Heart Rate and Blood Pressure Centile Curves and Distributions by Age of Hospitalized Critically Ill Children

Abstract

Heart rate (HR) and blood pressure (BP) form the basis for monitoring the physiological state of patients. Although norms have been published for healthy and hospitalized children, little is known about their distributions in critically ill children. The objective of this study was to report the distributions of these basic physiological variables in hospitalized critically ill children. Continuous data from bedside monitors were collected and stored at 5-s intervals from 3,677 subjects aged 0-18 years admitted over a period of 30 months to the pediatric and cardiac intensive care units at a large quaternary children's hospital. Approximately 1.13 billion values served to estimate age-specific distributions for these two basic physiological variables: HR and intra-arterial BP. Centile curves were derived from the sample distributions and compared to common reference ranges. Properties such as kurtosis and skewness of these distributions are described. In comparison to previously published reference ranges, we show that children in these settings exhibit markedly higher HRs than their healthy counterparts or children hospitalized on in-patient wards. We also compared commonly used published estimates of hypotension in children (e.g., the PALS guidelines) to the values we derived from critically ill children. This is a first study reporting the distributions of basic physiological variables in children in the pediatric intensive care settings, and the percentiles derived may serve as useful references for bedside clinicians and clinical trials.

Keywords: big data; blood pressure; critical care; heart rate; pediatrics; percentiles.

Figure 1

Analysis flow. (A) Example distributions of heart rate (HR) for 20 individual patients in the 12–18 months age group. (B) Sample non-smoothed distributions for HR for children in the three age groups depicted on the right. (C) Sample systolic blood pressure (BP) distributions for children in the same age groups as in (B). (D) Sample smoothed distributions for HR for children in the same three age groups as in (B). (E) Sample mean BP distributions for the same age groups as in (B). Note the narrower and skewed distribution shape. (F,G) Sample cumulative distribution functions for HR and mean BP for the 12–18 months age group with comparison to several fitted standard distributions with percentile lines overlaid.

Figure 2

Heart rate (HR) centile curves as a function of age. (A) HR centile curves for children aged 0–30 days depicting the 5th to 95th percentiles. (B) HR centile curves for children aged 0–18 years.

Figure 3

Arterial blood pressure (BP; systolic, mean, and diastolic) centile curves by age depicting the 5th to 95th percentiles. (A–C) Children aged 0–30 days. (D–F) children aged 0–18 years.

Figure 4

(A) Heart rate (HR) distribution by age—each vertical band represents one age group, with the probability of each HR value coded by the color map detailed in the legend. Overlaid in white lines are the cutoff ranges as published in the PALS guidelines. (B) Systolic blood pressure (sBP) distributions by age with a similar representation scheme as in (A). Overlaid in a white line are the lower cutoff values as published in the PALS guidelines. (C) The fraction of values outside (below or above) the PALS suggested cutoff ranges for HR (dark gray) or below the PALS suggested cutoff for sBP (light gray). The dotted line marks the 0.05 line.

Figure 5

Comparison of the heart rate (HR) percentiles by age derived in this study with those derived for hospitalized children by Bonafide et al. (2) (dotted lines).