Should the Current Approach to the Evaluation and Treatment of High Blood Pressure in Children Be Changed?

In this issue of The Journal, Collins and Alpert propose changes to the current approach to elevated blood pressure (BP) in pediatric patients. In their opinion: 1) the diagnosis of all forms of childhood hypertension, including pre-hypertension and white coat hypertension (WCH), need to be given more attention; 2) therapy, in particular lifestyle interventions, should be initiated sooner rather than later; and 3) greater research efforts are needed to determine physiologically appropriate cutoff levels for diagnosis on the basis of long-term outcomes and cardiovascular (CV) risk.1 The first opinion is in agreement with pediatric sub-specialists and clinical investigators who have a contributed substantially to the current state of knowledge on childhood hypertension. Significant new data have become available in particular for WCH, and a more concerted approach to this condition does seem warranted. The second opinion is also in agreement with expert opinion plus that of primary care clinicians and those in public health, although better methods to achieve lifestyle changes in the young are clearly needed. The third opinion may be the most interesting and raises the question of whether we now have enough data to abandon the 95th percentile as the cutoff point to diagnose elevated BP in pediatric patients.

Before the First Task Force report in 1977,2 little was known about pediatric hypertension. BP was not routinely measured in children, there were no consistent measurement methods, and pediatric BP reference values were limited. In the 1960s, Londe measured BP in healthy children and noted an increase in BP relative to age, growth, and development.3 He further analyzed the BP data to determine the range of systolic and diastolic BP stratified by age and, with a statistical approach described by Master et al,4 selected the 90th percentile to define elevated BP. Predictably, this led to a described prevalence of elevated BP (on the basis of a single determination) slightly above 10%. He also noted regression to the mean with repeated measurement and that the proportion of children with BP consistently >95th percentile was 1.9%,3 a figure remarkably similar to more contemporary data that encompass far larger numbers of children.

With the exception of extreme BP elevations, hypertension in childhood is generally asymptomatic and does not result in the same morbidity seen in adults with hypertension. Furthermore, there are no data available that directly link childhood BP with “hard” CV end points (stroke, heart failure, renal failure, death) in adulthood. Therefore, in the absence of mortality or morbidity end points, the statistical method was chosen to define childhood hypertension and continues to be followed.5 Collins and Alpert challenge this approach and propose development of “physiologically appropriate cutoff levels for diagnosis.”1 However, the fundamental question that remains unanswered is, at what BP level does the pathology begin? And, is it an absolute BP level or a BP percentile? If hypertension develops in 40% to 50% of adults eventually, why are we using only the top 5% of the pediatric BP distribution to define hypertension?

Part of the problem in attempting to link childhood data with adult outcomes has been the reliability of early data on BP in childhood. As described by Din-Dietzman et al,6 the very early childhood BP distribution curves (before the 1987 Second Task Force Report) are markedly higher than our current normative data, so that the 95th percentile BP level today is much lower than these early reference data. However, there has been little change in the pediatric BP distribution curves since those reported in the1987 Task Force Report,7 suggesting that the currently available normative data are quite reliable. Although it is unlikely that pediatric BP data obtained on children in the 1980s could be linked with subsequent adult events, it is possible that some longitudinal data may be available on intermediate markers of heightened CV risk, such as cardiac hypertrophy or carotid intimal-medial thickness.

It is now recognized, in adults, that hypertension is not simply a category limited to BP >140/90 mm Hg. More recent reports of longitudinal epidemiologic data in adults demonstrate that the BP risk for hypertension-related events begins to rise at a point well below the adult cutoff point of 140/90 mm Hg.8 The risk with increasing BP is continuous and begins to rise at approximately 120/80 mm Hg. This evidence led to the designation of a “pre-hypertension” category in adults having BP between 120/80 and 139/89 mm Hg. It is also recognized that BP levels <140/90 mm Hg confer heightened risk in the presence of other conditions such as diabetes mellitus, dyslipidemia, and metabolic syndrome. This shift in the approach to the management of hypertension incorporates assessment of multiple risk factors within individual patients and preventive interventions when risk is elevated.

The Fourth Report extended this concept to pediatrics, with heightened emphasis on primary prevention and attention to associated risk factors for children with BP >90th percentile or >120/80 mm Hg in adolescents.5 Therefore, the proportion of children considered to be at high risk for hypertension is increased to the top 10% or more in adolescence, when 120/80 is <90th percentile. Subsequently, new data, both cross-sectional and longitudinal, have been published on the association of childhood BP with intermediate markers of possible pathology, as mentioned by Collins and Alpert. With these emerging data, it is timely to re-examine the evidence to determine whether the definitions of hypertension are adequate to capture children with high-risk BP and especially to identify children who already have BP-related pathology.

Another question raised by Collins and Alpert is whether 3 elevated BP measurements are necessary to establish a diagnosis of hypertension in childhood. Why not use the average of 2 measurements, as in adults,8 rather than 3 measurements? The Fourth Report, and the earlier reports, are guidelines and not protocol. Clearly, when a child has stage 2 hypertension or has symptoms, it would not be appropriate to schedule repeat visits before initiating evaluation and treatment. Repeat measurements are recommended to avoid over-diagnosis and premature pharmacologic treatment. However, McNiece et al9 have demonstrated that in adolescents with initial BP measurements >95th percentile, some will remain with >95th percentile measurements, and many will regress to an average BP <95th percentile; however, those who regress usually remain in the pre-hypertensive range. Thus, adolescents with elevated BP on initial BP measurement tend to remain in a high-risk BP stratum, and opinion 2 applies: “lifestyle interventions should be initiated sooner rather than later.”

A similar case can be made for WCH. Recent studies have indicated that this condition may be present in a substantial number of children who undergo ambulatory BP monitoring for further evaluation of elevated office BP. What should be done once the diagnosis of WCH is made? The Fourth Report is silent on this point, but it should be noted that there are no pediatric consensus recommendations that state that children with WCH can be ignored. Indeed, data cited by Collins and Alpert from a number of recent cross-sectional studies support the concept that WCH likely represents a pre-hypertensive state, with early changes in target organs such as the heart and blood vessels detectable with careful study. Particularly noteworthy is the study of Lande et al, who demonstrated that although children with WCH did not have left ventricular hypertrophy per se, their left ventricular mass was elevated compared with that in control subjects matched for age, sex, and body mass index.10 Although further study is needed to better define the proper diagnostic thresholds for WCH and to identify the short- and long-term cardiovascular risks associated with this condition, these intermediate outcome data support the concept that early therapy in the form of lifestyle intervention is warranted in children with WCH. Repeat ambulatory BP monitoring after a period of follow-up to detect changes in BP pattern with time may also be appropriate in such children, but this needs to be established by proper longitudinal studies.

So, is it time to change the approach to high BP in children? Yes, and no. Collins and Alpert are correct in pointing out that there is a growing body of data published since the Fourth Report that supports the concept that elevated BP in childhood is not benign and is likely to have significant adverse consequences. It would be proper to systematically analyze these new data to determine whether they are sufficient to establish new diagnostic thresholds for pediatric hypertension. Perhaps it is time to convene a fifth Working Group on high BP in children and adolescents to accomplish this task. In the meantime, however, practitioners should continue to follow established consensus recommendations to carefully evaluate each child or adolescent with elevated BP in the context of their overall CV risk and to institute appropriate treatment—lifestyle modifications or drug therapy—on the basis of the outcome of such assessment.