Objectives. To describe the current prevalence of pediatric hypertension and the relationships between gender, ethnicity, overweight, and blood pressure.
Methods. School-based screening was performed in 5102 children (13.5 ± 1.7 years) from May through November 2002. Age, gender, ethnicity, weight, and height were ascertained, and body mass index (BMI) was calculated as weight (kg)/height (m2). Overweight was defined as BMI ≥95th percentile. Students with blood pressure >95th percentile on the first screening underwent a second screening 1 to 2 weeks later, and then a third screening if blood pressure was >95th percentile at the second screening.
Results. Ethnicity distribution was 44% white, 25% Hispanic, 22% African American, and 7% Asian. Overall, overweight prevalence was 20%, which varied significantly by ethnicity (31% Hispanic, 20% African American, 15% white, and 11% Asian). The prevalence of elevated blood pressure after first, second, and third screenings was 19.4%, 9.5%, and 4.5%, respectively. Elevated blood pressure on first screening was highest among Hispanics (25%) and lowest among Asians (14%). Ethnic differences in the prevalence of hypertension (elevated blood pressure on 3 screenings) were not significant after controlling for overweight. The prevalence of hypertension increased progressively as the BMI percentile increased from ≤5th percentile (2%) to ≥95th percentile (11%). After adjustment for gender, ethnicity, overweight, and age, the relative risk of hypertension was significant for gender (relative risk: 1.50; confidence interval: 1.15, 1.95) and overweight (relative risk: 3.26; confidence interval: 2.50, 4.24).
Conclusions. These results confirm an evolving epidemic of cardiovascular risk in youth, as evidenced by an increase in the prevalence of overweight and hypertension, notably among ethnic minority children.
The prevalence of childhood overweight has increased markedly over the last 2 decades.1 In concert with this increasing prevalence, anecdotal evidence suggests that pediatric hypertension may also have become more prevalent than previously reported. This increase reflects an epidemiologic shift from secondary hypertension (most often caused by renal disease) to primary (ie, essential) hypertension as the main cause of hypertension in the pediatric age range. Once considered relatively rare, primary hypertension in children has become increasingly common in association with other cardiovascular risk factors that include overweight, insulin resistance, and dyslipidemia. The association between overweight and hypertension in children has been reported in a variety of ethnic and racial groups, with virtually all studies finding higher blood pressures and/or higher prevalences of hypertension in overweight compared with lean children.2–9 The largest study in children to investigate the interaction between race, overweight, and hypertension reported that the odds of elevated blood pressure was significantly higher for the upper compared with the lower decile of body mass index (BMI), with few systematic racial differences identified.10 However, this study is inadequate to describe current ethnicity-specific trends in pediatric hypertension, because the analysis included only African American and white children and consisted of data collected only through 1991.
These limitations are highly relevant for several reasons. Hispanics are the fastest growing ethnic group in the United States, in whom the overweight epidemic has been particularly notable. Among Mexican American adolescents, the prevalence of overweight increased >10 percentage points between 1988 and 1994 and 1999 and2000.11 The National Longitudinal Survey of Youth reported that, in 1998, overweight prevalence had increased in all major ethnic groups, to 21.5% among African Americans, 21.8% among Hispanics, and 12.3% among non-Hispanic whites.12 Although large community-based screening studies have reported the prevalence of pediatric hypertension in previous decades,13–17 these studies have limited contemporary relevance because of limitations that include use of hypertension definitions that preceded the task-force reports, narrow age ranges, limited ethnic distributions, and reliance on a single set of blood pressure measurements. This last limitation is particularly important, because the task force recommends that elevated blood pressure be confirmed on at least 3 separate occasions before diagnosing hypertension. No recent studies have reported the prevalence of persistent hypertension using this recommendation. To describe the current prevalence of pediatric hypertension using the task-force-recommended definition and to determine whether ethnicity is a modifier of hypertension risk independent of overweight, we performed systematic hypertension and overweight screening in multiethnic, school-aged children in public schools in the Houston, Texas metropolitan area.
School-based hypertension and overweight screening was performed in 5102 students aged 10 to 19 years in 8 Houston public schools from May 2002 to November 2002. Schools were selected for screening with the intent of providing an overall ethnic distribution comparable to recent Houston census data. Because of the noninvasive nature of the protocol consisting only of measurements considered routine for school-entry/sports-participation examinations (ie, height, weight, and blood pressure), a “passive consent” process was used. Parents were notified in advance by letter sent by from each school that hypertension and overweight screening would be performed at the school. Parents were provided with a form that was to be signed and returned if the parents did not wish their child to participate. The study was approved by the University of Texas Health Science Center (Houston, TX) Committee for the Protection of Human Subjects.
Age was determined based on the date of the first screening measurements. Gender and ethnicity were self-described. For the purpose of analysis within the database, self-described ethnicity was categorized as white, African American, Hispanic, Asian, or other. Weight and height were measured in each student, and BMI was calculated as weight (kg)/height (m2). The most-current pediatric anthropometric reference data, specific for gender and age, were used to establish height, weight, and BMI percentile for each student.18 Z scores of BMI were generated from equations provided by the Centers for Disease Control and Prevention, and exact BMI percentiles were calculated for each student. BMI percentile categories of ≤5th, 10th, 25th, 50th, 75th, 90th, and ≥95th were generated by rounding the exact BMI percentile to the nearest categorical threshold. “Overweight” was defined as BMI ≥95th percentile.
At each school screening, 3 seated blood pressure and heart rate measurements were made at least 1 minute apart after ∼3 minutes of rest by using SpaceLabs oscillometric monitors (SpaceLabs, Inc, Redmond, WA). Students found to have an average systolic blood pressure (SBP) or diastolic blood pressure (DBP) greater than or equal to the gender, age, and height-percentile-specific 95th percentile blood pressure value as defined by ref 19 were contacted to undergo a second set of blood pressure measurements 1 to 2 weeks later. Students found to have blood pressure >95th percentile at the second screening were contacted again to undergo a third set of blood pressure measurements an additional 1 to 2 weeks later. Because not every student found to be hypertensive at first or second screening was able to be rescreened as described in the protocol, the prevalence of hypertension at the second and third screenings was calculated by dividing the number of students with elevated blood pressure at that screening by the sum of 1) the number of normotensive students at the previous screening(s) and 2) the number of students with elevated blood pressure at the previous screening(s) who were rescreened per protocol. This approach provides the most-reliable extrapolation of the hypertension prevalence to the entire study population.
Descriptive statistics are presented as percentages, means, and standard deviations (SDs). Two sample Student’s t tests were performed for between-group comparisons of continuous variables. Simple linear regression analysis was used to determine the pairwise association between continuous variables. Multiple linear regression was used to determine the relationship between multiple continuous variables. Multiple logistic regression on covariates was used to determine the probability of having hypertension based on different clinical variables. Relative risk of hypertension (unadjusted and adjusted) was calculated by binomial regression with a log link function. A P value <.05 was used to indicate statistical significance.
Demographic and clinical variables of students in the school-based screening are shown in Table 1. The most-common ethnicity was white, followed by Hispanic, African American, Asian, and other. Overall, the prevalence of overweight (BMI ≥95th percentile) was 20%. The distribution of BMI percentile categories for each of the 3 major ethnicities in the study (white, African American, and Hispanic) is shown in Fig 1. Males had a higher prevalence of overweight than females (24% vs 16%; P < .001). Ethnicity was a significant determinant of overweight status (P < .001). The prevalence of overweight was highest among Hispanics (31%), followed by African Americans (20%), whites (15%), and Asians (11%).
The overall results of the screening for elevated blood pressure are shown schematically in Fig 2. The prevalence of elevated blood pressure on first screen was 19.4%. Among those students found to have elevated blood pressure at first screen, 85% underwent a second set of measurements (842 of 990). Some students failed to be rescreened due to various reasons including school absenteeism and conflicting school activities/responsibilities. No systematic differences were found between students with elevated blood pressure that were rescreened as compared with those that were not. The prevalence of elevated blood pressure among students who underwent follow-up second-screening measurements was 56% (472 of 842). The estimated prevalence of elevated blood pressure after 2 screenings was calculated as the number of students with elevated blood pressure at second screening (n = 472) divided by the number of students who had normal blood pressure at first screening (n = 4112) plus the number of students who had elevated blood pressure at first screening and were rescreened successfully (n = 842). Using this method, the estimated prevalence of elevated blood pressure after 2 screenings was 9.5% (472 of 4954). Among those students found to have elevated blood pressure after the second screen, 86% underwent a third set of measurements (407 of 472). The prevalence of persistently elevated blood pressure among students who underwent follow-up third screening was 54% (221 of 407). The prevalence of hypertension after 3 screenings was calculated as the number of students with elevated blood pressure at third screening (n = 221) divided by the number of students who had normal blood pressure at first or second screening (n = 4482) plus the number of students who had elevated blood pressure at second screening and were rescreened successfully (n = 407). Using this method, the overall estimated prevalence of hypertension (using the task-force definition of 3 sets of elevated measurements) was 4.5% (221 of 4889).
The effects of gender (male versus female), ethnicity (African American/Hispanic/Asian versus white), and BMI percentile (continuous) on the presence of elevated blood pressure after first screening and after third screening were evaluated by simple logistic regression (Table 2). Males, as compared with females, had a higher prevalence of elevated blood pressure on first screening (23% vs 16%; P < .001) and on third screening (5.7% vs 3.4%; P < .001). An overall significant ethnic difference was found in the prevalence of elevated blood pressure on first screening (P < .001). Specifically, the prevalence of elevated blood pressure on first screening was highest among Hispanics (25%), followed by whites (18%), African Americans (17%), and Asians (16%). Systematic ethnic differences in the prevalence of elevated blood pressure after 3 sets of measurements did not persist. Hypertension was found among Hispanics (5.6%), African Americans (4.6%), whites (4.1%), and Asians (3.7%) with relatively equal frequency.
The strongest determinant of hypertension was BMI percentile (Table 2). A greater than fourfold increase in the prevalence of elevated blood pressure at first screening was observed as BMI percentile increased from ≤5th percentile (9%) to ≥95th percentile (38%). As expected, the overall prevalence of hypertension after third screening was lower for each BMI percentile category as compared with the first screening (Fig 3). A greater than fivefold increase in the prevalence of hypertension was found as BMI increased from the ≤5th to ≥95th percentile. The relationship of SBP and DBP to BMI percentile was also determined for the range of categorical BMI percentiles. SBP showed a progressive increase with each increase in BMI percentile (P < .001), whereas DBP showed no association (Fig 4).
The unadjusted and adjusted relative risk of hypertension was determined for all variables found to be associated with elevated blood pressure at first screening by univariate analysis (Fig 5 a). Significant unadjusted relative risk was found for overweight (2.54; confidence interval [CI]: 2.28–2.83), male gender (1.46; CI: 1.31–1.64), and Hispanic ethnicity (1.38; CI: 1.21–1.57). Relative risk by ethnicity was determined based on the comparison with whites (ie, white = 1.0). Age did not alter the risk of elevated blood pressure significantly. After simultaneous adjustment for gender, ethnicity, overweight, and age, the relative risk for elevated blood pressure at first screening remained significant for overweight (2.38; CI: 2.13–2.65), male gender (1.32; CI: 1.18–1.47), and Hispanic ethnicity (1.16; CI: 1.02–1.32). The same relative risk analysis was performed for hypertension confirmed at third screening (Fig 5 b). As for first screening, significant unadjusted relative risk was found for overweight (3.49; CI: 2.70–4.51), male gender (1.68; CI: 1.29–2.19), and Hispanic ethnicity (1.38; CI: 1.01–1.88). After simultaneous adjustment for gender, ethnicity, overweight, and age, a significant relative risk remained for overweight (3.26; CI: 2.50–4.24) and male gender (1.50; CI: 1.15–1.95). However, no significant relative risk by ethnic group was found in the adjusted analysis.
The extent and severity of childhood overweight has been reported recently to be increasing, particularly in minority populations.11 The results of the current study are confirmatory of recent estimates of overweight in children and also highlight how this high prevalence of overweight is contributing to a worsening cardiovascular risk profile in affected children. Overall, the prevalence of overweight and hypertension among children was found to be substantially higher than that reported in previous decades. Of school-aged children in the Houston area (age range 10–19 years), 20% had BMI ≥95th percentile. Within this population of >5000 children, the prevalence of overweight among Hispanic children was more than twice that of white children and 3 times that of Asian children. These prevalences are similar to those reported in the National Longitudinal Survey of Youth from 199812 but show an even higher prevalence of overweight in the Hispanic children (30% vs 22%). This higher prevalence may reflect the trend noted in this previous study that overweight was more prevalent in southern regions of the United States and indicate the progressive nature of the overweight epidemic. It is noteworthy that the ethnic distribution in the current study closely follows that reported from the year 2000 Houston census data, in which the ethnic distribution of Houston was reported to be 46% white, 30% Hispanic, 17% African American, and 5% Asian. As shown in Table 1, the ethnic distribution of the screened children closely resembles this overall ethnic distribution in Houston and thereby supports the generalizability of the findings to children of this region.
Historically, the reported prevalence of pediatric hypertension in the United States has varied greatly depending, in part, on the way in which hypertension is defined and the number of occasions on which blood pressure is measured.13–17 The prevalence of pediatric hypertension using the task-force guideline of 3 sets of confirmatory measurements has been estimated to be ∼1%. However, no recent studies have performed systematic screening for hypertension following this task-force guidance. In the current study, the overall prevalence of hypertension after 3 sets of measurements was 4.5%, more than fourfold higher than this 1% estimate. This higher overall prevalence was clearly and strongly determined by the skewed distribution to higher BMI percentiles in the study population. Children with BMI <85th percentile had a hypertension prevalence of only 2.6%. In comparison, children with BMI ≥95th percentile had a hypertension prevalence of 10.7%. Given the overall prevalence of overweight in the total study population of 20%, the tendency toward overweight was clearly the major contributing factor for the overall higher prevalence of hypertension found in the current study.
These results also illustrate that the classification of weight status into dichotomous categories of “overweight” or “nonoverweight,” although clinically useful, is somewhat limited. These arbitrary percentile-based categories of body habitus neglect to capture the continuous relationship between adiposity and blood pressure. Almost without exception, the prevalence of elevated blood pressure and/or hypertension rose with each successive increase in BMI percentile, even within the presumably normal range of BMI. These trends were observed at each of the first, second, and third screenings. Because both the BMI percentiles and the definition of elevated blood pressure account for normal, gender-specific, age-related increases in BMI and blood pressure, this trend seems to reflect a relationship that is independent of normal physical maturation. Thus, the relationship between BMI and hypertension cannot be characterized by a threshold effect but instead represents a continuous relationship, suggesting that increasing BMI percentile translates to incremental cardiovascular risk
The current study also provides strong evidence that ethnicity is not an independent determinant of hypertension risk in children and adolescents. This finding is noteworthy, because this is the only study to include all major ethnic groups in the analysis. Recent previous studies investigating systematic differences in the prevalence of hypertension based on ethnicity have been contradictory. Although Dekkers et al20 found higher blood pressure in African American as compared with white children even when controlling for adiposity, Rosner et al10 found no such systematic racial difference. Although Hispanics had a significantly higher unadjusted relative risk of elevated blood pressure after each set of screening measurements in the current study, no ethnic differences were found in the prevalence of hypertension after 3 sets of screening measurements after adjusting for overweight. These results further indicate that the primary contributing factor to hypertension among all ethnic groups, particularly Hispanics, was overweight.
The current study also highlights the importance of confirming elevated blood pressure on multiple occasions before labeling a child as hypertensive. Even among students with 2 sets of elevated blood pressure measurements, only 54% were persistently hypertensive at the third screening. These results raise the question of how long a particular child should be observed before considering antihypertensive therapy. Berenson et al21 showed in a pediatric antihypertensive medication study that blood pressure continued to fall over a 6-month observation period in a parallel, untreated control group despite 5 confirmatory blood pressure measurements over a 4-month observation period before study entry. Similarly, Sorof et al22 found in a recent antihypertensive trial that 17% of subjects initially labeled as persistently hypertensive normalized blood pressure during a 2-week placebo screening phase. If additional screening measurements had been performed on subsequent occasions in the current study, it is almost certain that the overall prevalence of hypertension would have continued to decrease. Although the current recommendation for diagnosing hypertension in children is to confirm that blood pressure is >95th percentile on each of 3 different occasions, it should be noted that failure to confirm hypertension on all occasions does not necessarily translate to low risk. Because of blood pressure variability, some children may have repeat blood pressure just below the 95th percentile, thereby removing them from the hypertensive category on that measurement. Nonetheless, many of these children will continue to have blood pressure in the high normal range, thus placing them at substantial risk for future hypertension.
Methodological issues may also have contributed to the high prevalence of hypertension in the current study. Automated oscillometric monitors, as opposed to auscultatory mercury manometers, were used for blood pressure measurement. Although oscillometric monitors are designed to reproduce mercury measurements accurately, the reliance on pressure oscillations as opposed to Korotkoff sounds inevitably results in deviations from the gold standard of indirect auscultatory blood pressure measurement. Unlike mercury manometers, the accuracy of which is determined primarily by observer technique, the accuracy of oscillometric monitors is determined by the validity of the manufacturers’ proprietary algorithms used to derive the SBP and DBP values. The type of monitor used in the current study has undergone independent validation, the results of which have been contradictory. A study of adults reported excellent reliability as compared with manual auscultatory measurements,23 whereas other studies in children have reported overestimation of SBP by ∼4 to 5 mm Hg.24,25 These issues regarding the validation of oscillometric monitors have become increasingly important for pediatric clinical practice, because hospitals and clinics have begun to mandate the removal of all mercury-containing equipment from patient care areas, and fewer care providers are undergoing formal training and certification in the manual measurement of blood pressure. In most patient care settings, oscillometric monitors are used almost exclusively for blood pressure measurement in children. Given that perfect agreement between oscillometric and mercury measurements will never be possible, future efforts in the field may require development of oscillometric normative data analogous to the task-force data set.
The current study confirms the presence of an evolving epidemic of cardiovascular risk in youth. The prevalence of overweight and hypertension in children has increased dramatically, most notably among ethnic minorities. Although the unadjusted relative risk of hypertension was higher in some ethnic groups, adjustment for the higher prevalence of overweight showed no ethnic predisposition to hypertension. Nonetheless, the fact that ethnicity was not an independent risk factor does not alter the fact that hypertension was more prevalent in Hispanics because they were disproportionately overweight. Other complications of overweight not assessed specifically in this study that are associated with cardiovascular disease, such as dyslipidemia, insulin resistance, glucose intolerance, type II diabetes mellitus, microalbuminuria, left ventricular hypertrophy, and pulmonary hypertension due to obstructive sleep apnea, are likely also highly prevalent in this population. Unless successful interventions and prevention strategies can be instituted at the local and national level, these observations suggest that the trend of decreasing cardiovascular disease in adults observed over the past 50 years may be reversed as the current population of overweight children and adolescents become adults.
This work was supported by grants from the National Heart, Lung, and Blood Institute (K23 HL04217-01A1 “End-Organ Injury in Hypertensive Children”), King Pharmaceuticals, Inc, and Wyeth Pharmaceuticals.
- Received March 4, 2003.
- Accepted June 6, 2003.
- Reprint requests to (J.M.S.) Division of Pediatric Nephrology and Hypertension, University of Texas Medical School, 6431 Fannin St, Rm 3.124, Houston, TX 77030. E-mail:
- ↵Ogden CL, Troiano RP, Briefel RR, Kuczmarski RJ, Flegal KM, Johnson CL. Prevalence of overweight among preschool children in the United States, 1971 through 1994. Pediatrics.1997;99(4) . Available at: www.pediatrics.org/cgi/content/full/99/4/e1
- Guillaume M, Lapidus L, Beckers F, Lambert A, Bjorntorp P. Cardiovascular risk factors in children from the Belgian province of Luxembourg. The Belgian Luxembourg Child Study. Am J Epidemiol.1996;144 :867– 880
- Chu NF, Rimm EB, Wang DJ, Liou HS, Shieh SM. Clustering of cardiovascular disease risk factors among obese schoolchildren: the Taipei Children Heart Study. Am J Clin Nutr.1998;67 :1141– 1146
- Freedman DS, Dietz WH, Srinivasan SR, Berenson GS. The relation of overweight to cardiovascular risk factors among children and adolescents: the Bogalusa Heart Study. Pediatrics.1999;103 :1175– 1182
- ↵Rosner B, Prineas R, Daniels SR, Loggie J. Blood pressure differences between blacks and whites in relation to body size among US children and adolescents. Am J Epidemiol.2000;151 :1007– 1019
- ↵Fixler DE, Laird WP, Fitzgerald V, Stead S, Adams R. Hypertension screening in schools: results of the Dallas study. Pediatrics.1979;63 :32– 36
- Rames LK, Clarke WR, Connor WE, Reiter MA, Lauer RM. Normal blood pressure and the evaluation of sustained blood pressure elevation in childhood: the Muscatine study. Pediatrics.1978;61 :245– 251
- ↵Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, et al. CDC growth charts: United States. Adv Data.2000;(314) :1– 27
- Update on the 1987 Task Force Report on High Blood Pressure in Children and Adolescents: a working group report from the National High Blood Pressure Education Program. National High Blood Pressure Education Program Working Group on Hypertension Control in Children and Adolescents. Pediatrics1996;98 :649– 658
- Copyright © 2004 by the American Academy of Pediatrics