Re: Blood pressure centiles for Great Britain – Can they be safely applied to clinical practice?
We thank Gandhi and Ozun for their thoughtful comments. We agree with them that blood pressure in children has been a neglected area of clinical practice, and that the lack of normative data to assist in interpretation of blood pressure measurements has been a major hindrance.
However, Gandhi and Ozun are concerned that the blood pressure centiles seem very high, particularly in older males, in whom 23% fulfilled the definition of hypertension based on British Hypertension Society (BHS) criteria .
They reflect that, in their clinical practice, they rarely see the levels of blood pressure that would be deemed high or high-normal by our suggested definitions, even in children they are confident have hypertension.
In response, we make the point that the centiles are derived from cross-sectional, representative samples from a number of health surveys in Great Britain. Measurements from more than 22,000 children and young adults were used to derive the centiles, using a reliable and established method .
Thus, we are confident that the centiles are indeed correct, for blood pressure measurements taken on a single occasion, using an automated oscillometric technique. However, the original blood pressure analyses were undertaken in the mid-1990s, and will, in due course, need updating and revising, to retain their relevance to contemporary children and young adults.
Centile charts are an adjunct to clinical assessment and decision- making. Our blood pressure centiles define expected ranges for blood pressure in children and young adults. However, centile charts are not infallible. A child with an evolving brain tumour may have growth hormone deficiency, and yet lie within the normal range for height. It is always important to bear in mind the complete clinical picture.
We also contend that children who have conditions which will very commonly result in adult hypertension – such as operated aortic coarctation – may have normal blood pressure in childhood. However, in this situation, the clinician may choose lower thresholds for intervention than would normally be the case, just as in adults with diabetes, in whom lower blood pressure targets are appropriate because of their much higher cardiovascular risk . Accordingly, for such children, it may be appropriate or desirable to, for example, define the 75th centile as the upper limit before intervention is considered.
In contrast to the concerns of Gandhi and Ozun, it has also been suggested to us that our blood pressure centiles are too low, thus causing a greater than expected number of children to fall into the high-normal or high blood pressure range (Peter Betts, personal communication).
Gandhi & Ozun comment that as blood pressure “tracks” from childhood into adult life, our definitions of high and high-normal blood pressure will fail to recognise all who may become hypertensive as adults. They appear to be arguing that the definition of hypertension in childhood should be based on progression to adult hypertension. However, the most common condition giving rise to hypertension – arteriosclerosis – is a progressive disease, aggravated by other risk factors such as smoking, obesity, salt intake, etc; and it is by no means self-evident that a person diagnosed with hypertension at age 30 or 40y should have been identifiable, based on blood pressure measurements in childhood. Moreover the centiles are not designed to predict future blood pressure (although they may be useful in this regard); they describe the normal range of blood pressures seen in children and young adults.
Paradoxically, Gandhi and Ozun go on to complain that our centiles may lead to over-investigation for causes of hypertension, based on high numbers of children and young adults fulfilling the BHS diagnostic criteria for hypertension in adults – particularly adolescent and young adult males. However, in our paper and in the commercially produced Blood Pressure charts, (Harlow Printing, South Shields, UK), we make the point that a single measurement of elevated blood pressure is usually insufficient for the “diagnosis” of high blood pressure for age, and that ideally the measurement should be repeated several times. Using repeated measurements only about a fifth of children, with high blood pressure on a single occasion, are ultimately determined truly to have high blood pressure . Thus, we disagree that proper application of the centiles would lead to inappropriate or unnecessary investigation.
Gandhi & Ozun object to our use of the standard 9 centile format used for all commonly used centile charts in Great Britain , although they comment, in their introductory paragraph, that the nature of the charts makes them readily comprehensible to all health professionals working in Britain. The principal objection offered is that the centile charts are different from contemporary charts in Europe and the USA. In order to facilitate international comparisons, we have however produced charts with an additional 95th centile.
Our object in producing blood pressure centiles for Great Britain was to inform British practice. We have already alluded to the methodological rigour used to create the centiles. Moreover, no such rigour can be claimed for centiles in common use elsewhere. We contend that for British practice, it is clearly relevant and appropriate to use centiles derived from British children and young adults. We find it surprising that referring to centile charts for blood pressure should be thought time- consuming and onerous, or that it might be thought preferable to use a calculated approximation to the 1987 US centiles  as an alternative.
Lastly, we reject the statement that having normal blood pressure based on our centiles should preclude appropriate lifestyle and other interventions for those children and young adults for whom it is appropriate.
1. Williams B, Poulter NR, Brown MJ, et al. British Hypertension Society guidelines for hypertension management 2004 (BHS-IV): summary. BMJ 2004;328:634–40. 2. Cole TJ, Green PJ. Smoothing reference centile curves: the LMS method and penalized likelihood. Stat Med 1992;11:1305–19. 3. Vijan S, Hayward RA. Treatment of hypertension in type 2 diabetes mellitus: blood pressure goals, choice of agents, and setting priorities in diabetes care. Ann Intern Med. 2003 Apr 1;138:593-602. 4. Hornsby JL, Mongan PF, Taylor AT, et al. ‘White coat’ hypertension in children. J Fam Pract 1991;33:617–23. 5. Cole TJ. Do growth chart centiles need a face lift? BMJ 1994;308:641–2.
Authors' Reply to Blood pressure charts in UK children
We thank Tullus and Brennan for their acknowledgement of the value of our recently published Blood Pressure Centiles for Great Britain.
They have compared our centiles with those published in the USA , and make a number of comments based on these observations, which we seek to address below. However, we would like to preface our responses with a comment. Few, if any, would feel it appropriate to use US references for height and weight, or other anthropometric measures in British children. Inevitably, our centiles, even if presented in the same way as the US data, would differ significantly. In marked contrast to the US data, the blood pressure measurements in the subjects included in our analysis were based on nationally representative samples of British children, obtained using the same type of blood pressure monitor throughout (Dinamap 8100, Critikon, Tampa, Florida, USA), to a very rigorous protocol. As such, we believe these centiles are the most accurate characterisation of normal blood pressure in children and young adults in any country yet published.
Tullus and Brennan ask why our centiles are based on age, rather than height. We examined this relationship in detail in our paper. Current US centiles are height- rather than age-based. However, it should be borne in mind that the studies from which the US centiles were derived were non- representative studies, using a range of methodologies. In contrast to what was observed in the US data, we found a negative effect of height on blood pressure after adjusting for weight. In other words, at any given weight, a taller child would have lower blood pressure. This effect was most pronounced for systolic blood pressure, but was also seen, albeit weakly, in diastolic blood pressure. We have reanalysed our data to clarify the relative importance of height and weight. For systolic blood pressure the percentage variation explained by weight and height together was 8.1%, while with weight alone (ignoring height) it was 7.9%, i.e. almost the same. Conversely, controlling for height (and ignoring weight) explained only 2.4% of the variation, so by far the largest effect on blood pressure comes from weight not height. It would be possible in principle to construct centiles adjusted both for age and weight, but we are unclear how useful they would be. We would very much welcome a debate on how to depict the data to best clinical advantage. Either way we are clear that, based on the British data, it would be quite inappropriate to express centiles by height.
Tullus and Brennan ask whether methodological flaws or rising obesity rates might contribute to the observed strong positive association between weight and blood pressure. The effect of weight on blood pressure applies across the weight spectrum, not just in the small proportion of obese subjects. The mean BMI SDS of the sample was +0.3 to +0.4 with an SD of 1.1, confirming a wide range of values. It is unlikely that obesity has contributed more than a small amount to the strong weight-blood pressure relationship.
The methodology used in the health surveys was rigorous and consistent, with arm circumference being measured to determine the appropriate cuff-size. Accordingly we do not accept that incorrect cuff size played any significant part in determining the observed relationship between weight and blood pressure.
Secondly, Tullus and Brennan criticise our adoption of the 98th centile to define high blood pressure for age. US blood pressure centiles define those above the 95th centile for blood pressure as having hypertension. Other US centile charts also use the 5th and 95th centiles to define normality. We do not feel that an arbitrary centile-based definition of hypertension in childhood is appropriate. Such a definition would inevitably define significant numbers of normal healthy children as having hypertension. We note that Tullus and Brennan are themselves unsure as to the appropriateness of using the 95th centile as the upper limit of normality. We prefer the terms high and high-normal blood pressure for age. We additionally suggest that the term hypertension be reserved for children with high blood pressure and evidence of pathological cause or effect.
Ultimately, a centile chart is a tool to inform clinical assessment and decision-making. In keeping with other centile charts in use in Great Britain, we chose to adopt the same standard nine-centile format ranging from the 0.4th to 99.6th centiles . In clinical practice, the difference between the 95th and 98th centiles is of little practical consequence. However, as requested by Tullus and Brennan, and acknowledging the potential clinical value to clinicians following internationally accepted treatment protocols, we here include charts amended to include the 95th centile for systolic and diastolic blood pressure.
On the third point raised by Tullus and Brennan, we concur that blood pressure measurements using automated oscillometric devices are commonly a little higher  than those seen using manual techniques, and this may account for some of the differences between the US and British centiles. For this reason we enjoin caution when referencing blood pressures derived manually, to the published centiles. Nevertheless, we strongly believe that our centiles are much more relevant for use in Great Britain, particularly when automated oscillometric blood pressure measurements are increasingly the norm .
1. National Blood Pressure Education Working Group on High Blood Pressure in Children and Adolescents. Fourth report on the diagnosis, evaluation and treatment of high blood pressure in children and adolescents: a working group report from the National High Blood Pressure Education Program. Pediatrics 2004;114:555-76.
2. Freeman JV, Cole TJ, Chinn S, et al. Cross sectional stature and weight reference curves for the UK, 1990. Arch Dis Child 1995;73:17- 24.
3. O'Brien E, Petrie J, Littler W, et al. The British Hypertension Society protocol for the evaluation of automated and semi- automated blood pressure measuring devices with special reference to ambulatory systems. J Hypertens 1990;8:607-19
4. O'Brien E. Demise of the mercury sphygmomanometer and the dawning of a new era in blood pressure measurement. Blood Press Monit 2003;8:19-21.
Text to accompany modiefied Blood Pressure figures:
Fig.1a-d Centiles for systolic and diastolic blood pressure in children and young adults 4-24y of age, modified to include 95th centile.
Fig 1a Male Systolic Blood Pressure Fig 1b Male Diastolic Blood Pressure Fig 1c Female Systolic Blood Pressure Fig 1d Female Diastolic Blood Pressure
Legend applies to all figures.
Re: a definiton of childhood prehypertension is also necessary
The definition of hypertension in adult life is arbitrary. In reality, there is a continuum of risk of adverse consequences of hypertension. Consequently, as Dr Jolobe says, those with “high-normal” blood pressure, or prehypertension, do have higher cardiovascular morbidity, particularly if associated with other comorbidities such as obesity or diabetes.
In our paper, we purposely chose not to define hypertension per se. Instead, we refer to high and high-normal blood pressure for age. We suggest that the term hypertension should be reserved for those with high blood pressure with evidence of pathological cause or effect. Given the lack of clarity about what constitutes hypertension in childhood, we would question the use of the term “prehypertension” in children. However, we do suggest that individuals above the 91st centile be considered as having “high-normal blood pressure for age”, especially as many, if not most, of these individuals will go on to have hypertension by accepted definitions in adult life.
The adverse consequences of hypertension are predominantly seen after the age of 40y, so we would not feel it appropriate to recommend screening for high- and high normal blood pressure in children or young adults, but agree with Dr Jolobe that opportunistic case-finding is valuable in this age range.
Re: Blood pressure charts in UK children
It was my pleasure to read important notes on blood pressure in UK children. I tried to see the chart but couldn't locate it. May I request you to provide the chart for our reference?
Thank you very much.
With best regards,
Sincerely, Dr. Narayan Bahadur Basnet
Blood pressure centiles for Great Britain – Can they be safely applied to clinical practice?
The study by Jackson et al(1) attempts to fill a gap in our knowledge in a very vital area. Paediatricians in the United Kingdom have traditionally not included a blood pressure measurement as part of routine clinical assessment, as hypertension is not generally considered to be a common paediatric problem. Those who did check BP had to rely on normal values derived from European and North American studies. Blood pressures centiles for GB are therefore a big step in the right direction and warmly welcomed. The data have been pooled from large representative samples and the methodology appears to be robust. The authors have chosen the well tested traditional nine-centile system, which all British health professionals are familiar with. It does however raise a number of issues.
Firstly the observed blood pressure appears to be remarkably high in a significant proportion of paediatric population. This is most obvious in the pubertal boys, nearly quarter of whom would be labeled as hypertensive as per the definition suggested by the British Hypertension Society(BHS). In fact the BHS classification of blood pressure level states that the optimal BP for adults is a value of < 120 mm Hg systolic and < 80 mm Hg diastolic. Although < 130 mm and < 85 may be accepted as normal, any value above 130/85 is at least high normal if not hypertensive(2). This is not concordant with the international definition of high blood pressure as suggested by World Health Organisation and International Society of Hypertension. In our own cardiology practice we struggle to see such high blood pressure values even in operated patients with coarctation of the aorta! More over if the author’s suggested definition of hypertension (BP above 98th centile) is applied, many children currently labeled as hypertensive would fall in the category of high normal/normal blood pressure. For any clinician this is a challenging conundrum. One has to ask if it is wise to label these children as normotensive when clearly a few years down the line they would be classified as hypertensive by our adult physician colleagues. Does accepting this new definition of hypertension inevitably mean that we are choosing to ignore an opportunity to identify and influence an important risk factor for future coronary heart disease? There is a growing body of evidence to suggest that future risk for coronary artery disease may start as early as in fetal life. Tireless efforts by professional bodies to prevent risk factors for ischaemic heart disease have lent its support to achieve even lower values of BP in adults. On the contrary, adopting higher normal blood pressure values in adolescence is going to be difficult to justify and is likely to lead to confusion let alone reduction of future risk of coronary artery disease.
Secondly the BHS guidelines for management of hypertension recommend that younger patients (age<20 years) should not be presumed to have essential hypertension and should be investigated for an underlying cause. In the light of the current data set this would mean that a quarter of British pubertal males need investigations for an underlying problem, and if not are we choosing to ignore a potential renal/ reno-vascular condition.
Thirdly by adapting a new centile system for defining normal and high blood pressure we are choosing to differ from both our American and European counterparts. This is at a time when there is universal agreement on the definition of hypertension in adults. The blood pressure centiles in the North American population are based on more recent data (1999-2000 National Health and Nutrition Examination Survey) and in view of the ongoing obesity epidemic much lower cut off for defining hypertension was recommended(3). They also recommend re-labeling high normal blood pressure as Pre-hypertension in order to promote preventive measures such as healthy diet and activity, which is an indication for life style changes. Admittedly these centiles are somewhat labour intensive and time consuming to use in routine clinical practice. In fact for the busy clinician the formula suggested by Somu et al(4), may prove to be an easier and quicker tool to identify children with hypertension without compromising from scientific normalcy.
Incorporating the new British blood pressure centiles in clinical practice effectively translates into ignoring a substantial number of children who would otherwise be a target for lifestyle and perhaps medical interventions. This is contrary to the recommendations made by British Hypertension Society and endorsed by National Institute for Clinical Excellence(5). We do not therefore feel comfortable in adopting the new Blood pressure centiles and definition of normal and high blood pressure values in children. We call for an open debate regarding the right way forward.
1. Jackson LV, Thalange NKS, Cole TJ. Blood pressure centiles for Great Britain. Arch Dis Child 2007;92:298-303.
2. Wlliams B, Poulter NR, Brown MJ et al. British Hypertension Society guidelines for hypertension management 2004 (BHS_IV): summary. BMJ 2004;328:634-40.
3. The fourth report on the Diagnosis, Evaluation and Treatment of High Blood Pressure in Children and Adolescents. Pediatrics 2004;114:555-576.
4. Somu S, Sundaram B, Kamalanatham AN. Early detection of hypertension in general practice. Arch Dis Child 2003;88:302.
5. Hypertension: management of hypertension in adults in primary care. NICE clinical guideline 34. Issue date June 2006.
Blood pressure charts in UK children
The study by Jackson et al on blood pressure centiles for Great Britain (1) provides us with valuable information and insight into children’s blood pressure centiles measured on automated monitors. This is the first time we have been able to see normative data for such large numbers of children in the UK.
There are, however, a few things in this publication that we would like to discuss. It is generally accepted that blood pressure in children depend not only on sex and age but also on the height of the child. The correlation with height is physiologic and needs to be taken into consideration when evaluating the child. In e.g. two-year-old girls with height between the 5th or 95th centile this difference can be 7 mm Hg. We were, thus, surprised to see only a weak correlation with height and disappointed to see that the author’s extensive data were not integrated with height.
The correlation with weight, found by the authors, is more to do with the increasing and often pathologic blood pressure in obesity. We also wondered if this strong correlation had anything to do with the size of the blood pressure cuff. The difference between a cuff that encircles 80% or 100% of the circumference of the arm can be significant especially in obese children.
Secondly we were surprised to see the authors had redefine hypertension to be above the 98th centile compared to the commonly used 95th without any explanation. The definition of hypertension is clearly much more complex in children compared to in adults. Children, so far, lack long-term prospective outcome data showing which blood pressure is optimal for each age and the definition is thus strictly statistical. We do not dispute that the 98th centile might well be a better definition then the 95th however this is an international agreement protocol that is followed by most doctors treating children with hypertension.
Thirdly the blood pressure values found in the new graphs are clearly much higher than those commonly used (2) even if they are difficult to compare as different centiles are given. As an example a 17-year-old boy of median height would be defined as hypertensive at 136mm Hg in the old charts and at 143-144 in the new. This is also a clinically very significant difference. One reason to this could be the well-known difference between manual and automated blood pressure measurements.
We would strongly suggest that the authors use their important data to make reference levels outlining the 95th centile for age and height centiles in children. Such graphs would be invaluable in clinical practice particularly for children where automated machine are the only available option for monitoring children blood pressure.
Consultant Paediatric Nephrologist
Great Ormond Street Hospital for Children
Great Ormond Street
London WC1N 3JH
Competing interests: none declared
1. Jackson LV, Thalange NKS, Cole TJ. Blood pressure centiles for Great Britain. Arch Dis Child 2007;92:298-303.
2. The fourth report on the diagnosis, evaluation, and treatment of high blood pressure in children and adolescents. Pediatrics 2004;114(2 Suppl):555-76.
a definiton of childhood prehypertension is also necessary
The recognition that childhood hypertension, with its attendant risk of cardiovascular disease, tracks into adulthood(1)(2), should be an impetus, not only to a satisfactory definition of hypertension in childhood(3), but also an impetus to a definition of its precursor, prehypertension. In the United States, amongst those aged 20 or more, an estimated 41.9 million men and 27.8 million women have prehypertension(4), and those with prehypertension have an increased risk of developing cardiovascular disease relative to those with optimal levels of blood pressure(5). The association with cardiovascular disease is more pronounced among those with diabetes and those with high Body Mass Index(5). Accordingly, the recognition that greater attention to blood pressure early in life potentially generates long-term benefits on cardiovascular health should prompt policy makers to be more pro-active about detection of hypertension in the primary care setting regradless of whether the chosen modality for case finding is opportunistic or sytematic(6). What is not acceptable is the current situation whereby an authoritative body such as the National Institute for Clinical Excellence(NICE) does not even have a policy statement on either opportunistic case finding or on systematic screening for hypertension(7)
(1) Whincup PH., Cook DG., Geleinjsne JM A life course approach to blood pressure In Kuh D., Ben-Shlomo Y eds A life course approach to chronic disease epidemiology Oxford University Press 2004:218-39.
(2) Celermajer DS., Ayer JGJ Childhood risk factors for adult cardiovascular disease and primary prevention in childhood Heart 2006:92:1701-6.
(3) Jackson LV., Thalonge NKS., Cole TJ Blood pressure centiles for Great Britain Archives of Disease in Childood 2007:92:298-303.
(4) Qureshi AI., Suri MFK., Kirmani J., Divani AA Prevalence and trends of prehypertension and hypertension in United States: National Health and Nutrition Examination Surveys 1976 to 2000 Meical Science Monitor 2005:11:CR403-409.
(5)Kshirsagar AV., Carpenter M., Bang H et al Blood pressure usually considered normal is associated with an elevated risk of cardiovascular disease American Journal of Medicine 2006:119:133-141.
(6)Maryon-Davis A., Press V on behalf of the Cardiovascular Health Working Group of the Faculty of Public Health and national Heart Forum Easing the pressure: Tackling hypertension A toolkit for developing a local strategy to tackle high blood pressure Developing a local hypertension strategy 2005:pages 45-47.
(7) National Institute for Clinical Excellence, 2004(expected review date June 2010) Management of hypertension in adults in primary care. Clinical Guideline. 18. London: National Institute of Excellence