Pharmacological interventions for hypertension in children

  • Review
  • Intervention

Authors


Abstract

Background

Hypertension is a major risk factor for stroke, coronary artery disease and kidney damage in adults. There is a paucity of data on the long-term sequelae of persistent hypertension in children, but it is known that children with hypertension have evidence of end organ damage and are at risk of hypertension into adulthood. The prevalence of hypertension in children is rising, most likely due to a concurrent rise in obesity rates. In children with hypertension, non-pharmacological measures are often recommended as first-line therapy, but a significant proportion of children will eventually require pharmacological treatment to reduce blood pressure, especially those with evidence of end organ damage at presentation or during follow-up. A systematic review of the effects of antihypertensive agents in children has not previously been conducted.

Objectives

To determine the dose-related effects of different classes of antihypertensive medications, as monotherapy compared to placebo; as combination therapy compared to placebo or a single medication; or in comparisons of various doses within the same class, on systolic or diastolic blood pressure (or both) in children with hypertension.

Search methods

We searched the Cochrane Hypertension Group Specialised Register, the Cochrane Central Register of Controlled Trials (CENTRAL) (2013, Issue 9), Ovid MEDLINE (1946 to October 2013), Ovid EMBASE (1974 to October 2013) and bibliographic citations.

Selection criteria

The selection criteria were deliberately broad due to there being few clinical trials in children. We included randomised controlled trials (RCTs) of at least two weeks duration comparing antihypertensive agents either as monotherapy or combination therapy with either placebo or another medication, or comparing different doses of the same medication, in children with hypertension. Hypertension was defined as an average (over a minimum of three readings) systolic or diastolic blood pressure (or both) on the 95th percentile or above for age, height and gender. 

Data collection and analysis

Two authors independently selected relevant studies, extracted data and assessed risk of bias. We summarised data, where possible, using a random-effects model. Formal assessment of heterogeneity was not possible because of insufficient data.

Main results

A total of 21 trials evaluated antihypertensive medications of various drug classes in 3454 hypertensive children with periods of follow-up ranging from three to 24 weeks. There were five RCTs comparing an antihypertensive drug directly with placebo, 12 dose-finding trials, two trials comparing calcium channel blockers with angiotensin receptor blockers, one trial comparing a centrally acting alpha blocker with a diuretic and one trial comparing an angiotensin-converting enzyme inhibitor with an angiotensin receptor blocker. No randomised trial was identified that evaluated the effectiveness of antihypertensive medications on target end organ damage. The trials were of variable quality and most were funded by pharmaceutical companies.

Among the angiotensin receptor blockers, candesartan (one trial, n = 240), when compared to placebo, reduced systolic blood pressure by 6.50 mmHg (95% confidence interval (CI) -9.44 to -3.56) and diastolic blood pressure by 5.50 mmHg (95% CI -9.62 to -1.38) (low-quality evidence). High dose telmisartan (one trial, n = 76), when compared to placebo, reduced systolic blood pressure by -8.50 (95% CI -13.79 to -3.21) but not diastolic blood pressure (-4.80, 95% CI -9.50 to 0.10) (low-quality evidence). Beta blocker (metoprolol, one trial, n = 140), when compared with placebo , significantly reduced systolic blood pressure by 4.20 mmHg (95% CI -8.12 to -0.28) but not diastolic blood pressure (-3.20 mmHg 95% CI -7.12 to 0.72) (low-quality evidence). Beta blocker/diuretic combination (Bisoprolol/hydrochlorothiazide, one trial, n = 94)when compared with placebo , did not result in a significant reduction in systolic blood pressure (-4.0 mmHg, 95% CI -8.99 to -0.19) but did have an effect on diastolic blood pressure (-4.50 mmHg, 95% CI -8.26 to -0.74) (low-quality evidence). Calcium channel blocker (extended-release felodipine,one trial, n = 133) was not effective in reducing systolic blood pressure (-0.62 mmHg, 95% CI -2.97 to 1.73) or diastolic blood pressure (-1.86 mmHg, 95% CI -5.23 to 1.51) when compared with placebo. Further, there was no consistent dose response observed among any of the drug classes. The adverse events associated with the antihypertensive agents were mostly minor and included headaches, dizziness and upper respiratory infections.

Authors' conclusions

Overall, there are sparse data informing the use of antihypertensive agents in children, with outcomes reported limited to blood pressure and not end organ damage. The most data are available for candesartan, for which there is low-quality evidence of a modest lowering effect on blood pressure. We did not find evidence of a consistent dose response relationship for escalating doses of angiotensin receptor blockers, calcium channel blockers or angiotensin-converting enzyme inhibitors. All agents appear safe, at least in the short term.

Résumé scientifique

Interventions pharmacologiques pour l'hypertension chez l'enfant

Contexte

L'hypertension est un important facteur de risque d'AVC, de coronaropathie et de lésions rénales chez l'adulte. Peu de données sont disponibles sur les séquelles à long terme de l'hypertension persistante chez l'enfant, mais il est connu que les enfants souffrant d'hypertension montrent des signes de lésions des organes cibles et sont à risque d'hypertension à l'âge adulte. La prévalence de l'hypertension chez l'enfant est en augmentation, probablement en raison d'une augmentation parallèle des taux d'obésité. Chez l'enfant atteint d'hypertension, les mesures non pharmacologiques sont souvent recommandées en traitement de première intention, mais une proportion significative des enfants nécessiteront en fin de compte un traitement pharmacologique pour réduire la pression artérielle, en particulier ceux montrant des signes de lésions aux organes cibles à la présentation ou pendant le suivi. Une revue systématique des effets des agents antihypertenseurs chez l'enfant n'a pas encore été réalisée.

Objectifs

Déterminer les effets dose-dépendants de différentes classes de médicaments antihypertenseurs, en monothérapie par rapport à un placebo ; en association par rapport à un placebo ou à un seul traitement ; ou en comparant différentes doses dans une même classe, sur la pression artérielle systolique ou diastolique (ou les deux) chez l'enfant souffrant d'hypertension.

Stratégie de recherche documentaire

Nous avons effectué des recherches dans le registre spécialisé du groupe Cochrane sur l'hypertension artérielle, le registre Cochrane des essais contrôlés (CENTRAL) (2013, numéro 9), Ovid MEDLINE (de 1946 à octobre 2013), Ovid EMBASE (de 1974 à octobre 2013) et les références bibliographiques.

Critères de sélection

Les critères de sélection ont été délibérément larges en raison du peu d'essais cliniques existant chez l'enfant. Nous avons inclus des essais contrôlés randomisés (ECR) d'une durée d'au moins deux semaines comparant des agents antihypertenseurs en monothérapie ou en association par rapport à un placebo ou à un autre médicament, ou comparant différentes doses d'un même médicament, chez l'enfant souffrant d'hypertension. L'hypertension était définie comme une pression artérielle systolique ou diastolique (ou les deux) moyenne (sur un minimum de trois mesures) égale ou supérieure au 95e percentile pour l'âge, la taille et le sexe.

Recueil et analyse des données

Deux auteurs ont sélectionné indépendamment les études pertinentes, extrait les données et évalué le risque de biais. Nous avons résumé les données, lorsque cela était possible, en utilisant un modèle à effets aléatoires. Une évaluation formelle de l'hétérogénéité n'a pas été possible en raison de données insuffisantes.

Résultats principaux

Un total de 21 essais évaluaient des médicaments antihypertenseurs de différentes classes chez 3 454 enfants hypertendus avec des périodes de suivi allant de trois à 24 semaines. Ces essais incluaient cinq ECR comparant un antihypertenseur directement avec un placebo, 12 essais de dose, deux essais comparant des inhibiteurs des canaux calciques avec des inhibiteurs des récepteurs de l'angiotensine, un essai comparant un alpha-bloquant agissant centralement avec un diurétique et un essai comparant un inhibiteur de l'enzyme de conversion de l'angiotensine avec un inhibiteur des récepteurs de l'angiotensine. Aucun essai randomisé n'a été identifié qui ait évalué l'efficacité des médicaments antihypertenseurs sur les lésions des organes cibles. Les essais étaient de qualité variable et la plupart étaient financés par des laboratoires pharmaceutiques.

Parmi les inhibiteurs des récepteurs de l'angiotensine, le candésartan (un essai, n = 240), en comparaison avec un placebo, a réduit la pression artérielle systolique de 6,50 mmHg (intervalle de confiance (IC) à 95 % -9,44 à -3,56) et la pression artérielle diastolique de 5,50 mmHg (IC à 95 % -9,62 à -1,38) (preuves de faible qualité). Le telmisartan à haute dose (un essai, n = 76), en comparaison avec un placebo, a réduit la pression artérielle systolique de -8,50 (IC à 95 % -13,79 à -3,21), mais pas la pression diastolique (-4,80, IC à 95 % -9,50 à 0,10) (preuves de faible qualité). Le bêta-bloquant (métoprolol, un essai, n = 140), en comparaison avec un placebo, a réduit la pression artérielle systolique de façon significative de 4,20 mmHg (IC à 95 % -8,12 à -0,28), mais pas la pression diastolique (-3,20 mmHg, IC à 95 % -7,12 à 0,72) (preuves de faible qualité). L'association bêta-bloquant/diurétique (bisoprolol/hydrochlorothiazide, un essai, n = 94) en comparaison avec un placebo, n'a pas entraîné de réduction significative de la pression artérielle systolique (-4,0 mmHg, IC à 95 % -8,99 à -0,19) mais a eu un effet sur la pression artérielle diastolique (-4,50 mmHg, IC à 95 % -8,26 à -0,74) (preuves de faible qualité). L'inhibiteur des canaux calciques (félodipine à libération prolongée, un essai, n = 133) n'était pas efficace pour réduire la pression artérielle systolique (-0,62 mmHg, IC à 95 % -2,97 à 1,73) ou diastolique (-1,86 mmHg, IC à 95 % -5,23 à 1,51) par rapport à un placebo. Par ailleurs, aucune relation dose-effet constante n'a été observée pour aucune classe de médicaments. Les événements indésirables associés aux agents antihypertenseurs ont été principalement mineurs et ont inclus céphalées, étourdissements et infections des voies respiratoires hautes.

Conclusions des auteurs

Dans l'ensemble, il existe des données éparses sur l'utilisation d'agents antihypertenseurs chez l'enfant, rapportant des résultats limités à la pression artérielle sans données sur les lésions des organes cibles. Le plus de données sont disponibles pour le candésartan, pour lequel il existe des preuves de faible qualité d'un effet modeste sur la diminution de la pression artérielle. Nous n'avons pas trouvé de preuve d'une relation dose-effet constante pour des doses croissantes d'inhibiteurs des récepteurs de l'angiotensine, des canaux calciques ou de l'enzyme de conversion de l'angiotensine. Tous les agents semblent sûrs, du moins à court terme.

Plain language summary

Medications for high blood pressure in children

Hypertension (high blood pressure) is known to increase the risk of heart disease, stroke and kidney failure. The prevalence of hypertension in children is rising. A significant proportion of children with hypertension require medication to reduce blood pressure and medication use has increased significantly over the past several years. 

This systematic review includes 21 trials, involving 3454 children, which evaluated different medications to lower blood pressure among children with hypertension. This evidence is up to date as of October 2013. Most trials were of very short duration with the average being seven weeks. The studies were of variable quality and mostly industry funded. Not all studies compared the effect of medication on blood pressure lowering to a placebo. Only a few classes of the commonly prescribed drugs have been evaluated and most had a modest effect on blood pressure, but it is uncertain whether this results in improved long-term outcomes for children. Higher doses of medication did not result in greater reduction of blood pressure. All of the drugs studied were safe for use, at least in the short term.

Résumé simplifié

Médicaments pour une pression artérielle élevée chez les enfants

L'hypertension (une pression artérielle élevée) est connue pour augmenter le risque de maladie cardiaque, d'accident vasculaire cérébral et d'insuffisance rénale. La prévalence de l'hypertension chez les enfants est en augmentation. Une proportion significative des enfants souffrant d'hypertension nécessitent un traitement médicamenteux pour réduire la pression artérielle et l'utilisation de médicaments a augmenté significativement ces dernières années.

Cette revue systématique inclut 21 essais, portant sur 3 454 enfants, qui évaluaient différents médicaments pour abaisser la tension artérielle chez les enfants souffrant d'hypertension. Les preuves sont à jour en octobre 2013. La plupart des essais étaient de très courte durée, la moyenne étant de sept semaines. Les études étaient de qualité variable et le plus souvent financées par l'industrie. Toutes les études ne comparaient pas l'effet du médicament sur la diminution de la pression artérielle à un placebo. Seulement quelques classes de médicaments couramment prescrits ont été évaluées et la plupart avaient un effet modeste sur la pression artérielle, mais on ignore si cela entraîne une amélioration des résultats à long terme pour les enfants. Des doses plus élevées de médicament n'ont pas apporté de plus grande réduction de la pression artérielle. Tous les médicaments étudiés étaient sûrs à utiliser, du moins à court terme.

Notes de traduction

Traduit par: French Cochrane Centre 15th June, 2014
Traduction financée par: Financeurs pour le Canada : Instituts de Recherche en Santé du Canada, Ministère de la Santé et des Services Sociaux du Québec, Fonds de recherche du Québec-Santé et Institut National d'Excellence en Santé et en Services Sociaux; pour la France : Ministère en charge de la Santé

Laički sažetak

Lijekovi za visoki krvni tlak kod djece

Hipertenzija (visoki krvni tlak) povećava rizik od srčanih bolesti, moždanog udara i zatajenja bubrega. Učestalost hipertenzije kod djece se povećava. Značajni udio djece s hipertenzijom zahtijeva lijekove kako bi se smanjio krvni tlak te se lijekovi kod djece posljednjih godina značajno više koriste. 

Ovaj Cochrane sustavni pregled uključio je 21 studiju koje su uključivale 3454 djece, koje su procjenjivale različite lijekove za snižavanje krvnog tlaka u djece s hipertenzijom. Dokazi se temelje na literaturi objavljenoj do listopada 2013. Većina studija su bile jako kratke s prosječnim trajanjem od sedam tjedana. Studije su bile varijabilne kvalitete i većinom ih je financirala industrija. Nisu sve studije uspoređivale učinak lijekova u usporedbi s placebom. Ispitano je samo nekoliko skupina uobičajeno prepisivanih antihipertenziva i većina je imala skromni utjecaj na krvni tlak, ali nije sigurno poboljšavaju li dugoročne ishode u djece. Više doze lijekova nisu uzrokovale veće sniženje krvnog tlaka. Upotreba svih lijekova bila je sigurna, barem kratkoročno.

Bilješke prijevoda

Hrvatski Cochrane
Preveo: Adam Galkovski
Ovaj sažetak preveden je u okviru volonterskog projekta prevođenja Cochrane sažetaka. Uključite se u projekt i pomozite nam u prevođenju brojnih preostalih Cochrane sažetaka koji su još uvijek dostupni samo na engleskom jeziku. Kontakt: cochrane_croatia@mefst.hr

Streszczenie prostym językiem

Stosowanie leków w wysokim ciśnieniu tętniczym u dzieci

Wiadomo, że nadciśnienie (wysokie ciśnienie tętnicze krwi) zwiększa ryzyko chorób serca, udaru mózgu i niewydolności nerek. Częstość występowania nadciśnienia u dzieci zwiększa się. Znacząca liczba dzieci z nadciśnieniem wymaga leczenia farmakologicznego w celu obniżenia ciśnienia tętniczego a stosowanie leków w ciągu ostatnich kilku lat istotnie się zwiększyło. 

Niniejszy przegląd systematyczny obejmuje 21 badań, o łącznej liczbie 3454 dzieci, w których oceniano stosowanie różnych leków w obniżaniu ciśnienia tętniczego krwi u dzieci z nadciśnieniem. Dane są aktualne do października 2013 r. Większość badań trwała krótko a średni czas obserwacji wyniósł siedem tygodni. Badania charakteryzowały się niejednorodną jakością i w większości finansowane były przez przemysł. Nie we wszystkich badaniach porównywano skuteczność stosowania leku obniżającego ciśnienie tętnicze względem placebo. Tylko kilka klas (grup) leków, spośród tych najczęściej przepisywanych, zostało ocenionych i większość miała niewielki wpływ na ciśnienie tętnicze kwi, jednakże nie jest pewne czy skutkuje to poprawą odległych wyników zdrowotnych u dzieci. Podawanie leków w większych dawkach leków nie skutkowało większym obniżeniem ciśnienia tętniczego krwi. Wszystkie oceniane leki były bezpieczne, przynajmniej w krótkim okresie obserwacji.

Uwagi do tłumaczenia

Tłumaczenie Magdalena Koperny Redakcja Joanna Zając

Summary of findings(Explanation)

Summary of findings for the main comparison. Antihypertensive drug class compared to placebo for hypertension in children: systolic blood pressure reduction
Antihypertensive drug class compared to placebo for hypertension in children: systolic blood pressure reduction
Patient or population: children (from 1 to 18 years of age) with hypertension from primary or secondary causes
Settings: outpatient
Intervention: antihypertensive drug class
Comparison: placebo
Outcomes

Magnitude of reduction in systolic blood pressuremmHg in treatment group minus placebo response

Mean difference with 95% CI

No of participants
(studies)
Quality of the evidence
(GRADE)
Comments

Angiotensin receptor blocker (any dose)

Mean duration of 4 weeks

-6.50 (-9.44 to -3.56)240
(1 study)
⊕⊝⊝⊝
Very low
Refer 1 , 2 , 4 and 5 below.
Beta-blocker (any dose)
Mean duration of 4 weeks
-4.20 (-8.12 to -0.28)140
(1 study)

⊕⊝⊝⊝

Very low

Refer 1 to 5 below.
Calcium channel blocker (any dose)
Mean duration of 3 weeks
-0.62 (-2.97 to 1.73)133
(1 study)
⊕⊝⊝⊝
Very low
Refer 1 to 5 below.

Beta-blocker/thiazide combination (any dose)

Mean duration of 10 weeks

-4.40 (-8.99 to 0.19)94
(1 study)
⊕⊝⊝⊝
Very low
Refer 1 to 5 below.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: confidence interval

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Comments:

1. Evidence based on only one study.

2. “Industry-funded study. Industry-funded studies are at risk of over-estimating effect size.”

3. “Wide confidence interval.”

4. “High risk of publication bias.”

5. “Unclear method of randomisation and allocation concealment”.

Summary of findings 2 Antihypertensive drug class compared to placebo for hypertension in children: diastolic blood pressure reduction

Summary of findings 2. Antihypertensive drug class compared to placebo for hypertension in children: diastolic blood pressure reduction
Antihypertensive drug class compared to placebo for hypertension in children: diastolic blood pressure reduction
Patient or population: children (from 1 to 18 years of age) with hypertension from primary or secondary causes
Settings: outpatient
Intervention: antihypertensive drug class
Comparison: placebo
Outcomes

Magnitude of reduction in diastolic blood pressure mmHg in treatment group minus placebo response

Mean difference with 95% CI

No of participants
(studies)
Quality of the evidence
(GRADE)
Comments

Angiotensin receptor blocker (any dose)

Mean duration of 4 weeks

-5.50 (-9.62 to -1.38)240
(1 study)
⊕⊝⊝⊝
Very low
Refer to 1, 2, 4 and 5 below.
Beta-blocker (any dose)
Mean duration of 4 weeks
-3.20 (-7.12 to 0.72)140
(1 study)
⊕⊝⊝⊝
Very low
Refer to 1 to 5 below.
Calcium channel blocker (any dose)
Mean duration of 3 weeks
-1.86 (-5.23 to 1.51)133
(1 study)
⊕⊝⊝⊝
Very low
Refer to 1 to 5 below.

Beta-blocker/thiazide combination (any dose)

Mean duration of 10 weeks

-4.5 (-8.26 to -0.74)94
(1 study)
⊕⊝⊝⊝
Very low
Refer to 1 to 5 below.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: confidence interval

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Comments:

1. Evidence based on only one study.

2. “Industry-funded study. Industry-funded studies are at risk of over-estimating effect size.”

3. “Wide confidence interval.”

4. “High risk of publication bias.”

5. “Unclear method of randomisation and allocation concealment”.

Summary of findings 3 Angiotensin receptor blocker for hypertension in children: dose response in systolic blood pressure

Summary of findings 3. Angiotensin receptor blocker for hypertension in children: dose response in systolic blood pressure
Angiotensin receptor blocker for hypertension in children: dose response in systolic blood pressure
Patient or population: children (from 1 to 18 years of age) with hypertension from primary or secondary causes
Settings: outpatient
Intervention: angiotensin receptor blocker at a particular dose
Comparison: angiotensin receptor blocker at a different dose
Outcomes

Magnitude of reduction in systolic blood pressure mmHg in treatment group at various dose comparisons

Mean difference with 95% CI

No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Angiotensin receptor blocker >/= 6 years
Angiotensin receptor blocker (high- versus low-dose)-4.16 (-5.47 to -2.86)418 (5 studies)⊕⊕⊝⊝
Low
Refer to 1 and 3 below.
Angiotensin receptor blocker (high- versus medium-dose)-0.46 (-2.44 to 1.53)237 (3 studies)⊕⊝⊝⊝
Very Low
Refer to 1 to 3 below.
Angiotensin receptor blocker (medium- versus low-dose)-3.13 (-5.43 to -0.83)160 (3 studies)⊕⊝⊝⊝
Very Low
Refer to 1 and 3 below.
Angiotensin receptor blocker < 6 years
Angiotensin receptor blocker (high- versus low-dose)-3.01 (-8.79 to 2.76)67 (2 studies)

⊕⊝⊝⊝

Very Low

Refer to 1 to 3 below.
Angiotensin receptor blocker (high- versus medium-dose)-1.76 (-4.80 to 1.29)67 (2 studies)

⊕⊝⊝⊝

Very Low

Refer to 1 to 3 below.
Angiotensin receptor blocker (medium- versus low-dose)-1.32 (-4.54 to 1.90)50 (2 studies)

⊕⊝⊝⊝

Very Low

Refer to 1 to 3 below.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: confidence interval

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Comments:

1. “Industry-funded study. Industry-funded studies are at risk of over-estimating effect size.”

2. “Wide confidence interval.”

3. “Unclear method of randomisation and allocation concealment in some studies”.

Summary of findings 4 Angiotensin receptor blocker for hypertension in children: dose response in diastolic blood pressure

Summary of findings 4. Angiotensin receptor blocker for hypertension in children: dose response in diastolic blood pressure
Angiotensin receptor blocker for hypertension in children: dose response in diastolic blood pressure
Patient or population: children (from 1 to 18 years of age) with hypertension from primary or secondary causes
Settings: outpatient
Intervention: angiotensin receptor blocker at a particular dose
Comparison: angiotensin receptor blocker at a different dose
Outcomes

Magnitude of reduction in diastolic blood pressuremmHg in treatment group at various dose comparisons

Mean difference with 95% CI

No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Angiotensin receptor blocker >/= 6 years
High- versus low-dose-3.48 (-5.00 to -1.95)418 (5 studies)⊕⊕⊝⊝
Low

Evidence based on 5 reasonable quality studies.

Refer to comments 1 and 3.

High- versus medium-dose-0.59 (-2.49 to 1.32)237 (3 studies)⊕⊝⊝⊝
Very low
Refer to comments 1 to 3.
Medium- versus low-dose-3.04 (-5.67to -0.40)160 (3 studies)⊕⊝⊝⊝
Very low
Refer to comments 1 and 3.
Angiotensin receptor blocker < 6 years
High- versus low-dose-2.85 (-8.63 to 2.92)67 (2 studies)

⊕⊝⊝⊝

Very low

Refer to comments 1 to 3.
High- versus medium-dose-0.56 (-4.42 to 3.29)67 (2 studies)

⊕⊝⊝⊝

Very low

Refer to comments 1 to 3.
Medium- versus low-dose-1.31 (-3.74 to 1.12)50 (2 studies)

⊕⊝⊝⊝

Very low

Refer to comments 1 to 3.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: confidence interval

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Comments:

1. “Industry-funded study. Industry-funded studies are at risk of over-estimating effect size.”

2. “Wide confidence interval.”

3. “Unclear method of randomisation and allocation concealment in some studies”.

Summary of findings 5 Angiotensin converting enzyme inhibitor for hypertension in children: dose response in systolic blood pressure

Summary of findings 5. Angiotensin converting enzyme inhibitor for hypertension in children: dose response in systolic blood pressure
Angiotensin converting enzyme inhibitor for hypertension in children: dose response in systolic blood pressure
Patient or population: children (from 1 to 18 years of age) with hypertension from primary or secondary causes
Settings: outpatient
Intervention: angiotensin-converting enzyme inhibitor at a particular dose
Comparison: angiotensin-converting enzyme inhibitor at a different dose
Outcomes

Magnitude of reduction in systolic blood pressure mmHg in treatment group at various dose comparisons

Mean difference with 95% CI

No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Angiotensin-converting enzyme inhibitor    
High- versus low-dose-5.20 (-10.46 to 0.06)187 (3 studies)

⊕⊝⊝⊝

Very Low

Refer to 1 to 3 below.
High- versus medium- dose-2.72 (-5.83 to 0.40)187 (3 studies)

⊕⊝⊝⊝

Very Low

Refer to 1 to 3 below.
Medium- versus low-dose-2.01 (-6.07 to 2.05)139 (3 studies)

⊕⊝⊝⊝

Very Low

Refer to 1 to 3 below.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: confidence interval

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Comments:

1. “Industry-funded study. Industry-funded studies are at risk of over-estimating effect size.”

2. “Wide confidence interval.”

3. “Unclear method of randomisation and allocation concealment in some studies”.

Summary of findings 6 Angiotensin converting enzyme inhibitor for hypertension in children: dose response in diastolic blood pressure

Summary of findings 6. Angiotensin converting enzyme inhibitor for hypertension in children: dose response in diastolic blood pressure
Angiotensin converting enzyme inhibitor for hypertension in children: dose response in diastolic blood pressure
Patient or population: children (from 1 to 18 years of age) with hypertension from primary or secondary causes
Settings: outpatient
Intervention: angiotensin-converting enzyme inhibitor at a particular dose
Comparison: angiotensin-converting enzyme inhibitor at a different dose
Outcomes

Magnitude of reduction in diastolic blood pressuremmHg in treatment group at various dose comparisons

Mean difference with 95% CI

No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
High- versus low-dose-5.81 (-11.87 to 0.26)187 (3 studies)

⊕⊝⊝⊝

Very Low

Refer comments 1 to 3.
High- versus medium-dose-4.31 (-8.59 to -0.03)187 (3 studies)

⊕⊝⊝⊝

Very Low

Refer comments 1 to 3.
Medium- versus low-dose-0.46 (-2.19 to 1.27)139 (3 studies)

⊕⊝⊝⊝

Very Low

Refer comments 1 to 3.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: confidence interval

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Comments:

1. “Industry-funded study. Industry-funded studies are at risk of over-estimating effect size.”

2. “Wide confidence interval.”

3. “Unclear method of randomisation and allocation concealment in some studies”.

Summary of findings 7 Calcium channel blocker for hypertension in children: dose response in systolic and diastolic blood pressure

Summary of findings 7. Calcium channel blocker for hypertension in children: dose response in systolic and diastolic blood pressure
Calcium channel blocker for hypertension in children: dose response in systolic and diastolic blood pressure
Patient or population: children (from 1 to 18 years of age) with hypertension from primary or secondary causes
Settings: outpatient
Intervention: calcium channel blocker at a particular dose
Comparison: calcium channel blocker at a different dose
Outcomes

Magnitude of reduction in systolic blood pressure mmHg in treatment group at various dose comparisons

Mean difference with 95% CI

No of participants
(studies)
Quality of the evidence
(GRADE)
Comments

High vs low dose

Systolic blood pressure

-1.7 (-4.43 to 1.03)141 (1 study)

⊕⊝⊝⊝

Very Low

Refer to 1 to 5 below.

High vs low dose

Diastolic blood pressure

-0.70 (-2.63 to 1.29)141 (1 study)

⊕⊝⊝⊝

Very Low

Refer comments from 1 to 5.

Only 2 doses were studied. Dose was not based on child's weight.

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: confidence interval

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Comments:

1. Evidence based on only one study.

2. “Industry-funded study. Industry-funded studies are at risk of over-estimating effect size.”

3. “Wide confidence interval.”

4. “High risk of publication bias.”

5. “Unclear method of randomisation and allocation concealment”.

Background

Description of the condition

Hypertension is a major risk factor for cardiovascular disease and increases the risk of myocardial infarction, stroke, congestive heart failure, peripheral vascular disease and end-stage renal disease in adults (Chobanian 2003). In children, hypertension is defined as an average systolic or diastolic blood pressure (or both) that is on the 95th percentile or above for age, gender and height on at least three separate occasions (National High Blood Pressure Group 2004). Average systolic or diastolic blood pressure levels that are at or above the 90th percentile but below the 95th percentile have been designated as pre-hypertension (National High Blood Pressure Group 2004). The prevalence of paediatric pre-hypertension and hypertension is approximately 10% and 4%, respectively (Din-Dzietham 2007) and is rising. This is probably due to the increased prevalence of obesity and overweight globally (Bovet 2006; Din-Dzietham 2007; Ogden 2006). 

The natural history of hypertension in children is relatively uncertain, compared with that in adulthood. Overt cardiovascular events such as stroke and acute myocardial syndromes are rare in the majority of hypertensive children. Consequently, recent publications have focused on markers of hypertensive end organ damage including left ventricular mass index, increased carotid intima-media thickness, hypertensive retinopathy and early renal injury (Assadi 2007; Daniels 1991; Laird 1981; Lande 2006; McNiece 2007; Sorof 2003). 

The National High Blood Pressure Education Program Working group on High Blood Pressure in Children and Adolescents (The Fourth Report) has provided guidelines for the diagnosis, evaluation and treatment of hypertension) (National High Blood Pressure Group 2004). Non-pharmacological measures including weight reduction, regular physical activity and dietary modification have been recommended as first-line therapy for children with primary hypertension. These conservative measures have been shown to reduce blood pressure by 1 to 3 mmHg (Lipszyc 2011). Comprehensive programmes incorporating therapeutic lifestyle changes can be successful in reducing weight and elevated blood pressure  (Couch 2008; Denzer 2004; Lipszyc 2011). However, due to the variable success of non-pharmacological measures, a significant proportion of children still require antihypertensive medications (Liberman 2009). The indications for pharmacological intervention include secondary hypertension due to conditions such as renal artery stenosis, hypertension in children with established end organ damage or symptomatic hypertension, and failure of non-pharmacological measures.  

Why it is important to do this review

A systematic review of randomised controlled trials evaluating the benefits and harms of antihypertensive agents currently in use in children, compared to placebo or active comparators, has not been performed. The information derived from this review will assist physicians caring for children with hypertension in their decision-making regarding the choice and dose of medication.

Objectives

Primary objectives

To determine the dose-related effects of different classes of antihypertensive medications, as monotherapy compared to placebo; as combination therapy compared to placebo or a single medication; or in comparisons of various doses within the same class, on systolic or diastolic blood pressure (or both) in children with hypertension.

Secondary objectives

  1. To determine the effect of different classes of antihypertensive medications on target end organ disease, such as left ventricular hypertrophy, cardiac function, nephropathy and retinopathy.

  2. To determine the adverse events leading to medication withdrawal and adverse biochemical effects of antihypertensive drug therapy on serum potassium, uric acid, creatinine, glucose and lipid profile.

  3. To determine the adherence with different antihypertensive medications measured by self report, a pill count method or any other method used in the trial.

Methods

Criteria for considering studies for this review

Types of studies

Only randomised controlled trials (including either parallel-group or cross-over study designs) of different classes of antihypertensive agents or differing doses among a drug class in children with hypertension, with a minimum two-week follow-up, were eligible for inclusion.        

Types of participants

Inclusion criteria

Trials with children (from one to 18 years of age) with hypertension from primary or secondary causes, such as chronic kidney disease, were included. Hypertension was defined as an average systolic or diastolic blood pressure (or both) that is on the 95th percentile or above for age, gender and height on at least three separate occasions.

Exclusion criteria

We specifically excluded trials which included secondary hypertension associated with a kidney transplant, nephrotic syndrome treated with calcineurin inhibitors or high-dose steroids, children with malignant hypertension or children needing perioperative control of blood pressure.

Types of interventions

All pharmacological interventions were potentially eligible. Interventions which we considered included:

  • blood pressure-lowering medications versus placebo;

  • combination drugs versus placebo;

  • combination drugs versus a single medication;

  • comparison between different doses of same antihypertensive medications versus placebo;

  • comparison of two different classes of antihypertensive medications (either as an indirect comparison of each drug with placebo or a direct comparison, if one trial used two drugs from two different classes with a placebo control group as well);

  • comparison of two different medications within the same antihypertensive class (either as an indirect comparison of each drug within the same drug class with placebo or a direct comparison if one trial used two drugs from the same class with a placebo control group as well).

Types of outcome measures

Primary outcomes
  1. Change in office measurement of systolic blood pressure.

  2. Change in office measurement of diastolic blood pressure.

  3. Change in 24-hour ambulatory blood pressure monitoring (ABPM) of systolic blood pressure.

  4. Change in 24-hour ambulatory blood pressure monitoring (ABPM) of diastolic blood pressure.

Secondary outcomes
  1. Change in night time dipping and blood pressure load on ABPM.

  2. Regression of left ventricular hypertrophy or left ventricular wall thickness.

  3. Effect on progression of chronic kidney disease or kidney failure or proteinuria or albuminuria.

  4. Change in retinopathy.

  5. Change in heart rate, cardiac output and systemic vascular resistance as measured by standardised bio-impedance methods.

  6. Change in levels of serum creatinine, serum potassium, serum uric acid, blood glucose, serum cholesterol and triglycerides.

  7. The number of medication withdrawals due to adverse events.

  8. Average per cent adherence to medications during the trial as reported in the trials.

Search methods for identification of studies

We searched the Database of Abstracts of Reviews of Effectiveness (DARE) and the Cochrane Database of Systematic Reviews for related reviews.

We searched the Cochrane Hypertension Group Specialised Register, the Cochrane Central Register of Controlled Trials (CENTRAL) (2013, Issue 9), Ovid MEDLINE (1946 to October 2013), Ovid EMBASE (1974 to October 2013) and ClinicalTrials.gov for primary studies.

We searched the electronic databases using a strategy combining a variation of the Cochrane Highly Sensitive Search Strategy for identifying randomised trials in MEDLINE: sensitivity-maximising version (2008 revision) with selected MeSH and free-text terms relating to children and hypertension. We used no language restrictions. We translated the MEDLINE search strategy (Appendix 1) for EMBASE (Appendix 2), CENTRAL (Appendix 3) and the Hypertension Group Specialised Register (Appendix 4) using the appropriate controlled vocabulary as applicable.

The Hypertension Group Specialised Register contains over 22,000 randomised controlled trials identified from searches of MEDLINE, EMBASE, AGRICOLA, Allied and Complementary Medicine (AMED), BIOSIS, CAB Abstracts, CINAHL, Food Science and Technology Abstracts (FSTA), Global Health, International Pharmaceutical Abstracts (IPA), ISI Web of Science, LILACS, ProQuest Dissertations and Theses, and Scirus.

We also searched the following other sources:

  • International Clinical Trials Registry Platform (WHO ICTRP);

  • OpenSIGLE (System for Information on Grey Literature in Europe);

  • reference lists of nephrology textbooks, review articles and all papers and relevant reviews identified.

We also:

  • contacted authors of relevant papers regarding any further published or unpublished work;

  • contacted authors of trials reporting incomplete information to provide the missing information;

  • searched ISI Web of Science for papers which cite studies included in the review.

Data collection and analysis

Selection of studies

We performed the initial search of all databases to identify citations with potential relevance. Two review authors (SC and DL) initially screened abstracts and excluded articles whose titles or abstracts (or both) were clearly irrelevant. We retrieved the full text of the remaining articles. We searched the bibliographies of pertinent articles, reviews and texts for additional citations. SC and DL independently assessed the eligibility of the trials using a trial selection form. There were no major disagreements. Any minor differences in opinion regarding study inclusion were resolved after discussion with RP.

Data extraction and management

SC and DL extracted data independently using a standardised template. 

Assessment of risk of bias in included studies

We assessed risk of bias using the standard Cochrane 'Risk of bias' tool (Higgins 2011).                

Measures of treatment effect

The position of the patient during blood pressure measurement may affect the blood pressure-lowering effect. When blood pressure measurements were available in more than one position, the sitting blood pressures were the first preference. In order not to lose valuable data, if blood pressures were taken in any other position (lying or standing), we collected data from that position. We reported on the mean difference of the drug compared to placebo. However, in trials with no placebo, we compared the mean change among various doses (and termed these changes non-placebo corrected).

Dealing with missing data

In case of missing information in the included studies, we made efforts to contact investigators by e-mail to obtain this. We were able to get missing information for the Hazan 2010a, Hazan 2010b, Li 2004 and Trachtman 2008 studies and this has been included in the study results.

Assessment of heterogeneity

We assessed heterogeneity using the Chi2 test and the I2 statistic. 

Data synthesis

We used the random-effects model to obtain summary statistics for estimates of effect. We carried out data synthesis and analysis using the Cochrane Review Manager Software, RevMan 5.2 (RevMan 2012). We combined data for changes in blood pressure (systolic and diastolic) using the mean difference (MD) or the standardised mean difference (SMD) if different scales were used.   

Subgroup analysis and investigation of heterogeneity

We had planned to investigate age, gender, pubertal status, primary versus secondary cause of hypertension, ethnicity and weight as sources of heterogeneity using subgroup analysis and meta-regression but this was not possible because of insufficient data.  

Results

Description of studies

See: Characteristics of included studies; Characteristics of excluded studies.

Results of the search

The search strategy identified 4635 citations, with another study identified from a reference list search of retrieved studies. We screened a total of 4636 abstracts and titles electronically and excluded 4602 abstracts as they did not fulfil the inclusion criteria (Figure 1). Of the 33 articles obtained for a full-text review, 20 RCTs met the inclusion criteria and were included in the systematic review. One more included study was available only as an abstract. Hazan 2010a and Hazan 2010b (one study with stratification) included two cohorts (cohort A, 62% white; cohort B, 100% black). Cohort A was termed Hazan 2010a and Cohort B was termed Hazan 2010b for the purpose of the analysis in this review. Some of the reasons for exclusion were: non-randomised nature of the study, experimental nature of the therapy, participants not meeting the inclusion criteria, unavailability of important details such as the number of patients per group and the dose of antihypertensive used. Reasons for excluding each trial are listed in the Characteristics of excluded studies table. 

Figure 1.

Study flow diagram.

Included studies

We included 21 studies (one with two cohorts shown as two reports): Batisky 2007; BMS 2005; Falkner 1983; Flynn 2004; Flynn 2008; Gartenmann 2003; Hazan 2010a; Hazan 2010b; Li 2004; Li 2010; Schaefer 2010; Schaefer 2011; Shahinfar 2005; Soffer 2003; Sorof 2002; Trachtman 2003; Trachtman 2008; Webb 2010; Webb 2013; Wells 2002; Wells 2010; Wells 2011.

Trial designs were variable. Only five trials directly compared an antihypertensive drug to placebo in a parallel study design. The majority of the remaining trials used a design in which an initial screening/washout phase was followed by an active drug phase during which all children were treated with different doses of the same drug (phase one). This was immediately followed by a placebo withdrawal phase (phase two) during which study participants were re-randomised to either continue on the same dose of the study drug or receive a placebo. The carry-over effect of the study drug from phase one to phase two could not be ruled out as there was no intervening washout phase. Thus, only data from phase one were included in the analyses.

Risk of bias in included studies

Overall risk of bias across all included studies is displayed in Figure 2. The risk of bias for each included study is displayed in Figure 3. Trial quality was variable.  Information regarding adequate sequence generation was only available in 30% of studies and description of allocation concealment was only available in 15%. Only 55% of trials reported the results for all of the pre-defined outcomes.

Figure 2.

'Risk of bias' graph: review authors' judgements about each methodological quality item presented as percentages across all included studies.

Figure 3.

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Effects of interventions

See: Summary of findings for the main comparison Antihypertensive drug class compared to placebo for hypertension in children: systolic blood pressure reduction; Summary of findings 2 Antihypertensive drug class compared to placebo for hypertension in children: diastolic blood pressure reduction; Summary of findings 3 Angiotensin receptor blocker for hypertension in children: dose response in systolic blood pressure; Summary of findings 4 Angiotensin receptor blocker for hypertension in children: dose response in diastolic blood pressure; Summary of findings 5 Angiotensin converting enzyme inhibitor for hypertension in children: dose response in systolic blood pressure; Summary of findings 6 Angiotensin converting enzyme inhibitor for hypertension in children: dose response in diastolic blood pressure; Summary of findings 7 Calcium channel blocker for hypertension in children: dose response in systolic and diastolic blood pressure

Effect on blood pressure

The included studies used either auscultatory or oscillometric methods to measure blood pressure throughout the study. One of the included studies (Schaefer 2011) used an ambulatory blood pressure monitor (ABPM) in a subset of patients.

Antihypertensive drug versus placebo

Five of the included trials compared an antihypertensive agent directly with placebo (Batisky 2007; Sorof 2002; Trachtman 2003; Trachtman 2008; Wells 2010). The five different drugs studied included: candesartan 0.05 mg/kg/day to 0.4 mg/kg/day (Trachtman 2008); telmisartan 1 to 2 mg/kg/day (Wells 2010); extended-release metoprolol 0.2 mg/kg to 2.0 mg/kg (Batisky 2007); a combination drug comprising bisoprolol and hydrochlorothiazide (bisoprolol either 2.5 mg or 10 mg in combination with hydrochlorothiazide 6.25 mg) (Sorof 2002); and extended-release felodipine in either a 2.5 mg, 5 mg or 10 mg dose (Trachtman 2003). See Analysis 1.1 (Figure 4); Analysis 1.2 (Figure 5); Table 1. There were no trials comparing an angiotensin-converting enzyme inhibitor directly to a placebo. Candesartan (one trial, n = 240), when compared to placebo, reduced systolic blood pressure by 6.50 mmHg (95% confidence interval (CI) -9.44 to -3.56) and diastolic blood pressure by 5.50 mmHg (95% CI -9.62 to -1.38) (low-quality evidence). High dose telmisartan (one trial, n = 76), when compared to placebo, reduced systolic blood pressure by -8.50 (95% CI -13.79 to -3.21) but not diastolic blood pressure (-4.80, 95% CI -9.50 to 0.10) (low-quality evidence). Metoprolol, when compared with placebo (one trial, n = 140), significantly reduced systolic blood pressure by 4.20 mmHg (95% CI -8.12 to -0.28) but not diastolic blood pressure (-3.20 mmHg 95% CI -7.12 to 0.72) (low-quality evidence). Bisoprolol/hydrochlorothiazide combination, compared with placebo (one trial, n = 94), resulted in a significant reduction in systolic blood pressure (-4.50 mmHg, 95% CI -8.26 to -0.74) but did have an effect on diastolic blood pressure (-4.50 mmHg, 95% CI -8.26 to -0.74) (low-quality evidence). Extended-release felodipine (one trial, n = 133) was not effective in reducing systolic blood pressure (-0.62 mmHg, 95% CI -2.97 to 1.73) or diastolic blood pressure (-1.86 mmHg, 95% CI -5.23 to 1.51) when compared with placebo.

Figure 4.

Forest plot of comparison: 1 Antihypertensive drug class versus placebo, outcome: 1.1 Systolic blood pressure.

Figure 5.

Forest plot of comparison: 1 Antihypertensive drug class versus placebo, outcome: 1.2 Diastolic blood pressure.

Table 1. Effect on systolic and diastolic blood pressure in paediatric trials of antihypertensive agents versus placebo
Study n Systolic blood pressure:
mean change from baseline
SD 95% CIDiastolic blood pressure:
mean change from baseline
SD 95% CI
Angiotensin receptor blocker
Trachtman 2008 (candesartan)
Low (0.05 mg/kg)69-8.508.00-10.39 to -6.61-5.309.1-7.45 to -3.15
Medium (0.2 mg/kg)68-10.809.60-13.08 to -8.52-7.6010.2-10.02 to -5.18
High (0.4 mg/kg)68-11.3010.80-13.87 to -8.73-7.009.9-9.35 to -4.65
Placebo35-3.807.80-6.38 to -1.22-1.3011.5-5.11 to 2.51
Wells 2010 (telmisartan)
Low (1 mg/kg)28-9.007.80-6.10 to -11.90-7.709.60-4.10 to -11.30
High (2 mg/kg)31-13.008.70-9.90 to -16.10-7.808.10-4.90 to -10.70
Placebo14-4.7010.70+0.90 to -10.30-2.907.60+1.10 to -6.90
Beta-blocker
Batisky 2007 (metoprolol)
Low (0.2 mg/kg)45-5.208.70-7.70 to -2.60-3.108.90-5.70 to -0.50
Medium (1 mg/kg)23-7.708.60-11.30 to -4.00-4.908.90-8.60 to -1.30
High (2 mg/kg)49-6.309.00-8.70 to -3.80-7.508.90-10.00 to -5.00
Placebo23-1.908.90-5.50 to 1.80-2.108.80-5.70 to 1.50
Combination
Sorof 2002 (bisoprolol/hydrochlorothiazide)
Bisoprolol 10 mg/hydrochlorothiazide 6.25 mg62-9.3011.81-12.24 to -6.36-7.209.45-9.55 to -4.85
Placebo32-4.9010.18-8.43 to -1.37-2.708.49-5.64 to 0.24
Calcium channel blocker
Trachtman 2003 (felodipine extended-release)  Placebo corrected   Placebo corrected  
Low (2.5 mg)33-0.7111.99-4.80 to 3.38-2.0713.94-6.82 to 2.69
Medium (5 mg)34-0.0611.80-4.60 to 3.33-4.6413.79-9.18 to 0.09
High (10 mg)31-1.1311.86-5.30 to 3.05+1.3113.73-3.56 to 6.11
CI: confidence interval; SD: standard deviation. All doses mentioned are: per day
Calcium channel blocker versus placebo, dose response

One of the included trials compared the different doses (2.5 mg, 5 mg and 10 mg once daily) of extended-release felodipine with a placebo and found no significant dose response relationship (Trachtman 2003) (Analysis 1.1 - Figure 4; Analysis 1.2 - Figure 5; Table 1).

Angiotensin receptor blockers, dose response, with or without a placebo control

Nine of the included trials examined different angiotensin receptor blockers, such as losartan (Shahinfar 2005; Webb 2013), candesartan (Schaefer 2010; Trachtman 2008), valsartan (Flynn 2008; Wells 2011), olmesartan (Hazan 2010a; Hazan 2010b), telmisartan (Wells 2010) and irbesartan (BMS 2005) (Table 2). Only unpublished data were available for one study (BMS 2005). Another study was only available as an abstract (Webb 2013).

Table 2. Dose response effects of angiotensin receptor blockers on systolic and diastolic blood pressure in paediatric trials stratified by age
Study n Systolic blood pressure:
mean change from baseline
SD 95% CIDiastolic blood pressure:
mean change from baseline
SD 95% CI
≥ 6 years of age
Shahinfar 2005 (losartan)
Low (0.07 mg/kg)70-4.407.60-6.20 to -2.60-6.007.60-7.80 to -4.20
Medium (0.75 mg/kg)40-10.009.10-12.90 to -7.10-11.709.10-14.60 to -8.80
High (1.44 mg/kg)64-8.609.50-11.00 to -6.30-12.208.90-14.40 to -10.00
Wells 2011 (valsartan)
Low (0.4 mg/kg)102-7.9010.40-9.92 to -5.88-4.6011.00-6.72 to -2.48
Medium (1.3 mg/kg)52-9.609.10-12.07 to -7.13-5.808.90-8.22 to -3.38
High (2.7 mg/kg)105-11.5011.20-13.64 to -9.36-7.409.50-9.22 to -5.58
Hazan 2010a* (olmesartan)
Low (2.5/5 mg)95-7.809.18-9.66 to -5.94-5.508.06-7.15 to -3.89
High (20/40 mg)95-12.6010.15-14.65 to -10.55-9.509.76-11.47 to -7.53
Hazan 2010b* (olmesartan)
Low (2.5/5 mg)56-4.7011.48-7.71 to -1.69-3.508.84-1.18 to -5.82
High (20/40 mg)56-10.709.26-13.13 to -8.27-7.608.17-5.46 to -9.74
BMS 2005 (irbesartan)
Low (0.5 mg/kg)106-11.7---3.80--
Medium (1.5 mg/kg)106-9.30---3.20--
High (4.5 mg/kg)106-13.30---6.60--
< 6 years of age
Schaefer 2010 (candesartan)
Low (0.05 mg/kg)29-6.009.40-9.42 to -2.58-5.206.70-7.64 to -2.76
Medium (0.2 mg/kg)32-8.909.20-12.09 to -5.71-7.9012.90-12.37 to -3.43
High (0.4 mg/kg)32-12.108.30-14.98 to -9.22-11.109.20-14.29 to -7.91
Flynn 2008 (valsartan)
Low (5/10 mg)37-8.408.44-11.12 to -5.68-5.506.06-7.45 to -3.55
Medium (20/40 mg)18-8.307.63-11.82 to -4.78-6.404.23-8.35 to -4.45
High (40/80 mg)35-8.607.55-11.10 to -6.10-5.508.47-8.31 to -2.69
Webb 2013 (losartan)
Low (0.1 mg/kg)33-7.30---8.20--
Medium (0.3 mg/kg)34-7.60---5.10--
High (0.7 mg/kg)32-6.70---6.70--
*Hazan 2010 study stratified study population into 2 cohorts (cohort A, 62% white; cohort B, 100% black). Cohort A was termed Hazan 2010a and Cohort B was termed Hazan 2010b. All doses mentioned are per day. CI: confidence interval; SD: standard deviation

The six trials in children older than six years of age reported a non-placebo corrected dose response in systolic and diastolic blood pressure. Changes in mean sitting systolic and diastolic blood pressure in the low-, medium- and high-dose groups of the different angiotensin receptor blockers are summarised in Table 2.

On pooling data from these six trials, there was a significant dose response of angiotensin receptor blockers when comparing the high- versus low-dose groups for systolic (-4.2 mmHg, 95% confidence interval (CI) -5.6 to -2.9) and diastolic blood pressure (-3.5 mmHg, 95% CI -5.0 to -2.0). Similarly there was a significant dose effect apparent in the medium- versus low-dose groups in systolic (-3.1 mmHg, 95% CI -5.4 to -0.8) and in diastolic blood pressure (-3.0, 95% CI -5.7 to -0.4). There was, however, no significant dose effect in the high- versus medium-dose groups in systolic (-0.5 mmHg, 95% CI -2.4 to 1.5) or in diastolic blood pressure (-0.6 mmHg, 95% CI -2.5 to 1.3) (Analysis 2.1 - Figure 6; Analysis 2.2 - Figure 7) .

Figure 6.

Forest plot of comparison: 2 Drug different doses, outcome: 2.1 SBP.

Figure 7.

Forest plot of comparison: 2 Drug different doses, outcome: 2.2 DBP.

Three trials of angiotensin receptor blockers in children less than six years old reported a non-placebo corrected dose response in systolic and diastolic blood pressure (Flynn 2008; Schaefer 2010; Webb 2013). The doses of angiotensin receptor blockers used were: valsartan 0.4 to 3.4 mg/kg/day, candesartan 0.05 to 0.4 mg/kg/day and losartan 0.1 to 0.7 mg/kg/day

Changes in mean sitting systolic and diastolic blood pressure in the low-, medium- and high-dose groups of valsartan, candesartan and losartan are summarised in (Analysis 2.1; Analysis 2.2; Table 2). On pooling the data from the two trials of angiotensin receptor blockers in younger children, no significant dose effect was apparent for systolic and diastolic blood pressure (Analysis 2.1 - Figure 6; Analysis 2.2 - Figure 7) . One study could not be pooled as no standard deviations (SD) were available (Webb 2013).

Angiotensin-converting enzyme inhibitor, dose response, no placebo control

Three of the included trials examined the dose response relationships of three different angiotensin-converting enzyme inhibitors, namely enalapril (Wells 2002), fosinopril (Li 2004) and lisinopril (Soffer 2003). The enalapril and lisinopril trials had a similar dose response design with the primary endpoint of change in diastolic blood pressure. These two studies used the study drug in a suspension form with doses ranging from 0.08 mg/kg/day up to 0.6 mg/kg/day (Wells 2002; Soffer 2003). The fosinopril study used a tablet form of the drug and the doses ranged between 0.1 mg/kg/day and 0.6 mg/kg/day. The primary endpoint of the fosinopril study was the change in systolic blood pressure from baseline (Li 2004). All three angiotensin-converting enzyme inhibitors resulted in significant lowering of mean sitting systolic and diastolic blood pressure in the low-, medium- and high-dose groups compared to baseline but did not always show a dose response with higher doses (Analysis 2.1 - Figure 6; Analysis 2.2 - Figure 7; Table 3).

Table 3. Dose response effect of other antihypertensive drugs on systolic and diastolic blood pressure in paediatric trials
  1. BP: blood pressure; CI: confidence interval; SD: standard deviation

Study n Systolic blood pressure:
mean change from baseline
SD 95% CIDiastolic blood pressure:
mean change from baseline
SD 95% CI
Angiotensin-converting enzyme inhibitors
Wells 2002 (enalapril)
Low (0.02 mg/kg)30-7.109.30-10.60 to -3.60-6.307.80-9.20 to -3.40
Medium (0.08 mg/kg)29-6.6010.10-10.40 to -2.80-8.908.70-12.20 to -5.60
High (0.58 mg/kg)50-12.5011.70-15.80 to -9.20-14.908.70-17.40 to -12.40
Li 2004 (fosinopril)
Low (0.1 mg/kg)73-10.9010.25-13.20 to -8.60-4.506.84-6.00 to -3.00
Medium (0.3 mg/kg)86-11.3010.25-13.40 to -9.10-4.206.49-5.70 to -2.80
High (0.6 mg/kg)79-11.909.78-14.20 to -9.70-5.107.11-6.60 to -3.60
Soffer 2003 (lisinopril)
Low (0.02 mg/kg)33-5.409.90-9.00 to -1.90-7.609.30-10.90 to -4.30
Medium (0.07 mg/kg)24-12.109.10-16.00 to -8.30-9.308.70-13.00 to -5.70
High (0.61 mg/kg)58-15.2012.10-18.40 to -12.00-16.4011.70-19.50 to 13.30
Selective aldosterone receptor blocker
Li 2011 (eplerenone)
Low (25 mg)58-7.66---3.80--
Medium (50 mg)62-7.84---2.70--
High (100 mg)184-7.99---3.05--
Calcium channel blocker
Flynn 2004 (amlodipine)
Low (2.5 mg)127-7.311.4-9.3 to -5.3-3.78.3-5.1 to -2.3
High (5 mg)141-9.011.4-10.9 to -7.1-4.48.3-5.8 to -3.0
*Angiotensin receptor blocker versus angiotensin-converting enzyme inhibitor
Schaefer 2011
Valsartan (80/160/320 mg)148-15.40---9.40--
Enalapril (10/20/40 mg)148-14.10---8.50--
Calcium channel blockers versus angiotensin receptor blockers
Webb 2010 (amlodipine versuslosartan)
Amlodipine (0.05 to 0.2 mg/kg)30-0.10---0.80--
Losartan (0.7 to 1.4mg/kg)30-5.50---3.40--
Gartenmann 2003 (amlodipine versusirbesartan) Median change from baseline Range  Median change from baseline Range 
Amlodipine (20 to 40 kg: 5 to 10 mg; > 40 kg: 10 to 20 mg)1112.0010.00 to 14.00 7.005.00 to 10.00 
Irbesartan (20 to 40 kg: 75 to 150 mg; > 40 kg: 150 to 300 mg)1313.009 to 16 9.007.00 to 11.00 
Alpha blocker versus diuretic, Falkner 1981 (clonidine versus hydrochlorothiazide)
Clonidine Mean BP SD  Mean BP SD 
Pre1514616.00138 to 154966.0093 to 99
Low (0.1 mg twice) 13613.00129 to 143898.0085 to 93
High (0.2 mg twice) 1368.00132 to 140885.0085 to 91
Hydrochlorothiazide      
Pre151456.00142 to 148978.0093 to 101
Low (25 mg twice) 13910.00134 to 144946.0091 to 97
High (50 mg twice) 13511.00129 to 141938.0089 to 97
All doses mentioned are per day
Angiotensin-converting enzyme inhibitor versus angiotensin receptor blocker

One of the included trials compared an angiotensin-converting enzyme inhibitor with an angiotensin receptor blocker in 296 children aged six to 17 years (Schaefer 2011). The study compared enalapril (≥ 18 to < 35 kg, 10 mg; ≥ 35 to < 80 kg, 20 mg; ≥ 80 to ≤ 160 kg, 40 mg) with valsartan (≥18 to < 35 kg, 80 mg; ≥ 35 to < 80 kg, 160 mg; ≥ 80 to ≤ 160 kg, 320 mg) for a period of 12 weeks. Both drugs were equally effective in reducing blood pressure compared to baseline with a mean systolic/diastolic blood pressure reduction of -15.4/-9.4 mmHg for valsartan compared to -14.1/-8.5 mmHg for enalapril (Table 3). Among the subset of patients who had ABPM done (n = 56), valsartan resulted in a greater reduction in mean 24-hour systolic blood pressure (valsartan -9.8 mmHg versus enalapril -7.2 mmHg; P = 0.03). Limited data were presented for this trial, including a lack of standard deviations or standard errors, preventing the calculation of confidence intervals.

Calcium channel blockers, dose response, no placebo control

One of the included trials assessed the antihypertensive efficacy of amlodipine (Flynn 2004). The study compared two different doses of amlodipine in a phase lasting four weeks, followed by an additional four-week placebo withdrawal phase. The data from the first four-week phase were included in this review. The dose response was not significant for either systolic or diastolic blood pressure (Analysis 2.1 - Figure 6; Analysis 2.2 - Figure 7; Table 3).

Calcium channel blocker versus angiotensin receptor blockers

Two of the included trials compared calcium channel blockers with angiotensin receptor blockers. One trial compared amlodipine (median dose 0.19 to 0.33 mg/kg/day) with irbesartan (median dose 2.9 to 4.8 mg/kg/day) (Gartenmann 2003), while the other compared amlodipine 0.1 to 0.2 mg/kg/day and losartan 0.7 to 1.4 mg/kg/day (Webb 2010). 

In the trial comparing amlodipine with irbesartan, both drugs were equally effective in reducing blood pressure compared to baseline (Table 3). The trial comparing amlodipine with losartan found a mean change in systolic blood pressure compared to baseline of -0.1 mmHg and -5.5 mmHg, respectively. When compared with baseline, treatment with losartan led to a decrease in diastolic blood pressure by 3.8 mmHg. However, there was an increase of 0.8 mmHg (SD 12.4) in the amlodipine group (Table 3). Confidence intervals were not available for this trial as standard deviations/standard errors for the change in blood pressures were not reported (Webb 2010).

Selective aldosterone antagonist, dose response, no placebo control

There was only one trial of eplerenone, a competitive aldosterone antagonist (Li 2010). The study compared three different doses (25 mg, 50 mg and 100 mg/day) of eplerenone in phase A which lasted for six weeks, followed by an additional four-week placebo withdrawal phase. The data from first six weeks were included in the review (Table 3). There was a mean reduction in systolic blood pressure compared to a baseline of -7.7 mmHg, -7.8 and -8.0 in the low-, medium- and high-dose groups, respectively. There was a mean reduction in diastolic blood pressure compared to a baseline of -3.8 mmHg, 2.7 mmHg and 3.1 mmHg in the low-, medium- and high-dose group, respectively. Confidence intervals were not available for this trial as standard deviations were not reported.

Alpha blocker versus diuretic, no placebo control

One of the included trials compared clonidine, a centrally acting alpha blocker, with hydrochlorothiazide, a diuretic (Falkner 1983). Active therapy was initiated at a low dose for the first 12 weeks (low-dose period, clonidine 0.1 mg twice daily and hydrochlorothiazide 25 mg twice daily) and increased to a high dose (clonidine 0.2 mg twice daily and hydrochlorothiazide 50 mg twice daily) if blood pressure goals were not achieved. Both clonidine and hydrochlorothiazide effectively reduced systolic and diastolic blood pressure (P < 0.05) (Table 3).

Progression of chronic kidney disease

None of the included trials reported the effect of blood pressure control on the progression of chronic kidney disease alone.

Adverse drug effects

The adverse events associated with the antihypertensive agents were mostly minor and occurred in approximately 3% to 40% of all patients studied (Table 4). The most common adverse events included headaches, dizziness and upper respiratory infections.

Table 4. Adverse effects of antihypertensive medications
Drugs Common side effects Less common side effects
Angiotensin-converting enzyme inhibitor

Headache (20.1%), nasopharyngitis (9.6%), cough (9.1%) with fosinopril
Headache (3.4%), cough (< 1%) with lisinopril. 2 in high-dose lisinopril group had abdominal pain, diarrhoea, nausea and vomiting and 2 in medium-dose group experienced dizziness. Only 1 event of cough was reported in the lisinopril study and thought to be unrelated to the drug

Dizziness was the most common side effect in the enalapril study (4 patients, followed by cough (3 patients) and headache (2 patients). No angioedema reported in any of the 3 trials

Low-dose lisinopril - 1 patient developed leucopenia and another patient had elevated creatinine. High-dose lisinopril: 1 patient developed hyperkalaemia and another had elevated creatinine

In the fosinopril study: elevated creatinine was the most frequent lab abnormality (drug versus placebo 8.9% versus 4.1%). All elevations were transient and did not exceed the upper limit of normal

None of the patients in the enalapril study experienced renal failure or hyperkalaemia 

Angiotensin receptor blocker in > 6 years

Headache was reported in 33.2% of children in the valsartan study and in 5.4% to 14.7% of children in the olmesartan study

The most common side effects in the candesartan trial were headache, upper respiratory infection, dizziness, cough and sore throat

The olmesartan study reported headache (7.4%) and pharyngo-laryngeal pain (6.3%) in the low-dose group and headache (14.7%) and dizziness (9.5%) in the high-dose group among the mixed racial cohort. Among the black patients, headache (5.4%) in the low-dose and headache (8.9%) and toothache (3.6%) in high-dose group were reported

In the valsartan study, 10% of patients had elevated creatinine and 4% patients had hyperkalaemia at some point during the study

In the candesartan trial, 8 patients discontinued treatment. The reasons for discontinuation were hypotension, arm fracture, dizziness, headache, low white cell count and progression of underlying renal disease (2 patients)

2 patients in the mixed racial group in the olmesartan study discontinued treatment (1 patient because of hypertension and other due to hypoaesthesia)

Angiotensin receptor blocker < 6 yearsAngiotensin receptor blocker was well tolerated in this age group with a low incidence of side effects and no adverse effect on growth

2 significant adverse events were reported in the candesartan study (Schaefer 2010). 1 event was a death related to acute on chronic renal failure provoked by diarrhoea-induced hypovolaemia and the second event was a vaso-occlusive event which occurred in a child with nephrotic syndrome. The authors report that both events were not directly attributed to angiotensin receptor blocker use

1 participant each had pruritus and rash, decreased appetite and blurred vision in phase 1 of the valsartan study. In phase 2, 1 patient had headache and 1 hypertriglyceridaemia

Beta-blockerHeadache (30%), upper respiratory infection (10%) and cough (19%) in the metoprolol studyFatigue (9%) and dizziness (6%). 5 patients discontinued the metoprolol study in the open-label period because of an adverse event (1 patient each for fatigue, nightmares, anxiety, dizziness and asthma)
Beta-blocker + diureticNo significant adverse effects reported  
Calcium channel blockerHeadache (33%), respiratory infection (12%) and nausea (10%) were the common adverse events among children on extended-release felodipinePedal oedema was noted in 2% of patients in the felodipine study and in ˜3.1% to 4.3% patients on amlodipine
Centrally acting α blocker versus diuretic14 of the 15 hydrochlorothiazide-treated participants had a reduction in serum potassium. Symptomatic hypokalaemia in 4 participants requiring potassium supplementation in the hydrochlorothiazide groupDrowsiness and dry mouth in 1 patient on clonidine

Six trials tested the safety of angiotensin receptor blockers in six to 16-year olds (BMS 2005; Hazan 2010a; Hazan 2010b; Schaefer 2011; Shahinfar 2005; Trachtman 2008; Wells 2010; Wells 2011). Two of these trials had an extended 52-week, open-label phase after the short randomised phase (Trachtman 2008; Wells 2011).

Angiotensin receptor blockers were well tolerated with headache being the most commonly observed adverse event, affecting up to a third of participants. In the valsartan dose response study, 10% of patients had elevated creatinine and 4% of patients had hyperkalaemia at some point during the study (Wells 2011). Gastroenteritis (one patient) and hyperkalaemia (one patient) were the two serious adverse events considered to be related to valsartan (Wells 2011). In the patient with hyperkalaemia, serum potassium returned to normal two days after valsartan was discontinued (Wells 2011). In another trial comparing valsartan with enalapril, a total of seven patients (4.6%) in the valsartan group discontinued treatment because of adverse effects (Schaefer 2011). The reasons for discontinuation were nausea in two patients and arrhythmia, dehydration, dizziness, headache, vomiting and hyperkalaemia in one patient each. In the candesartan trial, three patients discontinued treatment in the dose-ranging phase and five during the open-label phase due to adverse events (Trachtman 2008). These adverse events included hypotension, arm fracture, dizziness, headache, low white blood cell count and progression of underlying renal disease (two patients) (Trachtman 2008).

The most common clinical adverse events reported in the losartan study were headache and upper respiratory tract infection (Shahinfar 2005). Only one patient discontinued the study because of a drug-related clinical adverse event, namely mild reversible hypotension. Increased serum creatinine was reported in one patient in the medium-dose group and hyperkalaemia was reported in one patient in the low-dose group (Shahinfar 2005).

Most of the adverse events reported in the olmesartan study were mild to moderate in intensity, with more adverse events reported among non-black patients compared to black patients (Hazan 2010a; Hazan 2010b). The most common adverse events were headache, pharyngo-laryngeal pain and toothache. One patient discontinued treatment in the low-dose group due to moderate hypoaesthesia, possibly related to drug therapy. Haematology, serum chemistry and urine analysis did not change significantly from baseline (Hazan 2010a; Hazan 2010b).

Three trials tested the safety of angiotensin receptor blockers (candesartan, losartan and valsartan) in children under six years old (Flynn 2008; Schaefer 2010; Webb 2013). The candesartan study was of four weeks duration whereas the valsartan study had a four-week double-blind phase followed by a 52-week open-label phase. Both studies reported that angiotensin receptor blockers were well tolerated in this age group with a low incidence of side effects and no adverse effect on growth. The candesartan trial reported two significant adverse events. One event was a death related to acute on chronic renal failure provoked by diarrhoea-induced hypovolaemia and the second event was a vaso-occlusive event which occurred in a child with nephrotic syndrome (Schaefer 2010). The authors report that both events were not directly attributed to angiotensin receptor blocker medication use.

Four trials tested the short-term safety of angiotensin-converting enzyme inhibitors in children aged six to 16 years (Li 2004; Schaefer 2011; Soffer 2003; Wells 2002). One study using fosinopril reported the incidence of drug-related cough as 9.1% during the 52-week long open-label phase (Li 2004). None of the patients receiving lisinopril and only one out of 110 patients receiving enalapril were reported to have drug-related cough (Soffer 2003; Wells 2002). In the valsartan versus enalapril trial, the incidence of serious adverse effects was low (1.7%) and similar in both groups (Schaefer 2011). Headache, cough and nasopharyngitis were the most common adverse events. There was no discontinuation secondary to adverse events in the enalapril group.

No cases of angioedema were reported in these four angiotensin-converting enzyme inhibitor trials. Small increases in serum creatinine and blood urea nitrogen were observed but renal failure was not reported in any of the four trials. Importantly, the biochemical changes with angiotensin-converting enzyme inhibitors, including hyperkalaemia and rise in creatinine, were usually transient and did not lead to discontinuation of treatment in most cases.

One trial examined the effect of the novel aldosterone antagonist eplerenone in four to 16-year old children and reported adverse events in 50% or less of the participants (Li 2010). Most of these were of a mild nature and included headache and upper respiratory infections. Four children discontinued treatment, three of which were considered treatment-related (hypotension, hypertension and fatigue). Serious adverse events were reported in eight (2.6%) of the children and included diarrhoea, sleep apnoea, syncope, pericarditis, pneumonia, arthritis, sepsis and pleural effusion. None of these events were considered related to the study drug. Eosinophilia was noted in 13% to 17% of children. The majority of these children had abnormal readings at baseline and were from developing countries where parasitic diseases may have a role to play.

Headache (33%), respiratory infection (12%) and nausea (10%) were the common adverse events among children on extended-release felodipine (Trachtman 2003). Pedal oedema was noted in 2% of patients in the felodipine study and in ˜3.1% to 4.3% patients on amlodipine (Flynn 2004; Trachtman 2003). There were no serious biochemical abnormalities reported with calcium channel blockers.

In the study using extended-release metoprolol, no significant adverse events were noted during the initial double-blind, four-week period (Batisky 2007). As expected, the mean heart rate was slowed by extended-release metoprolol (maximum mean change -6.5 beats per minute at 1.0 mg/kg). The most common adverse events noted during the open-label, 52-week follow-up were headache (30%), upper respiratory infection (10%), cough (19%), fatigue (9%) and dizziness (6%). Five patients discontinued the study during this period because of an adverse event such as fatigue, nightmares, anxiety, dizziness or asthma. With the beta-blocker/thiazide combination, no significant drug-related adverse events were noted compared to placebo (Sorof 2002). A drop in serum potassium levels was seen in 14/15 participants receiving hydrochlorothiazide (Falkner 1983). Four of these required potassium supplementation.

Adherence

Information on compliance was available for only seven of the 21 trials (30%), with four reported in the publication (Batisky 2007; Trachtman 2003) and three received via personal communication (Hazan 2010a; Hazan 2010b; Li 2004; Trachtman 2008). Information regarding compliance was gathered by pill count or volume of suspension remaining. Compliance was reported to be over 80% in these trials. Compliance was reported to be equal across all dose groups in Batisky 2007 and Li 2004. One study reported higher compliance (94%) in the placebo group compared to the felodipine treatment groups (84% to 91%) (Trachtman 2003). This was a negative study with multiple design issues and differential compliance may have contributed to the negative result. Similarly, compliance was higher in the placebo group (97%) compared to the active treatment group (93%) in Trachtman 2008. In the Hazan 2010 study, compliance was better among the mixed racial group population compared to the black population. In the mixed racial group (Hazan 2010a), compliance was 94% or more, whereas in the black cohort (Hazan 2010b) compliance was 85% or more for both dose groups (low- and high-dose olmesartan), respectively. 

Discussion

Summary of main results

Twenty-one trials evaluated antihypertensive medications of various drug classes in a total of 3454 hypertensive children, with short periods of follow-up ranging from three to 24 weeks (mean 7.2 weeks).

Drugs blocking the renin-angiotensin system, calcium channel blockers, beta-blockers, alpha blockers and diuretics are widely used forthe management of hypertension in children but there have been few trials (n = 5) comparing the antihypertensive agent directly against placebo. The majority of the remaining studies were dose-finding and included three trials of angiotensin-converting enzyme inhibitors (lisinopril, enalapril and fosinopril), seven trials of angiotensin receptor blockers (valsartan, candesartan, losartan, irbesartan and olmesartan), one trial of a calcium channel blocker (amlodipine) and one trial which used a selective aldosterone antagonist, eplerenone. Four trials compared different classes of antihypertensive medications: two compared calcium channel blockers to angiotensin receptor blockers, one compared a centrally acting alpha blocker with a diuretic and one compared an angiotensin-converting enzyme inhibitor with an angiotensin receptor blocker.

We report that the use of candesartan led to a significant reduction in systolic (mean difference -6.50 mmHg, 95% CI -9.44 to -3.56) and diastolic blood pressure (mean difference -5.50 mmHg, 95% CI -9.62 to -1.38) when compared to a placebo. High dose telmisartan reduced systolic (mean difference -8.5, 95% CI -13.79 to -3.21) but not diastolic blood pressure (-4.80, 95% CI -9.50 to 0.10). There was no consistent dose response relationship, however, escalating doses of angiotensin receptor blockers are consistent with pharmacokinetic studies demonstrating a flattening of the dose response curve at higher doses in adults (Taddei 2011). Absence of an expected linear dose response could also have resulted from trial design issues such as the use of small dose ranges. Three trials assessed angiotensin receptor blockers in children under six years and demonstrated that angiotensin receptor blockers were modestly effective in reducing systolic and diastolic blood pressure and generally well tolerated in young children.

Angiotensin-converting enzyme inhibitors demonstrated a good reduction in systolic and diastolic blood pressure compared to baseline, although a consistent dose response relationship was again not evident when pooling the data. Absence of an expected linear dose response relationship may be due to a number of issues including trial design, with a narrow range of medication dosing among the three groups, imprecise mg/kg dosing, targeting systolic versus diastolic hypertension and relatively small sample sizes. Additionally, the pharmacokinetics of younger children may require more frequent or higher doses based on metabolism and dosing regimens were not specifically different than the reported adult regimens (Benjamin 2008).  

Beta-blockers are also commonly prescribed antihypertensive medications for children with known cardiac conditions and they are also used in neonatal intensive care units. However, they are not typically prescribed as first-line agents for lowering blood pressure in those with primary or secondary hypertension. The drugs primarily reduce blood pressure by dampening beta effects such as lowering heart rate, decreasing sympathetic overactivity and vasodilatation. All effects are also beneficial in those with obesity, obstructive sleep apnoea and in children with cardiac disease. There is also a significant benefit of carvedilol in children with heart failure (Flynn 2011). We found very limited data, however, on their antihypertensive efficacy and safety in children. Only one study used a beta-1 selective agent such as metoprolol (Batisky 2007). Metoprolol reduced systolic blood pressure by ˜4 mmHg but not diastolic blood pressure compared to placebo. In the beta-blocker/thiazide (bisoprolol/hydrochlorothiazide) trial (Sorof 2002), the combination drug did not significantly reduce systolic blood pressure. It did significantly lower diastolic blood pressure by ˜4.5 mmHg. The efficacy of beta-blockers in children is less than in adults which may reflect the concerns of increasing doses to effectively cause beta blockade and reduce the heart rate. Heart rate was reduced and medication was not escalated to achieve lower heart rates with adequate blockade. In addition, the small sample size and short duration of studies may have resulted in negative results. For example, in the metoprolol study only 49 children were assigned to the high-dose group (2 mg/kg/day) and the full dose was achieved only for two weeks even though it was a four-week study. Additionally, the combination beta-blocker/thiazide study had a high drop-out rate of 27% limiting the data for blood pressure changes. Only 25 patients in this study were escalated to the highest dose levels and did demonstrate a linear dose response. The authors also analysed the response based on the severity of initial hypertension and reported better response in those with severe hypertension at enrolment (systolic or diastolic blood pressure 5 mmHg or more above the 95th centile).  

Even though calcium channel blockers are very commonly prescribed antihypertensive agents in paediatric patients, the evidence for their blood pressure-lowering efficacy is limited. The blood pressure-lowering efficacy of extended-release felodipine was tested in one trial, and it failed to demonstrate a clinically significant reduction in either systolic or diastolic blood pressure (Trachtman 2003). In the amlodipine dose response trial, there was also no significant dose response for either systolic or diastolic blood pressure. The lack of blood pressure-lowering response with calcium channel blockers, despite the widespread clinical use of these medications, could reflect the individual trial designs and potential methodological issues in the trials presented. For example, the patients on the extended-release felodipine were only exposed to the highest dose of 10 mg for one week. There was also differential compliance in different dose groups which could impact on the overall decline in blood pressure from baseline to follow-up. Patient selection included patients with both systolic and diastolic hypertension; however, only diastolic pressures were included in the primary endpoint. This may have under-powered the study to demonstrate the antihypertensive efficacy of extended-release felodipine in diastolic blood pressure reduction as more patients had elevated systolic blood pressure at study entry. The amlodipine study only had two dose groups of 2.5 and 5 mg so the dose range was very small to compare differences in effect. Additionally, this study did not use individual participant-based dosing but gave all participants in the low-dose arm 2.5 mg and all participants in the higher-dose arm 5 mg, leading to inappropriate weight-based dosing (Benjamin 2008).

There were very few studies which measured end organ damage outcomes such as proteinuria. Only two of the included studies reported the effect of angiotensin-converting enzyme inhibitor/angiotensin receptor blockers on proteinuria (Schaefer 2010; Webb 2010), which were effective in reducing this. None of the included trials reported other surrogate markers of outcome such as the change in left ventricular hypertrophy, change in retinopathy or change in cardiac output, and systemic vascular resistance as measured by standardised bio-impedance methods.

The ESCAPE trial (effect of strict blood pressure control and angiotensin-converting enzyme inhibition on the progression of chronic renal failure in paediatric patients) did not fit the inclusion criteria of this review of the primary effect of the medication. It deserves a mention, however, since this is the only study to date which has looked at the impact of blood pressure control on the progression of kidney disease in children aged three to 18 years old with glomerular filtration rates (GFRs) of 15 to 80 ml/minute over a five-year period (Wühl 2009). In this study, patients were randomly assigned to intensified blood pressure control (with a target 24-hour mean arterial pressure below the 50th percentile) or to conventional blood pressure control (mean arterial pressure in the 50th to 95th percentile) using ramipril (6 mg/m2 body surface area/day) and additional medications to keep the blood pressure at target. A total of 385 children were randomised to the two groups. Fewer patients in the intensified control arm than in the conventional group reached the primary endpoint of a 50% decline in GFR or progression to end-stage renal disease (29.9% versus 41.7%; hazard ratio = 0.65; 95% CI 0.44 to 0.94). This study demonstrated that the intensified blood pressure control is required to reduce end organ damage and, also, that it would take on average three medications to achieve the controlled blood pressure. We speculate that the amount of blood pressure reduction may be a more important determinant of improved outcomes in hypertensive children rather than the choice of antihypertensive agent per se. It has been shown in adult trials that better blood pressure control results in better risk reduction irrespective of drug class (ALLHAT 2002; Hansson 1999; Hansson 2000). Therefore it may be useful to compare the long-term efficacy of different classes of drugs in reducing end organ damage using surrogate outcomes such as carotid intimal medial thickness, left ventricular hypertrophy, retinopathy and renal disease progression in hypertensive children.

Quality of the evidence

The main limitation of this review is the small number of trials conducted in children that could be included. Some studies were of sub-optimal study quality and only a few compared a drug directly with placebo. Additionally, many of these studies were of short duration and there was paucity of data regarding the impact of blood pressure control on target organ damage. Furthermore, only one of the included studies used ABPM to guide therapy in the treatment arms in a subset of children (Schaefer 2011).

The results of the analyses are not robust enough to provide firm recommendations for first-line agents in children with hypertension from primary or secondary causes. There is a suggestion, however, that those medications that act through the renin-angiotensin pathway may lower blood pressure more than other pharmacological interventions.

Authors' conclusions

Implications for practice

  1. This review provides low-quality evidence that angiotensin receptor blockers reduce systolic and diastolic blood pressure by at least 5 mmHg when compared to placebo.

  2. This review provides low-quality evidence that angiotensin receptor blockers in children under six years result in a modest reduction in systolic and diastolic blood pressure and are generally safe to use for a short duration.

  3. Angiotensin receptor blockers, however, did not display a consistent dose response relationship with escalating doses.

  4. Based on limited data, angiotensin-converting enzyme inhibitors demonstrate a reduction in systolic and diastolic blood pressure compared to baseline, although a consistent dose response relationship for escalating doses was not evident.

  5. Based on two small-sized studies, beta-blockers modestly reduce systolic blood pressure by 4 mmHg but have no significant effect on diastolic blood pressure compared to placebo.

  6. Even though calcium channel blockers are very commonly prescribed antihypertensive agents in paediatric patients, the evidence for their blood pressure-lowering efficacy is limited. Further, there was no consistent dose response displayed in one study comparing a high- versus low-dose calcium channel blocker.

  7. The most commonly used antihypertensive agents were safe to use in short-term studies.

Implications for research

The data are limited as it is often considered unethical or unacceptable for parents and physicians to withhold medications if clinically indicated in children. Thus, the use of placebo-controlled trials is limited in children. There is a need for more long-term trials to determine the efficacy and safety of antihypertensive agents in reducing target organ damage. Furthermore, there are classes of medications that are known to potentiate blood pressure-lowering, such as diuretics with an angiotensin-converting enzyme inhibitor or beta-blocker, which have not been systematically studied in children. In addition, there is a need for evaluation of the safety and reliability of long-term monitoring of therapy with home blood pressure monitoring and ambulatory blood pressure monitoring. The ESCAPE trial (Wühl 2009) also addressed the effect of intensive therapy with a targeted blood pressure lower than that prescribed in current guidelines and this should be considered for clinical trials in children. Future studies should be rigorous in considering the blood pressure goals achieved and in using a standardised medication with dose escalation to determine the most efficacious medications in children. 

Acknowledgements

The authors would like to acknowledge the help provided by the Cochrane Hypertension Group and Ms Miranda Cumpston from the Australasian Cochrane Centre.

Data and analyses

Download statistical data

Comparison 1. Antihypertensive drug class versus placebo
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Systolic blood pressure5 Mean Difference (Random, 95% CI)Totals not selected
1.1 ARB (any dose)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
1.2 ARB (high-dose versus placebo)2 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
1.3 ARB (medium-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
1.4 ARB (low-dose versus placebo)2 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
1.5 BB (any dose)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
1.6 BB (high-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
1.7 BB (medium-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
1.8 BB (low-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
1.9 CCB (any dose)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
1.10 CCB (high-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
1.11 CCB (medium-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
1.12 CCB (low-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
1.13 BB/thiazide combination (any dose) versus placebo1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
2 Diastolic blood pressure5 Mean Difference (Random, 95% CI)Totals not selected
2.1 ARB (any dose)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
2.2 ARB (high-dose versus placebo)2 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
2.3 ARB (medium-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
2.4 ARB (low-dose versus placebo)2 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
2.5 BB (any dose)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
2.6 BB (high-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
2.7 BB (medium-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
2.8 BB (low-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
2.9 CCB (any dose)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
2.10 CCB (high-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
2.11 CCB (medium-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
2.12 CCB (low-dose versus placebo)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
2.13 BB/thiazide diuretic (any dose)1 Mean Difference (Random, 95% CI)0.0 [0.0, 0.0]
Analysis 1.1.

Comparison 1 Antihypertensive drug class versus placebo, Outcome 1 Systolic blood pressure.

Analysis 1.2.

Comparison 1 Antihypertensive drug class versus placebo, Outcome 2 Diastolic blood pressure.

Comparison 2. Antihypertensive drug dose response
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Systolic blood pressure12 Mean Difference (IV, Random, 95% CI)Subtotals only
1.1 ARB (high- versus low-dose), age ≥ 6 years6837Mean Difference (IV, Random, 95% CI)-4.16 [-5.47, -2.86]
1.2 ARB (high- versus medium-dose), age ≥ 6 years3397Mean Difference (IV, Random, 95% CI)-0.46 [-2.44, 1.53]
1.3 ARB (medium- versus low-dose), age ≥ 6 years3401Mean Difference (IV, Random, 95% CI)-3.13 [-5.43, -0.83]
1.4 ARB (high- versus low-dose), age < 6 years2133Mean Difference (IV, Random, 95% CI)-3.01 [-8.79, 2.76]
1.5 ARB (high- versus medium-dose), age < 6 years2117Mean Difference (IV, Random, 95% CI)-1.76 [-4.80, 1.29]
1.6 ARB (medium- versus low-dose), age < 6 years2116Mean Difference (IV, Random, 95% CI)-1.32 [-4.54, 1.90]
1.7 ACEI (high- versus low-dose)3323Mean Difference (IV, Random, 95% CI)-5.20 [-10.46, 0.06]
1.8 ACEI (high- versus medium-dose)3326Mean Difference (IV, Random, 95% CI)-2.72 [-5.83, 0.40]
1.9 ACEI (medium- versus low-dose)3275Mean Difference (IV, Random, 95% CI)-2.01 [-6.07, 2.05]
1.10 CCB (high- versus low-dose)1268Mean Difference (IV, Random, 95% CI)-1.70 [-4.43, 1.03]
2 Diastolic blood pressure12 Mean Difference (IV, Random, 95% CI)Subtotals only
2.1 ARB (high- versus low-dose), age ≥ 6 years6837Mean Difference (IV, Random, 95% CI)-3.48 [-5.00, -1.95]
2.2 ARB (high- versus medium-dose), age ≥ 6 years3397Mean Difference (IV, Random, 95% CI)-0.59 [-2.49, 1.32]
2.3 ARB (medium- versus low-dose), age ≥ 6 years3401Mean Difference (IV, Random, 95% CI)-3.04 [-5.67, -0.40]
2.4 ARB (high- versus low-dose), age < 6 years2133Mean Difference (IV, Random, 95% CI)-2.85 [-8.63, 2.92]
2.5 ARB (high- versus medium-dose), age < 6 years2117Mean Difference (IV, Random, 95% CI)-0.56 [-4.42, 3.29]
2.6 ARB (medium- versus low-dose), age < 6 years2116Mean Difference (IV, Random, 95% CI)-1.31 [-3.74, 1.12]
2.7 ACEI (high- versus low-dose)3323Mean Difference (IV, Random, 95% CI)-5.81 [-11.87, 0.26]
2.8 ACEI (high- versus medium-dose)3326Mean Difference (IV, Random, 95% CI)-4.31 [-8.59, -0.03]
2.9 ACEI (medium- versus low-dose)3275Mean Difference (IV, Random, 95% CI)-0.46 [-2.19, 1.27]
2.10 CCB (high- versus low-dose)1268Mean Difference (IV, Random, 95% CI)-0.70 [-2.69, 1.29]
Analysis 2.1.

Comparison 2 Antihypertensive drug dose response, Outcome 1 Systolic blood pressure.

Analysis 2.2.

Comparison 2 Antihypertensive drug dose response, Outcome 2 Diastolic blood pressure.

Appendices

Appendix 1. MEDLINE search strategy

Database: Ovid MEDLINE(R) 1946 to present with daily update
Search date: 4 October 2013
--------------------------------------------------------------------------------
1 exp hypertension/
2 hypertens$.tw.
3 exp blood pressure/
4 ((elevat$ or high$ or rais$) adj3 (diastolic or systolic or arterial or blood) adj pressure).tw.
5 bloodpressure.tw.
6 or/1-5
7 drug therapy/
8 ((drug or pharmaco$ or combination) adj (treatment$ or intervention$ or therap$)).tw.
9 exp antihypertensive agents/
10 (antihypertensive$ or anti hypertensive$).tw.
11 pharmacotherap$.tw.
12 (medicat$ or dose$ or dosing or dosage).tw.
13 or/7-12
14 exp child/
15 (child$ or schoolchild$ or school age$ or preschool$ or pre-school$ or kid or kids or toddler$).mp.
16 exp adolescent/
17 adoles$.tw.
18 (teen$ or boy$ or girl$).mp.
19 exp puberty/
20 (pubert$ or pubescen$ or prepubescen$).mp.
21 exp pediatrics/
22 (pediatric$ or paediatric$ or peadiatric$).tw.
23 or/14-22
24 exp adolescent/ and exp adult/
25 or/14-15,17-22
26 24 not 25
27 23 not 26
28 randomized controlled trial.pt.
29 controlled clinical trial.pt.
30 randomi?ed.ab.
31 placebo.ab.
32 dt.fs.
33 randomly.ab.
34 trial.ab.
35 groups.ab.
36 or/28-35
37 (animal$ not (human$ and animal$)).mp.
38 36 not 37
39 6 and 13 and 27 and 38

Appendix 2. EMBASE search strategy

Database: EMBASE <1974 to 2013 October 03>
Search date: 4 October 2013
--------------------------------------------------------------------------------
1 exp hypertension/
2 hypertens$.tw.
3 exp blood pressure/
4 blood pressure.mp.
5 bloodpressure.tw.
6 or/1-5
7 drug therapy/
8 ((drug or pharmaco$ or combination) adj (treatment$ or intervention$ or therap$)).tw.
9 exp antihypertensive agent/
10 (antihypertensive$ or anti hypertensive$).tw.
11 pharmacotherap$.tw.
12 (medicat$ or dose$ or dosing or dosage).tw.
13 or/7-12
14 exp child/
15 (child$ or schoolchild$ or school age$ or preschool$ or pre-school$ or kid or kids or toddler$).mp.
16 exp adolescent/
17 adoles$.tw.
18 (teen$ or boy$ or girl$).mp.
19 exp puberty/
20 (pubert$ or pubescen$ or prepubescen$).mp.
21 exp pediatrics/
22 (pediatric$ or paediatric$ or peadiatric$).tw.
23 or/14-22
24 exp adolescent/ and exp adult/
25 or/14-15,17-22
26 24 not 25
27 23 not 26
28 randomized controlled trial/ (359812)
29 crossover procedure/
30 double-blind procedure/
31 randomi?ed.tw.
32 (crossover$ or cross-over$).ab.
33 randomly.ab.
34 placebo$.tw.
35 (doubl$ adj blind$).tw.
36 allocat$.ab.
37 comparison.ti.
38 or/28-37
39 (exp animal/ or animal.hw. or nonhuman/) not (exp human/ or human cell/ or (human or humans).ti.)
40 38 not 39
41 6 and 13 and 27 and 40

Appendix 3. Cochrane CENTRAL search strategy

Database: EBM Reviews - Cochrane Central Register of Controlled Trials <October 2013>
Search date: 4 October 2013
--------------------------------------------------------------------------------
1 exp hypertension/
2 hypertens$.tw.
3 exp blood pressure/
4 ((elevat$ or high$ or rais$) adj3 (diastolic or systolic or arterial or blood) adj pressure).tw.
5 bloodpressure.tw.
6 or/1-5
7 drug therapy/
8 ((drug or pharmaco$ or combination) adj (treatment$ or intervention$ or therap$)).tw.
9 exp antihypertensive agents/
10 (antihypertensive$ or anti hypertensive$).tw.
11 pharmacotherap$.tw.
12 (medicat$ or dose$ or dosing or dosage).tw.
13 or/7-12
14 exp child/
15 (child$ or schoolchild$ or school age$ or preschool$ or pre-school$ or kid or kids or toddler$).mp.
16 exp adolescent/
17 adoles$.tw.
18 (teen$ or boy$ or girl$).mp.
19 exp puberty/
20 (pubert$ or pubescen$ or prepubescen$).mp.
21 exp pediatrics/
22 (pediatric$ or paediatric$ or peadiatric$).tw.
23 or/14-22
24 exp adolescent/ and exp adult/
25 or/14-15,17-22
26 24 not 25
27 23 not 26
28 randomized controlled trial.pt.
29 controlled clinical trial.pt.
30 randomi?ed.ab.
31 placebo.ab.
32 dt.fs.
33 randomly.ab.
34 trial.ab.
35 groups.ab.
36 or/28-35
37 (animal$ not (human$ and animal$)).mp.
38 36 not 37
39 6 and 13 and 27 and 38

Appendix 4. Cochrane Hypertension Group Specialised Register search strategy

Database: Cochrane Hypertension Group Specialised Register
Search date: 4 October 2013
--------------------------------------------------------------------------------
1 (adolesc* and elevated blood pressure) [af] not adult [af]
2 (adolesc* and high blood pressure) [af] not adult [af]
3 (adolesc* and hypertens*) [af] not adult [af]
4 (boys and elevated blood pressure) [af] not adult [af]
5 (boys and high blood pressure) [af] not adult [af]
6 (boys and hypertens*) [af] not adult [af]
7 (child* and elevated blood pressure) [af] not adult [af]
8 (child* and high blood pressure) [af] not adult [af]
9 (child* and hypertens*) [af] not adult [af]
10 (girls and elevated blood pressure) [af] not adult [af]
11 (girls and high blood pressure) [af] not adult [af]
12 (girls and hypertens*) [af] not adult [af]
13 (paediatric* and elevated blood pressure) [af] not adult [af]
14 (paediatric* and high blood pressure) [af] not adult [af]
15 (paediatric* and hypertens*) [af] not adult [af]
16 (pediatric* and elevated blood pressure) [af] not adult [af]
17 (pediatric* and high blood pressure) [af] not adult [af]
18 (pediatric* and hypertens*) [af] not adult [af]
19 (preschool or pre-school) and hypertens* [af] not adult [af]
20 (pubert* and elevated blood pressure) [af] not adult [af]
21 (pubert* and high blood pressure) [af] not adult [af]
22 (pubert* and hypertens*) [af] not adult [af]
23 or/1-22

Contributions of authors

Swasti Chaturvedi formulated the idea and developed the basis for the protocol. She took a lead role in searching, identifying and assessing studies, data extraction and analysis, and writing up the final review.

Deborah Lipszyc helped in identifying and assessing studies, double-checking the data extraction and the 'Risk of bias' assessment. She also participated in writing up the final review.

Christoph Licht participated in writing up the review.

Jonathan Craig helped in resolving any disagreement and completion ofthe review.

Rulan Parekh participated in developing the protocol, assessment of studies, data extraction and analysis, and in completing the final manuscript.

Declarations of interest

None known.

Sources of support

Internal sources

  • Hospital for Sick Children, Toronto, Canada.

External sources

  • No sources of support supplied

Differences between protocol and review

Changed protocol from pharmacological interventions for primary hypertension to include pharmacological interventions for both primary and secondary causes of hypertension, given the paucity of trials for primary hypertension.

Deborah Lipszyc was added as an author. She helped in identifying and assessing studies, double-checking the data extraction and the 'Risk of bias' assessment. She also participated in writing up the final review.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Batisky 2007

Methods

4-week, double-blind, multicentre, randomised, placebo-controlled, parallel-group study (period 1) after a 1- to 2-week single blind, placebo run-in period. Period 2 was a 54-week open-label protocol aimed primarily at collecting safety and tolerability information

Sites: 28 centres in the US

Participants

144 patients with sitting systolic blood pressure and/or sitting diastolic blood pressure ≥ 95th percentile adjusted for age, sex and height

Age range: 6 to 16 years

Interventions

4 treatment groups:

Placebo (n = 23)

Extended-release metoprolol succinate 0.2 mg/kg (n = 45)

Extended-release metoprolol succinate 1.0 mg/kg (n = 23)

Extended-release metoprolol succinate 2.0 mg/kg (n = 49)

Outcomes

Primary outcome:

1. Change from baseline to the end of the treatment period trough value (24 ± 2 hours post-dosing) in sitting systolic blood pressure

Secondary outcome:

2. Sitting diastolic blood pressure

3. Standing blood pressures

Antihypertensive response (trough sitting systolic and diastolic blood pressure < 95th percentile)

NotesFunding: AstraZeneca LP
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk"randomised in 1:2:1:2 ratio". Method of randomisation was not described
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk“Double blind”. Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Low risk97% (140/144 patients) included in efficacy analysis. 2 patients were randomised incorrectly and had no baseline data, 2 had no post-baseline BP measurements. 98% (142/144 patients) included in the safety analysis
Selective reporting (reporting bias)Low riskResults for all pre-defined objectives reported
Other biasHigh riskFunding: AstraZeneca LP. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

BMS 2005

Methods

3-week, double-blind, randomised, multicentre, dose response study followed by a placebo-controlled 2-week withdrawal period. The last period was a 6-month, open-label protocol aimed primarily at collecting safety information

Sites: sites in the United States, Eastern and Western Europe and the former Soviet Union

Participants

318 patients with sitting systolic blood pressure and/or sitting diastolic blood pressure ≥ 95th percentile adjusted for age, sex and height

Age range: 6 to 16 years

Interventions

Phase A (n = 318):

Irbesartan low-dose (n ˜103)

Irbesartan medium-dose (n ˜106)

Irbesartan high-dose (n ˜106)

Outcomes

Primary outcome:

1. Change from baseline to the end of the treatment period in sitting systolic blood pressure

Secondary outcomes:

2.Change from baseline to the end of the treatment period in sitting diastolic blood pressure

NotesFunding: Bristol Myers Squibb
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskMethod of randomisation was not described
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Double blind". Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNot described
Selective reporting (reporting bias)Unclear riskLimited information available from the abstract
Other biasHigh riskFunding: Bristol Myers Squibb. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Falkner 1983

Methods

24-week, double-blind, randomised study after a 4- week period with placebo

Country: US

Participants

30 patients with blood pressure > 95th percentile

Age range: 13 to 19 years

Interventions

2 treatment groups :

Clonidine 0.1 mg (n = 15)

Hydrochlorothiazide 25 mg (n = 15)

*After 12 weeks of therapy, if blood pressure < 90th percentile, continued on dose as above. In those who had not achieved treatment goals, the dosage was increased to clonidine 0.2 mg (bid) or hydrochlorothiazide 50 mg (bid)

Outcomes

Primary outcome:

1. Change in systolic blood pressure and diastolic blood pressure

Secondary outcome:

2. Cardiovascular response to mental stress

3. Pre-post stress catecholamines

NotesFunding: not reported
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskMethod of randomisation was not described
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Double blind". Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo report of drop-outs in study
Selective reporting (reporting bias)Low riskResults for all pre-defined objectives reported
Other biasUnclear riskSponsorship or funding and conflict of interest were not declared by the authors in the publication

Flynn 2004

Methods

4-week, double-blind, multicentre study (phase 1) after a 2-week washout period. During phase 2 participants were randomly assigned to continue receiving amlodipine or to be withdrawn to placebo for an additional 4 weeks

Sites: US, Canada, Argentina and Brazil

Participants

268 patients with sitting systolic blood pressure ≥ 95th percentile adjusted for age, sex and height on 3 occasions

Age range: 6 to 16 years

Interventions

Phase 1

2 treatment groups:

Amlodipine 2.5 mg (n = 127)

Amlodipine 5 mg (n = 141)

Phase 2

3 treatment groups:

Placebo (n = 90)

Amlodipine 2.5 mg (n = 84)

Amlodipine  5 mg (n = 94)

Outcomes

Primary outcome:

1. Change in systolic blood pressure at the end of phase 2 compared with the baseline blood pressure at random assignment

Secondary outcomes:

1. Reduction in diastolic blood pressure

2. Magnitude of blood pressure reduction according to dose received (in mg/kg)

3. Relation of blood pressure reduction to Tanner Stage

4. Changes in blood pressure from baseline to end of phase 1 and from end of phase 1 to end of phase 2

Notes

Funding: Pfizer Pharmaceuticals

Only phase 1 data can be used

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk“...were randomly assigned to receive either 2.5 or 5 mg of amlodipine per day”
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Low risk"Double blind". Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk

95% (256/268 patients) included in efficacy analysis

Unclear why 10 patients discontinued participation and to which group they had belonged

Selective reporting (reporting bias)Low riskResults for all pre-defined objectives reported
Other biasHigh riskFunding: Pfizer Pharmaceuticals. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Flynn 2008

Methods

2-week, double-blind, randomised, multicentre, dose response study (phase 1) after a 1-week single-blind placebo screening. During phase 2, participants were randomly reassigned to either continue receiving their phase 1 dose or to be withdrawn to placebo for an additional 14 days. The final phase was a 52-week, open-label treatment phase to achieve systolic blood pressure < 95th percentile

Sites: multiple sites in North America, South America, Asia and Africa

Participants

90 patients with sitting systolic blood pressure ≥ 95th percentile adjusted for age, sex and height

Age range: 1 to 5 years

Interventions

Phase 1

3 treatment groups (n = 90):

Valsartan low-dose (n = 37)

Valsartan medium-dose (n = 18)

Valsartan high-dose (n = 35)

Phase 2

2 treatment groups (n = 87):

Placebo (n = 43)

Valsartan 2.5 mg (n = 44)

Outcomes

Primary outcome:

1.Change in mean sitting systolic blood pressure from baseline to end of phase 1

2.Change in mean sitting systolic blood pressure from end of phase 1 to end of phase 2

Secondary outcomes:

1.Change in mean sitting systolic blood pressure from baseline to end of phase 2

2.Change in mean sitting diastolic blood pressure from baseline to end of phase 1

3.Change in mean sitting diastolic blood pressure from end of phase 1 to end of phase 2

4.Change in mean sitting diastolic blood pressure from baseline to end of phase 2

Notes

Funding: Novartis Pharmaceuticals

Only phase 1 data can be used

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk"randomly assigned to..... groups in a 2:1:2 ratio". Method of randomisation was not reported
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Double blind". Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Low riskA total of 87 participants (97%) completed phase 1, of whom 83 (95%) went on to complete phase 2. Reasons for discontinuation provided in Figure 1
Selective reporting (reporting bias)Low riskResults for all pre-defined objectives reported
Other biasHigh riskFunding: Novartis Pharmaceuticals. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Gartenmann 2003

Methods

16-week, open-label, randomised study after a 6-week washout and run-in period. The final phase was a 1-year, open-label treatment

Sites: 18 sites in the United States/Puerto Rico and 20 sites in Europe

Participants

26 patients with established non-diabetic chronic kidney disease (> 6 months), normal or slightly increased plasma creatinine (< 177 µmol/L), urinary protein concentration > 300 mg/L on early morning urine, untreated (> 6 weeks) arterial hypertension (systolic values 5 to 30 mmHg and diastolic values 1 to 15 mmHg above the 95th centile for age, length and gender), stable (> 12 weeks) immunosuppressive treatment and weight < 20 kg

Age range: 6 to 18 years

Interventions

2 treatment groups:

Amlodipine (n = 13) initial dose 5 mg (> 20, < 40 kg) and 10 mg (≥ 40 kg)

Irbesartan (n = 13) initial dose 75 mg (> 20, < 40 kg) and 150 mg (≥ 40 kg)

Outcomes

1. Blood pressure

2. Heart rate

3. Body weight

4. Blood chemistry including haemoglobin, sodium, potassium, creatinine, uric acid, albumin, aminotransferase and creatine kinase

5. Written questionnaire for the presence of abdominal pain, back pain, constipation, cough, diarrhoea, oedema, fatigue, headache, insomnia, orthostasis and rash

NotesFunding: not reported
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskMethod of randomisation was not described
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
High riskOpen-label
Incomplete outcome data (attrition bias)
All outcomes
Low risk2 patients in the amlodipine group withdrew due to the occurrence of severe oedema and headache
Selective reporting (reporting bias)Low riskResults for all pre-defined objectives reported
Other biasUnclear riskSponsorship or funding and conflict of interest were not declared by the authors in the publication

Hazan 2010a

Methods

3-week, double-blind, randomised, multicentre, dose-ranging study (period 1) after a 2-week screening/washout period. Period 2 was a 2-week, placebo-controlled, double-blind withdrawal period

Patients were enrolled in 1 of 2 cohorts based on race. Cohort A represented all races and cohort B was an all-black patient population. For the purpose of this meta-analysis the 2 cohorts are considered as 2 different studies. Cohort A will be represented in this chart and Cohort B in Hazan 2010b

Sites: 61 sites (not listed)

Participants

190 patients with sitting systolic blood pressure ≥ 95th percentile adjusted for age, sex and height, weighing ≥ 20 kg

Age range: 6 to 16 years

Interventions

Period 1:

2 treatment groups:

Olmesartan medoxomil 2.5 mg (> 20 kg, < 35 kg) or 5.0 mg (≥ 35 kg) (n = 95)

Olmesartan medoxomil 20 mg (> 20 kg, < 35 kg) or 40 mg (≥ 35 kg) (n = 95)

Period 2:

2 treatment groups (n = 182)

Placebo (n = 89)

Olmesartan same dose (n = 93)

Outcomes

Primary outcome:

1. The change from baseline in trough sitting systolic blood pressure and sitting diastolic blood pressure to the end of period 1

Secondary outcomes:

1. To compare olmesartan medoxomil with the placebo for period 2

Notes

Funding: Daiichi Sankyo, Inc.

Only phase 1 data can be used

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskNot described in the paper. Personal communication with Dr Heyrman. The randomising of participants was accomplished centrally using an Interactive Voice Recognition System (IVRS). The randomisation schedule was prepared by ClinPro, Inc. and participants were randomised to treatment sequences
Allocation concealment (selection bias)Low riskNot described in the paper. Personal communication with Dr Heyrman. To maintain the double-blind conditions during period II (cohorts A and B) and period III (all cohorts) of the study, an independent, unblinded study pharmacist compounded each suspension based on the dose required for each participant for that period according to the randomisation scheme. Placebo and olmesartan compounded suspensions were identical in appearance and taste; therefore, neither the participant nor the investigator knew the treatment being administered. In addition, the compounded suspensions were prepared at 2 different concentrations, 0.5 mg/mL and 4 mg/mL, which blinded the investigators and participants to the dosage group (i.e., low-dose olmesartan or high-dose olmesartan) by providing the same volume of liquid for administration regardless of dose. Only the pharmacist in charge of compounding the study medication had access to the randomisation schedule
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-blind. Full details provided by personal communication
Incomplete outcome data (attrition bias)
All outcomes
Low risk

182 patients/190 completed period A. Reasons for drop-out were included

179 patients/182 completed period B. Reasons for drop-out were included

Selective reporting (reporting bias)Low riskResults for all pre-defined objectives reported
Other biasHigh riskFunding: Daiichi Sankyo, Inc. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Hazan 2010b

Methods

3-week, double-blind, randomised, multicentre, parallel-group, dose-ranging study (period 1) after a 2-week screening/washout period. Period 2 was a 2-week, placebo-controlled, double-blind withdrawal period

Patients were enrolled in 1 of 2 cohorts based on race. Cohort A represented all races and cohort B was an all-black patient population. For the purpose of this meta-analysis the 2 cohorts are considered as 2 different studies. Cohort B will be represented in this chart and Cohort A in Hazan 2010a

Sites: 61 sites (not listed)

Participants

112 patients with sitting systolic blood pressure ≥ 95th percentile adjusted for age, sex and height, weighing ≥20 kg

Age range: 6 to 16 years

Interventions

Period 1:

2 treatment groups:

Olmesartan medoxomil low-dose 2.5 mg (> 20 kg, < 35 kg) or 5.0 mg (≥ 35 kg) (n = 56)

Olmesartan medoxomil high-dose 20 mg (> 20 kg, < 35 kg) or 40 mg (≥ 35 kg) (n = 56)

Period 2:

2 treatment groups:

Placebo (n = 54)

Olmesartan same dose (n = 53)

Outcomes

Primary outcome:

1. The change from baseline in trough sitting systolic and sitting diastolic blood pressure to the end of period 1

Secondary outcomes:

1. To compare olmesartan medoxomil with the placebo for period 2

Notes

Funding: Daiichi Sankyo, Inc.

Only phase 1 data can be used

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskNot described in the paper. Personal communication with Dr Heyrman. The randomising of participants was accomplished centrally using an Interactive Voice Recognition System (IVRS). The randomisation schedule was prepared by ClinPro, Inc. and participants were randomised to treatment sequences
Allocation concealment (selection bias)Low riskNot described in the paper. Personal communication with Dr Heyrman. To maintain the double-blind conditions during period II (cohorts A and B) and period III (all cohorts) of the study, an independent, unblinded study pharmacist compounded each suspension based on the dose required for each participant for that period according to the randomisation scheme. Placebo and olmesartan compounded suspensions were identical in appearance and taste; therefore, neither the participant nor the investigator knew the treatment being administered. In addition, the compounded suspensions were prepared at 2 different concentrations, 0.5 mg/mL and 4 mg/mL, which blinded the investigators and participants to the dosage group (i.e., low-dose olmesartan or high-dose olmesartan) by providing the same volume of liquid for administration regardless of dose. Only the pharmacist in charge of compounding study medication had access to the randomisation schedule
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-blind. Full details provided above by personal communication
Incomplete outcome data (attrition bias)
All outcomes
Low risk

107 patients/112 completed period A. Reasons for drop-out were included

104 patients/107 completed period B. Reasons for drop-out were included

Selective reporting (reporting bias)Low riskResults for all pre-defined objectives reported
Other biasHigh riskFunding: Daiichi Sankyo, Inc. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Li 2004

Methods

4-week, double-blind, randomised, multicentre, dose-ranging study (period B) after a 10-day screening period followed by a fosinopril test dose of 0.1 mg/kg (period A). Period C was a 2-week placebo withdrawal phase. The final phase was a 52-week open-label safety phase (period D)

Sites: 45 sites in the United States, 4 in Israel and 11 in Russia

Participants253 patients with either hypertension (3 sequential readings of sitting systolic blood pressure and/or sitting diastolic blood pressure > 95th percentile adjusted for age, sex and height) or high-normal blood pressure (systolic or diastolic blood pressure > 90th percentile but < 95th percentile) with an associated clinical condition
Age range: 6 to 16 years
Interventions

Period B

3 treatment groups:

Fosinopril 0.1 mg/kg (n = 83)

Fosinopril 0.3 mg/kg (n = 87)

Fosinopril 0.6 mg/kg (n = 83)

Period C

2 treatment groups:

Fosinopril any dose (n = 106)

Placebo (n = 115)

Outcomes

Primary outcome:

1. Change in trough systolic blood pressure from baseline in patients treated with low-, medium- and high-dose fosinopril in period B

Secondary outcomes:

1. Change in trough diastolic blood pressure from baseline in patients treated with low-, medium- and high-dose fosinopril after period B

2. Change in trough systolic blood pressure or diastolic blood pressure from the end of period B to the end of period C in fosinopril-treated patients (all 3 levels combined) compared with placebo-treated patients

3. Per cent of patients with both systolic blood pressure and diastolic blood pressure < 90th percentile at the end of period B

4. Safety evaluation of adverse events and laboratory abnormalities

NotesFunding: Bristol-Myers Squibb. Only period B data can be used
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskNot described in article. Personal communication with Dr. Jennifer Li: randomisation was performed centrally through an integrated voice response system using a commercial vendor. Study sites had access to the system 24 hours a day and 7 days per week with technical assistance
Allocation concealment (selection bias)Low riskNot described in article. Personal communication with Dr Jennifer Li: allocation concealment was assured through central randomisation to a particular drug kit (placebo versus study drug). There were matching placebos to the study drug
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-blind. Full details provided through personal communication
Incomplete outcome data (attrition bias)
All outcomes
Low risk

238 patients/253 completed period B. Reasons for drop-out were included

221 patients/235 completed period C. Reasons for drop-out were included

Selective reporting (reporting bias)Low riskResults for all pre-defined objectives reported
Other biasHigh riskFunding: Bristol-Myers Squibb. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Li 2010

Methods

6-week, double-blind, randomised, multicentre, parallel-group, dose-ranging study (phase A) after a screening/washout period of unspecified duration. Phase B was a 4-week, placebo-controlled, double-blind withdrawal period

Sites: 43 centres in the US, India, South Africa, Russia and the Dominican Republic

Participants

304 patients with sitting systolic blood pressure ≥ 95th percentile adjusted for age, sex and height

Age range: 4 to 16 years

Interventions

Phase A (n = 304):

Eplerenone low-dose (n = 58)

Eplerenone medium-dose (n = 62)

Eplerenone high-dose (n = 184)

 

Phase B (n = 277):

Eplerenone low-dose (n = 26)

Eplerenone medium-dose (n = 28)

Eplerenone high-dose (n = 86)

Placebo (n = 137)

Outcomes

Primary outcome:

1. Change from baseline of phase B to the end of the treatment period in sitting systolic blood pressure

Secondary outcomes:

1. Change from baseline of phase B to the end of the treatment period in sitting diastolic blood pressure

2. Evaluate the effect of eplerenone on SBP and DBP as a function of dose and body size

3. Safety evaluation

Notes

Funding: Pfizer, Inc.

Only phase A data can be used

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk"randomised' method was not described
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Double blind". Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNot all reasons for drop-out were included
Selective reporting (reporting bias)Low riskAll pre-determined outcomes were reported
Other biasHigh riskFunding: Pfizer, Inc. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Schaefer 2010

Methods

4-week, double-blind, randomised, multicentre, dose-ranging study after a 1-week, single-blind placebo period. The final phase was a 1-year open-label treatment

Sites: 18 sites in the United States/Puerto Rico and 20 sites in Europe

Participants

93 patients with sitting systolic blood pressure and/or sitting diastolic blood pressure > 95th percentile adjusted for age, sex and height

Age range: 1 to 6 years

Interventions

3 treatment groups:

Candesartan 0.05 mg/kg (n = 29)

Candesartan 0.20 mg/kg (n = 32)

Candesartan 0.4 mg/kg (n = 32)

Outcomes

Primary outcome:

1. Change in sitting systolic blood pressure from baseline

Secondary outcomes:

1. Change in sitting diastolic blood pressure from baseline

2. Median percentage change in urine albumin/creatinine ratio

NotesFunding: AstraZeneca LP
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskRandomised. No further description provided
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Double blind". Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Low riskAll 93 patients were included in the intention-to-treat analysis
Selective reporting (reporting bias)Low riskResults for all pre-defined objectives reported
Other biasHigh riskFunding: AstraZeneca LP. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Schaefer 2011

Methods

12-week, double-blind, multicentre, randomised, parallel-group, active-controlled study after a single-blind, placebo run-in period (4 to 28 days)

Country: 60 study centres worldwide

Participants

300 patients with sitting systolic blood pressure ≥ 95th percentile adjusted for age, sex and height

Age range: 6 to 17 years

Interventions

2 treatment groups:

Valsartan (n = 151) 40/80/160 mg (week 1) to 80/160/320 mg (weeks 2 to 12)

Enalapril (n = 149) 5/10/20 mg (week 1) to 10/20/40 mg (weeks 2 to 12)

Outcomes

Primary outcome:

1. Change in sitting systolic blood pressure from baseline

Secondary outcomes:

1.Change in sitting diastolic blood pressure from baseline

2. Percentage of patients achieving systolic blood pressure control

NotesFunding: Novartis Pharma AG, Basel
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk“…randomised study” method was not described
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk“Double blind”. Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Low risk281/300 patients completed the study. Reasons for drop-out were included
Selective reporting (reporting bias)Low riskResults for all pre-defined objectives were reported
Other biasHigh riskFunding: Novartis Pharma AG, Basel. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Shahinfar 2005

Methods

3-week, double-blind, multicentre, randomised study (phase 1) after a 2 to 7-day washout period. During phase 2 participants were randomly assigned to continue receiving losartan or to be withdrawn to placebo for additional 2 weeks

Sites: 14 US sites and 29 international sites in Africa, Europe and South America

Participants

177 patients with sitting diastolic blood pressure > 95th percentile adjusted for age, sex and height

Age range: 6 to 16 years

Interventions

Phase 1

3 treatment groups stratified by weight (< 50 kg and ≥ 50 kg):

Losartan 2.5 or 5 mg (n = 70)

Losartan 25 or 50 mg (n = 40)

Losartan 50 or 100 mg (n = 65)

Phase 2

4 treatment groups:

Placebo (n = 87)

Losartan  2.5 or 5 mg (n = 33)

Losartan 25 or 50 mg (n = 15)

Losartan 50or 100 mg (n = 29)

*Only data from phase 1 included in the review

Outcomes

Primary outcome:

1.To demonstrate a positive dose response effect of 3 doses of losartan

Secondary outcome:

1. Mean change in diastolic blood pressure from baseline

Notes

Funding: Merck Pharmaceuticals

Only phase A data can be used

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk“...randomised study”. No further description provided
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Double blind". Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Low risk174/175 patients included in efficacy analysis. 1 patient was lost to follow-up
Selective reporting (reporting bias)Low riskResults for all pre-defined objectives reported
Other biasHigh riskFunding: Merck Pharmaceuticals. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Soffer 2003

Methods

2-week, double-blind, multicentre study after a 1-week washout period. During phase 2 participants were randomly assigned to continue receiving lisinopril or to be withdrawn to placebo for an additional 2 weeks

Sites: multiple centres in North America, South America and Europe

Participants

115 patients with sitting diastolic blood pressure > 95th percentile adjusted for age, sex and height on 2 confirmatory measurements

Age range: 6 to 16 years

Interventions

Phase 1

3 treatment groups stratified by weight (< 50 kg and > 50 kg):

Lisinopril 0.625 or 1.25 mg (n = 33)

Lisinopril 2.5 or 5 mg (n = 24)

Lisinopril 20 or 40 mg (n = 58)

Phase 2

4 treatment groups:

Placebo (n = 51)

Lisinopril  2.5 or 5 mg (n = 15)

Lisinopril 25 or 50 mg (n = 11)

Lisinopril 50 or 100 mg (n = 27)

*Only data from phase 1 included in the review

Outcomes

Primary outcome:

1. To demonstrate a dose-response relationship for sitting diastolic blood pressure

Secondary outcome:

1. The blood pressure P difference between the lisinopril and the placebo-treated arm in phase 2

Notes

Funding: Merck and Zeneca Pharmaceuticals

Only phase 1 data can be used

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk“Randomised study.” Method of randomisation was not described
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Double blind". Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIn period 2 a total of 11 patients who discontinued in period 1 were excluded from the analysis". Unclear why these 11 patients discontinued the study and which group they belonged to
Selective reporting (reporting bias)Low riskResults for all pre-defined objectives reported
Other biasHigh riskFunding: Merck and Zeneca Pharmaceuticals. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Sorof 2002

Methods

10-week, double-blind, multicentre study after a 2-week washout period. This was followed by a 2-week dose-tapering phase

Sites: 22 centres in USA and Brazil

Participants

94 patients with sitting systolic and/or diastolic blood pressure > 95th percentile adjusted for age, sex and height

Age range: 6 to 17 years

Interventions

2 treatment groups:

Bisoprolol fumarate/hydrochlorothiazide (n = 62)

Placebo (n = 32)

Outcomes

Primary outcome:

Comparisons between treatment group of absolute and percentage reduction of blood pressure from baseline

NotesFunding: Wyeth-Ayerst and Procter and Gamble Pharmaceuticals
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskRandomised study. Method of randomisation was not described
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Double blind". Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
High risk94/140 included in full analysis
Selective reporting (reporting bias)Low riskResults for all pre-defined objectives reported
Other biasHigh risk

Funding: Wyeth-Ayerst and Procter and Gamble Pharmaceuticals

Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Trachtman 2003

Methods

3-week, double-blind, multicentre, randomised, placebo-controlled study after a 1- to 3-week screening period to confirm reproducible high blood pressure levels, and washout of antihypertensive medications. Phase 2 was an optional 14-week open-label extension phase designed primarily to collect safety information

Country: US

Participants

133 patients with sitting systolic blood pressure and/or sitting diastolic blood pressure > 95th percentile adjusted for age, sex and height

Age range: 6 to 16 years

Interventions

4 treatment groups (od):

Placebo (n = 35)

Extended-release felodipine 2.5 mg (n = 33)

Extended-release felodipine 5 mg (n = 34)

Extended-release felodipine 10 mg (n = 31)

Outcomes

Primary outcome:

1. Change in trough sitting diastolic blood pressure from baseline to the end of 3-week, double-blind treatment phase

Secondary outcome:

2. Changes in trough sitting systolic blood pressure at week 3

3. Changes in trough standing and supine blood pressures

4. Proportion of responders (i.e. patients whose trough sitting systolic blood pressure and diastolic blood pressure fell < 90th percentile for age, sex and height) after the 3-week treatment period

NotesFunding: unclear
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk"an interactive voice response system was utilized....." (p. 549)
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Double blind". Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Low risk96.2% (128/133 patients) included in efficacy analysis. Reasons for discontinuation described on p.1136
Selective reporting (reporting bias)Low riskResults for all pre-defined objectives reported
Other biasUnclear riskSponsorship or funding and conflict of interest were not declared by the authors in the publication

Trachtman 2008

Methods

4-week, double-blind, multicentre study after a 1- week washout period. This was followed by a 1-year, open-label trial

Sites: multiple centres in the US and Europe

Participants

240 patients with sitting systolic blood pressure ≥ 95th percentile adjusted for age, sex and height

Age range: 6 to 17 years

Interventions

4 treatment groups stratified by weight (< 50 kg and ≥ 50 kg):

Candesartan 2 or 4 mg (n = 69)

Candesartan 8 or 16 mg (n = 68)

Candesartan 16 or 32 mg (n = 68)

Placebo (n = 35)

Outcomes

Primary outcome:

1. Placebo-corrected change in sitting systolic blood pressure

Secondary outcome:

1. Placebo-corrected change in sitting diastolic blood pressure

2. Pair-wise comparisons in ANCOVA models with baseline blood pressure as a covariate and treatment group as a factor

NotesFunding: AstraZeneca Pharmaceuticals
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskPersonal communication. "The study statistician specifies the randomisation instructions to a separate AstraZeneca randomisation group. They generate the randomisation code for the study drug packaging unit and for the central randomisation office. As study sites identified eligible patients they contacted the central randomisation office where an interactive voice response system assigned the study drug package number".
Allocation concealment (selection bias)Low riskPersonal communication. The active study drug and placebo were identical in appearance and the blinding to treatment/dose level was within weight panel. A double-dummy packaging scheme was employed so that the study drug packages were indistinguishable across doses
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-blind. Full details are provided above by personal communication
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk

 "The study investigators randomised a total of 240 patients, all of whom were included in the intent to treat analysis group. For 9 patients, a week 4 blood pressure value was imputed with the last available observation. Eleven patients discontinued treatment in the dose-ranging study: 3 for study eligibility violations, 3 for adverse events, and 5 for other reasons."

Unclear to which group these patients belonged

Selective reporting (reporting bias)Low riskResults for all pre-defined outcomes available (reported and personal correspondence)
Other biasHigh riskFunding: AstraZeneca Pharmaceuticals. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Webb 2010

Methods

12-week, double-blind, multicentre, randomised, parallel-group study after a 4-week washout period during which all patients received amlodipine suspension

Sites: multiple centres in North America, South America, Europe and Asia

Participants

306 patients with proteinuria (urine protein-creatinine ratio ≥ 0.3g/g) in 3 first-morning spot urine collections at baseline, with or without hypertension

Hypertensive stratum: 1) systolic or diastolic blood pressure > 90th percentile in patients on ACE-I/ARBs 2) systolic or diastolic blood pressure > 95th percentile in patients not receiving an ACE-I/ARB for proteinuria or 3) patients with documented hypertension on any antihypertensive medication

Age range: 6 to 17 years

Interventions

Hypertensive stratum

2 treatment groups:

Amlodipine 0.05 to 2 mg/kg/day (n = 30)

Losartan (n = 30)

Normotensive stratum

2 treatment groups:

Placebo (n = 124)

Losartan 1.4 mg/kg (n = 122)

*Only data from hypertensive stratum are included in the review

Outcomes

Primary outcome:

1. The change in urine protein-creatinine ratio

Secondary outcome:

1. Drop in blood pressure

NotesFunding: Merck Pharmaceuticals
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskPatients were assigned an allocation number according to a computer-generated randomised allocation schedule
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Double blind". Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk98.6% (302/306) included in full analysis. Unclear whether the 4 participants excluded from analysis belong to the hypertensive or normotensive stratum. In the hypertensive stratum 10% (6/60) participants dropped out, 1 from the losartan group due to adverse reaction and 5 from the amlodipine group (2 adverse reaction, 3 other). No further description of adverse reaction or "other"
Selective reporting (reporting bias)Low riskResults for all pre-defined outcomes reported
Other biasHigh riskFunding: Merck Pharmaceuticals. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Webb 2013

Methods12-week, open-label, randomised, parallel-group study. Patients were randomised to 1 of 3 losartan dose groups
Participants

101 patients

Age range: 6 months to 6 years

Interventions

3 treatment groups:

Losartan 0.1 mg/kg/day (n = 33)

Losartan 0.3 mg/kg/day (n = 34)

Losartan 0.7 mg/kg/day (n = 32)

Outcomes

Primary outcome:

1. Placebo-corrected change in sitting systolic blood pressure after 3 weeks

Secondary outcome:

1. Placebo-corrected change in sitting diastolic blood pressure after 3 weeks

NotesMerck Pharmaceuticals
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskMethod of randomisation was not described
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
High riskOpen-label
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskData available only for 99 children; reason for drop-out of 2 children unclear
Selective reporting (reporting bias)Unclear riskStudy available in abstract form only - unable to judge
Other biasUnclear riskFunding: Merck Pharmaceuticals. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Wells 2002

Methods

2-week, double-blind, multicentre, randomised study after a 1-week washout period. During phase 2 participants were randomly assigned to continue receiving enalapril or to be withdrawn to placebo for an additional 2 weeks

Sites: multiple centres in North America and South America

Participants

110 patients with sitting diastolic blood pressure ≥ 95th percentile adjusted for age, sex and height

Age range: 6 to 16 years

Interventions

Phase 1

3 treatment groups stratified by weight (< 50 kg and ≥ 50 kg):

Enalapril 0.625 or 1.25 mg (n = 30)

Enalapril 2.5 or 5 mg (n = 29)

Enalapril 20 or 40 mg (n = 50)

Phase 2

4 treatment groups:

Placebo (n = 50)

Enalapril  2.5 or 5 mg (n = 12)

Enalapril 25 or 50 mg (n = 14)

Enalapril 50 or 100 mg (n = 25)

*Only data from phase 1 included in the review

Outcomes

Primary outcome:

1. To demonstrate a dose-response relationship for sitting diastolic blood pressure

Secondary outcome:

2. The blood pressure difference between the enalapril and the placebo-treated arm in phase 2

Notes

Funding: Merck Pharmaceuticals

Only phase 1 data can be used

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk"Randomised study". Method of randomisation was not described
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Double blind". Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk99% (109/110) patients included in intention-to-treat analysis. 5 patients excluded from per-protocol analysis - unclear to which group they belonged
Selective reporting (reporting bias)Low riskResults for all pre-defined outcomes reported
Other biasUnclear riskFunding: Merck Pharmaceuticals. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Wells 2010

Methods

4-week, double-blind, multicentre, randomised, parallel-group, placebo-controlled study after a 2-week washout period

Country: 16 centres in the US, Brazil and Mexico

Participants

77 patients with sitting systolic blood pressure ≥ 95th percentile adjusted for age, sex and height

Age range: 6 to 17 years

Interventions

3 treatment groups:

Telmisartan 'low-dose' group: 1 mg/kg (n = 30)

Max dose (ages 6 to 12 years) = 60 mg

Max dose (ages 12 to 18 years) = 120 mg

Telmisartan 'high-dose' group: 2 mg/kg (n = 31)

Max dose (ages 6 to 12 years) = 120 mg

Max dose (ages 12 to 18 years) = 160 mg

Placebo (n = 16)

Outcomes

Primary outcome:

1. Change in sitting systolic blood pressure from baseline

Secondary outcomes:

Change in sitting diastolic blood pressure from baseline

Notes"The author(s) received no financial support for the research and/or authorship of this article"
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"interactive voice randomization system”
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Double blind". Used but not described
Incomplete outcome data (attrition bias)
All outcomes
High risk115 participants signed consent, only 77 randomised (unclear reason); 64/77 patients completed the study. Reasons for drop-out were included
Selective reporting (reporting bias)Low riskResults for all pre-defined outcomes reported
Other biasLow riskNone identified

Wells 2011

  1. a

    ACE: angiotensin-converting enzyme
    ARB: angiotensin receptor blocker
    bid: twice a day
    BP: blood pressure
    DBP: diastolic blood pressure
    od: once a day
    SBP: systolic blood pressure

Methods

2-week, double-blind, randomised, multicentre, dose-ranging study (phase 1) after a screening/washout period lasting up to 7 days. Phase 2 was a 2-week randomised, placebo-controlled, double-blind withdrawal period. The final phase was a 52-week open-label extension phase to assess long-term safety

Sites: 55 centres in 9 countries

Participants

261 patients with sitting systolic blood pressure ≥ 95th percentile adjusted for age, sex and height

Age range: 6 to 16 years

Interventions

Phase 1 (n = 261):

Valsartan low-dose (n = 103)

Valsartan medium-dose (n = 53)

Valsartan high-dose (n = 105) 

Phase 2 (n = 245):

Valsartan  (n = 123)

Placebo (n = 122)

Outcomes

Primary outcome:

1. Slope of dose response curve for change from baseline in sitting systolic blood pressure in phase 1

2. Change in sitting systolic blood pressure from end of phase 1 to the end of phase 2

Secondary outcomes:

1. Mean sitting DBP

Notes

Funding: Novartis Pharmaceuticals

Only phase 1 data can be used

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk"Randomised". Method not described
Allocation concealment (selection bias)Unclear riskAllocation concealment was not described
Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Double blind". Full details not provided
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskTotal of 261 children randomised. 2 patients did not take study medications. 16 patients (6.1%) did not complete phase 1: 3 patients had protocol violation, 4 patients withdrew consent, 2 patients due to adverse reaction, 2 patients because of administrative problems, 4 due to perceived lack of response and 1 was lost to follow-up. Unclear to which group these 16 patients belonged. Overall 259 children included in intention-to-treat analysis
Selective reporting (reporting bias)Low riskResults for all pre-defined outcomes reported
Other biasHigh riskFunding: Novartis Pharmaceuticals. Industry-sponsored trials are shown to bias towards overestimate of effect size compared to non-industry-sponsored trials

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
  1. a

    RCT: randomised controlled trial

Bachmann 1984Not a RCT
Berenson 1990Not a RCT
Chan 2012Did not meet the inclusion criteria
Cichocka 1993The number of children in each group and dose of drug were not reported so unable to assess medication efficacy
Ellis 2004Not a RCT
Feig 2008Experimental nature of the therapy
Friedman 1980Not a RCT
Meyers 2011Post-hoc analysis of Wells 2011 study
Moncica 1995Study participants did not meet the review inclusion criteria
Morsi 1992Follow-up less than 2 weeks
Rogan 2000Study participants did not meet review inclusion criteria
Sinaiko 1977Not a RCT
Sinaiko 1983Only 6 patients met the review inclusion criteria
White 2003This double-blind, randomised, cross-over trial compared the anti-proteinuric effects of enalapril and losartan in 6 children with proteinuria and renal injury. Only 2 patients had hypertension. Results are not reported separately
Wühl 2009Did not meet the review inclusion criteria.