Hypertension and End‐Organ Damage in Children––Is the Picture Less Fuzzy Now?

Ambulatory blood pressure(BP) monitoring (ABPM) is a well‐tolerated1, 2 and effective technology3, 4 that has transformed the understanding and management of hypertension in the past decades,3 including in children.5 Recent papers in this Journal testify to the evolving significance of ABPM with regard to white‐coat hypertension, nocturnal nondipping,3, 6 BP variability,7, 8, 9 and the relatively newly recognized entity of masked hypertension.10, 11 However, there is much to be learned and to understand about pediatric hypertension, the scope of the disease, and the impact of it on children in the 21st century. Children are different from adults as they are less likely to have comorbidities “contaminating” the clinical picture. Conceptually, a pediatric cohort of a disease may represent a “pure model” for the illness, not affected by the multiple coexisting disease processes of a middle‐aged or elderly cohort. Despite their smaller sizes, pediatric cohorts may demonstrate biomedical signals with less ambiguity than those of adults. To state it differently, an adult cohort may not be ideal to separate the effect of hypertension from those of the confounding comorbidities. Accordingly, it would be helpful to understand and define the earliest damages taking place during early‐onset hypertension in a pediatric cohort. Perhaps the scenario is analogous to acute kidney injury in critical illness, where the adverse effect of volume overload was first demonstrated in children,12, 13, 14 well before adult studies were forthcoming.15, 16

The paper by Conkar and colleagues17 published in this issue of The Journal of Clinical Hypertension 17 set out to examine the association of ABPM‐derived BP and BP phenotypes with end‐organ damage in a moderately sized cohort of 82 newly diagnosed pediatric hypertensive children. The paper was a methodologically sound investigation that included obtaining a sleep‐wake diary during ABPM monitoring and gathering multiple clinically meaningful measurements of end‐organ damage. The patients were recruited over a relatively large age range, 6 to 17 years, with a presumed new diagnosis of hypertension. Nonetheless, the presence of hypertensive retinopathy in approximately one third of the patients argues for some degree of chronicity already in place. In fact, the duration of suspected hypertension before the referral was not known for these children and teenagers. Enrolled patients were all taller than 120 cm and were examined very carefully with multiple technologies: cardiac echocardiography to assess left ventricular mass, carotid ultrasound to measure intima‐media thickness (cIMT), technology to measure vascular stiffness (pulse wave velocity [PWV] and aortic augmentation index [AIx]). Additionally, they underwent 24‐hour urine collections for measurement of albumin excretion and a dilated eye exam with an ophthalmologist. In doing so, the authors performed a careful, multi‐focused assessment of end‐organ damage, using both older (eye examination) and newer technologies (PWV, AIx), focusing primarily on structural parameters, rather than surrogate laboratory parameters. While overall BP was not markedly elevated (BP 118.7±12/70.6±11.2 mm Hg) in these children, the cohort had a surprisingly large burden of presumed end‐organ damage (68%), which should have necessitated ruling out major causes of secondary hypertension such as renal artery stenosis. According to the authors’ statement, obesity and secondary causes of hypertension were excluded by history, physical examination, urinalysis, serum chemistries, renal ultrasonography, and other tests as indicated, according to guidelines from the Working Group on High Blood Pressure in Children and Adolescents. However, no actual data on plasma renin activity or imaging studies of the renal artery were presented in the paper. It would have been interesting to see the actual results of chemistry parameters, such as serum potassium, uric acid, or perhaps lead level in these patients. Hypokalemia may also provide a hint of Liddle syndrome,18 and the extent of hypertensive retinopathy (35.3%) suggested a history of elevated BP predating the diagnosis in these patients. Uric acid, an another important emerging risk factor,8, 19 may also be useful during routine chemistry but was not assessed in this paper. If elevated uric acid was detected, measurement of the serum lead level would have been appropriate. Another important limitation of the current paper is the authors’ definition of renal end‐organ damage. The National Kidney Foundation's definition of chronic kidney disease (CKD) calls for repeated demonstration of albuminuria >30 mg/g of creatinine at least 3 months apart to diagnose CKD.20 However, in the current paper albuminuria was assessed only on a single occasion and may have led to potential overdiagnosis of CKD as end‐organ damage.

The key finding of the study was that most of the end‐organ damage in this cohort was associated with systolic BP load. Most of the cohort were nondippers during the overnight portion of BP monitoring. However, nondipper status was not predictive of end‐organ damage. This appears to be an important distinction from adult cohorts.3, 6 Another key finding of the paper is the lack of association of daytime systolic BP (>25% over the 95th percentile) with left ventricular hypertension (LVH), vs the positive association with PWV, AIx, and cIMT. Further, there was a similar association maintained when the cohort was dichotomized according to systolic BP load >50 vs <50%, once again correlating with features of premature atherosclerosis and vascular stiffness (PVW, P=.031; cIMT, P=.004) in a graded manner. However, the study's main findings were not further confirmed on multivariate analysis or at least adjusted for the main demographic variables (age, sex, weight)––an obvious shortcoming of the paper. In this context, it should be noted that LVH was assessed by calculating left ventricular mass index, adjusting for height2, 7 and thus mitigated interference from age and weight.

One possible explanation for the study's main finding is that systolic (rather than diastolic) BP appears to be associated with atherosclerosis and stiffening of the arteries, similar to that in elderly patients. The alternative explanation is, of course, that these children's hypertension was more genetically determined (hence, an earlier presentation in childhood) and may represent a different hypertension and hypertension phenotype than adult illness. Nonetheless, the low burden of established hypertension in the parents (approximately 15%, as shown in Table 1 in the paper by Conkar and colleagues22) argues against excessive genetic influence in this cohort.

The incidence of LVH in a much larger cohort of children with hypertension has been shown to be much lower.21, 22 A recent study conducted during 2006 to 2011, studied an average of 1.3 million patients aged 2 to 18 years per year. A total of 16,322 met the definition of hypertension (2.6 of 1000). Among patients with hypertension, 5585 (34%) underwent echocardiography. In patients with echocardiography, 8.0% had LVH or dysfunction.22 The high rates of LVH in children described by the current authors (approximately 15%) have been rarely reported from most population cohorts with “primary” hypertension. In adult studies, LVH correlates with ABPM‐derived and office‐based BP, and office‐based BP does not add to the predictive value.23 Variability of the ABPM‐derived BP itself is an emerging adverse marker for the severity of hypertension.8, 9 Performing home BP measurement in the same cohort is yet another untapped path open for the authors' exploration. A shorter period of BP monitoring up to 6 hours was shown to have excellent correlation with full ABPM‐derived values24 and it remains to be seen whether the same holds true for assessing end‐organ damage.

While selection criteria for the study's enrollment were clear, we do not know how many children were screened in the community or the size of the population serving as the referral basis for the Ege University Medical Center. We know very little about baseline frequency of hypertension in the underlying cohort and delays in recognition of hypertension caused by lack of regular BP measurements in referring communities may have introduced a frank selection bias into the study. Additionally, only a single measurement of ABPM was performed and the monitoring was performed for 24 hours only. Repeated or longer ABPM measurements may have further fine‐tuned the authors’ findings with regard to LVH, PVW, cIMT, and AIx. The main limitation of the study is, however, its cross‐sectional nature: hence, it cannot imply causality, only association. While the study enrollment had evidently abnormal ABPM studies, we do not have a comparator cohort. That is, it may have been interesting to see the degree of findings of potential end‐organ damage evaluation among those children referred for ABPM for suspected high BP but who passed ABPM or in normal controls from the community. Somewhat along the same line, potential future investigations may evaluate the validity or clinical utility of a repeated assessment for these children, who may already have end‐organ damage (eg, LVH or affirmative ophthalmoscopy findings) but passed the initial ABPM testing. It may be interesting to know, whether the cohort members’ main demographic characteristics (age, sex, body mass index) were different when grouped according to each end‐organ damage listed. Again, a control group of age‐ and weight‐matched children with normal BP from the local community may have further strengthened the study's findings. Hypertension in children and adolescents are likely to be influenced by both ethnic background and environment, rising commensurately with obesity––up to 3.5% in North America.25

Perhaps one final moral of the paper by Conkar and colleagues17 is that one does not have to have the newest and fanciest test or biomarkers available to ask meaningful questions; rather, one can ask and answer meaningful questions by organizing the technology and skills available in a meaningful and coherent manner. In doing so, it is not much different than performing good patient care––which is, not coincidentally, what physicians do.