Pre-hypertension and Hypertension in Pediatrics: Don't Let the Statistics Hide the Pathology

Blood Pressure Measurement 

Error and variability in BP measurements are well known. When there is variance in the accuracy of a measurement method, such as that seen in BP measurements,14 variability in the measured value occurs with multiple measurements—some measurements will be elevated and some will be normal or even decreased. The mean value of these multiple BP measurements in a given patient will most closely approximate the ambulatory BP, a finding demonstrated by Pearce et al.15 This is known as regression to the mean and is a phenomenon that was described in the Fourth Report on the Diagnosis, Evaluation, and Treatment of High Blood Pressure in Children and Adolescents (Fourth Report) to justify the requirement that, to identify a child as having hypertension, BP measurements in the hypertensive range must be obtained on ≥3 occasions.16 Investigators using ambulatory BP monitoring (ABPM) have demonstrated that BP measured during the first 2 hours of a 48-hour measurement are higher than those measured 24 hours later,17 a finding not demonstrated with measurements made on non-consecutive days. This has been referred to as the accommodation effect and is the basis for recommending BP measurements on >1 occasion.16

Sustained Hypertension 

As stated in the Fourth Report, “the long-term health risks for hypertensive children and adolescents can be substantial; therefore, it is important that clinical measures be taken to reduce these risks and optimize health outcomes.”16 However, the exact level and duration of BP elevation that causes target organ damage in the young has not been established.16 In the Fourth Report, HTN was defined as an average SBP, diastolic BP, or both that is ≥95th percentile for sex, age, and height on ≥3 occasions. Unfortunately, these guidelines lack significant scientific validity; they are based on percentile cutoffs from normative distributions of BP in children presumed healthy,16 as opposed to data demonstrating association with long-term CV risk. The definition of HTN in pediatric patients has been “based not upon risk data but upon clinical experience and consensus.”18 We contend that this is a major deficiency in the current approach to the diagnosis and management of HTN in pediatrics. In a disease with a very far-reaching impact, we believe that a more evidence-based approach to the diagnosis, such as that used by the adult scientific community, should be developed.

Current data estimate that the prevalence of HTN in children and adolescents is 1.0% to 4.5%.19, 20, 21, 22 In adults aged 20 to 34 years, HTN has been described to be present in 11.4% of men and 6.4% of women.2 If we trust that the prevalence of HTN in the late teen years is similar to that stated throughout childhood, as demonstrated by McNiece et al,22 it makes little sense that the incidence of HTN would suddenly increase 3- to 5-fold in early adulthood. The lack of consistency between the pediatric and adult prevalences must be related in part to the difference in the definition of HTN in adults versus in pediatrics; adult HTN is defined by 1 measurement cutoff point (>140/90), as opposed to the statistically derived measurements (95% for age, height, and sex) used to define pediatric HTN. However, another possible cause of the seeming differences in the prevalence of HTN from childhood to adulthood has to do with BP measurement. If only 2 BP measurements had been used for the McNiece cohort instead of 3, the incidence of HTN would have been 9%. This assumes no sex differences in the prevalence of HTN. In adult studies, consistent sex differences have been found. However, the 9% composite prevalence would fall between the 4.5% prevalence documented for children by Sorof et al and the 11.4% found in men,2 an expected result on the basis of the age-related trends of HTN prevalence. To our knowledge, no data exist that demonstrate that 2 hypertensive BP measurements are less accurate in diagnosing HTN in pediatric patients than are 3 measurements. Perhaps the use of a minimum of 3 elevated BP readings, as opposed to a minimum of 2 recommended in adults,23 allows patients that “truly” have hypertension to escape detection, a practice that is not ideal in a disease with such significant long-term consequences. Further, there is a need for studies evaluating BP that include both adolescents and adults as a means to bridge the childhood to adulthood transition.

Ethnic differences in the prevalence of HTN highlight the weakness of using normative distribution data for defining a disease. It is well known that in adults there are significant ethnic differences in the prevalence of HTN (it is more prevalent in the African American population than in the Caucasian population, for example),2, 22 and associated CVDs.2 Similar differences in arterial stiffness, a marker of CVD, have been demonstrated in ethnic groups in normotensive adolescent and adult populations.6, 24 The normative data on which the diagnosis of HTN in pediatrics is based do not contain information on ethnic distributions. With the increasing numbers of ethnic minorities in the United States25 and differences in CV risk by ethnic group, it makes sense that, if normative data are to be used to define HTN, ethnicity-specific normative distributions should be developed. However, if a Gaussian distribution were used in such groups with a higher prevalence of HTN, as would be the case if ethnicity-specific norms were devised, the diagnosis of HTN would be made at higher levels in those minority groups. Such a diagnostic approach would serve to decrease the percentage of patients who would receive lifestyle or pharmacologic interventions, and thereby increase the long-term CV risk in these groups, clearly an undesirable outcome. Because those ethnic groups have worse outcomes, a more aggressive diagnostic and therapeutic approach may need to be used to reduce long-term CV morbidity and mortality. This again clearly highlights the significant weakness in the use of normative distributions for the determination of HTN and highlights that large cohort outcomes, such as those performed by the investigators of the Bogalusa and Muscatine studies,5, 26 are required to evaluate the role of ethnicity adequately in long-term HTN-related CV outcomes.

White Coat Hypertension 

Until recently, WCH, defined as having a BP level consistently >95% for age in a physician's office or clinic, but having normotensive readings outside the clinical setting,16 has been dismissed as being of little concern in pediatric patients. Although the Fourth Report discusses the diagnostic measures used in WCH, the pathologic nature of the disease was not addressed, largely because of the absence of compelling pediatric data when the report was published. The absence of outcome data has prevented development of an evidence-based definition of WCH in children. Inadequate normative data are available and there is controversy on which cutoff points to use.27 In their study, Sorof et al demonstrated that the validity of using the mean ABP >95th percentile to determine HTN is questionable because the Task Force percentiles are based exclusively on auscultatory daytime BP measured at rest.27 Further, Sorof et al surmised that using Task Force limits to interpret ambulatory BP may lead to over-diagnosis of ambulatory hypertension in children.27

Unfortunately, the pathologic nature of WCH has been minimized by several earlier investigations. Ben-Dov showed that patients with WCH on treatment had decreased all-cause mortality versus other treated hypertensive groups.28 However, they did not analyze all-cause mortality in normotensive control subjects, nor did they analyze all-cause mortality in non-treated patients with WCH. In another study, by Kouidi et al,29 410 athletes were examined, and 10% were found on their first measurement to have elevated BP consistent with HTN. After repeat measurements, 18 subjects had elevated BP. Those 18 subjects wore 24-hour ABPM, and only 2 were found to have sustained HTN. The results of echocardiograms and electrocardiograms in the subjects with WCH were normal. In a study using a low threshold value to define WCH, Verdecchia et al reported normal left ventricular mass.30 For treatment of WCH, Waeber et al demonstrated that treating patients with WCH before they developed HTN on ABPM may not confer benefit, but rather result in over-treatment of HTN.31

The preponderance of extant literature is replete with evidence that WCH is anything but a normal response to a routine doctor visit. Several cross-sectional studies in adults have shown that WCH is a form of HTN that is intermediate between normotension and sustained hypertension.10, 11 These findings were reported in a long-term study by Pickering et al.32 A study by Alderman et al found that adult patients with hypertension who had a higher BP when taken by a doctor than a nurse were at a higher risk of myocardial infarction during a 14-year follow-up period.33 Glen et al found in adults that left ventricular function and arterial compliance, elasticity, and stiffness were abnormal both in subjects with WCH and in subjects with sustained hypertension, when compared with normotensive subjects.10 Further, Manning et al have shown that some adult patients with only small elevations of BP already have evidence of left ventricular hypertrophy.34 Other investigators have shown that, compared with normotensive subjects, carotid intimal-medial thickness was greater at baseline in WCH, increased more during follow-up, and was not different from patients with sustained HTN.12 Kavey et al demonstrated that exercise BP was exaggerated and left ventricular mass index was increased in pediatric patients with WCH.35 Recently, Lande et al have shown that left ventricular mass index in patients with WCH, although less than in patients with hypertension, was increased in comparison to normotensive pediatric patients,13 a finding previously shown in multiple adult populations.10, 11 These findings demonstrate that WCH is a pathologic entity that should be managed as such. Most recently, the American Heart Association has highlighted these findings in a scientific statement paper by Urbina et al.36

Our concern is that WCH is frequently used as a means by which practitioners can avoid “labeling” a patient as hypertensive. This sort of approach should be discontinued. Multiple studies have shown that WCH is a pathologic entity resulting in significant end-organ physiologic changes. Although the use of ABPM in patients with suspected WCH may be appropriate, the physician should not disregard these abnormal BP measurements. With a reported prevalence of 16% to 52%,20, 27, 35 WCH represents a significant disease entity that is all-too-often disregarded.

Pre-hypertension 

After the Seventh Report by the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC7)23 introduced the concept of PHTN, the Fourth Report included a similar category for pediatric patients.16 The Fourth Report recommendations for PHTN are an amalgamation of JNC7 recommendations and the earlier pediatric guidelines. In the Fourth Report, PHTN in children is defined as average SBP or diastolic BP levels that are ≥90th percentile, but <95th percentile; however, when the BP in an adolescent patient is >120/80mm Hg, then the patient is considered to have PHTN by JNC7 standards. Careful scrutiny of the data in the Fourth Report reveals that the 50th percentile for SBP in a 17-year-old male patient at the 75th percentile for height is 120 mm Hg, the adult cutoff for PHTN. Without the introduction of PHTN in JNC7, such patients would have been deemed to have normal BP. This again highlights the significant limitations and risks of using statistically derived normative cutoffs to define this disease.

The prevalence of PHTN in the United States population ≥20 years of age is 37.4%.2 This includes nearly 42 million men and 28 million women. When the normative data from the Fourth Report are used to diagnose PHTN, the expected incidence of PHTN in adolescents would be approximately 2% to 4%, an incidence markedly below that in adults. However, in large pediatric cohorts, the incidence of PHTN has been demonstrated to be 16% to 17%.20, 22 This discrepancy again demonstrates a deficiency in the current statistically based diagnostic approach to PHTN in pediatrics.

Current Attitudes Toward Hypertension 

Currently, we think there is an overly cautious approach to the diagnosis of HTN in pediatrics. Although the desire to proceed carefully has merit, it can ultimately become a disservice to patients and families when it delays therapeutic interventions demonstrated to decrease long-term CV morbidity and mortality. This guarded approach to the diagnosis of HTN in pediatrics is, at least in part, responsible for the recommendation in the Fourth Report for ≥3 BP measurements in the hypertensive range before the diagnosis of HTN can be made, in contrast to recommendations in adults.23 It is clear from existing morbidity and mortality data that elevated BP, whether in PHTN, WCH, or HTN, is physiologically detrimental in both adults and children. It is also well known that early treatment of adult HTN significantly reduces subsequent CV morbidity and mortality.37

Recommendations 

We believe that a number of changes in the approach to the diagnosis and management of pediatric HTN urgently need to be made. These changes should be addressed in a revised (Fifth) Report on the Diagnosis, Evaluation, and Treatment of High Blood Pressure in Children and Adolescents. First, the definition of HTN must be revised. The new definition of HTN in youth should move away from a Gaussian statistical determination and instead be based on CV risk as defined by long-term epidemiologic evidence from investigations such as the Muscatine and Bogalusa studies.526

Second, a new approach to treatment and monitoring must be undertaken. The sooner a diagnosis is made and addressed with lifestyle modifications, and, when needed, medical therapies, the better the long-term prognosis for the patient. Although some might contend that there are no data to demonstrate that earlier diagnosis and treatment affect long-term outcomes in pediatrics, the argument is short-sighted. The long-term effects of chronic essential HTN, such as renal failure, stroke, myocardial infarction, and death, are rarely seen in pediatric patients. Only with long-term studies will the true magnitude of the impact of HTN be determined. Further, because the first step in treatment recommendations for primary HTN is lifestyle modification, little or no harm could be done; benefit is more likely.38 PHTN should be managed in a similar manner, with aggressive lifestyle modifications. Although no studies have evaluated and determined the evidence-based management approach for WCH, because of the increasing body of evidence that demonstrates the pathologic nature of WCH, we believe that patients who fall into this diagnostic category should be treated with lifestyle modifications and should undergo closer surveillance to verify that they are not progressing to sustained HTN. Further, we recommend that such surveillance should be performed to include yearly ABPM.

Finally, as aforementioned, more long-term studies in large populations are required to define accurately the BP thresholds during childhood at which subsequent CV morbidity and mortality risks are increased. Although such studies certainly require substantial time, effort, and financial resources, their effect on the outcomes of the leading cause of death in western societies cannot be disregarded. Through such studies, important deficits in our current knowledge could be resolved, including the questions about the number of BP readings required to diagnose HTN accurately in youth, ethnic differences in pediatric HTN, including subsequent morbidity in adulthood, and appropriate therapeutic approaches, including pharmacologic and non-pharmacologic modalities. In the interim, although further longitudinal studies are being conducted, short-term studies should be performed to begin to evaluate these questions.