Objective: To produce representative cross-sectional blood pressure reference centiles for children and young people living in Great Britain. Design: Analysis of blood pressure data from seven nationally representative surveys: Health Surveys for England 1995-8, Scottish Health Surveys 1995, and National Diet & Nutrition Survey 1997. Methods: Blood pressure was measured using the Dinamap 8100 with the same protocol throughout. Weight and height were also measured. Data for 11 364 males and 11 537 females aged 4-23 years were included in the analysis, after excluding 0.3% missing or outlying data. Centiles were derived for systolic, diastolic, mean arterial and pulse pressure using the latent moderated structural (LMS) equations method. Results: Blood pressure in the two sexes was similar in childhood, rising progressively with age and more rapidly during puberty. Systolic pressure rose faster and was appreciably higher in adult men than in adult women. After adjustment for age, blood pressure was related more to weight than height, the effect being stronger for systolic blood pressure. Pulse pressure peaked at 18 years in males and 16 years in females. Conclusions: These centiles increase our knowledge of blood pressure norms in contemporary British children and young people. High blood pressure for age should be defined as blood pressure above the 98th centile, and high-normal blood pressure for age as blood pressure between the 91st and 98th centiles. The centiles identify children and young people with increased blood pressure, and will be of benefit to both clinical practice and research.T here is no satisfactory definition of hypertension in children.1 As a result, blood pressure is often not measured in paediatric clinical practice, and understanding the clinical significance of blood pressure readings in children is hampered by the lack of satisfactory reference data with which to interpret them.Reference blood pressure centiles should therefore improve the understanding of blood pressure variation in childhood. In Britain and worldwide, there have been many studies of childhood blood pressure, but all are of limited use in Great Britain owing to the use of non-representative populations, limited age ranges and mixed methodologies for blood pressure measurement. Accordingly, we have developed representative cross-sectional blood pressure references for children and young people living in Great Britain. METHODSBlood pressure data from seven national health and social surveys carried out between 1995 and 1998 were obtained from the UK Data Archive (http://www.data-archive.ac.uk/)( The survey samples were obtained by stratified multistage sampling techniques to ensure that there was a proportional representation of the population at large by sex, age, geographical region and social class.2 In brief, the demographic characteristics of a geographical area are known from census and other data. Using this information, a representative sample of individuals from the target age groups for each survey was obtained....