Prevalence of Secondary Hypertension in Otherwise Healthy Youth with a New Diagnosis of Hypertension: A Meta-Analysis

Abstract

Objective

To estimate the prevalence of secondary hypertension among otherwise healthy children with hypertension diagnosed in the outpatient setting.

Study design

MEDLINE, PubMed Central, Embase, Web of Science, and Cochrane Library were systematically searched for observational studies reporting the prevalence of secondary hypertension in children who underwent evaluation for hypertension and had no known comorbidities associated with hypertension at the time of diagnosis. Two authors independently extracted the study-specific prevalence of secondary hypertension in children evaluated for hypertension. Prevalence estimates for secondary hypertension were pooled in a random effects meta-analysis.

Results

Nineteen prospective and 7 retrospective studies including 2575 children with hypertension were analyzed, with a median of 65 (min-max range, 9–486) participants in each study. Studies conducted in primary care or school settings reported a lower prevalence of secondary hypertension (3.7% [95% CI: 1.2–7.2%]) than studies conducted in referral clinics (20.1% [95% CI: 11.5–30.3%]). When stratified by study setting, there were no significant subgroup differences according to study design, country, participant age range, hypertension definition, blood pressure device, or study quality. Although studies applied different approaches to diagnose secondary hypertension, diagnostic evaluations were at least as involved as the limited testing recommended by current guidelines.

Conclusions

The low prevalence of secondary hypertension among children with a new diagnosis of hypertension identified on screening reinforces clinical practice guidelines to avoid extensive testing in the primary care setting for secondary causes in most children with hypertension.

Hypertension is one of the most common chronic diseases of childhood, affecting 4% of children, including 15% of children with obesity.¹ Hypertension in most children has historically been attributed to secondary causes,^{2, 3} with intrinsic kidney disease or renovascular disease reported in up to 77% to 97% of children with secondary hypertension.^4–6 In cohorts of children with hypertension followed in pediatric nephrology or pediatric hypertension clinics, 45% to 85% of cases have secondary hypertension.^4–10

The high prevalence of secondary hypertension observed in subspecialty clinics may not reflect the prevalence of secondary hypertension in children found to have hypertension in the primary care setting.¹¹ Studies have found that secondary hypertension is uncommon in asymptomatic children with hypertension detected on screening examinations.^12–15 The United States Preventive Services Task Force has identified the lack of accurate prevalence estimates of secondary hypertension in asymptomatic children as an important gap in pediatric research.¹⁶ Defining the pretest probability of secondary hypertension in a healthy child with a new diagnosis of hypertension may help inform the shared decision-making between providers and families considering further diagnostic workup and minimizing the harms and costs of extensive testing.

The purpose of this systematic review and meta-analysis is to estimate the pooled prevalence of secondary hypertension among children undergoing evaluation for hypertension in the ambulatory setting. We hypothesized that the prevalence of secondary hypertension in otherwise healthy children diagnosed with hypertension will be lower than the high prevalence observed among all children with hypertension followed in referral clinics.

Methods

The systematic review protocol has been registered in the International Prospective Register of Systematic Reviews database (registration number: CRD42021229313). The systematic review and meta-analysis were conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement.¹⁷

Eligibility Criteria

We included observational studies describing youth and young adults ≤20 years old diagnosed with hypertension in the outpatient setting who underwent further evaluation for secondary hypertension. Secondary hypertension was defined as an identifiable cause of hypertension due to kidney, renovascular, cardiac, endocrine, or genetic disorders or other environmental or medication exposure. Studies were included if the underlying cause of hypertension in each child was not known at the time of the diagnostic evaluation. Studies were excluded if they included children with chronic diseases associated with hypertension, such as chronic kidney disease, that were known at the time of the initial hypertension evaluation. We included studies that diagnosed hypertension based on at least two outpatient blood pressure measurements. Although the objective of some studies may not have been to report the prevalence of secondary hypertension among children with hypertension, studies were included if this information was available from the results.

Search Strategy

MEDLINE, PubMed Central, Embase, Web of Science, and the Cochrane Library were searched on January 7, 2021 from inception for studies reporting the prevalence of secondary hypertension among youth with a diagnosis of hypertension. Searches were limited to human studies. No date or language limits were imposed on the search. Non-English articles were translated to English using Google Translate.¹⁸ An experienced medical librarian was consulted on methodology and performed a medical subject heading (MeSH) analysis of known key articles [mesh.med.yale.edu]. Scoping searches were performed in each database and an iterative process was used to translate and refine the searches. To maximize sensitivity, the formal search used controlled vocabulary terms and synonymous free-text words to capture the concepts of “children” and “secondary hypertension.” The search strategies for each database are shown in Table I (available at www.jpeds.com). Reviewers checked for additional relevant cited and citing articles using included studies.

Table 1 Online.

Search Strategies

MEDLINE

1 ((secondary or moderate or mild) adj2 hypertension).tw,kf.

2 exp pediatrics/ or exp child/ or exp adolescent/ or (adolescence or adolescent or adolescents or child or children or juvenile or juveniles or kid or kids or paediatric or paediatrics or pediatric or pediatrics or pubescence or pubescent or teen or teens or teenager or teenagers or youth or youths or preadolescent).tw,kf.

3 1 and 2

4 3 not (Animals/ not (Animals/ and Humans/))

Embase

1 exp pediatrics/ or exp child/ or exp adolescent/ or (adolescence or adolescent or adolescents or child or children or juvenile or juveniles or kid or kids or paediatric or paediatrics or pediatric or pediatrics or pubescence or pubescent or teen or teens or teenager or teenagers or youth or youths or preadolescent).tw,kw.

2 ((secondary or moderate or mild) adj2 hypertension).tw,kw.

3 1 and 2

4 3 not ((exp animal/ or nonhuman/) not exp human/)

Web of Science

Indexes=SCI-EXPANDED, SSCI, A&HCI, CPCI-S, CPCI-SSH, BKCI-S, BKCI-SSH, ESCI, CCR-EXPANDED, IC Timespan=All years TOPIC: (adolescence or adolescent or adolescents or child or children or juvenile or juveniles or kid or kids or paediatric or paediatrics or pediatric or pediatrics or pubescence or pubescent or teen or teens or teenager or teenagers or youth or youths or preadolescent) AND TOPIC: ((secondary or moderate or mild) NEAR/2 hypertension)

Cochrane

((adolescence or adolescent or adolescents or child or children or juvenile or juveniles or kid or kids or paediatric or paediatrics or pediatric or pediatrics or pubescence or pubescent or teen or teens or teenager or teenagers or youth or youths or preadolescent)):ti,ab,kw AND (((secondary or moderate or mild) NEAR/2 hypertension)):ti,ab,kw

PubMed Central

(((“secondary hypertension”[Body - Key Terms] OR “moderate hypertension”[Body - Key Terms] OR “mild hypertension”[Body - Key Terms]))) AND (((adolescence[Title] OR adolescent[Title] OR adolescents[Title] OR child[Title] OR children[Title] OR juvenile[Title] OR juveniles[Title] OR kid[Title] OR kids[Title] OR paediatric[Title] OR paediatrics[Title] OR pediatric[Title] OR pediatrics[Title] OR pubescence[Title] OR pubescent[Title] OR teen[Title] OR teens[Title] OR teenager[Title] OR teenagers[Title] OR youth[Title] OR youths[Title] OR preadolescent[Title])) OR (adolescence[Abstract] OR adolescent[Abstract] OR adolescents[Abstract] OR child[Abstract] OR children[Abstract] OR juvenile[Abstract] OR juveniles[Abstract] OR kid[Abstract] OR kids[Abstract] OR paediatric[Abstract] OR paediatrics[Abstract] OR pediatric[Abstract] OR pediatrics[Abstract] OR pubescence[Abstract] OR pubescent[Abstract] OR teen[Abstract] OR teens[Abstract] OR teenager[Abstract] OR teenagers[Abstract] OR youth[Abstract] OR youths[Abstract] OR preadolescent[Abstract]))

Data Abstraction

The final search retrieved 4846 references, which were pooled in EndNote and deduplicated [www.endnote.com]. This set of search results was uploaded to Covidence [www.covidence.org] for screening. Covidence identified more duplicates, leaving 3039 studies for screening. Every title and abstract were screened by two independent authors. Any title and abstract identified for inclusion by at least one author was reviewed in the full text stage. Two independent authors screened each full text article for inclusion. For studies that met inclusion criteria in the systematic review, two authors independently extracted the following data: first author, publication year, study design, study years, study setting, study country, age group, definition of hypertension, workup for secondary hypertension, number of hypertensive cases with secondary hypertension, and total number of cases with hypertension. Disagreements in article selection or data extraction were resolved by discussion to consensus.

Risk of Bias in Individual Studies

Studies were critically appraised using a checklist adapted from the Joanna Briggs Institute quality assessment tool for systematic reviews of prevalence^{19, 20} (Figure 1; available at www.jpeds.com). The critical appraisal checklist compared each study’s characteristics with those of a high quality study measuring prevalence of secondary hypertension in children who underwent evaluation for hypertension. Study sample size was assessed according to the Joanna Briggs Institute tool²⁰ using a hypothesized baseline risk of secondary hypertension of 20% in otherwise healthy children with hypertension.^{21, 22}

Figure 1.

Critical Appraisal Tool

Data Synthesis

Prevalence estimates for secondary hypertension were pooled in a random effects meta-analysis using the DerSimonian and Laird approach.²³ Confidence intervals for individual studies were calculated using the score method.²⁴ Individual study variances were stabilized by Freeman-Tukey double arcsine transformation and study weights are represented by the inverse of the variance of the transformed proportion.²⁵ The pooled transformed proportion and its confidence interval were then back-transformed to the pooled prevalence estimate.^{24, 25} The risk of bias assessment was not used for weighting of effect estimates.^{26, 27} Between-study statistical heterogeneity was assessed by the I² statistic. Due to suspected heterogeneity between study populations in primary care and school settings compared with referral clinics, we stratified by study setting before conducting additional subgroup analyses. We used bivariate random effects meta-regression to explore the association between the prevalence of secondary hypertension and the following variables: study design (prospective vs retrospective), country (United States vs non-U.S. studies), method of blood pressure measurement (in-office measurement vs 24-hour ambulatory blood pressure monitoring [ABPM]), participant age range (only adolescent patients aged ≥10 years vs general pediatric population), diagnostic criteria for hypertension, and study quality. Because only children with complete follow-up blood pressure data who underwent diagnostic testing were included in the pooled prevalence estimates, a sensitivity analysis was performed in which the pooled prevalence was calculated after excluding studies with significant loss to follow-up. To determine if our results were influenced by the method of transforming individual study proportions, we conducted an additional sensitivity analysis in which we fit a generalized linear mixed model to the data using the logit transformation of proportions.²⁸ We evaluated for publication bias through visual inspection of funnel plots and Eggers tests.²⁹

Statistical analyses were performed in Stata/SE version 17.0 (StataCorp, College Station, TX).

Results

Study Characteristics

The PRISMA flow diagram for search results is shown in Figure 2 (available at www.jpeds.com). Of 3039 unique titles and abstracts screened, 141 full text articles were assessed for eligibility, and 26 studies met inclusion criteria for the meta-analysis. The most common reasons for excluding full text articles were wrong study design (n=33), inclusion of children whose cause of hypertension was already known (n=19), or that the study did not specify the number of children with primary or secondary hypertension (n=13).

Figure 2.

Flow Diagram for the Selection of Studies Estimating the Prevalence of Secondary Hypertension Among Youth Evaluated for Hypertension

The median number of patients with hypertension that were included in each study was 65 (min-max range: 9–486). Table II describes the individual study methods, study populations, exclusion criteria, and hypertension definitions. In 18 prospective cohort studies,^{12–15, 30–43} children with hypertension were initially identified through blood pressure screening in school or during primary care visits, then underwent further evaluation for secondary causes after verification of hypertension on more than one measurement. The remaining seven retrospective cohort studies^{21, 22, 44–48} and one prospective cohort study⁴⁹ describe a referral population of children with hypertension evaluated in pediatric nephrology or pediatric hypertension clinics. Two studies confirmed hypertension by 24-hour ABPM,^{47, 48} and the other studies diagnosed hypertension based on office measurement alone. For the 23 studies that described the method of office blood pressure measurement, 19 studies used auscultatory measurement exclusively, 3 studies used a combination of auscultatory and automated measurement, and 1 study used auscultatory measurement for all patients plus Doppler technique for infant measurements. The age of included patients varied, with seven studies including only adolescent patients aged ≥10 years compared with 19 studies including a broad general pediatric age range. Twelve studies defined hypertension as a blood pressure greater than the 95^th percentile, five studies used a higher percentile cutoff, two studies used a combination of the 95^th percentile cutoff or an absolute blood pressure threshold of 140/90, five adolescent studies used an absolute threshold alone, one study used the 90^th percentile as the threshold, and one study did not specify the diagnostic criteria. The seven studies that incorporated absolute blood pressure thresholds in their diagnostic criteria for adolescents used higher thresholds than the 130/80 cutoff for stage 1 hypertension defined in the most recent AAP guideline.⁵⁰ For the 20 studies that used blood pressure percentiles to diagnose hypertension based on office measurement, one study⁴⁷ used the percentiles from the 2017 AAP guideline;⁵⁰ six studies^{21, 22, 41, 42, 45, 48} used the 2004 National Heart, Lung, and Blood Institute guidelines;⁵¹ two studies^{39, 46} used the 1987 Task Force Report;⁵² two studies^{38, 44} used the 1977 Task Force Report;⁵³ six studies^{12, 14, 34, 36, 40, 43} used the distribution of their own study cohort’s blood pressures to calculate percentiles; one study¹³ used a previously published blood pressure distribution derived from a cohort of children measured at the same site;⁵⁴ and two studies^{15, 35} did not specify the reference population. Of the eight specialty clinic studies, seven studies defined hypertension according to the existing clinical practice guideline at the time of publication,^50–53 and one study⁴⁹ did not specify a definition. All included studies verified hypertension on multiple measurements however the number of visits varied between studies. The results of the quality assessment of individual studies are shown in Table III (available at www.jpeds.com). Quality scores ranged from 3 to 9 on a 10-point scale.

Table 2.

Study Characteristics for Articles Included in the Meta-Analysis

Source	Country	Study Years	Study Type	Setting	Age Range (Min-Max)	No. of BP Measurements	Hypertension Definition	BP Technique	Study Population	Exclusions	$\frac{\text{\#}\text{Secondary}\text{Cases}}{\text{Total}\text{\#}\text{with}\text{HTN}}$
Londe, 1971	USA	NS	P	Primary care clinic	4–18y	Repeatedly high BP for ≥1 year (no. measurements not specified)	a >90th %ile repeatedly, occasionally >95th %ile; >140/90 for >15yo	Auscultatory	74 asymptomatic children diagnosed with HTN during routine exam.	NS	5/74
Kilcoyne, 1974	USA	NS	P	School-based screening	14–19y	≥2 visits with 7–10 day interval	≥140/90	Auscultatory	3537 screened for HTN.	Did not consent to diagnostic evaluation.	0/27
Silverberg, 1975	Canada	NS	P	School-based screening	15–20y	2 visits	≥150/95	Auscultatory	15,594 screened for HTN. 156/350 with HTN on initial screen had follow up BP. 19/156 had persistent HTN.	NS	2/19
Pistulková, 1976	Czech Republic	NS	P	School-based screening	11–15y	2 measurements on same visit, 1 hour apart	≥135/80	Auscultatory	8168 screened for HTN. 185/226 with HTN had diagnostic evaluation.	NS	27/185
Levine, 1976	USA	NS	P	School-based screening	14–19y	2 visits	≥140/90	Auscultatory and/or automated	1863 screened for HTN. 101/110 with HTN on initial screen had follow up BP. 46/101 had persistent HTN. 28/46 consented.	Did not consent to admission for workup.	0/28
Rames, 1978	USA	1971–1974	P	School-based screening	5–18y	2 visits	b ≥95th %ile or ≥140/90	Auscultatory	6622 screened for HTN.	NS	5/41
De Cesaris, 1980	Italy	NS	P	School-based screening	6–14y	3 visits over 1 month	b ≥97th %ile	Auscultatory	1500 screened for HTN.	NS	0/99
Környei, 1981	Hungary	1976–1978	P	School-based screening	7–14y	≥3 visits and sustained over 1.5 years	c >95th %ile	Auscultatory	1886 screened for HTN. 22/25 with HTN consented.	Parents did not consent to workup.	3/22
Berenson, 1983	USA	1979–1980	P	School-based screening	8–18y	≥4 visits	b ≥90th %ile	Auscultatory and automated	1604 screened for HTN. 94/100 with persistent HTN consented.	Did not consent to evaluation.	4/94
Belova, 1986	Bulgaria	1972–1983	P	School-based screening	6–15y	2 visits with 7–10 day interval	c ≥97th %ile	NS	9015 screened for HTN.	NS	38/250
Andriska, 1986	Hungary	1975–1985	P	Subspecialty clinic	0–18y	3 measurements on same visit	NS	Auscultatory	NS	NS	26/105
Ogborn, 1987	Canada	1979–1984	R	Pediatric nephrology clinic	4w–18y	≥2 visits	d >95th %ile	Auscultatory; Doppler for infants	Consecutive referrals	Known comorbidity associated with HTN	31/103
Michaud, 1989	Switzerland	NS	P	School-based screening	16–19y	≥3 visits	≥140/90	Auscultatory	3386 screened for HTN. 102/113 with HTN on initial screen had follow up BP. 43/102 had sustained HTN.	Did not follow-up for repeat measurements	0/43
Gupta, 1991	India	1985–1986	P	School-based screening	5–15y	≥7 visits	b >2 SD above mean	Auscultatory	3194 screened for HTN. 16/210 with HTN on initial screen had sustained HTN for 6 months. 9/16 had diagnostic investigation.	NS	0/9
Gregoric, 1992	Slovenia	1978–1981	P	School-based screening	7–19y	3 visits with ≥1 week interval	d >95th %ile	Auscultatory	8583 screened for HTN.	NS	7/103
Horn, 1994	UK	1988–1991	P	School-based screening	10–15y	≥3 visits	e >95th %ile	Auscultatory	14570 screened for HTN. 23/117 with HTN on initial screen had sustained HTN.	NS	0/23
Kania, 1994	USA	1992–1994	R	Pediatric hypertension clinic	8m–17y	≥3 visits	e ≥95th %ile	Auscultatory and/or automated	Consecutive referrals by pediatrician or school nurse due to HTN on >3 occasions	Symptomatic or signs of acute disease (e.g. hematuria)	3/39
Mukhopadhya, 1996	India	1990–1991	P	School-based screening	5–15y	≥4 visits	b >2SD above mean	Auscultatory	1170 screened for HTN.	NS	0/9
Savitha, 2007	India	2006	P	School-based screening	10–16y	≥3 visits with 1 week intervals	f ≥95th %ile	Auscultatory	503 screened for HTN.	Known comorbidity or medication associated with HTN	0/31
Wiesen, 2008	USA	2002–2007	R	Pediatric hypertension clinic	Mean 13.3y (SD +/− 4.4y)	NS	f ≥95th %ile up to 95 th %ile + 20/10 mmHg	NS	249 healthy children referred for mild to moderate hypertension. 220/249 had records available for review.	Severe HTN, acute complaints, evidence of kidney disease	24/220
Kapur, 2010	USA	2003–2004	R	Pediatric nephrology clinic	5–18y	≥3 visits	f ≥95th %ile	Auscultatory	246 referred for repeatedly high BP. 166/246 had HTN confirmed in referral clinic.	Known comorbidity or medication associated with HTN	15/166
Baracco, 2012	USA	2002–2005	R	Pediatric nephrology clinic	5–19y	≥3 visits	f ≥95th %ile	Auscultatory	Referred by general pediatrician.	Known comorbidity or medication associated with HTN	57/167
Valent-Morić, 2012	Croatia	2006y2009	R	Pediatric nephrology clinic	4–19y	3 visits	Office: f >95th %ile ABPM: SBP load ≥25%	Office: NS Home: ABPM	Referred by primary care due to BP >95th %ile on ≥3 visits.	NS	Office: 18/139; ABPM: 18/107
Kumar De, 2013	India	2010–2011	P	School-based screening	5–15y	≥6 visits at monthly intervals	b >2 SD above mean	Auscultatory	9961 screened for HTN.	Any chronic disease	0/37
Karmakar, 2019	India	NS	P	School and primary care screening	5–10y	3 visits	f >95th %ile	Auscultatory	1356 screened for HTN.	Any chronic disease; parents did not consent; absent on screening day.	6/56
Çakıcı, 2020	Turkey	2013–2018	R	Pediatric nephrology clinic	5–19y	3 measurements on same visit	Office: g ≥95th %ile ABPM: Mean 24h BP ≥95th %ile and SBP or DBP load ≥25%	Office: auscultatory; Home: ABPM	553 consecutive referrals for HTN. 486/553 had HTN as measured in referral clinic.	Taking medication associated with HTN; height <120cm.	Office: 186/486; ABPM: 186/383

Min, Min, minimum; Max, maximum; NS, not specified; R, retrospective; P, prospective; y, year; w, weeks; SD, standard deviation; BP, blood pressure; ABPM, ambulatory blood pressure monitor; SBP, systolic blood pressure; DBP, diastolic blood pressure; HTN, hypertension.

^aPercentiles were derived from a previously published blood pressure distribution¹ of 1473 children at the same site as the included study.

^bPercentiles and/or mean and standard deviation were derived from the distribution of blood pressures in the study cohort.

^cDoes not specify a reference population for percentile calculations.

^dPercentiles were derived from the 1977 Task Force Report.²

^ePercentiles were derived from the 1987 Task Force Report.³

^fPercentiles were derived from the 2004 Task Force Report.⁴

^gPercentiles were derived from the 2017 American Academy of Pediatrics clinical practice guideline.⁵

¹Londe S. Blood Pressure in Children. Clin Pediatr (Phila). 1966;5(2):71–8.

²Blumenthal S, Epps RP, Heavenrich R, Lauer RM, Lieberman E, Mirkin B, et al. Report of the task force on blood pressure control in children. Pediatrics. 1977;59:I-ii, 797–820.

³National Heart, Lung, and Blood Institute. Report of the Second Task Force on Blood Pressure Control in Children—1987. Pediatrics. 1987;79:1–25.

⁴National High Blood Pressure Education Program Working Group on High Blood Pressure in Children and Adolescents. The fourth report on the diagnosis, evaluation, and treatment of high blood pressure in children and adolescents. Pediatrics. 2004;114:555–76.

⁵Flynn JT, Kaelber DC, Baker-Smith CM, Blowey D, Carroll AE, Daniels SR, et al. Clinical Practice Guideline for Screening and Management of High Blood Pressure in Children and Adolescents. Pediatrics. 2017;140:e20171904.

Table 3 Online.

Results of Critical Appraisal Checklist for Studies Included in Meta-Analysis

	Was the sample representative of the target population? (0–1 stars)	Were study participants recruited in an appropriate way? 0–1 stars	Was the sample size adquate? (0–2 stars)	Were study subjects and setting described in detail? (0–1 stars)	Selection bias (0–2 stars)	Were objective, standard criteria used for measurement of the condition? (0–1 stars)	Was the condition measured in a standard, reliable way? (0–1 stars)	Adequacy of response rate (0–1 stars)
Londe, 1971		*	*	*	**	*	*	*
Kilcoyne, 1974		*		*	**	*	*	*
Silverberg, 1975		*		*	**	*	*	*
Pistulková, 1976		*	*		*	*
Levine, 1976		*		*	*	*	*	*
Rames, 1978		*		*	*	*	*	*
De Cesaris, 1980		*		*			*
Környei, 1981	*	*		*	**	*	*	*
Berenson, 1983		*	*		**	*	*	*
Belova, 1986		*	**	*	**	*		*
Andriska, 1986		*	*	*	**		*	*
Ogborn, 1987	*		*	*	**	*		*
Michaud, 1989		*			**	*	*	*
Gupta, 1991		*		*	*	*	*
Gregoric, 1992	*	*	*	*	**		*	*
Horn, 1994	*	*		*	**		*
Kania, 1994	*			*	*	*	*	*
Mukhopadhya, 1996		*		*	**	*	*	*
Savitha, 2007	*	*		*	**	*	*	*
Wiesen, 2008	*		*	*	*	*
Kapur, 2010	*		*	*	**	*	*	*
Baracco, 2012	*		*	*	**	*	*
Valent-Morić, 2012	*		*	*	**	*	*	*
Kumar De, 2013		*		*	**	*	*	*
Karmakar, 2019	*	*		*			*
Çakıcı, 2020	*		**	*	**	*	*	*

Loss-to-follow-up affected several studies. In one school-based screening study,³³ 55% of children with hypertension detected on initial screening never had a repeat blood pressure checked; and in 4 other school-based studies,^{30–32, 36} more than 15% of children with hypertension did not undergo any diagnostic workup. Workup for secondary causes of hypertension and etiologies of secondary hypertension for each study are shown in Table IV (available at www.jpeds.com). Although studies applied different approaches, the 26 studies appear to have performed diagnostic evaluations at least as involved as the limited laboratory testing recommended in the AAP guideline with measurement of serum creatinine, electrolytes, and urinalysis.⁵⁰ In the 24 studies that describe the specific etiology for each case of secondary hypertension, 68% of secondary cases were due to intrinsic kidney disease or structural abnormalities and 9% were due to renovascular causes.

Table 4 Online.

Secondary Hypertension Workups and Etiologies According to Individual Studies

Source	Secondary Hypertension Workup	Etiologies for Cases of Secondary Hypertension
Londe, 1971	Serum electrolytes, urea clearance, BUN, sCr, 24h urine CrCl, 24h urine catecholamines, 24h urine aldosterone, 24h urine Na, 24h urine K, EKG, CXR, hemoglobin, PRA, IVP	Total 5 cases: all renal
Kilcoyne, 1974	UA, UCx, endogenous creatine clearance, rapid-sequence pyelography, CXR, EKG, electrolytes, fasting BG, cholesterol, T4, cortisol, latex fixation (for collagen vascular disease), lupus erythematosus preparations, ANA, hemoglobin electrophoresis, peripheral PRA, 24h urine CrCl, uNa, PRA, urine VMA, +/− renal arteriography and renal biopsy	0 cases
Silverberg, 1975	sCr, K, uric acid, cholesterol, BG, UA, UCx, EKG, CXR, IVP, +/− urine metanephrines, renal angiogram or renin-aldosterone studies	Total 2 cases: 1 RAS, 1 elevated sCr of unknown cause
Pistulková, 1976	UA, Addis sediment (assessment of urine sediment), clearance, EKG, CXR, fundoscopy, excretory urography, isotope renography, +/− vasorenography	Total 27 cases: 15 abnormal isotope renography, 6 PDA, 5 nephropathy, 1 coarctation
Levine, 1976	Electrolytes, Ca, Phos, total protein and albumin-globulin ratio, BUN, sCr, cholesterol, protein electrophoresis, lipoprotein electrophoresis, PRA, serum thyroxine, plasma FSH and LH, 17-beta-ol-androgens, urine SG and osm after overnight thirst, CrCl, urine metanephrines, 17-ketosteroid, 17-hydroxysteroid, aldosterone excretion, cardiac series, bone age, EKG, rapid sequence IVP	0 cases
Rames, 1978	Electrolytes, BUN, UA for all; additional studies for patients who consented to admission for workup: sCr, UCx, lipids, CBC, blood glucose, complement, IVP, 24h urine electrolytes, UPC, 24h urine steroids and catecholamines, 24h urine aldosterone	Total 5 cases: 2 coarctation, 1 RAS, 1 hydronephrosis, 1 oral contraceptive
De Cesaris, 1980	UA, additional workup not specified	0 cases
Környei, 1981	UA, UCx, 12-hour urine concentration test, EKG, electrolytes, sCr, lipid panel, rapid sequence IV urography, 24h urine CrCl, urine protein, urine VMA, urine 17-ketosteroid	Total 3 cases: 2 chronic GN, 1 chronic pyelonephritis
Berenson, 1983	Serum and urine electrolytes, serum and urine creatinine, hemoglobin, EKG, urine sediment, urine protein, urine reducing substances, urine blood, urine culture, 24h urine collection (Na, K, Cr)	Total 4 cases: 1 coarctation, 1 UTI with UPJ obstruction and hydronephrosis, 1 RAS, 1 elevated sCr
Belova, 1986	EKG, echo, CXR, urography, renal isotope studies, aortography, PRA, fundoscopy. Primary hypertension diagnosis required lack of clinical symptoms, positive family history of hypertension, and “labile” BP on 3 hourly measurements.	Total 38 cases: causes not specified
Andriska, 1986	CBC, UA, BG, electrolytes, cholesterol, lipoproteins, BUN, CrCl, 24h urine catecholamines, CXR, IV urogram, EKG, fundoscopy	Total 26 cases: 16 coarctation, 4 DM, 2 pyelonephritis, 2 GN, 2 not specified
Ogborn, 1987	UA, CXR, renal technetium Tc 99m pentetate, EKG, 24h urine protein and CrCl, plasma electrolytes, BUN, sCr, Ca, Mg, Phos, ALP, TFT, supine and standing renin values, plasma aldosterone, plasma cortisol profile, 24h urine catecholamines, +/− IVP, VCUG, angiography, renal biopsy, CT scan	Total 31 cases: 6 GN, 4 renal dysplasia, 4 renal hypoplasia, 4 obstructing uropathy and urogenital anomalies, 3 pyelonephritis, 3 asphyxia or ischemia, 2 neurofibromatosis, 2 Williams syndrome, 1 retroperitoneal sarcoma, 1 polycystic kidney, 1 medication-related
Michaud, 1989	Blood count, BUN, urine, additional workup not specified	0 cases
Gupta, 1991	UA, urine VMA, BUN, sCr, electrolytes, cholesterol, CXR, abdominal XR, IVP, EKG, fundoscopy	0 cases
Gregoric, 1992	3-day hospitalization for workup (workup not specified)	Total 7 cases: 4 chronic pyelonephritis, 2 coarctation, 1 MPGN
Horn, 1994	All had UA, UPC; some had renal u/s, sCr, peripheral vein renin levels, lipids, electrolytes	0 cases
Kania, 1994	UA, renal u/s, echo, serum biochemistry, fasting lipids, 24h ABPM if able to tolerate	Total 3 cases: 1 renal hypoplasia, 1 RAS, 1 Graves disease
Mukhopadhya, 1996	UA, BUN, sCr, cholesterol, CXR, abdominal XR, EKG, fundoscopy	0 cases
Savitha, 2007	UA, UCx, BUN, sCr, electrolytes, renal u/s, CBC	0 cases
Wiesen, 2008	UA, fasting complete metabolic and lipid profiles, renal u/s, echo, 24h ABPM (if >10y)	Total 24 cases: 10 renal, 4 renovascular, 10 cardiac
Kapur, 2010	Per institution protocol (including CBC, BUN, sCr, electrolytes, PRA, serum aldosterone, TFT, UA, UPC, urine catecholamines, renal u/s w/ or w/o Doppler, echo, renal arteriography, DMSA scan, MRI/MRA)	Total 15 cases: 8 renal scarring, 5 RAS, 1 pheochromocytoma, 1 juxtaglomerular tumor
Baracco, 2012	CBC, electrolytes, BUN, sCr, PRA, serum aldosterone, TFT, UA, spot urine catecholamines, renal u/s, echo, DMSA scan	Total 57 cases: 45 renal scarring, 7 RAS, 1 single kidney, 1 coarctation, 1 juxtaglomerular tumor, 1 pheochromocytoma, 1 hyperthyroidism
Valent-Morić, 2012	Standardized, comprehensive diagnostic evaluation as recommended by the European Society of Hypertension1 (CBC, electrolytes, BUN, sCr, fasting BG, lipids, fasting triglycerides, UA for microalbuminuria and proteinuria, renal u/s, CXR, EKG, echo, +/− additional labs and imaging as needed)	Total 18 cases: causes not specified
Kumar De, 2013	UA, BUN, sCr, cholesterol, EKG, renal u/s, IVP	0 cases
Karmakar, 2019	Not specified	Total 6 cases: 5 renal, 1 endocrine
Çakıcı, 2020	UA, sCr, BUN, uric acid, electrolytes, renal u/s, echo, fundoscopy; with additional workup based on history, exam, and initial studies	Total 186 cases: 68 renal scarring, 21 obstructive uropathy, 23 renal cystic disease, 20 renal parenchymal disease, 5 single kidney, 17 renovascular disease, 13 coarct, 6 primary hyperaldosteronism, 5 hyperthyroidism, 2 pheochromocytoma, 2 AVM, 2 neural tube defect, 2 Arnold-Chiari malformation, 1 Gordon syndrome

BUN, blood urea nitrogen; sCr, serum creatinine; CrCl, creatinine clearance; 24h, 24-hour; EKG, electrocardiogram; CXR, chest x-ray; PRA, plasma renin activity; IVP, intravenous pyelogram; UA, urinalysis; UCx, urine culture; BG, blood glucose; T4, thyroxine; ANA, antinuclear antibody; uNa, urine sodium; VMA, vanillylmandelic acid; K, potassium; Ca, calcium; Phos, phosphorus; FSH, follicle-stimulating hormone; LH, luteinizing hormone; SG, specific gravity; osm, osmolality; CBC, complete blood count; UPC, urine protein to creatinine ratio; IV, intravenous; Mg, magnesium; ALP, alkaline phosphatase; TFT, thyroid function test; VCUG, voiding cystourethrogram; CT, computed tomography; XR, x-ray; u/s, ultrasound; echo, echocardiogram; ABPM, ambulatory blood pressure monitor; DMSA, dimercaptosuccinic acid; MRI/MRA, magnetic resonance imaging/angiography; RAS, renal artery stenosis; PDA, patent ductus arteriosus; GN, glomerulonephritis; UTI, urinary tract infection; UPJ, ureteropelvic junction; DM, diabetes mellitus; MPGN, membranoproliferative glomerulonephritis; AVM, arteriovenous malformation

¹Lurbe E, Cifkova R, Cruickshank JK, et al. Management of high blood pressure in children and adolescents: recommendations of the European Society of Hypertension. J Hypertens. 2009;27:1719–1742.

Prevalence of Secondary Hypertension

Among the 2575 youth undergoing evaluation for hypertension, there were 457 cases of secondary hypertension. Random effects modeling of these 26 studies yielded a pooled prevalence of secondary hypertension among otherwise healthy youth with hypertension of 8.0% (95% CI: 4.0–13.0%). There was high between-study heterogeneity (I² 93.0%). Studies conducted in primary care or school settings reported a lower pooled prevalence of secondary hypertension (3.7% [95% CI: 1.2–7.2%], I² 78.9%) (Figure 3) compared with studies conducted in referral clinics (20.1% [95% CI: 11.5–30.3%], I² 94.6%) (Figure 4). When stratified by study setting, there were no significant subgroup differences in the prevalence of secondary hypertension according to prospective or retrospective study design, study country, participant age range, number of visits required for hypertension diagnosis, blood pressure threshold, method of blood pressure measurement, or study quality (Table V). For the stratified analysis, the funnel plots and Eggers tests did not suggest any publication bias (Figure 5; available at www.jpeds.com). In a sensitivity analysis excluding the five studies^{30–33, 36} with significant loss to follow-up from the pooled prevalence estimate in the primary care and school setting, the prevalence remained similar (3.8% [95% CI: 0.9–7.9%], I² 80.3%) to the primary analysis. In the two referral clinic studies that used ABPM in addition to office blood pressure measurement to diagnose hypertension, the prevalence of secondary hypertension was 17%⁴⁸ and 49%⁴⁷ among youth with hypertension confirmed by ABPM, compared with 13%⁴⁸ and 38%,⁴⁷ respectively, among youth with hypertension diagnosed by office measurements alone. In the additional sensitivity analysis using a generalized linear mixed model, the results were consistent with the primary analysis (Figure 6; available at www.jpeds.com).

Figure 3. Pooled Prevalence of Secondary Hypertension Among Youth with Hypertension Identified in a School or Primary Care Screening.

Case count is number of children with secondary hypertension divided by number of children with hypertension who underwent evaluation for secondary causes. ES, effect size or prevalence in percent with 95% confidence interval. Study weights are represented by the inverse of the variance of the transformed proportion.

Figure 4. Pooled Prevalence of Secondary Hypertension Among Youth with Hypertension Referred to a Subspecialty Clinic.

Case count is number of children with secondary hypertension divided by number of children with hypertension who underwent evaluation for secondary causes. ES, effect size or prevalence in percent with 95% confidence interval. Study weights are represented by the inverse of the variance of the transformed proportion.

Table 5.

Prevalence of Secondary Hypertension Among Otherwise Healthy Youth with Hypertension Using Random Effects Meta-Analysis

	No. Studies	Total No. Subjects with Hypertension	No. Cases with Secondary Hypertension	Prevalence, % (95% CI)	I2, %a	p-value ((Egger’s test)b	p-value (Subgroup difference)
Overall	26	2575	457	8.0 (4.0–13.0)	93.0	0.002	NA
Setting
School or Primary Care	18	1150	97	3.7 (1.2–7.2)	78.9	0.21	<0.001
Referral Clinic	8	1425	360	20.1 (11.5–30.3)	94.6	0.28
Subgroup Analyses for School or Primary Care Setting
Study Design
Prospective	18	1150	97	3.7 (1.2–7.2)	78.9	0.21	NA
Study Country
United States	5	264	14	3.9 (0.8–8.5)	50.0	0.87	0.56
Non-U.S.	13	886	83	3.7 (0.6–8.4)	82.9	0.25
Study Setting
School-based Screening Only	16	1020	86	3.1 (0.5–6.9)	81.2	0.21	0.99
Primary Care or Hybrid of Primary Care and School	2	130	11	8.3 (4.0–13.9)	NA	0.83
Study Age Group
Adolescent Only (≥10y)	7	356	29	1.8 (0–8.3)	81.4	0.25	0.94
General Pediatrics	11	794	68	5.0 (1.5–9.8)	79.2	0.49
No. Visits Required for Hypertension Diagnosis
<3 visits	6	550	72	8.2 (3.4–14.6)	73.0	0.38	0.13
≥3 visits	12	600	25	2.0 (0.2–5.0)	61.3	0.87
Hypertension Threshold
≥90th %ile	1	94	4	4.3 (1.7–10.4)	NA	NA	0.67
≥95th %ile or ≥130/80 if ≥13y	17	1056	93	3.6 (1.0–7.4)	79.9	0.20
Study Quality Score
0–5	4	349	33	4.5 (0.0–17.3)	90.7	0.60	0.82
6–10	14	801	64	3.5 (0.9–7.2)	73.1	0.27
Subgroup Analyses for Referral Clinic Setting
Study Design
Prospective	1	105	26	24.8 (17.5–33.8)	NA	NA	0.86
Retrospective	7	1320	334	19.4 (10.0–30.9)	95.4	0.28
Study Country
United States	4	592	99	14.6 (5.0–27.8)	93.2	0.70	0.22
Non-U.S.	4	833	261	26.1 (14.9–39.2)	92.5	0.20
Method of BP Measurement
Office	6	800	156	18.6 (10.8–28.9)	91.3	0.89	0.19
ABPMc	2	490	204	41.0 (36.7–45.4)	NA	0.03
Study Quality Score
0–5	1	220	24	10.9 (7.4–15.7)	NA	NA	0.33
6–10	7	1205	336	21.7 (12.5–32.5)	93.8	0.12

y, years; AAP, American Academy of Pediatrics; BP, blood pressure; ABPM, ambulatory blood pressure monitoring

^aThe I² statistic measures between-study statistical heterogeneity, with I²>75% considered high heterogeneity.

^bThe Egger’s test is a quantitative assessment for publication bias or small-study effects, with p<0.05 indicating significant publication bias or small-study effects.

^cThe 2 studies that provide prevalence estimates for hypertension by office measurement and by ABPM are categorized as ABPM for the subgroup analysis.

Figure 5. Funnel Plot by Study Setting.

Visual inspection shows funnel plot symmetry, which suggests that major publication bias was not present. Quantitatively, the non-significant p-values for Egger’s tests in each setting also suggest that publication bias was not present.

Figure 6. Random Effects Meta-Analysis Using Generalized Linear Mixed Model to Estimate Pooled Prevalence of Secondary Hypertension Among Youth with Hypertension, Overall and Stratified by Study Setting.

Case count is the number of children with secondary hypertension divided by the total number of children diagnosed with hypertension who underwent evaluation for secondary causes. ES, effect size or prevalence in percent with 95% confidence interval.

Discussion

This systematic review and meta-analysis found that most youth with hypertension who underwent evaluation for secondary causes were determined to have primary hypertension. The prevalence of secondary hypertension was significantly lower in prospective studies in which youth with hypertension were identified on screening examinations in school or primary care settings (3.7% [95% CI: 1.2–7.2%]) compared with the predominantly retrospective studies of youth with hypertension referred to pediatric nephrology or hypertension clinics (20.1% [95% CI: 11.5–30.3%]).

The literature documenting high prevalence of secondary hypertension among children with hypertension may not be generalizable to apparently normal children found to have hypertension during routine examinations because these past studies include children with known chronic kidney disease and other comorbidities that cause hypertension,^{5, 6, 8, 11} define hypertension using severely increased blood pressure cutoffs instead of age-based norms from the AAP clinical practice guidelines,^{55, 56} or describe inpatient populations.^57–59 Consistent with the hypothesis that most otherwise healthy children ≥6 years old with hypertension will not have a secondary cause, the yield of diagnostic laboratory and imaging investigations in children with mild to moderate hypertension is low.⁴⁵ In two retrospective cohorts of children referred to pediatric hypertension clinics for further evaluation, none of the children had a clinically relevant abnormality on their basic metabolic panel or urinalysis and 5%⁴⁶ to 8%⁴⁵ had a contributory abnormality identified on kidney ultrasound.

Our findings reinforce two aspects of the 2017 AAP hypertension clinical practice guideline: namely, that most children ≥6 years old do not require an extensive evaluation for secondary causes of hypertension unless there are history or physical examination features suggestive of a secondary cause; and that ABPM should be performed for confirmation of hypertension in children and adolescents.⁵⁰ Clinical characteristics like obesity^{5, 60, 61} or a family history of primary hypertension in older adults^{5, 8, 47} may also reduce the likelihood that a child’s hypertension is due to a secondary cause and thereby obviate the need for extensive workup.⁵⁰ Alternatively, younger age^{4, 8, 62} and history of prematurity^{8, 47} may increase the probability that a child’s hypertension is due to a secondary cause and warrant further diagnostic investigation.

Given a prevalence of white coat hypertension in up to half of children diagnosed with in-office hypertension,^63–65 application of 24-hour ABPM to diagnose hypertension may be a cost-effective strategy⁶⁵ to prevent unnecessary testing for secondary causes in children with white coat hypertension. In addition, certain findings on ABPM, like non-dipping status^{47, 66} and elevated diastolic load,^{66, 67} are more common in children with secondary hypertension than primary hypertension and may help identify which children have the highest pretest probability of a secondary cause. As only two studies included in the meta-analysis used ABPM to diagnose hypertension, additional data are needed on the prevalence of secondary hypertension in children with a new diagnosis of hypertension confirmed on ABPM. Nevertheless, our findings support completing 24-hour ABPM in a child with hypertension who is well-appearing and otherwise asymptomatic before performing extensive diagnostic testing.

Although it is critical to consider secondary etiologies of hypertension in the differential diagnosis when evaluating a child with hypertension, the perception that secondary hypertension is the predominant form of hypertension among otherwise healthy children may be one barrier to effective hypertension management in the primary care setting. In a survey of general pediatricians regarding their approach to children with hypertension, 82% refer to a specialist and 40% felt uncomfortable evaluating and treating hypertension, partly due to the risk of missing a case of secondary hypertension.⁶⁸ Qualitative interviews with primary care pediatric providers have shown that pediatricians feel their training was focused on the initial workup of hypertension and subspecialty referral but not on lifestyle management of primary hypertension.⁶⁹ In contrast, studies of pediatric cardiologists’ and nephrologists’ perspectives have found that these subspecialists may want primary care providers to assume a larger role in the management of children with hypertension, particularly in adolescents, as subspecialty referrals for all children with hypertension may not be necessary.^{70, 71} Awareness of the low prevalence of secondary hypertension among children with hypertension detected on screening examinations may help mitigate primary care providers’ concerns about managing hypertension in an otherwise healthy child and encourage pediatricians to trial lifestyle management prior to referral.⁵⁰

Our meta-analysis has several limitations. Most of the studies conducted in referral clinics were retrospective. Exclusion of patients from the pooled prevalence estimate due to lack of a secondary workup in these retrospective studies may introduce selection bias such that children with hypertension more likely to have a secondary cause may have been more likely to undergo diagnostic testing. Only two studies used 24-hour ABPM to diagnose hypertension, which likely resulted in some children with white coat hypertension in the other studies undergoing negative diagnostic workups for secondary causes. Although all studies verified hypertension on multiple measurements, including studies that measured blood pressure on <3 occasions may have resulted in the inclusion of children with more mild hypertension who were less likely to have a secondary cause. The lower prevalence of secondary hypertension among children evaluated in the school or primary care setting emphasizes the need to confirm the diagnosis of hypertension in the community with at least three independent measurements using appropriate technique before proceeding with any additional diagnostic evaluation. Blood pressure thresholds for diagnosing hypertension also varied between studies; however, all diagnostic thresholds were at least as high as the absolute blood pressure cutoffs recommended in the AAP guideline⁵⁰ or the 95^th percentile derived from the blood pressure norms at the time of publication or from the study-specific blood pressure distribution, except in one study that used a 90^th percentile cutoff.⁴³ We found no difference in the prevalence of secondary hypertension in the sensitivity analysis excluding the study using the 90^th percentile cutoff. The absence of primary studies published since 2017 implies that the pooled prevalence estimates depend, in part, on older studies that used different blood pressure reference norms and may have used outdated laboratory and imaging tests. However, all studies appear to have incorporated the limited blood and urine testing recommended by current guidelines.⁵⁰ It is also possible that included studies may have misclassified some diagnoses as secondary causes that may not actually cause hypertension. In addition, studies included different age groups. Although we observed no difference in the pooled prevalence for studies that included adolescents only compared with studies that included a broad general pediatric age range, the data were not available to calculate a prevalence estimate for the subgroup of children <10 years old who may be most likely to have a secondary cause of hypertension. Individual study data did not allow for extraction of prevalence estimates according to the stage of hypertension so we were unable to evaluate guideline recommendations⁵¹ supporting more extensive workups for children with stage 2 hypertension. Lastly, pooled prevalence estimates may be sensitive to the method of transformation used to stabilize individual study variances, particularly with sparse data.²⁸ Nevertheless, in the additional sensitivity analysis using a generalized linear mixed model, the pooled prevalence estimates remained consistent with the primary analysis.

Abbreviations:

AAP

American Academy of Pediatrics

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

ABPM

ambulatory blood pressure monitoring

CI

confidence interval