Age-Based Percentiles of Measured Mean Arterial Pressure in Pediatric Patients in a Hospital Setting

Objectives: To develop and validate age-specific percentile curves of measured mean arterial pressure for children in a hospital setting. Design: Retrospective observational study of electronic records. Setting: Tertiary care, freestanding pediatric hospital in Seattle, WA. Patients: Nonpremature children, birth to 18 years old, evaluated in the emergency room, or admitted to either acute care or critical care units. Interventions: Oscillometric blood pressure data collected from February 2012 to June 2016 were examined for documentation of systolic, diastolic, and mean arterial pressure values. Quantile curves were developed using restricted cubic splines and validated with two sets of patient data. The effects of birth sex and behavioral state on the curves were examined. The frequency of values less than 5th percentile for mean arterial pressure within a population was compared with four published criteria for hypotension. Measurements and Main Results: Eighty-five-thousand two-hundred ninety-eight patients (47% female) provided 2,385,122 mean arterial pressure readings to develop and validate age-based distributions to create percentile curves and a reference table. The behavior state of patients affected the curves, with disturbed behavior state more prevalent in toddler-aged patients. There was no clinical difference between females and males within age brackets. Mean arterial pressure quantiles identified additional hypotensive episodes as compared with systolic blood pressure thresholds and predicted mean arterial pressure values. Code and data available at: https://osf.io/upqtv/. Conclusions: This is the first study reporting age-specific quantiles of measured mean arterial pressure in children in a hospital setting. The percentile curves may guide care in illnesses when perfusion pressure is critical and serve as parameter for bedside and electronic record-based response to clinical change. Future work to correlate threshold mean arterial pressure values with outcomes would be feasible based on quantile curves.