1 INTRODUCTION

Child/adolescent obesity strongly tracks into adulthood1, 2 and conveys an increased risk for morbidity and mortality of cardiovascular disease in adulthood, with the strength of the relationship to risk for cardiovascular disease increasing as the child ages.3, 4 Obesity is an established risk factor for hypertension in youth and risk increases with severity of obesity.5 A systematic review of 19 studies reported that youth with elevated blood pressure (BP) were more likely to have increased risk for markers of cardiovascular disease (such as high pulse wave velocity, high carotid intima-media thickness, left ventricular hypertrophy) in adulthood.6 Thus, recognition of elevated BP by standardized BP measurement, and interpretation and early management are important aspects of optimal clinical care for youth with obesity.

Using current guideline interpretations for BP,7 national samples report 27.5% of youth ages 12–19 years with obesity had BP above a normal range (ie, elevated, stage 1, or stage 2 classification; henceforth, high BP); and 30.6% of youth with severe obesity (defined as body mass index [BMI]-for-age ≥120% of the 95th percentile) had high BP.8 A recent report evaluated BP among national samples of youth ages 8–12 years and 13–17 years, reporting 2–3 times higher risk for high BP among youth with obesity as compared to youth of normal weight, but youth with severe obesity were not separately evaluated.9

Because treatment of high BP in childhood could potentially reduce long-term cardiovascular risks, it is of particular importance to understand prevalence and risks for high BP among youth with obesity in the clinical setting to plan for their evaluation and management needs. This study evaluates data obtained in the Pediatric Obesity Weight Evaluation Registry (POWER). POWER is a consortium of multicomponent pediatric weight management (PWM) programs across the United States which prospectively enrolls patients seeking obesity management care and tracks parameters of care in an aggregate database.10 Three-quarters of youth in POWER present with severe obesity.11 This study aims to evaluate data from youth enrolled in POWER to: (1) assess prevalence of high BP in the clinical setting among care-seeking youth with obesity; and (2) identify characteristics associated with high BP status at a baseline visit.

2 METHODS

POWER was established in 2014 with the aim to better understand the complex nature of PWM and to collectively evaluate and improve medical care strategies for children and adolescents with obesity and improve health outcomes.10 The Data Coordinating Center for POWER is located at Cincinnati Children's Hospital Medical Center. POWER leadership supports POWER's infrastructure, which include an aggregate database and collaborative work, such as research and educational webinars. Financial support for POWER is provided via enrollment fees paid biannually by individual sites, with oversight provided by POWER's administrative staff at Cincinnati Children's Hospital Medical Center.

The Cincinnati Children's Hospital Institutional Review Board approved this study. Additionally, Institutional Review Boards at each site approved the study and monitored the site-specific consent/assent process. The POWER study is registered with Clinical.Trials.gov (NCT02121132).

Patients evaluated were prospectively recruited from 35 participating sites in May 2014 through December 2017. Not all sites participated for the entire data period, as some joined POWER later during the study period.

Each site contributed data to a central database. Database elements included patient demographics (race, ethnicity, insurance type) and characteristics (sex, age, weight, height, BP). One BP measurement was entered for each patient visit; documentation of the BP measurement procedure (eg, electronic or manual measure, or if it was a report of an average of repeated BP measurements) was not obtained. Additionally, there was no information on BP measurement practices as protocols by site were not available; however, POWER did provide written BP assessment recommendations to sites based on Centers for Disease Control and Prevention (CDC) guidelines12 within a Data Definitions Document. The registry included optional data entry for patient diagnoses/conditions (eg, hypertension) existing at the time of the initial PWM visit and for medications/treatments (eg, antihypertensive medications) at the time of the initial PWM visit. However, these fields were not uniformly completed so were not considered in analyses. Thus, some youth included in these analyses had hypertension and may been taking antihypertensive medication before their initial PWM visit.

Data of patients aged 3–17 years with valid BP, height and weight measures, and body mass index (BMI) for age ≥95th percentile at an initial visit were evaluated. Because BP guidelines lack interpretations for patients at extreme height-for-age (height z-score [HtZ] > 3.9 or < - 3.9), patients with height-for-age extreme values at their initial visit were excluded from these analyses.

Among 8933 active POWER participants, 990 (11.1%) were excluded from analyses; initial visit data of 7943 patients (88.9%) were analyzed. There were 107 (1.2%) excluded due to age missing, age < 3 years, or age ≥18 years, 509 (5.7%) excluded due to missing height or weight or BP, 299 (3.3%) had BMI < 95th BMI percentile, and 75 (0.8%) had extreme height-for-age values.

Data were used to compute percentage of the 95th percentile for BMI (%BMIp95) and height z-score (HtZ) based on CDC 2000 references.13 BP interpretations were done using references in the 2017 guideline (SAS macro).7 Obesity class groups were defined as follows: class 1, 100% to < 120% %BMIp95; class 2, 120% to < 140% %BMIp95; class 3, ≥140% %BMIp95; classes 2 and 3 were considered severe obesity.14

2.1 Analyses

The frequency and percentage of patients overall, and for normal BP and high BP groups are presented. To evaluate associations with sex, age group, race/ethnicity, insurance group, age, %BMIp95, and HtZ, as appropriate, Pearson's chi-square test was used to compare binary or nominal categorical variables, and the Kruskal-Wallis test was used to compare continuous variables or ordinal categorical variables across BP group status. Additionally, for each BP interpretation category (normal, elevated, stage 1, stage 2) descriptive data for these same categorical and continuous variables are presented.

A mixed-effects logistic regression model, with site as a random effect to account for multi-site sampling, was used to evaluate the associations of factors with BP group status (normal BP vs. high BP group). Factors in the model included categorical variables of sex, age group, race/ethnicity, insurance group, and obesity class group. To further focus on youth with BP interpreted as stage 1 or stage 2 hypertension, a second mixed-effects logistic regression model applied the same methods to examine associations of factors with BP interpretation group status between the combined normal BP and elevated BP groups (normal/elevated group) versus the group of youth with stage 1 or stage 2 hypertension (stage 1/stage 2 group). Patients with missing race/ethnicity or insurance group were excluded from logistic regression analyses.

Frequency of BP interpretation groups by site are reported. All analyses were performed using SAS software version 9.4 (SAS Institute Inc., Cary, NC, USA).

3 RESULTS

Data on 7943 patients were analyzed; 54% were females. Patients were most often either Hispanic (32%) or White non-Hispanic (NH) (39%) and 59% had public health insurance (Table 1). Mean age was 11.7 years (SD 3.3); about half (47%) of patients were aged 12–18 years; a majority (73%) had severe obesity. There were 35 sites contributing data with a median of 131 patients per site in analyses (interquartile range [IQR], 56–223 patients; range, 2–1239 patients).

TABLE 1. Patient characteristics by blood pressure status group Grouped by BP Status Overall (N = 7943) Normal BP (N = 4061, 51.1%) High BPa (N = 3882, 48.9%) Variable Level n % n % n % p-valuec Categorical variables Age group, yearsd 3–5 367 4.6 192 52.3 175 47.7 <.0001 6–10 1296 16.3 711 54.9 585 45.1 9–11 2547 32.1 1380 54.2 1167 45.8 12–14 2197 27.7 1153 52.5 1044 47.5 15–17 1536 19.3 625 40.7 911 59.3 Sex Male 3668 46.2 1704 46.5 1964 53.5 <.0001 Female 4275 53.8 2357 55.1 1918 44.9 Race/ethnicity Black NH 1495 18.8 810 54.2 685 45.8 .0001 Hispanic 2503 31.5 1304 52.1 1199 47.9 White NH 3129 39.4 1557 49.8 1572 50.2 Other and multiracial NH 429 5.40 228 53.1 201 46.9 Unknown 387 4.9 162 41.9 225 58.1 Primary health insurance Private 2331 29.3 1155 49.6 1176 50.4 .0127 Public 4669 58.8 2381 51.0 2288 49.0 Self-pay/none 65 0.8 33 50.8 32 49.2 Unknown 878 11.1 492 56.0 386 44.0 Obesity Classb Class 1 2166 27.3 1350 62.3 816 37.7 <.0001 Class 2 2854 35.9 1543 54.1 1311 45.9 Class 3 2923 36.8 1168 40.0 1755 60.0 Continuous variablesd Age, years Median 7943 11.7 4061 11.5 3882 12.0 <.0001 25th 9.5 9.3 9.6 75th 14.3 13.8 14.8 Mean 11.7 11.5 12.0 STD 3.3 3.2 3.4 Min 3.0 3.0 3.0 Max 18.0 18.0 18.0 %BMIp95 Median 7943 132.1 4061 128.3 3882 136.9 <.0001 25th 118.8 115.9 122.3 75th 148.5 143.0 154.8 Mean 136.6 131.8 141.6 STD 25.4 22.5 27.2 Min 100.0 100.0 100.0 Max 505.2 505.2 376.8 Height-for-age z-score Median 7943 0.8 4061 0.8 3882 0.8 .4454 25th 0.1 0.1 0.1 75th 1.5 1.5 1.5 Mean 0.8 0.8 0.8 STD 1.1 1.1 1.1 Min -3.7 -3.7 -3.6 Max 3.7 3.7 3.7 All tests treat the column variable as nominal. Abbreviations: BP, blood pressure; NH, non-Hispanic; BMI, body mass index; %BMIp95, percentage of the 95th BMI percentile.

Nearly half of patients (3882/7943, 48.9%) had BP interpreted as high BP (18.9% elevated, 23.9% stage 1, and 6.0% stage 2) and 51.1% had normal BP. High BP was found for 60.0% of youth with class 3 obesity, 45.9% with class 2 obesity, and 37.7% with class 1 obesity, (Table 1). Highest frequency of high BP status occurred in older teens, males, and for White NH youth and those of unknown race/ethnicity status (Table 1). Descriptive data for continuous variables (age, %BMIp95, and HtZ) are also presented in Table 1. Table S1 (online) presents these data for four BP interpretation groups: normal BP (51.1%); elevated BP (18.9%); stage 1 hypertension (23.9%); and stage 2 hypertension (6.0%).

As a next step, the first multivariate analysis evaluated the 6752 patients with complete data; 1191 patients (15.0%) were excluded due to missing race/ethnicity and/or missing insurance group. The first analysis evaluated factors associated with being in the high BP (n = 3311, 49.0%) versus the normal BP group (n = 3441, 51.0%). Youth with severe obesity and males were more likely to have high BP (Table 2). Younger age groups (3–5, 6–8, 9–11, 12–14 years) were less likely to have high BP compared to youth ages 15–17 years. White NH youth were more likely to have high BP status as compared to Black NH youth, but were similar to Hispanic and Other/Multiracial NH groups. Insurance group did not significantly predict BP group status.

TABLE 2. Generalized linear multivariable model examining association of characteristics with being in the high BP group compared to normal BP group based on 2017 AAP CPG BP interpretations Univariate analysis+ Multivariate analysis† Characteristics OR (95% CI) p value OR (95% CI) p value Sex  Male 1.44 (1.30, 1.59) <.0001 1.41 (1.27, 1.56) <.0001  Female Ref Ref Age groups  3–5 years 0.69 (0.54, 0.89) .0042 0.65 (0.50, 0.85) .0012  6–8 years 0.57 (0.48, 0.68) <.0001 0.57 (0.48, 0.68) <.0001  9–11 years 0.60 (0.52, 0.69) <.0001 0.63 (0.54, 0.73) <.0001  12–14 years 0.64 (0.55, 0.74) <.0001 0.64 (0.55, 0.75) <.0001  15–17 years Ref Ref Race/ethnicity  Black NH 0.89 (0.77, 1.02) .0972 0.84 (0.72, 0.97) .0212  Hispanic 0.86 (0.75, 0.98) .0253 0.91 (0.79, 1.05) .1908  Other and multiracial NH 0.92 (0.73, 1.15) .4688 0.97 (0.77, 1.22) .7735  White NH Ref Ref Insurance type  Public 0.98 (0.88, 1.09) .6829 0.95 (0.85, 1.07) .4203  Self-pay/none 0.98 (0.58, 1.65) .9463 1.04 (0.61, 1.78) .8767  Private Ref Ref Obesity classa  Class 3 2.46 (2.16, 2.80) <.0001 2.39 (2.10, 2.73) <.0001  Class 2 1.43 (1.26, 1.62) <.0001 1.42 (1.25, 1.62) <.0001  Class 1 Ref Ref Abbreviations: BP, blood pressure; AAP CPG, American Academy of Pediatrics Clinical Practice Guideline7; OR, odds ratio; Ref, reference; NH, non-Hispanic.

The next multivariate analyses again evaluated the 6752 patients with complete data to examine factors associated with being in the stage 1/stage 2 BP group (n = 2021, 29.9%) as compared to the normal/elevated BP group (n = 4731, 70.1%). Youth with severe obesity and males were more likely to be in the stage 1/stage 2 BP group (Table 3). However, as compared to youth ages 15–17 years, only those in the 12-14-year age group were less likely to be in the stage 1/stage 2 BP group; with no significant differences found between other age groups compared to youth ages 15–17 years. Neither race/ethnicity nor insurance group significantly predicted BP group status.

TABLE 3. Generalized linear multivariable model examining association of characteristics with being in the Stage 1/Stage 2 BP group compared to the normal/elevated BP group based on 2017 AAP CPG BP interpretations Univariate analysis+ Multivariate analysis† Characteristics OR (95% CI) p value OR (95% CI) p value Sex  Male 1.31 (1.18, 1.46) <.0001 1.25 (1.12, 1.40) <.0001  Female Ref Ref Age groups  3–5 years 1.07 (0.81, 1.39) .6443 1.01 (0.77, 1.33) .9501  6–8 years 0.96 (0.80, 1.15) .6503 0.98 (0.81, 1.17) .7943  9–11 years 0.89 (0.76, 1.04) .1486 0.97 (0.82, 1.13) .6804  12–14 years 0.68 (0.58, 0.79) <.0001 0.69 (0.58, 0.81) <.0001  15–17 years Ref Ref Race/ethnicity  Black NH 0.95 (0.82, 1.11) .5515 0.87 (0.74, 1.02) .0855  Hispanic 0.83 (0.73, 0.95) .0049 0.91 (0.77, 1.06) .2228  Other and multiracial NH 1.02 (0.79, 1.30) .8976 1.05 (0.81, 1.35) .7298  White NH Ref Ref Insurance type  Public 1.05 (0.93, 1.18) .4558 0.99 (0.88, 1.13) .9205  Self-pay/none 0.91 (0.51, 1.62) .7404 0.95 (0.52, 1.71) .8531  Private Ref Ref Obesity class*  Class 3 2.77 (2.39, 3.21) <.0001 2.75 (2.37, 3.20) <.0001