapidpdf/adc.2006.110502v1 Is chronic fatigue syndrome (CFS/ME) heritable in children, and if so why does it matter? Esther Crawley(1) and George Davey Smith(2) 1 Corresponding author: Centre for Child and Adolescent health, Hampton House, Cotham Hill, Bristol, BS6 6JS, UK, Email: Esther.crawley@bristol.ac.uk, Tel: 0117 331 0892/3, Fax: 0117 331 1891 2 MRC centre for Causal Analyses in Translational Epidemiology (CaiTE), Department of Social Medicine, University of Bristol, UK ADC Online First, published on September 5, 2007 Abstract This article looks at the evidence for the heritability of Chronic Fatigue Syndrome or ME (CFS/ME). The role of genes and the environment and their potential interaction is discussed and the problems with studies in this area are explored. One of the main difficulties with studying CFS/ME is the fact that it is likely to be a heterogeneous condition. We need a clear definition of CFS/ME in children and sample sizes for genetic studies need to be much larger if we are to make an impact in this important area. In the meantime however, paediatricians shouldn't be surprised when they see a family where CFS/ME occurs in more than one person. Key words: CFS/ME fatigue genes heritable Chronic fatigue syndrome or ME (CFS/ME) is surprisingly common in children with a prevalence of between 0.19 and 2% based on telephone surveys in the UK and the USA 1 2 3. Lifetime prevalence (up to 30 years old) of self-reported CFS/ME, uncorroborated by a physician, of 0.8% has been reported from the 1970 British Birth Cohort4. Lifetime prevalence (age 8 to 17) of disabling fatigue of 3 months and 6 months of 2.34% and 1.29% has been reported from a longitudinal cohort of twins5. This means that almost all paediatricians reading this article will have seen and managed children with CFS/ME. Some paediatricians will have noted a family history of CFS/ME and may have wondered whether this was due to genetic heritability or an environmental factor. The causes of CFS/ME have long been debated which has not necessarily been helpful to the clinical management of children with CFS/ME6-8. This article examines the evidence base for the genetic heritability of CFS/ME. This is an important area for paediatricians to know about as it will inform our discussions with children, young people and their families. Familial aggregation The familial aggregation of CFS/ME was first described 16 years ago in Lyndonville, New York state. In this study, a questionnaire asking about symptoms of CFS/ME and possible risk factors (allergies, asthma, risk factors for infection and family history) were distributed to all 914 students at the Lyndonville Central School. Having a family member with CFS/ME was a strong predictor of CFS/ME, with a risk ratio of 35.9 9. It is not clear from this study whether the authors only asked about first degree relatives or "any" relative. At this stage, the authors concluded that this could be due to either the transmission of an infectious agent, the presence of genetic factors or some environmental factor yet to be defined. In another small study 3 years later, 50% of children with CFS/ME had a family history of CFS/ME10 however once again, it is not clear whether these were first degree or "any" relatives. Twin Studies Investigators have used twin methodology to determine how much of the family aggregation described is due to genetic factors and how much to the environment. Twin studies in adults have shown consistently higher concordance rates in monozygotic twins compared to dizygotic twins for CFS/ME with the monozygotic correlation usually at least twice that of the dizygotic correlation 11-14(for example 0.55 and 0.19 14 and 0.43 and 0.16 12. As environmental factors (for example infections or parental up bringing) during early life should have been similar within monozygotic and dizygotic twin pairs, the difference between the monozygotic twins who share 100% of the DNA and the dizygotic twins, who will on average share 50%, can be used to calculate the contribution made by genetic heritability. The difference in concordance rates increases with increasingly stringent case definition in adults 14 . This provides evidence that there is a core CFS/ME group among whom genetic factors play an important role. Repeating the studies in children is difficult for many reasons, not least the fact that there is currently no agreed criteria for CFS/ME in children. The Royal College of Paediatrics and Child Health do not recommend a minimum time frame before making a diagnosis in children but merely require that they have fatigue which is disabling and without another identifiable cause6. Having said that, twin studies investigating fatigue in children are consistent with those in adults. In the UK, carers of 670 twin pairs were questioned about disabling fatigue of more than a week and more than a month. In both cases the concordance was higher in monozygotic twins compared to dizygotic twins (0.81 Vs 0.59 for one week, 0.75 Vs 0.47 for one month)15. This suggests that the genetic contribution to the experience of disabling fatigue is high. This study was extended, length of time for fatigue was defined as a few days, more than a week, more than 1, 3 or 6 months and the relationship between fatigue and depression was investigated in 1468 pairs of twins16. For short duration fatigue, genetic heritability was high and no shared environmental contribution (from environmental factors shared by twins, as many background family-based factors would be expected to be) could be detected. However, for prolonged fatigue, there appeared to be substantial environmental influences and a modest genetic contribution. Is the heritability of fatigue due to depression? This begs the question as to whether the fatigue being measured is actually part of a mood disorder such as depression, also known to be heritable. In the paediatric study described above, for both the short and prolonged fatigue the majority of the genetic and environmental variance was shown to be specific to disabling fatigue and different from factors contributing to depression16. This is consistent with an adult study where depression, anxiety and psychological distress were measured in 1004 adult twin pairs 12. Structural equation modelling suggested that the familial aggregation for fatigue was largely due to additive genetic factors, which were not shared by the other measures of psychological distress. In other words, chronic fatigue is heritable and this heritability is aetiologically distinct from psychological distress. These studies are important as they suggest that fatigue is not just a variant of depression. CFS/ME therefore seems to at least partially obey Turkheimer's "laws of behavioural genetics" (see table 4)17. Many conditions that obey these laws are, however, also environmentally determined when changes in risk over time and between populations are considered. For example, obesity shows high heritability - typically around 70% - but at the same time we have witnessed a dramatic increase in prevalence, clearly implicating changing environmental factors. Estimates of genetic and environmental contributions to disease risk from twin studies cannot, of course, reflect the contribution of near-universal environmental exposures, or ones that are changing across all strata of society, such as the balance between energy expenditure and energy intake 18. Furthermore, the assumption that apparently shared exposures - such as parental divorce - have the same effect on two individuals is not necessarily justified, as prior experiences and individual characteristics may lead to such exposures producing different effects in different people17. Association studies As in many conditions with evidence of substantial heritability - and some without - much recent work has involved candidate gene association studies in population based case-control studies. The candidate genes have been chosen to reflect the postulated mechanisms through which CFS/ME arises, for example, variants in the hypothalamo-pituitary adrenal axis (HPA). If variants related to different HPA activity profiles were associated with risk of CFS/ME, this would provide good evidence that potentially modifiable environmental influences on HPA function are causally related to the risk of developing CFS/ME. The reasoning here is complex but important. If we continue to use the HPA axis as an example, we can see that potential environmental influences on a child's HPA function, say parental discord, or even direct measures of HPA function, will be associated with a large array of other potential causes of CFS/ME. Therefore the associations will be confounded by these factors and it is not possible to isolate the aetiological processes that could be modified to prevent or treat CFS/ME from spurious confounders. Furthermore, the disease process involved in CFS/ME may alter HPA function, rather than disturbed HPA function leading to CFS/ME, a situation referred to as reverse causation. Genotypes are not, however, associated with confounding factors, nor do disease processes alter germline genetic variants. Thus a genotype related to HPA function can be taken as a marker of such function that is neither confounded nor susceptible to reverse causation, and thus provides better evidence of causality. This approach to using genetic data has been referred to as Mendelian Randomization" and applied in other clinical settings19;20. This approach to genetic epidemiology utilises genetic data to understand potentially modifiable causes of disease, rather than to provide a basis for predictive genetic testing. Genetic association studies of CFS/ME to date have only been carried out in adults. The findings to date have been mixed, and are summarised below. HPA axis: Variants in the Corticosteroid Binding Globulin (CBG) gene have been described in an isolated kindred and the presence was associated with idiopathic chronic fatigue and CFS/ME (CDC defined)21. However when investigated in a case control study, there was no robust difference in CBG prevalence in a group of 248 patients with CFS/ME compared to 248 controls22. Further single nucleotide polymorphisms (SNPs) of interest in the HPA axis are described below. HLA genotype: Whether HLA genotypes are associated with CFS/ME or not is controversial. One study showed no significant associations compared to controls23 and one study suggested an increased frequency of HLA-DQA1*0124. Both studies were small (patient group 58 and 49 respectively) and in the second study, the association was not corrected for multiple tests. Cytokines: One study in Italian patients with CFS/ME investigated promoter polymorphisms of IL-10, IL-6 and IFN and demonstrated an increase in TNF genotypes compared to controls with a decrease in IFN low producers25. However this is only a single study and to date the findings have not been replicated. The bottom line regarding association studies in this field has to be that they are massively underpowered, thus it is not unexpected that no robust findings have emerged. Realistic assumptions about the effect sizes that might exist between common genetic variants and a complex non homogenous illness like CFS/ME would suggest that studies at least an order of magnitude larger than current studies are required if they are to be informative26. The effort is worth it, however, as they have the potential to inform us about modifiable risk factors for CFS/ME, through utilisation of the Mendelian Randomization approach discussed above. Furthermore, if adequately powered studies yield validated associations between genetic variants and risk of CFS/ME then this suggests that the CFS/ME patient group studied at the very least contacts a subgroup with a coherent diagnosis, since if the diagnosis is actually completely incoherent, no true associations could be found. In addition to studies of genetic variants and CFS/ME risk, there have been studies of differences in gene expression - i.e. levels of RNA transcription products - in patients with CFS/ME compared to controls. These studies are informative in that they tell us about potential mechanisms of disease development, but ultimately gene expression data are measures of bodily state and reaction, unlike genotype data. As patients who receive a diagnosis of CFS/ME are distinguished from those who don't by phenotypic differences, it is expected that differences in gene expression will exist. Problems arise however because sedentary activity itself, yet alone differences in medication or diet may well alter gene expression. Gene expression studies are therefore less informative about the ontology of CFS/ME than are genetic association studies, although the combination of gene expression and genetic variant studies can, when they provide congruent information, be particularly powerful 27. Three relatively small studies have described altered gene expression in patients with CFS/ME compared to controls28-30. Each study identified genes with differential expression patterns (see table 1). However, the associations identified have not been replicated and each study has described different biological pathways. This is likely to be due to the small sample size in what is likely to be a non-homogenous illness together with the problems discussed above. More recently, further genetic information has emerged from an elaborate investigation involving 227 residents of Wichita31, carried out by the Centers for Disease Control and Prevention. During a 2 day hospital stay, data were collected on the participants psychiatric status, sleep characteristics and cognitive function. Biological samples were collected to measure neuroendocrine status, autonomic nervous system function, systemic cytokines and peripheral blood gene expression. 20 investigators from the disciplines of medicine, mathematics, biology, engineering and computer science were then given the task of analysing and interpreting the data over 6 months. The results give a clearer understanding of CFS/ME but perhaps more importantly describe a novel way of analysing large datasets in a complex illness. In the first part of the study, the CDC used principle components methodology and latent class analysis (see table 3 for explanation) to divide a group of women with fatigue, chronic fatigue syndrome and controls into five separate groups which could be explained by their symptoms, clinical and laboratory findings 31;32. They then used several methods to analyse gene expression and came up with several different results (see table 2 for summary of gene expression results). As well as looking at gene expression, the CDC group also looked at genomic DNA. One group tested whether single nucleotide polymorphisms (SNPs, see table 3) could distinguish the five groups described above, whilst another group investigated whether single nucleotide polymorphism (SNP) profiles could be used to predict whether a patient had CFS/ME or not. In the first study, they found that three of the classes described on clinical and laboratory grounds were associated with genes involved in the HPA axis function or mood related neurotransmitter systems (monoamine oxidase A and B and tryptophan hydroxylase)33. In the second study, one particular combination of 5 SNPs was able to identify patients versus controls (OR 8.94, p<0.00001)34. The most important genes for this study coded for neuronal tryptophan hydroxylase (TPH2), catechol-o-methlyltransferase, and a glucocorticoid receptor, NR3C1. (TPH2 is the rate-limiting enzyme in the synthesis of serotonin, itself a precursor of melatonin). Genes for corticotropin releasing hormone receptors were also implicated. Although potentially interesting in terms of the biological pathways involved, these studies are heavily underpowered and replications are required before the findings are given any credence. In summary, there is now increasingly strong evidence that CFS/ME is heritable. It seems likely that the heritability contributes to the experience of fatigue as well as to the development of CFS/ME itself. It should therefore not be surprising that CFS/ME runs in families and paediatricians will see families where more than one person is affected. Although there is some agreement that a model involving the joint action of genes and environment is required, there is currently little agreement on the actual genes and environmental factors involved. The fact that CFS/ME is heterogenous hampers research in this area. However, understanding more about the different biological processes and the genes involved in CFS/ME is an important step in understanding and developing methods for preventing or treating this important condition. Genetic studies may forward this research and provide an anchorage point around which CFS/ME can be understood and the modifiable risk factors that influence risk can be determined. Acknowledgements Thanks to The Linbury Trust for funding Dr Crawley's research time. Competing interests: Dr Crawley is a medical advisor to the Association of Young people with ME (AYME) Funding: Dr Crawley and her research team, are in part, funded by The Linbury Trust and the Bath Unit for Research in Paediatrics (BURP). Licence for publication: The corresponding author has the right to grant on behalf of all authors and does grant on behalf of all authors, an exclusive licence on a worldwide basis to the BMJ publishing group ltd and its licensees to permit this article (if accepted) to be published in Archives of Disease in Childhood editions and any other BMJPGL products to exploit all subsidery rights, as set out in our licence (http://adc.bmjjournals.com/ifora/licence.dtl). Tables Table 1: Gene expression studies ------------------------------------------------------------------------------- Author (year) Patients and controls Number of genes Differentially tested expressed genes ------------------------------------------------------------------------------- Vernon 2002 30 5 patients, 17 controls 1764 genes CMRF35 antigen, IL-8, HD protein Powell 2003 28 7 patients, 4 controls Differential Cathepsin C and display MAIL1, TNFalpha, Moesin Kaushik 2005 29 25 patients, 25 controls 9522 genes ABCD4, PRKCL1, MRPL23, CD2BP2, KHSRP, BRMS1, GABARAPL1 (IL- 10RA| ) V ------------------------------------------------------------------------------- Table 2: Gene expression studies from CDC data ---------------------------------------------------------------------------------------------- Author Method Result ---------------------------------------------------------------------------------------------- Fang, Selected patients with most or 24 identified genes, 11 common pathways with 2006 35 least depression or fatigue. the following functions: immune responses and Principle components analysis apoptosis; metal ion binding, ion transport and to select genes which were ion channel activity; cascade, signal common to both pathways and transduction, cell-cell signalling, regulation of separated patients with cell growth; neuronal activity genes; a CFS/ME. photosensitive gene and DNA fragmentation and TNF signalling pathway Whistler Quantitative trait analysis was 839 genes identified which mapped to the 2006 36 used to identify correlation of following cellular functions: metabolism, gene expression with fatigue transcriptional regulation and cell signalling pathways. Carmel A computational approach was 32 and 26 genes which can discriminate 2006 37 developed to identify genes between the 5 and 6 class solution. Most that could discriminate the classes were distinguished by: ZNF350, classes identified using SLC1A6, FBXO7, VAC14 and some by PTH2 microarray expression data on and TCL1A 15,315 genes Broderick Applied principle components 39 genes, single most influential gene was 2006 38 analysis (PCA) to 59 variables Sestrin1, also, calcium channel activity and which made up CFS/ME and transciption peptidase MBTPS1, protein kinase then to rotated gene expression C-like 1 (PRKCL-1) and KHSRP. data to find gene expression that correlates with illness ---------------------------------------------------------------------------------------------- Table 3: Explanation of terms used in this paper ------------------------------------------------------------------------------- Term Definition or explanation ------------------------------------------------------------------------------- Principle components methodology A method used to find a few combinations of 39 variables, called components, that adequately explain the overall observed variation, and thus to reduce the complexity of the data Latent class analysis A statistical method for finding subtypes of related cases (latent classes) from multivariate categorical data Single Nucleotide polymorphism DNA sequence variation occurring when a single nucleotide A,T,C or G in a genome differs between members of a species ------------------------------------------------------------------------------- Table 4: Turkheimers laws of behavioural genetics17 ------------------------------------------------------------------------------- Law 1 Everything is heritable Law 2 The environment effect of being raised in the same family is substantially smaller than the genetic effect and is often close to zero Law 3 Most behavioural variability remains in the error term after genetic effects and the effects of being raised in the same family have been accounted for ------------------------------------------------------------------------------- Reference List 1 Chalder T, Goodman R, Wessely S, Hotopf M, Meltzer H. Epidemiology of chronic fatigue syndrome and self reported myalgic encephalomyelitis in 5-15 year olds: cross sectional study. BMJ. 2003; 327(7416):654-655. 2 Jordan KM, Ayers PM, Jahn SC, Taylor KK, Huang CF, Richman J et al. Prevelance of Fatigue and Chronic Fatigue Syndrome-Like Illness in Children and Adolescents. J Chronic Fatigue Syndrome. 2000; 6(1):3-21. 3 Jones JF, Nisenbaum R, Solomon L, Reyes M, Reeves WC. Chronic fatigue syndrome and other fatiguing illnesses in adolescents: a population-based study. J Adolesc Health. 2004; 35(1):34-40. 4 Viner R, Hotopf M. Childhood predictors of self reported chronic fatigue syndrome/myalgic encephalomyelitis in adults: national birth cohort study. BMJ. 2004; 329(7472):941. 5 Farmer A, Fowler T, Scourfield J, Thapar A. Prevalence of chronic disabling fatigue in children and adolescents. Br J Psychiatry. 2004; 184:477-481. 6 Royal College of Paediatrics and Child Health. Evidence Based Guideline for the Management of CFS/ME (Chronic Fatigue Syndrome/Myalgic Encephalopathy) in Children and Young People. 2004. London. Ref Type: Report 7 A report of the CFS/ME working group. A report to the chief medical officer of an independant working group. 2002. Ref Type: Report 8 Prins JB, van der Meer JW, Bleijenberg G. Chronic fatigue syndrome. Lancet. 2006; 367(9507):346-355. 9 Bell KM, Cookfair D, Bell DS, Reese P, Cooper L. Risk factors associated with chronic fatigue syndrome in a cluster of pediatric cases. Rev Infect Dis. 1991; 13 Suppl 1:S32-S38. 10 Bell DS, Bell KM, Cheney PR. Primary juvenile fibromyalgia syndrome and chronic fatigue syndrome in adolescents. Clin Infect Dis. 1994; 18 Suppl 1:S21-S23. 11 Hickie IB, Bansal AS, Kirk KM, Lloyd AR, Martin NG. A twin study of the etiology of prolonged fatigue and immune activation. Twin Res. 2001; 4(2):94-102. 12 Hickie I, Kirk K, Martin N. Unique genetic and environmental determinants of prolonged fatigue: a twin study. Psychol Med. 1999; 29(2):259-268. 13 Hickie I, Bennett B, Lloyd A, Heath A, Martin N. Complex genetic and environmental relationships between psychological distress, fatigue and immune functioning: a twin study. Psychol Med. 1999; 29(2):269-277. 14 Buchwald D, Herrell R, Ashton S, Belcourt M, Schmaling K, Sullivan P et al. A twin study of chronic fatigue. Psychosom Med. 2001; 63(6):936-943. 15 Farmer A, Scourfield J, Martin N, Cardno A, McGuffin P. Is disabling fatigue in childhood influenced by genes? Psychol Med. 1999; 29(2):279-282. 16 Fowler TA, Rice F, Thapar A, Farmer A. Relationship between disabling fatigue and depression in children: genetic study. Br J Psychiatry. 2006; 189:247-253. 17 Turkheimer E. Spinach and Ice cream: Why social science is so difficult. In: DiLally LF, ed. Behavior Genetics Principles -Perspectives in development, Personality and Psychopathology. Washington: American Psychological Association, 2004: 161-189. 18 Davey SG, Ebrahim S. Epidemiology--is it time to call it a day? Int J Epidemiol. 2001; 30(1):1-11. 19 Davey SG, Ebrahim S. 'Mendelian randomization': can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol. 2003; 32(1):1-22. 20 Davey SG, Ebrahim S. What can mendelian randomisation tell us about modifiable behavioural and environmental exposures? BMJ. 2005; 330(7499):1076-1079. 21 Torpy DJ, Bachmann AW, Grice JE, Fitzgerald SP, Phillips PJ, Whitworth JA et al. Familial corticosteroid-binding globulin deficiency due to a novel null mutation: association with fatigue and relative hypotension. J Clin Endocrinol Metab. 2001; 86(8):3692-3700. 22 Torpy DJ, Bachmann AW, Gartside M, Grice JE, Harris JM, Clifton P et al. Association between chronic fatigue syndrome and the corticosteroid-binding globulin gene ALA SER224 polymorphism. Endocr Res. 2004; 30(3):417-429. 23 Underhill JA, Mahalingam M, Peakman M, Wessely S. Lack of association between HLA genotype and chronic fatigue syndrome. Eur J Immunogenet. 2001; 28(3):425- 428. 24 Smith J, Fritz EL, Kerr JR, Cleare AJ, Wessely S, Mattey DL. Association of chronic fatigue syndrome with human leucocyte antigen class II alleles. J Clin Pathol. 2005; 58(8):860-863. 25 Carlo-Stella N, Badulli C, De SA, Bazzichi L, Martinetti M, Lorusso L et al. A first study of cytokine genomic polymorphisms in CFS: Positive association of TNF-857 and IFNgamma 874 rare alleles. Clin Exp Rheumatol. 2006; 24(2):179-182. 26 Colhoun HM, McKeigue PM, Davey SG. Problems of reporting genetic associations with complex outcomes. Lancet. 2003; 361(9360):865-872. 27 Schadt EE, Lamb J, Yang X, Zhu J, Edwards S, GuhaThakurta D et al. An integrative genomics approach to infer causal associations between gene expression and disease. Nat Genet. 2005; 37(7):710-717. 28 Powell R, Ren J, Lewith G, Barclay W, Holgate S, Almond J. Identification of novel expressed sequences, up-regulated in the leucocytes of chronic fatigue syndrome patients. Clin Exp Allergy. 2003; 33(10):1450-1456. 29 Kaushik N, Fear D, Richards SC, McDermott CR, Nuwaysir EF, Kellam P et al. Gene expression in peripheral blood mononuclear cells from patients with chronic fatigue syndrome. J Clin Pathol. 2005; 58(8):826-832. 30 Vernon SD, Unger ER, Dimulescu IM, Rajeevan M, Reeves WC. Utility of the blood for gene expression profiling and biomarker discovery in chronic fatigue syndrome. Dis Markers. 2002; 18(4):193-199. 31 Vollmer-Conna U, Aslakson E, White PD. An empirical delineation of the heterogeneity of chronic unexplained fatigue in women. Pharmacogenomics. 2006; 7(3):355-364. 32 Aslakson E, Vollmer-Conna U, White PD. The validity of an empirical delineation of heterogeneity in chronic unexplained fatigue. Pharmacogenomics. 2006; 7(3):365- 373. 33 Smith AK, White PD, Aslakson E, Vollmer-Conna U, Rajeevan MS. Polymorphisms in genes regulating the HPA axis associated with empirically delineated classes of unexplained chronic fatigue. Pharmacogenomics. 2006; 7(3):387-394. 34 Goertzel BN, Pennachin C, de Souza CL, Gurbaxani B, Maloney EM, Jones JF. Combinations of single nucleotide polymorphisms in neuroendocrine effector and receptor genes predict chronic fatigue syndrome. Pharmacogenomics. 2006; 7(3):475- 483. 35 Fang H, Xie Q, Boneva R, Fostel J, Perkins R, Tong W. Gene expression profile exploration of a large dataset on chronic fatigue syndrome. Pharmacogenomics. 2006; 7(3):429-440. 36 Whistler T, Taylor R, Craddock RC, Broderick G, Klimas N, Unger ER. Gene expression correlates of unexplained fatigue. Pharmacogenomics. 2006; 7(3):395-405. 37 Carmel L, Efroni S, White PD, Aslakson E, Vollmer-Conna U, Rajeevan MS. Gene expression profile of empirically delineated classes of unexplained chronic fatigue. Pharmacogenomics. 2006; 7(3):375-386. 38 Broderick G, Craddock RC, Whistler T, Taylor R, Klimas N, Unger ER. Identifying illness parameters in fatiguing syndromes using classical projection methods. Pharmacogenomics. 2006; 7(3):407-419. 39 Sterne J, Kirkwood BR. In: Goodgame F, ed. Essential Medical Statistics. Massachusetss: Blackwell Science, 2003: 98-106. -------- (c) 2007 BMJ Publishing Group Ltd. (c) 2007 Royal College of Paediatrics and Child Health.