Absolute height-specific thresholds to identify elevated blood pressure in children

Absolute height-specific thresholds to identify elevated blood pressure in children
of 5
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Copyright © Lippincott Williams Wilkins. Unauthorized reproduction of this article is prohibited. Absoluteheight-specificthresholdstoidentifyelevatedbloodpressureinchildren Arnaud Chiolero a , Gilles Paradis b , Giacomo D. Simonetti c , and Pascal Bovet a,d Objective:  Identification of children with elevated bloodpressure (BP) is difficult because of the multiple sex, age,and height-specific thresholds to define elevated BP. Wepropose a simple set of absolute height-specific BPthresholds and evaluate their performance to identifychildren with elevated BP in two different populations. Methods:  Using the 95th sex, age, and relative-height BPUS thresholds to define elevated BP in children (standardcriteria), we derived a set of (non sex- and non age-specific) absolute height-specific BP thresholds for 11height categories by 10cm increments. Using data fromlarge school-based surveys conducted in Switzerland( N  ¼ 5207; 2621 boys, 2586 girls; age range: 10.1–14.9years) and in the Seychelles ( N  ¼ 25759; 13048 boys,12711 girls; age range: 4.4–18.8 years), we evaluated theperformance of these height-specific thresholds to identifychildren with elevated BP. We also derived sex-specificabsolute height-specific BP thresholds and compared theirperformance. Results:  In the Swiss and the Seychelles surveys, theprevalence of elevated BP (standard criteria) was 11.4 and9.1%, respectively. The height-specific thresholds toidentify elevated BP had a sensitivity of 80 and 84%, aspecificity of 99 and 99%, a positive predictive value of 92and 91%, and a negative predictive value of 97 and 98%,respectively. Performance of sex-specific absolute height-specific BP thresholds was similar. Conclusion:  A simple table of height-specific BPthresholds allowed identifying children with elevated BPwith high sensitivity and excellent specificity. Keywords:  blood pressure, pediatric, predictive value,screening, sensitivity Abbreviations:  BP, blood pressure; NPV, negativepredictive value; PPV, positive predictive value INTRODUCTION T he identification of children with elevated bloodpressure (BP) is difficult notably because thethresholds to define elevated BP are sex, age, andheight-specific [1–3]. Consequently, there are numerousthresholds, one for each sex, age, and height strata. Thus,accounting for both sexes, 17 age categories, and sevenheight percentile categories, the BP references recom-mended in United States [1] and in European [2] guidelines include 476 specific thresholds for the 95th percentile of SBP and DBP of children aged 1–17 years.Furthermore,BPthresholdsaredefinedaccordingtothepercentile of height, not according to absolute height [1,2].Because the recommended BP references are based ondatafromUSchildren,theUSreferenceheightchartshouldbeusedtodeterminetheheightpercentileforagivenchild.This is problematic because health practitioners incountries outside of the United States are using specificnational height charts and height percentiles may substan-tially differ from the US height percentiles.These difficulties deter health professionals to assesselevated BP in children, threaten the validity of the BPassessment when it is done, and may lead to the under-diagnosis of hypertension in children [4]. Therefore, it would be useful for health professionals to have a simplertool to identify elevated BP in children [5–9]. In this study, we developed simple user-friendly BP thresholds basedsolely on absolute height and evaluated the performance of these thresholds to identify children with elevated BP inlarge school-based surveys in two different populations,-that is, Switzerland, a high-income European country [10,11], and the Seychelles, a middle-income Africancountry  [12,13]. METHODS Determination of height-absolute bloodpressure thresholds  We predefined 11 absolute height categories with 10cmincrements, i.e., less than 85cm, 85–94, 95–104, 105–114,115–124, 125–134, 135–144, 145–154, 155–164, 165–174,and at least 175cm, respectively. For each height category, we identified the corresponding US height percentile for Journal of Hypertension 2013, 31:1170–1174 a Institute of Social and Preventive Medicine (IUMSP), University Hospital Center,Lausanne,Switzerland,  b Department ofEpidemiology,Biostatistics, andOccupationalHealth, McGill University, Montreal, Canada,  c University Children’s Hospital, Insel-spital and University of Bern, Switzerland and  d Ministry of Health, Victoria, Mahe´,Republic of SeychellesCorrespondence to Arnaud Chiolero, MD, PhD, Institute of Social and PreventiveMedicine (IUMSP), University Hospital Center (CHUV/UNIL), Biopoˆle 2, Route de laCorniche10,1010 Lausanne,Switzerland.Tel:+4121 31472 72;fax:+4121 3147373; e-mail: Received  4 December 2012  Revised  8 January 2013  Accepted  15 February 2013J Hypertens 31:1170–1174    2013 Wolters Kluwer Health | Lippincott Williams &Wilkins.DOI:10.1097/HJH.0b013e32836041ff 1170  Volume 31    Number 6    June 2013 Original Article  Copyright © Lippincott Williams Wilkins. Unauthorized reproduction of this article is prohibited. each sex and age-strata of children. For each of these heightpercentiles, the corresponding US SBP and DBP 95th per-centile threshold was identified [1]. For a given absoluteheight category, a range of (systolic or diastolic) BP values was obtained and the mean value was used as the BPthreshold for the corresponding absolute height category.The SBP and DBP thresholds to define elevated BP for eachabsolute height category are shown in Table 1. Swiss survey The survey was described in details elsewhere [10]. Briefly,anthropometric data and BP were measured in all childrenof the sixth school grade of the Vaud canton (Switzerland)in 2005–2006. Measurements were performed in a quietand temperate room by trained clinical officers. Threemeasurements of BP were taken on the right arm, with acuff of appropriate size for the arm circumference, at 1-minintervals after a rest of at least 3min, in a seated position.Readings were obtained with clinically validated oscillo-metric devices (Omron M6) [14,15]. The average of the twolast BP readings was used. The survey was approved by theethical research committee of the Faculty of Biology andMedicine, UniversityofLausanne.Consentwassoughtfromdirectorsofallschools.Signed consent ofoneoftheparentsand of the child was obtained. Seychelles surveys The surveys were described in details elsewhere [12,13].Briefly, school-based surveys were conducted annually between 1998 and 2006 among all students in four schoolgrades, that is, kindergarten (K) and 4th (G4), 7th (G7), and10th(G10)yearsofcompulsoryschool.Datawerecollectedduring routine medical visits at school by trained schoolnurses. BP was measured according to a standardizedprotocol with clinically validated oscillometric automateddevices (Omron M5) [16]. BP was measured in the sittingposition after a rest of at least 5min on the right arm, with acuff of appropriate size for the arm circumference. Tworeadings were obtained at 1-min intervals and averaged.The surveys were approved by the Ministry of Health afterscientific and ethical reviews. Informed consent wasobtained from the parents or guardians, and children werefree to participate. Statistical analyses For all children in these surveys, Z-scores and correspond-ing percentiles of SBP and DBP were generated using theUS reference tables. Elevated SBP and elevated DBP weredefined as SBP and as DBP equal to or above the referentsex, age, and height-specific 95th percentile [1]. Thesedefinitions were considered as the US standard. Childrenhad elevated BP if they had elevated SBP, elevated DBP,or both.Children with BP at or above the height-specific BPthresholds were considered to have elevated BP, elevatedSBP, or elevated DBP, respectively. We assessed the per-formance of these height-specific BP thresholds to identify children with elevated BP, elevated SBP, and elevated DBPaccording to the US standard criteria. Sensitivity, specificity,positive predictive value (PPV), and negative predictive value (NPV) [17–19] were computed separately for childrenin the Swiss survey and children in the Seychelles survey.To assess whether the performance differed according toage and height of the children, we conducted subgroupanalyses by school grades, by height quartiles, and by height centiles category ( < 15th,   15th to  < 50th,   50thto < 85th, and  85th centiles, respectively) using data fromthe Seychelles survey.Finally, we conducted additional sensitivity analyses toassess the impact of choosing sex-specific values forabsolute height BP thresholds; for a given absolute heightthreshold, a range of BP values was obtained for each sexand the mean value was used as the BP cutoff for thecorresponding absolute height threshold; and to assess theimpact of choosing lower or higher values for absoluteheight BP thresholds, that is, the lowest and the highest BP values corresponding to the US 95th percentile within eachheight category, respectively. RESULTS Characteristics of participants in both surveys are shown inTable 2. In the Swiss survey, 5207 children participated(2621 boys, 2586 girls; age range: 10.1–14.9 years). In theSeychelles survey, 25759 children participated (13048boys; 12711 girls; age range: 4.4–18.8 years). The preva-lence of elevated BP (using the standard US BP thresholds) TABLE 2. Characteristics of participants in the Swiss and in theSeychelles surveysSurvey Characteristics  N   or mean SD Switzerland  N   (boys/girls) 5207(2621/2586)Age (year) 12.3 0.5Weight (kg) 44.3 9.4Height (cm) 153.5 7.6SBP (mmHg) 112.9 9.9DBP (mmHg) 65.8 7.1Seychelles  N   (boys/girls) 25759(13048/12711)Age (year) 10.6 4.0Weight (kg) 37.4 17.1Height (cm) 140.5 21.8SBP (mmHg) 103.6 13.1DBP (mmHg) 64.4 9.5 TABLE 1. Blood pressure (BP) thresholds for each absoluteheight categoriesHeightcategory (cm)Heightrange (cm)SBP(mmHg)DBP(mmHg) 80  < 85 104 6190 85–94 107 65100 95–104 110 68110 105–114 112 72120 115–124 114 76130 125–134 117 78140 135–144 120 80150 145–154 123 81160 155–164 128 83170 165–174 131 85180  > 175 136 87 Thresholds to identify elevated blood pressure in children  Journal of Hypertension  1171  Copyright © Lippincott Williams Wilkins. Unauthorized reproduction of this article is prohibited.  was 11.4% in Switzerland and 9.1% in the Seychelles,respectively. The prevalence of elevated systolic BP andof elevated DBP was 10.9 and 1.8% in Switzerland and 5.7and 5.4% in the Seychelles, respectively.In both populations, the performance of height-specificBP thresholds to identify elevated (systolic or diastolic) BP was good, with high sensitivity (Switzerland: 80%;Seychelles: 84%) and very high specificity (Switzerland:99%; Seychelles: 99%) (Table 3, panel A). The PPV washigh (Switzerland: 92%; Seychelles: 91%) and the NPV very high (Switzerland: 97%; Seychelles: 98%). With thesethresholds, 10.0% of Swiss children and 8.4% of Seychelleschildren had elevated BP (Fig. 1). A similar performance was observed for the identification of elevated SBP(Table 3, panel B) and of elevated DBP (Table 3, panel C) in both populations. In the Seychelles, the performanceof absolute-height specific BP thresholds was similar whenassessed separately for each school grade (online supple-ment: Table S1,, for eachheight quartile (online supplement: Table S2,, and for each height centiles category (online supplement: Table S3, A245).The performance of sex-specific absolute-height specificBPthresholdswasalsohighbutnotsubstantiallybetterthannonsex-specific absolute-height specific BP thresholds: thesensitivity ranged between 77 and 87%, the specificity between 99 and 100%, the PPV between 90 and 96%,and the NPV between 97 and 100%, respectively.UsingthelowestBPvaluesinsteadofthemeanBPvaluescorresponding to the US 95th percentile within eachheight category (online supplement: Table S4,, the sensitivity was improved (range:95–99%), the specificity remained high (95–99%) butthe PPV was much lower (59–68%), in both populations.The NPV remained very high (99–100%). With thesethresholds, 15.1% of Swiss children and 12.7% of Seychelleschildren had elevated BP. Using the highest BP valuesinstead of the mean BP values corresponding to the US95th percentile within each height category (online supple-ment: Table S5,, thesensitivity was much lower (52–72%), the specificity andPPV were higher (99–100%), and the NPV remained high(95–99%), in both populations. With these thresholds, 7.2%of Swiss children and 6.2% of Seychelles children hadelevated BP. TABLE 3. Sensitivity, specificity, positive (PPV) and negative predictive (NPV) values of the height-specific blood pressure (BP) thresholdsto identify children with elevated BP (standard US thresholds) in the surveys conducted in Switzerland and Seychelles A) Elevated SBP/DBP usingstandard US thresholdsNo Yes Total Sensitivity 80%Switzerland Elevated SBP/DBP usingheight-specific thresholdsNo 4569 118 4687 Specificity 99%Yes 43 477 520 PPV 92%Total 4612 595 5207 NPV 97%No Yes Total Sensitivity 84%Seychelles Elevated SBP/DBP usingheight-specific thresholdsNo 23213 372 23585 Specificity 99%Yes 204 1970 2174 PPV 91%Total 23417 2342 25759 NPV 98%B) Elevated SBP usingstandard US thresholdsNo Yes Total Sensitivity 80%Switzerland Elevated SBP using height-specific thresholdsNo 4594 114 4708 Specificity 99%Yes 44 455 499 PPV 91%Total 4638 569 5207 NPV 98%No Yes Total Sensitivity 81%Seychelles Elevated SBP using height-specific thresholdsNo 24140 274 24414 Specificity 99%Yes 139 1206 1345 PPV 90%Total 24279 1480 25759 NPV 99%C) Elevated DBP usingstandard US thresholdsNo Yes Total Sensitivity 77%Switzerland Elevated DBP using height-specific thresholdsNo 5107 22 5129 Specificity 100%Yes 4 74 78 PPV 95%Total 5111 96 5207 NPV 100%No Yes Total Sensitivity 85%Seychelles Elevated DBP using height-specific thresholdsNo 24247 209 24456 Specificity 100%Yes 116 1187 1303 PPV 91%Total 24363 1396 25759 NPV 99% Panel A: elevated SBP/DBP; panel B: elevated SBP; panel C: elevated DBP. Chiolero  et al. 1172  Volume 31    Number 6    June 2013  Copyright © Lippincott Williams Wilkins. Unauthorized reproduction of this article is prohibited. DISCUSSION  A simple table of BP thresholds based solely on a child’sabsolute height (instead of sex and age-specific percentilesof height) allowed the identification of children with elev-ated BP in two different populations with high sensitivity and excellent specificity. The performance did not differacross age or height-categories and was not substantially improved with sex-specific BP thresholds. Using lower values for BP thresholds increased the sensitivity butincreased substantially the number of children consideredas having elevated BP and decreased the positive predictive value. Using higher values for BP thresholds decreasedsubstantially the sensitivity. Also using the data from US BP thresholds referencetable, Kaelber  et al.  [6] have produced a simple table of BPthresholds to identify children with elevated BP. In thisstudy, BP thresholds corresponded to the sex and age-specific value for the 90th percentile of BP for the 5thpercentileof height,reducing thenumbersof BP thresholdsfrom 476 to 64. The authors did not assess the performanceof these thresholds but it is expected that the sensitivity  would be very high. However, it is also expected that theproportion of false positive cases will be high because thethresholds to detect elevated BP were set at the 90thpercentile of BP and at the lowest percentile of height.Mitchell  et al.  [8] have produced a table with only 10 BPthresholds, with increments of 5mmHg, accounting for3 years age categories, and regardless of sex or height.Usingthistableinauniversity-basedpediatricclinicleadstoan improved recognition of children with BP in the hyper-tension range, from 15% of cases before implementationto 77% after. This suggests that simple and user-friendly BPthresholds table can help identify children with elevatedBP.For each absolute height threshold, we selected themean of BP values corresponding to the US 95th BPpercentile of different sex, age, and height-categories. Asa consequence, the sensitivity of these thresholds wassuboptimal (80–84%) and some children with elevatedBP were not identified. However, children not identifiedas having elevated BP had slightly elevated BP, that is, a BPlevel relatively close to the 95 th percentile (Fig. 1). Impor-tantly, no children with very high level of BP were missed.Using lower BP thresholds, we had a higher sensitivity toidentify children with elevated BP but at a cost of a sub-stantial increase in the number of false-positive cases.Favoring specificity over sensitivity can be justified becausemodest elevation of BP is not a dangerous condition in theshort term and children with truly sustained elevated BP will be likely identified at further visits.One limitation of this study is that BP was measured atone visit and this does not allow recognition of children with sustained elevated BP [10]. Indeed, the diagnosis of hypertension requires multiple BP measurements at least atthree separate visits [3]. Nevertheless, while the use of asimple tool to identify children with elevated BP at one visitisappealing,standardmethods,inparticulartheassessmentof BP at different visits, are more adequate to confirm thepresence of hypertension. Also, because BP was measuredat one visit, the prevalence of elevated BP was relatively highinbothsurveyscomparedwithstudiesinwhichBPhasbeen measured on repeated occasions [10]. Because BPmay relate to pubertal status, the performance of theproposed thresholds may differ according to pubertal sta-tus. We had no information on pubertal status but thesensitivity analyses did not indicate that the performancediffered across height quartiles. Another limitation is that absolute height-specific BPthresholds were evaluated in school-age children but notin infants and toddlers and no children in both surveys hada height below 92cm. Application to these age groups or tochildren below this height would require further research. A more general limitation is that the US reference BPthresholds are not based on cardiovascular disease(CVD) risk and the use of the 95th percentile to defineelevated BP is arbitrary. There is a need for risk-baseddefinition of elevated BP in children, but it is inherently a difficult task because children have a low absolute risk of CVD. Finally, it would be better to derive height-specific BPthresholds directly from the srcinal dataset used to pro-duce the US BP reference table [1]. On the contrary, a majorstrength of ourstudyistheevaluationoftheperformanceof theproposedheight-basedBPthresholdsusinghighquality  (a) Switzerland(b) SeychellesBP <95th percentileBP ≥ 95th percentileIdentified as havingelevated BP FIGURE 1  Suppose 100 children randomly chosen within each of the Swiss and Seychelles surveys and ranked according to the level of blood pressure (BP), from thelowest to the highest value. Children with elevated BP (  95th sex-, age-, and height-specific US reference value) are shaded in grey. In Switzerland (panel A), 11 childrenhave elevated (systolic and/or diastolic) BP. Using absolute height-specific BP thresholds, nine (squared) out of these 11 children would be identified as having elevated BP,2 with slightly elevated BP would not. In the Seychelles (Panel b), nine children have elevated BP. Using absolute height-specific BP thresholds, seven (squared) out of thesenine children would be identified as having elevated BP, two with slightly elevated BP would not. Thresholds to identify elevated blood pressure in children  Journal of Hypertension  1173  Copyright © Lippincott Williams Wilkins. Unauthorized reproduction of this article is prohibited. data from large and school-based surveys conducted in twohighly different populations, that is, Switzerland [10,11] andin the Seychelles. [12,13].In conclusion, simple user-friendly tools to identify children with elevated BP are needed by clinicians. Weshow that using a simple table with a limited number of absolute-height specific BP thresholds allows the identifi-cation of children with elevated BP with confidence. Othermethods have been proposed to identify children withelevated BP, for example, using simple formulae account-ing for age [5,7] or the BP-to-height ratio [9]. New techno- logical tools could also help assess BP percentiles inchildren, including applications for smartphones [20,21]andwebonlinecalculators[21,22].Nevertheless,thesetoolsrequire the proper electronic technology or being online atthe time of the visit. A simple BP reference table may wellbecome a preferred instrument for the screening in every day clinical practice for clinicians heavily constrained by time. Further studies should be conducted to evaluate whether the proposed table can be used in different clinicalsettings and whether this tool can improve the detection of children with elevated BP. ACKNOWLEDGEMENTS No specific funding for this study. Conflicts of interest There are no conflicts of interest. REFERENCES 1. The fourth report on the diagnosis, evaluation, and treatment of highblood pressure in children and adolescents.  Pediatrics  . 2004; 114:555– 576.2. Lurbe E, Cifkova R, CruickshankJK, Dillon MJ, Ferreira I, Invitti C,  et al. Management of high blood pressure in children and adolescents:recommendations of the European Society of Hypertension.  J Hyper-tens   2009; 27:1719–1742.3. Chiolero A, Bovet P, Paradis G. Screening for elevated blood pressurein children. A critical appraisal.  JAMA Pediatric   2013; 167:266–273.4. Hansen ML, Gunn PW, Kaelber DC. Underdiagnosis of hypertension inchildren and adolescents.  JAMA  2007; 298:874–879.5. Somu S, Sundaram B, Kamalanathan AN. Early detection of hyperten-sion in general practice.  Arch Dis Child   2003; 88:302.6. KaelberDC, PickettF. Simpletable to identifychildrenandadolescentsneeding further evaluation of blood pressure.  Pediatrics   2009;123:e972–e974.7. Simonetti GD, Rizzi M, Bianchetti MG. Simple references for managingarterial hypertension in children with kidney disease.  J Hypertens   2010;28:1109.8. MitchellCK, TheriotJA,SayatJG, MuchantDG, FrancoSM. Asimplifiedtable improves the recognition of paediatric hypertension.  J Paediatr Child Health  2011; 47:22–26.9. Lu Q, Ma CM, Yin FZ, Liu BW, Lou DH, Liu XL. How to simplify thediagnostic criteria of hypertension in adolescents.  J Hum Hypertens  2011; 25:159–163.10. Chiolero A, Cachat F, Burnier M, Paccaud F, Bovet P. Prevalenceof hypertension in schoolchildren based on repeated measure-ments and association with overweight.  J Hypertens   2007; 25:2209– 2217.11. Chiolero A, Paccaud F, Bovet P. Prehypertension and hypertensionamong adolescents of Switzerland.  J Pediatr   2007; 151:e24–e25.12. Chiolero A, Madeleine G, Gabriel A, Burnier M, Paccaud F, Bovet P.Prevalenceofelevatedbloodpressureandassociationwithoverweightin children of a rapidly developing country.  J Hum Hypertens   2007;21:120–127.13. Chiolero A, Paradis G, Madeleine G, Hanley JA, Paccaud F, Bovet P.Discordant secular trends in elevated blood pressure and obesity inchildren and adolescents in a rapidly developing country.  Circulation 2009; 119:558–565.14. Topouchian JA, El Assaad MA, Orobinskaia LV, El Feghali RN, AsmarRG. Validationof twoautomaticdevices for self-measurementof bloodpressure according to the international protocol of the EuropeanSociety of Hypertension: the Omron M6 (HEM-7001-E) and the OmronR7 (HEM 637-IT).  Blood Press Monit   2006; 11:165–171.15. Stergiou GS, Yiannes NG, Rarra VC. Validation of the omron 705 itoscillometric device for home blood pressure measurement in childrenand adolescents: the Arsakion School Study.  Blood Press Monit   2006;11:229–234.16. El AssaadMA, TopouchianJA,AsmarRG. Evaluationof two devicesforself-measurement of blood pressure according to the internationalprotocol: the Omron M5-I and the Omron 705IT.  Blood Press Monit  2003; 8:127–133.17. Altman DG, Bland JM. Diagnostic tests. 1: sensitivity and specificity.  BMJ   1994; 308:1552.18. Altman DG, Bland JM. Diagnostic tests 2: predictive values.  BMJ   1994;309:102.19. MallettS,HalliganS,ThompsonM,CollinsGS,AltmanDG.Interpretingdiagnostic accuracy studies for patient care.  BMJ   2012; 345:e3999.20. Pediacents., [Accessed on 7 August 2012].21. MedCalc 3000 Cardiac for iPhone, iPod and Android. [Accessed on 7 August 2012].22. Age-based Pediatric Blood Pressure Reference Charts. [Accessed on 7 August 2012]. Reviewer’s Summary Evaluation Reviewer 1 Identification of children with elevated blood pressure isdifficult because of the multiple sex-, age-, and height-specific thresholds to define elevated blood pressure. Chil-dren are now commonly having their blood pressuremeasured, but clinicians need to have better tools to recog-nize and diagnose high-normal blood pressure andhypertension. Therefore, Chiolero and co-workers proposeabsolute height-specific thresholds to identify abnormalbloodpressurevalues.Thethresholdswerecalculatedusingonly one blood pressure measurement. However, the diag-nosis of hypertension should be based on multiple officeblood pressure measurements, taken on separate occasions.The problem of pediatric hypertension is among us, and thedifficultyofdefininghighbloodpressurethresholdvaluesinchildren and adolescents is once more reinforced. Chiolero  et al. 1174  Volume 31    Number 6    June 2013
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks