104 ORIGINAL ARTICLE New cross sectional stature, weight, and head circumference references for Down's syndrome in the UK and Republic of Ireland M E Styles, T J Cole, J Dennis, M A Preece . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Arch Dis Child 2002;87:104­108 Aim: To present a growth reference for children with uncomplicated Down's syndrome living in the UK See end of article for and Republic of Ireland. Data are available for height and weight in the age range 0­18 years, includ- authors' affiliations ing the first three months of life, and for head circumference in the first year. . . . . . . . . . . . . . . . . . . . . . . . Methods: The study sample was drawn from 16 discrete geographical areas and was representative Correspondence to: of children age 19 years of age or less who are now living in the UK and Republic of Ireland. Multiple Prof. M A Preece, Institute growth measurements for 1507 children were obtained retrospectively by case note search. Data from of Child Health, University children with significant cardiac or other major pathology were excluded from analysis. Data from pre- College London, 30 term babies were excluded up to age 2 years. Centile curves were constructed from 5913 selected Guilford Street, London WC1N 1EH, UK; measurements from 1089 children and were derived using Cole's LMS method. mpreece@ich.ucl.ac.uk Results: The resulting centiles differ substantially from those previously available in the UK, which were based on selective US data published in 1988. Accepted 21 November 2001 Conclusions: We propose that these charts should now be adopted as the standard UK/Republic of . . . . . . . . . . . . . . . . . . . . . . . Ireland reference. Shortstatureisarecognisedcharacteristicofmostpeople SUBJECTSANDMETHODS with Down's syndrome. Average height at most ages is Ethical approval was obtained from the South Thames Multi- around the 2nd centile for the general population. For the centre Research Ethics Committee and from local research majority the cause of growth retardation is not known.1 Some ethics committees. conditions leading to poor growth (congenital heart disease,2 3 sleep related upper airway obstruction,4 coeliac Data collection disease,5 6 thyroid hormone deficiency,7 8 and nutritional inad- Community paediatricians in the UK and Republic of Ireland equacy caused by feeding problems9 10) occur more frequently were approached through personal contact or by advertise- among those with the syndrome. There is also a high ment at national paediatric meetings, to see whether they had prevalence of overweight/obesity, particularly in adolescence adequate special needs registers to identify all the children and adult life.11 12 However, people with the syndrome are not with Down's syndrome in their geographical area. Sixteen necessarily overweight in relation to their height. As with the such areas were identified: one each in the Republic of Ireland, general population, weight is influenced by environmental12 13 Northern Ireland, and Scotland, and the remaining 13 in six of as well as biological factors, the eight geographically determined health regions in 14 and for most, preventive England. measures are both feasible and effective. All relevant medical records were searched by the local pae- Regular growth surveillance of children with Down's diatrician or one of the authors (MS). Children were given an syndrome should aid early identification both of pathological anonymous identification number. All records of height, causes of growth retardation and of incipient overweight/ weight, and head circumference were collected, together with obesity. Growth charts are recognised as a useful tool for the method and date of measurement, date of birth, sex, and monitoring the growth and wellbeing of children. However, gestation. Ethnicity was not consistently recorded in the where normal growth patterns differ from the general source records, but assuming that in such cases the child was population, it has been found useful and clinically important white, 94% of the sample was white. A record was also made to use syndrome specific growth charts.15 of any coexisting illness (for example, heart disease, bowel The Down's syndrome growth charts in current use in the disorder, or malignancy); the date and result of the most UK were produced by Cronk and colleagues,16 based on US recent thyroid function test was also recorded. In addition data published in 1988.17 As with the general population, data were incorporated from a cross sectional population height of those with Down's syndrome varies from country to based growth study in the Eastern Health Board area of the country. Those in the Netherlands are taller than those in the Republic of Ireland.20 To aid centile construction for the later teenage years, a further 27 young adults, aged 20­24, were USA and much taller than those in Sicily.18 19 In addition the recruited from a further education college, giving a total study US data were derived from five different clinic or research population of 1507. based samples. The study sample was therefore not represen- To avoid any child being over represented in the dataset, not tative of the total population. Furthermore no exclusions were more than two measures per year were included for children made on the basis of either coexistent major pathology (for over 1 year and not more than four measures in the first year. example, cardiac or thyroid disease) or gestational age at birth. A total of 8818 measurement episodes were entered into the We therefore identified a need for a contemporary UK growth dataset; 96% of children identified through the special condi- reference for children with uncomplicated Down's syndrome. tions registers provided data. www.archdischild.com Growth reference for Down's syndrome 105 Table 1 Frequencies of measurements by age and 180 sex Age (y) Male Female Total 99.6 160 98 0 810 673 1483 91 1 281 246 527 75 2 269 246 515 50 3 244 239 483 25 4 221 204 425 140 9 5 192 176 368 2 6 160 164 324 0.4 7 143 133 276 8 131 122 253 120 9 113 109 222 10 114 106 220 11 103 84 187 12 81 58 139 Height (cm) 100 13 70 68 138 14 55 58 113 15 44 38 82 16 37 27 64 17 20 17 37 80 18 11 12 23 19 5 2 7 20 6 6 12 21 3 9 12 60 22 2 0 2 24 0 1 1 Total 3115 2798 5913 400 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Age (y) Excluded children Figure 1 Girls' Down's syndrome height chart for the UK and Data for dead children still on the special needs register were Republic of Ireland. excluded. Data from those with coexistent major pathology and/or preterm birth (less than 37 completed weeks gestation) were excluded as follows. Those with no abnormality or a Centile curves were fitted to the data using Cole's LMS small atrioventricular or ventricular septal defect, or a patent method.21 Briefly, this assumes that the data can be ductus arteriosus that had already closed or was asympto- transformed to normality by a suitable power transformation matic were included in the study; those with a higher score or (L), and the distribution is then summarised by the median history of cardiac surgery were excluded. In addition children (M) and coefficient of variation (S). The values of L, M, and S with other multiple or major pathology were excluded. Meas- are constrained to change smoothly with age, and the fitted urements for children born preterm were excluded for the first values can be used to construct any required centile curves. two years of life, but later measurements were included. Cole and colleagues21 give more details. Measuring techniques Reported measuring techniques were variable over time both RESULTS within and between areas. In most areas staff had been A total of 407 children were excluded because of cardiac trained, including paediatricians, clinic nurses, health visitors disease or other major pathology. Of those remaining, 138 (more frequent in the under 5s), and school nurses (more were preterm; 11 of these had no data after age 2, hence were common in the over 5s). Infants were weighed naked; older excluded. After exclusions, the final sample consisted of 1089 children were measured without footwear in underwear or children (597 boys and 492 girls). There were 5913 occasions light indoor clothing. with at least one valid measurement, providing 5681 weights, Equipment varied between and within areas. Standard 4941 heights, and 2364 head circumferences. Table 1 gives the clinic equipment included regularly calibrated scales and sta- numbers of measurements by sex and year of age. A quarter of diometers. Community staff often used portable equipment all measurements were in the first year of life, while there that was less often calibrated. were relatively few points after 14 years of age. The data collected between age 1 and 5 years contained Some subjects were measured on more than one occasion; measures of both supine length and height, because the age at one third of subjects provided just one measurement, which children with Down's syndrome can stand unaided is representing 6% of all measurements. At the other extreme, 15 extremely variable children each provided 20 or more measurements. The 868 subjects with between one and nine measurements provided Statistical analysis half the total, the other half coming from the remaining 221 Data were cleaned in several stages. Bivariate plots of height, subjects with 10 or more measurements. weight, and head circumference were used to identify gross Centiles were fitted to the data using the LMS method. For disproportion. Data points so identified were scrutinised, height and head circumference the distribution was assumed going back to the source data if necessary, and transcription normal; for weight there was appreciable skewness and the errors corrected. If a value was deemed highly unlikely (more age varying power transformation was adjusted for it. Using than 5 SDs from the mean), but if there was no evidence of a these data, reference charts for girls and boys with Down's transcription error, the point was deleted. In addition, runs of syndrome in the UK and Republic of Ireland were constructed individual longitudinal data were scrutinised for evidence of using the nine centile curve format proposed by Cole22 and unlikely measurements, such as apparent loss of height, and used in the current UK 1990 growth reference23 (figs 1, 2, and where possible, edited. 3). www.archdischild.com 106 Styles, Cole, Dennis, et al 120 In addition to the main child and teenage data a small number of measures were taken of young adults in order to facilitate construction of the curves in the late teenage years. 99.6 Although the numbers were relatively small in later years the 100 centile curves throughout childhood, and particularly in ado- lescence and early adulthood, look sensible, with the height 98 and head circumference centiles flattening off as expected as 80 91 adulthood approaches. We believe that the curves reflect an accurate picture of the 75 UK and Irish populations as they are and have been over the 60 50 past 19 years. We do not propose that they are a standard to be 25 eight (kg) achieved, particularly for weight in the older age groups where 9 W 2 it is clear that a significant proportion of the population is 40 0.4 obese. The tendency to overweight in late childhood and the teenage years is notable. Our data reveal that 30% of those aged 10 or more have a body mass index (BMI) greater than the 91st and 20% greater than the 98th centile for the general 20 population. We have already mentioned that overweight is not inevitable in this population. For most, as in the general popu- lation, due attention to diet and exercise is effective in 0 controlling overweight/obesity. We have shaded the area above 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 the 75th weight centile in the 5­18 age range and referred the Age (y) user to the UK 1990 BMI chart which we have reproduced on Figure 2 Boys' Down's syndrome weight chart for the UK and the front cover. We recommend that for the time being, as in Republic of Ireland. the general population, those whose BMI lies above the 98th centile are significantly overweight (obese) and in need of further assessment and guidance, and those above the 91st 55 centile are overweight and should be carefully monitored. It is likely that we will update this recommendation as we evaluate 99.6 98 the usefulness of the charts in clinical practice. 50 91 75 Of those falling below the 2nd centile for height and/or 50 weight, some will have major pathology. There is evidence 25 9 that, as in the general population, those with major cardiac 45 2 0.4 disease will be over represented at the lower end of the charts, as will those with other major pathology. We do not have suf- ficient data to construct charts for preterm babies and, as in 40 the general population, advise that measurements for those born before 37 completed weeks gestation should not be plot- ted on the charts until the estimated due date is reached. 35 Thereafter they should be charted relative to estimated due Head circumference (cm) date for at least a year. Preliminary data suggest that many babies with the 30 syndrome do not regain birth weight until around 1 month.9 This is not reflected in the charts because of their cross sectional nature. This early failure to thrive is usually a result 25 of feeding difficulties, many of which resolve after the first few 0 1 2 3 4 5 weeks. However, the possibility of major gastrointestinal Age (y) pathology (duodenal narrowing, gastro-oesophageal reflux) Figure 3 should always be borne in mind. We further advise that from Girls' Down's syndrome head circumference chart for the UK and Republic of Ireland. 1 month, in the absence of major pathology, babies whose weight continues to fall through the centiles should have their dietary intake monitored, as ongoing feeding difficulties and DISCUSSION failure to thrive are not uncommon in this population.10 We have produced the first growth charts relevant to children Individual children may show growth spurts and plateaus with Down's syndrome in the UK and Republic of Ireland. that are more prolonged than in the general population but While not based on a dedicated sample survey we believe the which are not reflected in the reference charts. Similarly the distribution of the contributing centres to be broadly charts suggest an absence of the pubertal growth spurt. Those representative. The advantages of using retrospective data col- with the syndrome do have an adolescent growth spurt, but it lection rather than performing a prospective study are: firstly, is usually less vigorous than in the general population and in the short time required to collect the data; secondly, the ease some children may occur at an earlier age.25 When this of recruiting a whole population rather than a smaller selected happens it will impose a limiting effect on final height. group of volunteers; and thirdly, the low cost. Comparison with the Castlemead/US charts,16 which have The measuring techniques, equipment used, and personnel hitherto been the only Down's specific charts available in the involved varied over time, both within centres and between UK, reveals notable differences: centres. However, the variation we recorded reflects normal · The UK height centiles map fairly closely to the US charts practice and for this reason, and because of the large sample between 3 and 12 years, but subsequently the US centiles size, we consider that the reference is wholly applicable for use drift below the UK charts so that by age 17 the US 50th in day to day practice. We note too that Cotterill and centile for boys is below the UK 25th and the 25th around colleagues24 have shown that trained school nurses are as reli- the 5th (fig 4). Note the discontinuity in the US centiles at able as auxologists in assessing the height of schoolchildren. age 3. www.archdischild.com Growth reference for Down's syndrome 107 180 99.6 120 98 91 95 75 99.6 160 100 50 75 25 50 9 98 25 2 80 140 5 0.4 91 95 75 75 60 50 50 120 eight (kg) 25 25 W 5 9 40 2 Height (cm) 0.4 100 20 80 00 1 2 3 4 5 6 7 8 9 101112131415161718 Age (y) 60 Figure 5 Girls' Down's syndrome weight chart for the UK and Republic of Ireland, with the centiles of Cronk and colleagues16 superimposed. 400 1 2 3 4 5 6 7 8 9 101112131415161718 Age (y) clinical implications. We therefore urge that use of the US charts should now be discontinued and recommend that the Figure 4 Boys' Down's syndrome height chart for the UK and Republic of Ireland, with the centiles of Cronk and colleagues16 new charts are adopted as the standard reference for these two superimposed. countries. The charts are published in A4 format for clinic use and in A5 format for use within the Parent Held Personal · US teenage boys thus appear shorter than their UK Child Health Record. They are therefore available for use in counterparts (and also heavier, data not shown). everyday clinical practice. · The median weights for girls are similar on the two charts, but the US charts show far greater variance before puberty ACKNOWLEDGEMENTS and less subsequently (fig 5). The UK data thus support the We would like to thank the Down's Syndrome Medical Interest Group widespread clinical impression of considerable overweight/ for initiating this project and for support throughout, and the many obesity in later childhood and adolescence. community paediatricians who were involved in the time consuming task of data collection. We are also grateful to Tam Fry and the Child · The US charts have no data for the first three months. Past Growth Foundation for support, encouragement, and for covering this age the US and UK length centiles are similar, but the travel expenses (MS), and John Short and his team at Harlow Print- UK weight centiles are appreciably higher, reflecting the ing for steering through the final production of the charts. Copies of absence of preterm births and major pathology. For instance the A4 charts and the PCHR inserts are available from Harlow Print- at age 1 the US 5th girls weight centile falls on the UK 0.4th, ing, Maxwell Street, South Shields, Tyne and Wear NE33 4PU, UK. and the 25th on the 5th. The boys are similar. This means that if the US charts are used to assess babies with uncom- . . . . . . . . . . . . . . . . . . . . . plicated Down's syndrome, an appreciable number of those Authors' affiliations with failure to thrive (see above) will remain undiagnosed. M E Styles, Sussex Weald and Downs NHS Trust, Chichester, UK T J Cole, M A Preece, Institute of Child Health, University College London, UK The charts based on these data are printed for clinic (A4) J Dennis, Down's Syndrome Medical Interest Group and parental (A5, Personal Child Health Record format) use in the 9 centile format proposed by Cole.22 Following consultation REFERENCES with parents and doctors, the 0.4th, 50th, and 99.6th centiles 1 McCoy EE. Growth patterns in Down's syndrome. In: Lott IT, McCoy EE, from the 1990 UK growth reference are overprinted on the A4 eds. Down syndrome: advances in medical care. New York: Wiley-Liss, charts to facilitate comparison with children without Down's Inc., 1992. syndrome. 2 Cronk CE. Growth of children with Down's syndrome: birth to age 3 years. Pediatrics 1978;61:564­8. 3 Greenwood RD, Nadas AS. The clinical course of cardiac disease in Conclusion Down's syndrome. Pediatrics 1976;58:893­7. We present for the first time centile charts for weight, height, 4 Stebbens VA, Samuels MP, Southall DP, et al. Sleep related upper and head circumference appropriate for children with Down's airway obstruction in a cohort with Down's syndrome. Arch Dis Child 1991;66:1333­8. syndrome living now in the UK and Republic of Ireland. Those 5 Jansson J, Johansson C. Down syndrome and celiac disease. J Pediatr with the syndrome show different biologically determined Gastr Nutr 1995;21:443­5. patterns of growth from the general population; in addition, a 6 Csizmadia CGDS, Mearin ML, Oren A, et al. Accuracy and cost-effectiveness of a new strategy to screen for celiac disease in range of both major and minor pathologies may further com- children with Down syndrome. J Pediatr 2000;137:756­61. promise growth. Growth surveillance throughout childhood 7 Karlsson B, Gustafsson J, Hedov G, et al. Thyroid dysfunction in Down's will aid in the early identification of a range of pathologies and syndrome: relation to age and thyroid autoimmunity. Arch Dis Child in the prevention of overweight and obesity, and can be best 1998;79:242­5. 8 Sharav T, Collins RM, Baab PJ. Growth studies in infants and children achieved by using syndrome specific charts. The charts with Down's syndrome and elevated levels of thyrotropin. Am J Dis Child presented here are more informative and accurate than the 1988;142:1302­6. earlier US charts currently in use in the UK and Republic of 9 Chilvers M. Time for children with Down's syndrome to regain birth weight. Nottingham audit findings presented at DSMIG meeting, Ireland. Differences throughout the age range have significant September 1997. www.archdischild.com 108 Styles, Cole, Dennis, et al 10 Spender Q, Stein A, Dennis J, et al. An exploration of feeding difficulties 18 Cremers MJG, van der Tweel I, Boersma B, et al. Growth curves of in children with Down's syndrome. Dev Med Child Neurol Dutch children with Down's syndrome. J Intellect Dis Res 1996;38:681­94. 1996;40:412­20. 11 Chumlea WC, Cronk CE. Overweight among children with Trisomy 21. 19 Piro E, Pennino C, Cammarata M, et al. Growth charts of Down J Ment Defic Res 1981;25:275­80. syndrome children in Sicily: evaluation of 382 children 0­14 years of 12 Prasher VP. Overweight and obesity amongst Down's syndrome adults. J age. Am J Med Genet Suppl 1990;7:66­70. Intellect Disabil Res 1995;39:437­41 20 Harper J, Murphy J, Philip M, et al. Growth in children and adolescents 13 Sharav T, BonmanT. Dietary practices, physical activity and body mass with Down Syndrome in Ireland, 3 months to 18 years. Ir J Med Sci index in a selected population of Down's syndrome children and their 2000;169:200­25. siblings. Clin Paediatr 1992;31:341­4. 21 Cole TJ, Freeman JV, Preece MA. British 1990 growth reference centiles for weight, height, body mass index and head circumference fitted by 14 Luke A, Roizen NJ, Sutton M, Schoeller DA. Energy expenditure in maximum penalized likelihood. Stat Med 1998;17:407­29. children with Down syndrome: correcting metabolic rate for movement. J 22 Cole TJ. Do growth chart centiles need a facelift? BMJ Pediatr 1994;125:829­38. 1994;308:641­2. 15 Ranke MB. Disease specific growth charts, do we need them? Acta 23 Freeman JV, Cole TJ, Chinn S, et al. Cross-sectional stature and weight Paediatr Scand Suppl 1989;356:17­25. reference curves for the UK 1990. Arch Dis Child 1995;73:17­24. 16 Cronk CE, et al. Growth charts for children with Down's syndrome age 24 Cotterill AM, Majrowski WH, Hearn SJ, et al. Assessment of the 3 months to 18 years. Castlemead Publications, 12 Little Mundells, reliability of school nurse height measurements in an inner-city population Welwyn Garden City, Herts AL7 1EW, UK, 1989. (the Hackney Growth Initiative). Child Care Health Dev 17 Cronk C, Crocker AC, Pueschel SM, et al. Growth charts for children 1993;19:159­65. with Down syndrome: 1 month to 18 years of age. Pediatrics 25 Arnell H, Gustaffson J, Ivarsson SA, Anneren G. Growth and pubertal 1988;81:102­10. development in Down's syndrome. Acta Paediatr 1996;65:1102­6. POSTCARD FROM DOWN UNDER............................................................... Is this a Pom I see before me? Howdoyoudefinenationality?ForsomeIimaginethatthisis indeterminateamountofevolutiontomedicalschool,andhereyou fairly straightforward. I find it a little more challenging. I know have classification and generalisation raised to a fine art: "These chil- that I'm English, because I was born in England. In some parts dren develop asthma," " These children have lower IQs," and so on. of the world it is easiest to define oneself as English because being Trying to teach myself to spot generalisations has been much less British is apparently synonymous--and somehow less clear. The easy than I'd thought, because many are wrapped up in pseudo- phrase "Ah, English...Manchester United!" must be taught in schools intellectual justification or compliment. For example, any statement around the world. I know that being English makes me British, but I starting with: "Australians are . . ." has to be suspect unless it is also know that my country of origin--on official forms--is better followed by a strict statistical or geographical definition. Just described as UK, as this describes the United Kingdom of Great Brit- substitute it with "Black people are . . ." or "Jewish people are . . ." and ain and Northern Ireland. Besides, despite being happy to write Brit- perhaps you'll begin to see the flaw. At a more trivial level, I often find ish, I feel a certain pompousness about writing the "Great" in Great myself saying "Cyclists are . . ." or "Paediatricians are . . ." Of course, Britain: I can hear Cilla Black (an English 60s pop idol still going these statements are nowhere near as offensive as overtly racist com- ments, but at the same time they can easily include the same crass strong as a TV host) expressing the view that (referring to Margaret generalisations. Thatcher) "Maggie put the Great back in Great Britain" and I shudder (In case you are interested, to Australians: "Poms are . . .infrequent a little. Sometimes I look at my National Training Number and at my bathers." Apparently we get by on only one bath a week, which seems burgundy coloured passport, and at that point I know that I'm a like sheer indulgence to me. Pom, incidentally, is derived either from European, but I have really no idea what that means. pompous, or from Prisoner of His (Her) Majesty; take your pick). My colleagues and friends here are in no doubt that they are Aus- To begin to turn this full circle: An Australian can be wholeheart- tralian. The majority of them are Queenslanders too, of which they are edly proud to be an Australian, at the same time as disapproving of doubly and distinctly proud. Pride is a concept writ large in the Aus- Pauline Hanson--right wing leader of the One Nation political party, tralian psyche, and sometimes overused to the point of ridiculousness. mandatory sentencing, or the Australian Government's handling of For example, a fast food chain does not simply sponsor a children's the MV Tampa stand off. This leads to an oddly circular situation, cen- soccer tournament: it "Proudly Sponsors" it, failing to see--or perhaps tral to any nationalistic feeling, where the proud national makes a cynically exploiting--the strange juxtaposition between fast food and broad generalisation, whilst recognising that the generalisation is sport. immediately and profoundly flawed. There are many things of which I feel proud, but I hesitate to use the Where does that leave me and my Englishness/Britishness/ word, in the same way that I flinch from the "Great" in Great Britain. Europeanism? I don't think I'm any the wiser, and I suppose I will con- This time, instead of Cilla I see a racist thug--or worse, a racist tinue to define myself depending on the circumstances, while harbour- intellectual--saying that "We should be proud of our race," or some ing a certain jealousy for the folk who see things more simply. However, such meaningless garbage. I must leave this here, as there is an "a" in the month, and so it is time The urge to generalise from specifics is almost overwhelming. I've for my bath. done it at least once already here, arguably two or three more times, I D Wacogne and will do it again before I finish. Inferring a general rule from a spe- Dr Wacogne was on secondment at the Royal Children's Hospital, Brisbane cific interaction must have been essential for our survival: for example for two years and is now completing his SpR training at the North Staffordshire "These berries are good to eat. Those berries are not." Fast forward an Hospital, UK. www.archdischild.com