Ethics approval
The study was approved by the Institutional Review Board of the University of California Davis. An exemption for the need for written informed consent was granted by the Institutional Review Board of the University of California Davis due to the study's retrospective nature. All methods were performed in accordance with the relevant guidelines and regulations.
Subjects
Children < 3 years of age admitted to either the pediatric wards or intensive care unit (ICU) services between January 2017 and January 2019 with a discharge diagnosis of bronchiolitis were included in this study. An age and sex-matched control group, consisting of children admitted to the aforementioned services with one of the following medical conditions: skin and soft tissue infection, urinary tract infection, or seizure was included for comparison. These diagnoses were chosen for the control group as they are common reasons for admission in children < 3 years, do not affect respiratory status, and may be distressing in terms of pain and discomfort. The intent of this control group was to explore whether there was a specific link between bronchiolitis and BP abnormalities that may be attributed to the disease process itself rather than the agitation associated with an inpatient hospitalization or the discomfort of having respiratory distress unrelated to the underlying diagnosis (i.e., white coat HTN).
Exclusion criteria
Subjects needed to have at least 1 day of a minimum of two “legitimate” BP readings recorded in the flowsheet. Measurements were deemed “legitimate” if there was no nursing documentation that the patient was agitated or moving during the reading, or that the BP was measured on a lower extremity. Notably, those measurements not specifying from which extremity they were obtained were assumed to be taken from an upper extremity, as is standard practice in our hospital, and included in the analysis. Patients were excluded from the study if they had previously diagnosed HTN or were on antihypertensive medications prior to admission, or if they were intubated at any point during the hospitalization, as many sedation medications have vasoactive properties that can affect BP. Other exclusion criteria included a diagnosis of bronchopulmonary dysplasia, presence of acute kidney injury, electrolyte abnormalities, a comorbid acute process overlapping with one of the other inclusion diagnoses, and absence of a recorded height in children as this is needed for determination of BP percentiles.
Study design
This was a single center retrospective case–control study. HBP and HTN were defined as at least one inpatient day of either averaged diastolic BP (DBP) or averaged systolic BP (SBP) readings above the 90th and 95th percentile, respectively, determined by using age-appropriate BP charts. Standard published normative pediatric reference charts were used based on the age of the child [8,9,10]. The most recent clinical practice guideline for screening of BP, published in 2017, was used for children ≥ 1 year of age [8]. Since normative data on younger children was not a part of these guidelines, the report of the second task force on BP control in children was used for infants 1 month to 1 year of age [9], and data developed by Zubrow et al. [10] was used for infants < 44 weeks of corrected gestational age. In the cohort admitted for seizures, only those BP readings that were obtained after the child had regained normal sensorium were included in the study data. BP measurements were taken by trained pediatric nursing personnel using a validated oscillometric measuring device (Dinamap; Critikon Company LLC, FL, USA). Legitimate BP measurements were transcribed from flow sheets, and SBP and DBP were individually averaged over each inpatient day. These daily averages were then assigned percentiles using the appropriate reference charts. To quantify the severity of HTN, when present, the HTN index was calculated by dividing daily averaged SBP and DBP values by the respective reference 95th percentile values [11]. Data relating to patient demographics and clinical management that are known to, or could possibly, influence BP were collected: age, sex, body mass index (BMI), use of systemic corticosteroids or albuterol, length of stay, history of prematurity, and severity of illness as determined by wards/ICU admission status. Additional data abstracted included mention in the medical records of recognition of BP abnormalities, inpatient treatment of or outpatient referral for BP abnormalities, and outpatient BP measurements for patients who were followed in our health-system. Postdischarge outpatient BP measurements were obtained by reviewing all clinical encounters taking place within 3 months of hospital admission and accessible through the electronic health record. The primary outcome measure was the prevalence of BP abnormalities during the hospital stay—HBP, HTN, and the composite of both HBP and HTN. The secondary outcome measure was the severity of BP abnormalities, as quantified by the HTN index.
Statistical approach
Sample size calculations were based on our assumption that children admitted with bronchiolitis would have a tenfold higher prevalence of HTN (10%), compared to the control groups (1%). To achieve a power of 80% to detect this difference and an α of 0.05, we needed 100 children in each of the two groups. Multivariate logistic regression analyses were performed using Stata MP ver. 16 (Stata Corp., College Station, TX, USA). Regressions were estimated separately for the three different primary outcome measures: HTN, HBP, and the composite of HBP or HTN. Model specification was tested by performing linktests, comparing Bayesian information criteria, and computing the area under the model’s receiver operating characteristic curve. HTN and HBP were defined as the dependent variable outcomes of the following independent variables: sex, corrected gestational age, normative chart used, BMI > 90th percentile, level of care, length of stay, a diagnosis of bronchiolitis, prematurity, and use of albuterol or steroids as a part of inpatient treatment. Comparisons of HTN indices for bronchiolitis and control groups was performed using the Student t-test.