Diagnosis of hypertension: Ambulatory pediatric American Heart Association/European Society of Hypertension versus blood pressure load thresholds

Abstract The agreement between the traditionally‐used ambulatory blood pressure (ABP)‐load thresholds in children and recently‐recommended pediatric American Heart Association (AHA)/European Society of Hypertension (ESH) ABP thresholds for diagnosing ambulatory hypertension (AH), white coat hypertension (WCH), and masked hypertension (MH) has not been evaluated. In this cross‐sectional study on 450 outpatient participants, the authors evaluated the agreement between previously used ABP‐load 25%, 30%, 40%, 50% thresholds and the AHA/ESH thresholds for diagnosing AH, WCH, and MH. The American Academy of Pediatrics thresholds were used to diagnose office hypertension. The AHA threshold diagnosed ambulatory normotension/hypertension closest to ABP load 50% in 88% (95% CI 0.79, 0.96) participants (k 0.67, 95% CI 0.59, 0.75) and the ESH threshold diagnosed ambulatory normotension/hypertension closest to ABP load 40% in 86% (95% CI 0.77, 0.94) participants (k 0.66, 95% CI 0.59, 0.74). In contrast, the AHA/ESH thresholds had a relatively weaker agreement with ABP load 25%/30%. Therefore, the diagnosis of AH was closest between the AHA threshold and ABP load 50% (difference 3%, 95% CI ‐2.6%, 8.6%, p = .29) and between the ESH threshold and ABP load 40% (difference 4%, 95% CI ‐2.1%, 10.1%, p = .19) than between the AHA/ESH and ABP load 25%/30% thresholds. A similar agreement pattern persisted between the AHA/ESH and various ABP load thresholds for diagnosing WCH and MH. The AHA and ESH thresholds diagnosed AH, WCH, and MH closest to ABP load 40%/50% than ABP load 25%/30%. Future outcome‐based studies are needed to guide the optimal use of these ABP thresholds in clinical practice.


INTRODUCTION
The pediatric American Heart Association (AHA) and European Society of Hypertension (ESH) guidelines recommend the use of 24-h ambulatory blood pressure (ABP) monitoring (ABPM) to diagnose ambulatory hypertension (AH). [1][2][3] Based on office hypertension/normotension as per an office blood pressure (OBP) threshold, AH as per an ABP threshold diagnoses white coat hypertension (WCH) and masked hypertension (MH). [1][2][3][4][5] In children, ABP load (proportion of ABP readings higher than 95 th mean ABP percentile) has been traditionally used to diagnose AH. [6][7][8][9][10][11] However, the lack of consensus and paucity of outcome-based studies have led to the use of various ABP load thresholds, ranging from 25%, 8,11 30%, 12 40% 13 to 50%. 10 To establish a uniformity in the use of ABP thresholds and consistency with the use of mean-ABP thresholds in adults, 14 The agreement between various ABP-load thresholds and the AHA/ESH thresholds is not known, which makes it challenging to interpret the diagnosis of AH across the studies using either AHA/ESH or an ABP-load threshold. Therefore, we evaluated the agreement between the AHA/ESH thresholds and previously used ABP-load thresholds (25%, 30%, 40% and 50%) for diagnosing AH, WCH, and MH.

METHODS
This was a single center, retrospective cross-sectional study performed after approval by the University of Western Ontario research ethics board. The study involved a retrospective review of existing clinical data and was therefore exempted from the need for an individual informed consent. The records of children who underwent 24-h ABPM at a tertiary care outpatient hypertension clinic (London, Ontario, Canada) were collected. The participants were referred to our outpatient clinic because of suspected hypertension based on OBP assessments by the primary health care providers. In those with multiple ABPM assessments, their first ABPM recording was included for this analysis. The data was collected between January, 2003 and December, 2008 (n = 159) as a part of previous studies [16][17][18][19] and recently between January, 2018 and September, 2020 (n = 291). During both the study periods, there was a uniformity in the protocol regarding offering ABPM to patients older than 5 years and for evaluating secondary hypertension (Fourth Report guidelines 20  were instructed to continue with their regular daily activity, to avoid strenuous exercise, to keep arm still at the time of ABP recording by the equipment and to maintain a wake-sleep log for defining day and night ABP. The data was inspected to edit outliers. The adequacy of ABP recordings was established based on minimum one reading per hour during day and nighttime, and more than 40 readings in 24-h. 1 24-h, day and night systolic and diastolic ABP were analyzed by the 24-h, day and night mean 95 th systolic and diastolic ABP percentiles as per the normative data by Wuhl and coworkers 24 Systolic and diastolic ABP load over 24-h, day and night were calculated as the percentage of ABP measurements higher than respective mean 95 th systolic and diastolic ABP percentiles. 24

Outcomes
Our primary outcome was to evaluate the agreement between the AHA/ESH thresholds and various ABP-load thresholds in earlier studies for diagnosing AH. Our secondary outcome was to assess  Table 1).

Statistical methods
Normally distributed continuous variables were reported as mean and degree of skewness. 24 The agreement between the AHA/ESH thresholds and ABP-load thresholds was calculated by the accuracy (the proportion of ABP classified similarly by the two ABP thresholds) and kappa statistics. 26

Patient characteristics
In the initial screening, 544 participants who had ABPM studies dur-  Figure 1). The agreement between the AHA threshold and different ABP load thresholds remained consistent in sub-groups based on age, sex, primary hypertension, and those not taking blood pressure medications (Table 4).
Similar to the AHA threshold, the ESH threshold had a larger difference with ABP load 30% (difference 7%, 95% CI 1.2%, 12.8%, p = .01) and ABP load 25% (difference 12%, 95% CI 6.4%, 17.5%, p < .001) ( Table 4). The agreement pattern in the diagnosis of WCH between the AHA/ESH thresholds and the ABP load thresholds did not significantly change when analyzed in the sub-groups based on age, sex, primary hypertension and those not on blood pressure medications (Table 4).
The agreement pattern in the diagnosis of MH by the AHA/ESH thresholds and the ABP load thresholds did not significantly change when analyzed in the subgroups based on age, sex, primary hypertension, and those not on blood pressure medications (Table 4).

DISCUSSION
In the absence of pediatric literature on the agreement between various ABP-load thresholds (ABP load 25%, 30%, 40%, and 50%) used in previous studies [6][7][8][9][10][11]  represents the extent of ABP elevation whereas ABP load denotes the frequency of ABP elevation. Therefore, in patients with sustained high mean ABP, ABP load close to 100% cease to offer additional information, 31 whereas in patients with normal mean ABP, ABP load might quantify the degree of ABP fluctuations above normal limits. 6,7 Consequently, recent studies reevaluated the role of ABP load as an additive covariate to mean ABP, which showed no added increase in the prediction of target organ damage by mean ABP after accounting for ABP load. 8,27,28 In absence of an additive benefit of ABP load, the recognition of AH in clinical practice requires either a mean ABP or an ABP-load threshold. In this scenario, our findings provide a connecting link for a consistent interpretation of AH based on either AHA/ESH threshold or one of the ABP-load thresholds.

ACKNOWLEDGMENT
None declared.

AUTHOR CONTRIBUTIONS
It is to state that the work described has not been published before, and it is not under consideration for publication anywhere else. For the manuscript, APS conceived the idea, conducted the statistical analysis, drafted the manuscript and supervised overall development of