Global Prevalence of Hypertension in Children: A Systematic Review and Meta-analysis

Key Points

Question  What is the prevalence of hypertension in the general pediatric population?

Findings  In this systematic review and meta-analysis of 47 articles, the prevalence of childhood hypertension increased from 1994 to 2018 and the increase was associated with higher body mass index, with the pooled estimate being 4.00% among individuals 19 years and younger. In 2015, the prevalence of childhood hypertension ranged from 4.32% among children aged 6 years to 3.28% among those aged 19 years and peaked at 7.89% among those aged 14 years.

Meaning  The findings suggest that childhood hypertension is becoming more common in the general pediatric population, representing a considerable public health challenge worldwide.

Importance  Reliable estimates of the prevalence of childhood hypertension serve as the basis for adequate prevention and treatment. However, the prevalence of childhood hypertension has rarely been synthesized at the global level.

Objective  To conduct a systematic review and meta-analysis to assess the prevalence of hypertension in the general pediatric population.

Data Sources  PubMed, MEDLINE, Embase, Global Health, and Global Health Library were searched from inception until June 2018, using search terms related to hypertension (hypertension OR high blood pressure OR elevated blood pressure), children (children OR adolescents), and prevalence (prevalence OR epidemiology).

Study Selection  Studies that were conducted in the general pediatric population and quantified the prevalence of childhood hypertension were eligible. Included studies had blood pressure measurements from at least 3 separate occasions.

Data Extraction and Synthesis  Two authors independently extracted data. Random-effects meta-analysis was used to derive the pooled prevalence. Variations in the prevalence estimates in different subgroups, including age group, sex, setting, device, investigation period, BMI group, World Health Organization region and World Bank region, were examined by subgroup meta-analysis. Meta-regression was used to establish the age-specific prevalence of childhood hypertension and to assess its secular trend.

Main Outcomes and Measures  Prevalence of childhood hypertension overall and by subgroup.

Results  A total of 47 articles were included in the meta-analysis. The pooled prevalence was 4.00% (95% CI, 3.29%-4.78%) for hypertension, 9.67% (95% CI, 7.26%-12.38%) for prehypertension, 4.00% (95% CI, 2.10%-6.48%) for stage 1 hypertension, and 0.95% (95% CI, 0.48%-1.57%) for stage 2 hypertension in children 19 years and younger. In subgroup meta-analyses, the prevalence of childhood hypertension was higher when measured by aneroid sphygmomanometer (7.23% vs 4.59% by mercury sphygmomanometer vs 2.94% by oscillometric sphygmomanometer) and among overweight and obese children (15.27% and 4.99% vs 1.90% among normal-weight children). A trend of increasing prevalence of childhood hypertension was observed during the past 2 decades, with a relative increasing rate of 75% to 79% from 2000 to 2015. In 2015, the prevalence of hypertension ranged from 4.32% (95% CI, 2.79%-6.63%) among children aged 6 years to 3.28% (95% CI, 2.25%-4.77%) among those aged 19 years and peaked at 7.89% (95% CI, 5.75%-10.75%) among those aged 14 years.

Conclusions and Relevance  This study provides a global estimation of childhood hypertension prevalence based on blood pressure measurements in at least 3 separate visits. More high-quality epidemiologic investigations on childhood hypertension are still needed.

Hypertension, also known as elevated blood pressure (BP), is a well-recognized risk factor for cardiovascular diseases and chronic kidney disease worldwide.^1-3 Hypertension also substantially contributes to mortality and disability.³ Globally, more than 1 billion adults were living with hypertension in 2015, among whom most were in low- and middle-income countries.^2,4

Previous pathophysiologic and epidemiologic evidence has suggested that childhood hypertension is associated with essential hypertension in adulthood and detrimental lifelong cardiovascular events.^5-7 Compared with that of adulthood hypertension, the measurement of childhood hypertension is relatively complicated and unstable.^8,9 The prevalence of elevated BP in children, defined as a systolic BP (SBP) or a diastolic BP (DBP) greater than or equal to the 95th percentile by sex, age, and height, has been suggested to sustainably decrease by 53.7% in the second visit and by 77.7% in the third visit compared with the first visit.¹⁰ Therefore, the fourth report from the National High Blood Pressure Education Program (NHBPEP) Working Group in the United States has suggested that childhood hypertension be confirmed as a high BP on at least 3 separate occasions, and the cutoffs of high BP should simultaneously account for the variations of age, sex, and body size.⁹

From the public health perspective, reliable estimates of the prevalence of childhood hypertension serve as the basis for adequate prevention and treatment, as well as evidence-based health resource allocation and policy making. Despite the existence of a large volume of studies that have assessed the prevalence of hypertension in children and adolescents, to our knowledge, the prevalence estimates of childhood hypertension have rarely been synthesized at the global level.^11-13

To fill this gap of knowledge, we conducted a systematic review of studies that reported the prevalence of hypertension or elevated BP in children. We aimed to assess the prevalence of childhood hypertension, prehypertension, and stage 1 and stage 2 hypertension at the global level. When possible, the factors potentially associated with childhood hypertension were also explored.

This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline.¹⁴ The protocol of this study was not preregistered.

From inception to June 2018, 2 researchers (Y. Zhang and J.Y.) independently conducted a literature search in PubMed, MEDLINE, Embase, Global Health, and Global Health Library by using a combination of search terms related to hypertension (hypertension or high blood pressure or elevated blood pressure), children (children or adolescents), and prevalence (prevalence or epidemiology). Then a search of reference lists of the included studies in the first step was performed to complement our database searches. No language or time restrictions were applied. The full search strategies for different bibliographic databases are presented in eTable 1 in the Supplement.

To be included in this systematic review, studies needed to be primary investigations based on a generally representative sample of children or adolescents (≤19 years of age) and provide numerical prevalence estimates of hypertension, prehypertension, stage 1 hypertension, stage 2 hypertension, or different phenotypes of hypertension (systolic hypertension, diastolic hypertension, isolated systolic hypertension, isolated diastolic hypertension, or systolic-diastolic hypertension). Only studies that reported the prevalence of systematic hypertension (rather than intracranial or pulmonary hypertension) were included. For studies that were conducted for both adults and children, the prevalence data of hypertension had to be able to be disaggregated for the pediatric group. The adopted methods of measuring BP and definitions of hypertension had to be explicitly described. To avoid an overestimation, only studies that repeated BP measurements on at least 3 separate occasions were eligible.⁹ Furthermore, the diagnosis of hypertension should have been performed according to the distribution curves of SBP and DBP, observing the corresponding values at different percentiles.⁹ Studies that were confined to a subgroup of children who were not representative of the general pediatric population (eg, obese children, children with specific diseases, and young athletes) were excluded. For multiple articles that used data from the same investigation (duplicates), the one with the most comprehensive results or the largest sample size was kept. However, when different aspects or subgroups of the same investigation were separately reported in different articles, all those articles were kept.

After removing duplicates from different bibliographic databases, 2 researchers (Y. Zhang and J.Y.) independently screened the titles and abstracts of all retrieved records from the literature search. Then the same 2 researchers assessed the eligibility of potentially relevant articles in full text against the selection criteria. Consensus was reached for any disagreements through discussion.

In different articles, the term of elevated BP was not unified. An SBP or DBP of greater than or equal to the 90th percentile but less than the 95th percentile could be termed as high-normal or prehypertension; similarly, an SBP or DBP of greater than or equal to the 95th percentile could be high BP, elevated BP, or hypertension.^15-18 To ensure comparability among studies and our ability to synthesize prevalence data, the definition of hypertension in this study was prestandardized in accordance with the fourth report of the NHBPEP (Table 1).⁹

Data were independently extracted from the included articles by 2 researchers (Y. Zhang and M.Z.). The collected information included title, author(s), year of publication, year of investigation, study location (country, setting [urban vs rural], and region), study design, sampling strategy, diagnostic criteria, device for BP measurement (aneroid, oscillometric, and mercury), sample size, age range, and the number of participants affected by hypertension. The regions of study location were designated as African Region, Region of the Americas, Southeast Asia Region, European Region, Eastern Mediterranean Region, and Western Pacific Region according to the World Health Organization (WHO) criteria and as high-income countries and low- and middle-income countries according to the World Bank (WB) criteria. For studies in which the investigation date was not provided, we imputed the year of investigation by subtracting 4 years from the year of publication based on the mean time difference between the year of investigation and publication in which data were provided (eTable 2 in the Supplement).

We rated the quality of included articles according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline in 5 dimensions: sample population, sample size, participation rate, outcome assessment, and analytical methods (eTable 3 in the Supplement).¹⁹ The total score, ranging from 0 to 10, represented the overall bias risk of each article.

Before pooling prevalence estimates, the variance of the raw prevalence from each included study was stabilized by using the Freeman-Tukey double arc-sine transformation.²⁰ All estimates were presented after back transformation. We assessed heterogeneity of prevalence estimates among studies using the Cochran Q test and I² index.^21,22 For the Cochran Q test, P < .05 represented significant heterogeneity. For I² index, values of 25% or lower corresponded to low degrees of heterogeneity, 26% to 50% to moderate degrees of heterogeneity, and greater than 50% to high degrees of heterogeneity.^21-23 Because of high heterogeneity (as expected and observed), a random-effects (DerSimonian and Laird method) meta-analysis was used to calculate the overall pooled prevalence of hypertension with 95% CIs throughout this study.^23,24 To examine whether single studies had a disproportionally excessive influence, we applied a leave-1-out sensitivity analysis for each meta-analysis.²⁵ Publication bias in the meta-analysis was detected qualitatively by visual inspection of funnel plots and quantitatively by the Egger linear regression test and the Begg rank correlation test when more than 10 estimates were available in a single analysis.^26-28

For childhood hypertension, prehypertension, and stage 1 and stage 2 hypertension, we conducted subgroup meta-analyses to determine the potential sources of heterogeneity. As a rule, at least 3 studies should be available per subgroup.

For childhood hypertension, multiple data points (age- or sex-specific prevalence) were generally reported in a single study. To assess the associations of various sample characteristics and the prevalence of childhood hypertension, we first conducted a univariable meta-regression, followed by a multivariable meta-regression.^29,30 As a rule, at least 10 data points should be available for each variable in univariable meta-regression and 20 in multivariable meta-regression (the eMethods in the Supplement gives more details).^23,31 Data were analyzed using Stata, version 14.0 (StataCorp) and R, version 3.3.0 (R Foundation for Statistical Computing).

As outlined in Figure 1, our initial literature search identified a total of 9084 records. After applying the eligibility criteria, 47 articles were included in our quantitative synthesis, of which 47 articles provided prevalence data on hypertension, 16 on prehypertension, 6 on stage 1 hypertension, and 6 on stage 2 hypertension. The list of the 47 included articles is given in eTable 4 in the Supplement.

The detailed characteristics of the included articles can be found in eTable 4 in the Supplement. All the included articles were based on cross-sectional investigations and defined childhood hypertension in the prespecified standardized manner. A total of 32 of the 47 articles (68%) were published from 2010 onwards, and 22 (47%) were conducted in urban-rural mixed settings. In addition, 29 (62%) of the included articles reported the prevalence data for both boys and girls and 28 (60%) with a sample size greater than 2000. The most commonly used device for measuring BP was mercury sphygmomanometer (19 [40.4%]), followed by oscillometric sphygmomanometer (16 [34.0%]). Moreover, 13 of the 47 articles (28%) were conducted in the Region of the Americas (13 [28%]) or the European Region (13 [28%]) and in low- and middle-income countries (26 [55.3%]). All the included articles had a quality score of at least 6. The detailed quality assessments are presented in eTable 5 in the Supplement.

Table 2 gives the results of overall and subgroup meta-analyses. For childhood hypertension, the pooled prevalence was 4.00% (95% CI, 3.29%-4.78%) by using random-effects meta-analysis (eFigure 1 in the Supplement). The sensitivity analysis showed that the pooled prevalence of hypertension among children varied from 3.85% (95% CI, 3.17%-4.60%) to 4.10% (95% CI, 3.39%-4.88%) after removing a single study at 1 time (eFigure 2 in the Supplement), but no single study had an excessive influence on the pooled prevalence. No publication bias was found based on the funnel plot, Egger test, and Begg test (eFigure 3 in the Supplement). The pooled prevalence of different hypertension phenotypes was also estimated using random-effects models: 2.99% (95% CI, 1.92%-4.29%) for systolic hypertension, 1.87% (95% CI, 1.06%-2.91%) for diastolic hypertension, 1.50% (95% CI, 0.83%-2.36%) for isolated systolic hypertension, 0.73% (95% CI, 0.34%-1.24%) for isolated diastolic hypertension, and 1.25% (95% CI, 0.72%-1.92%) for systolic-diastolic hypertension. Table 2 also gives the prevalence of childhood hypertension according to sex, urban or rural setting, device, investigation period, body mass index (BMI), WHO region, and WB region. The prevalence of childhood hypertension did not differ significantly when stratified by sex, urban or rural setting, WHO region, and WB region. The prevalence of childhood hypertension was the highest when taken by an aneroid sphygmomanometer (7.23%; 95% CI, 3.83%-11.59%) compared with mercury (4.59%; 95% CI, 3.24%-6.15%) or oscillometric (2.94%; 95% CI, 2.37%-3.57%) sphygmomanometers. An upward secular trend in the prevalence of childhood hypertension was detected, by which the prevalence was the highest in the latest period of 2010 to 2014 (6.02%; 95% CI, 4.38%-7.91%) than during the 2000s (3.30%; 95% CI, 2.69%-3.97%) and 1990s (1.26%; 95% CI, 0.79%-1.84%). A difference in childhood prevalence was also noted in different BMI groups, by which obese (15.27%; 95% CI, 7.31%-25.38%) and overweight (4.99%; 95% CI, 2.18%-8.81%) children had substantially higher prevalence estimates than children with normal weight (1.90%; 95% CI, 1.06%-2.97%).

Regarding prehypertension in children, the pooled prevalence was estimated to be 9.67% (95% CI, 7.26%-12.38%) based on a random-effects meta-analysis (Table 2 and eFigure 4 in the Supplement). According to the leave-1-out sensitivity analysis (eFigure 5 in the Supplement), the pooled prevalence of childhood prehypertension ranged from 9.10% (95% CI, 6.80%-11.70%) to 10.46% (95% CI, 8.24%-12.90%) when removing 1 study at a time from the pooled analysis. No study disproportionately affected the overall result. The funnel plot, Egger test, and Begg test suggested no publication bias (eFigure 6 in the Supplement). The subgroup meta-analyses indicated no statistically significant difference in prehypertension prevalence among children by age group (6-9 years vs 10-19 years), sex (male vs female), setting (urban vs rural), BP measurement method (oscillometric vs mercury), investigation period (2004-2009 vs 2010-2014), BMI group (underweight vs normal weight vs overweight vs obese), WHO region (Region of the Americas vs European Region), or WB region (high-income countries vs low- and middle-income countries).

The pooled prevalence was 4.00% (95% CI, 2.10%-6.48%) for stage 1 childhood hypertension and 0.95% (95% CI, 0.48%-1.57%) for stage 2 childhood hypertension from random effects meta-analyses (Table 2 and eFigure 7 and eFigure 8 in the Supplement). Subgroup meta-analyses were only performed by sex and device type because of the availability of data sources. No statistically significant difference of prevalence rates was found between sexes, whereas studies that used mercury sphygmomanometers showed higher prevalence rates among children (stage 1 hypertension: 6.73%; 95% CI, 3.41%-11.06%; stage 2 hypertension: 1.74%; 95% CI, 1.11%-2.51%) than those using oscillometric sphygmomanometers (stage 1 hypertension: 2.04%; 95% CI, 1.37%-2.85%; stage 2 hypertension: 0.42%; 95% CI, 0.22%-0.68%).

For childhood hypertension, we conducted a multilevel mixed-effects meta-regression because of the availability of a substantial number of age- and sex-specific data points. To control for the association of different devices with prevalence estimates (as detected in the above subgroup meta-analyses), we chose only studies that used mercury sphygmomanometer for measuring BP, which had the largest data set (96 data points) compared with those that used an aneroid sphygmomanometer (9 data points) or oscillometric sphygmomanometer (29 data points). The association between age and hypertension prevalence among children is shown in eFigure 9 in the Supplement. Five variables with more than 10 data points (age, sex, investigation year, WHO region, and WB region), were first assessed in univariable meta-regression analyses (eTable 6 in the Supplement). The results of univariable meta-regression analyses demonstrated that age and investigation year were significantly associated with the prevalence of childhood hypertension. The final model for estimating the age-specific prevalence of hypertension in children aged 6 to 19 years for the years 2000, 2010, and 2015 is detailed in the eMethods in the Supplement.

As shown in Figure 2 and Table 3, the prevalence of hypertension (measured by mercury sphygmomanometer) increased from 4.32% (95% CI, 2.79%-6.63%) among children aged 6 years to 7.89% (95% CI, 5.75%-10.75%) among those aged 14 years and then decreased to 3.28% (95% CI, 2.25%-4.77%) among those aged 19 years in 2015. During the 15 years from 2000 to 2015, the increasing rates of childhood hypertension prevalence were similar across the whole age range (6-19 years), fluctuating at 75% to 79%.

This systematic review and meta-analysis comprehensively describes the prevalence of hypertension in children based on available data published from 1994 to 2018. The prevalence of hypertension among children varied significantly when measured by different devices. A positive secular trend of childhood hypertension prevalence was observed during the last 2 decades of the analysis. Overweight and obese children were more likely to have hypertension than their underweight or normal weight counterparts. On the basis of studies that measured BP by mercury sphygmomanometer, the age-specific prevalence of childhood hypertension from 2000 to 2015 was established. Between 2000 and 2015, the prevalence of childhood hypertension increased by 75% to 79% among children aged 6 to 19 years, among whom the prevalence continued to increase before the onset of puberty and during puberty, reached the peak level at the end of puberty, and steadily decreased until the beginning of adulthood.

Previous systematic reviews^11,32-34 have synthesized the prevalence of childhood hypertension in Africa, Nigeria, Brazil, and worldwide. However, none of those studies adopted the standardized BP measurement in children recommended by the NHBPEP, which states that the diagnosis of childhood hypertension should be confirmed on at least 3 occasions to avoid false-positive cases.⁹ To our knowledge, this study was the first systematic review and meta-analysis to explore the global prevalence of childhood hypertension based on BP measurements on at least 3 separate occasions.

In line with previous systematic reviews and individual investigations,^11,17,35,36 a positive association between the prevalence of childhood hypertension and BMI was observed in our study. This finding supports previous results showing that obesity may be a risk factor for hypertension and underlines the importance of weight control for hypertension management in the pediatric population.³⁶ Another key finding of this study is the pattern of hypertension prevalence according to age, by which the prevalence of childhood hypertension started to increase rapidly from the onset of puberty and reached the peak level at the end of puberty. In previous studies,^37,38 a higher level of BP during puberty than before or after it has been well documented, which might be associated with hormone change and rapid growth spurts.

Studies^39,40 in the United States have observed an increase in BP in children during the past decade, partially caused by an increase in childhood obesity, especially abdominal obesity. In this study, a significant temporal trend of increasing prevalence of childhood hypertension during the past 2 decades was also found at the global level, as revealed in subgroup meta-analysis and meta-regression. However, such a secular trend was not observed in Africa during the past 2 decades, as previously reported.¹¹ With the potential forthcoming epidemic of childhood obesity in developing countries, an increase in the prevalence of childhood hypertension may also transpire in these countries. In 2017, the new clinical practice guideline for screening and management of high BP in children and adolescents updated the normative pediatric BP table in the fourth report by NHBPEP by excluding data for overweight and obese children, according to which the global prevalence of childhood hypertension might be even higher.⁴¹ Considering the unfavorable health consequences of childhood hypertension, this finding highlights the need for global actions to prevent and manage childhood hypertension.^7,36

Strengths of this study include the comprehensive search strategies, a double review process, and stringent selection criteria. In our systematic review, we included only studies that were conducted in the general pediatric population so that the generalizability of our results could be well guaranteed. Moreover, the standardized definitions of hypertension and its subtypes reduced heterogeneity largely because of methodologic variability and made the synthesis of prevalence possible. Also, we were able to pool the prevalence of hypertension and its phenotypes, prehypertension, and stage 1 and stage 2 hypertension in children based on the available evidence, which allowed our systematic review and meta-analysis to provide a broad scope of the prevalence of childhood hypertension. For the first time, to our knowledge, in a systematic review and meta-analysis, we constructed age-specific prevalence of childhood hypertension and explored its secular trend after eliminating the effects of BP measurement devices.

Several intrinsic limitations of this study should also be recognized. First, although we unified the definitions of childhood hypertension and its subtypes before pooling the prevalence estimates, substantial heterogeneity was detected. Second, the limited number of included studies for prehypertension, stage 1 hypertension, and stage 2 hypertension in children increased the uncertainty of our pooled prevalence estimates, and the sources of heterogeneity could only be explored by subgroup meta-analysis in a limited set of groups. Third, we could not estimate the prevalence of childhood prehypertension, stage 1 hypertension, and stage 2 hypertension at the regional level. Even for childhood hypertension, for which the contributing data points successfully covered all the 6 WHO regions, the prevalence estimation at the regional level was not optimal given that more than half of the included studies were concentrated in only 2 regions (Region of the Americas and European Region).

Our overall pooled prevalence of childhood hypertension was lower than that in a previous systematic review of the worldwide prevalence (4.0% vs 11.2%).⁴² The large disparity might be explained mainly by the different numbers of visits for BP measurements in these 2 systematic reviews. In their study, the pooled prevalence of childhood hypertension was based on individual studies that had measured BP on a single occasion or on 2 occasions or more, which could lead to a higher prevalence estimate given that the prevalence of childhood hypertension could decrease with the increase of visit numbers.¹⁰

This study suggests that childhood hypertension represents a considerable public health challenge worldwide. Childhood hypertension was generally more common in adolescents undergoing puberty and children who were overweight or obese. An upward trend of hypertension prevalence in children during the past 2 decades was observed and may persist in the future. More high-quality epidemiologic investigations on childhood hypertension (ideally in accordance with the recommendations by NHBPEP) appear to be needed, especially for different subgroups of hypertension (prehypertension, stage 1 hypertension, and stage 2 hypertension) and within the Region of the Americas, Eastern Mediterranean Region, Southeast Asia Region, and Western Pacific Region.

Accepted for Publication: June 21, 2019

Corresponding Author: Yajie Zhu, PhD, The George Institute for Global Health, University of Oxford, Oxford OX1 2BQ, United Kingdom (yajie.zhu@georgeinstitute.ox.ac.uk).

Published Online: October 7, 2019. doi:10.1001/jamapediatrics.2019.3310

Author Contributions: Dr Zhu had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Song, Zhu, Rahimi, Rudan.

Acquisition, analysis, or interpretation of data: Song, Zhang, Yu, Zha, Zhu.

Drafting of the manuscript: Song.

Critical revision of the manuscript for important intellectual content: Zhang, Yu, Zha, Zhu, Rahimi, Rudan.

Statistical analysis: Song, Zhu.

Administrative, technical, or material support: Zhang, Yu, Zha, Rudan.

Supervision: Song, Zhu, Rudan.

Conflict of Interest Disclosures: Dr Rahimi reported receiving grants from National Institute for Health Research Oxford Biomedical Research Centre, British Heart Foundation, Economic and Social Research Council, Research Councils UK, and Oxford Martin School, University of Oxford, during the conduct of the study, and personal fees from PLOS Medicine and BMJ Heart outside the submitted work. No other disclosures were reported.