Review of the Literature
Textbooks
Nursing and medical textbooks often provide tables with normative data for VS; however, the tables were not consistent across texts. Although parameter ranges overlapped and were similar, some differences were notable. For instance, for children 3 years of age, Bowden and Greenberg (2008) listed a normal heart rate (HR) range for children 3 to 4 years of age as 80 to 120 beats/minute, whereas Van Hare and Dubin (2001) grouped 3-year-olds with younger counterparts and listed 89 to 152 beats/minute to be the normative HR range. Several nursing textbooks reproduced a table from a medical textbook that displayed HR ranges dependent of age and wakefulness (Gillette et al., 1989). However, Gillette et al. (1989) cited only a 1981 study that found 65% of healthy 7-to 11-year-olds (N = 104) experience sinus pauses and heart rates below 45 to 55 beats/minute during the night hours when monitored over 24 hours (Southall, Johnston, Shinebourne, & Johnston, 1981). Other than this reference in Gillette et al. (1989), all other textbooks cited other textbooks for HR or respiratory rate normative data.
When presenting BP normative charts, textbooks refer to or reproduce the Fourth Report on High Blood Pressure in Children and Adolescents. The Fourth Report was developed from previously obtained datasets (National High Blood Pressure Education Program [NHBPEP] Work ing Group on High Blood Pressure in Children and Adolescents, 2004). This report presents the values for the 50th, 90th, 95th, and 99th percentiles for systolic BP (SBP) and diastolic BP (DBP) based on gender, age in years, and height percentiles, yielding a complex table of 1,904 values. The primary purpose of this report was "to provide recommendations for diagnosis, evaluation, and treatment of hypertension based on available evidence" (NHBPEP Working Group on High Blood Pressure in Children and Adolescents, 2004, p. 555). The Fourth Report defines pre-hypertension and hypertension based on the 90th and 95th percentile respectively (NHBPEP Working Group on High Blood Pressure in Children and Adolescents, 2004). Yet, many values for the 90th percentile fall above the 120/80 adult guidelines for hypertension (Kaelber & Pickett, 2009; Krishna, PrasannaKumar, Desai, & Thennarasu, 2006). Kaelber and Pickett (2009) have proposed a simplified table based on age and gender only, but again, the purpose is aimed at identifying children needing further attention regarding hypertension and do not provide minimum BP values. While these tables are based on a large dataset, the lack of control in the procedures of obtaining the data is a limitation.
One series of nursing textbooks (Hockenberry & Wilson, 2011; Wilson & Hockenberry, 2012) cite Park and Menard (1989) for normative oscillometric BP tables for infants through 5 years of age. Park and Menard (1989) showed that oscillometric BP readings were higher than those obtained by auscultation. Again, these textbook tables are indicated for diagnosis of hypertension versus full normative ranges for these ages.
Research Establishing Normative Parameters
After this review of a sample of nursing and medical textbooks, the authors concluded that there was a lack of consistency across textbooks in the presentation of normative values for VS. Additionally, the given normative values lacked reference to empirical data. Although BP charts cited research, the authors questioned using guidance on hypertension for purposes of determining normative values in the acute setting. The following is a review of recent research into determining normative values. Table 1, Table 2 and Table 3 give additional information on these studies. Not all studies produced normal ranges of values.
Studies on Respiratory Rates (RR) in Children. The need for reference ranges prompted researchers in Italy to assess RR in healthy infants and young children (see Table 1) (Rusconi et al., 1994). Researchers measured RR by direct placement of stethoscope on the bare chest for 60 seconds. Rusconi et al. (1994) found evidence of significant difference in RR between awake and sleeping conditions. Further, RR measured by auscultation were significantly higher than by observation only in both awake and sleeping conditions (Rusconi et al., 1994). These findings are presented as smooth centile curves rather than as normal range of rates.
Wallis and associates completed two descriptive studies out of the United Kingdom and South Africa (Wallis, Healy, Undy, & Maconochie, 2005; Wallis & Maconochie, 2006). Both studies assessed fully clothed school children after 10 minutes of sitting. Researchers observed chest wall movement for 60-second observation. The authors maintained that this method of observation decreases child's awareness as compared to a pneumogram or auscultation. Although researchers were explicit in counting partial breaths as full breaths, these articles refer to only one 60-second period being observed, and there is no reference to inter-rater reliability. The possibility of measurement error is a limitation to these studies, and therefore, do not give strong evidence of the normal distribution of respiratory rates.
In a meta-analysis of 69 studies reporting on respiratory rates and heart rates in children from birth to 18 years of age, Fleming et al. (2011) calculated RR and HR centiles. Methods of measuring RR were reported as primarily manual; however, manual was not defined. Smooth curve graphs are presented showing the decrease in RR as the child ages, but normal range of rates are not given.
Studies on Heart Rates (HR) in Children. Studies aimed at determining normative HR in children have differed in the method of measurement and in the length of observation. Two studies by Wallis and associates also looked at defining the distribution of HR in healthy children (Wallis et al., 2005; Wallis & Maconochie, 2006). The procedures used a finger probe monitor for obtaining HR. The authors provided rationale using a finger probe, stating there was no evidence to suggest that HR would be altered by the presence of the probe (Wallis et al., 2005). However, the authors do not explain the reason for averaging the five-second intervals over a 60-second span. Both HR and RR normative ranges were derived from statistical transformations due to skewed distributions. These authors define normative HR and RR ranges by age based on the 2.5 and 97.5 percentile values from these samples (see Table 2).
Other studies have used 24-hour electrocardiogram (ECG) monitors to gather HR data. Massin, Bourguignont, and Gerard (2005) compared the HR and rhythms of healthy ambulatory children with hospitalized children. These authors provided data on minimum, maximum, and mean HR for groups; however, children 1 to 5 years of age were presented as one group. Another descriptive study of 616 children from birth to 20 years of age used Holter monitor recordings to establish age- and gender-based "limits" (Salameh et al., 2008). Authors reported mean minimum HR and mean HR by age group; again children 1 to 5 years of age were presented as one group. No maximum HR was reported because activity was not controlled, and therefore, no "normative" ranges were given.
Other factors may influence HR in children. Fleming et al. (2011) identified several factors that influ ence HR in children. In addition to the setting and methods of measurement, their review also found the level of development of the country and year of the study to be factors in children's HR.
Racial, ethnic, and gender disparities were also found. A descriptive study found African-American children 6 to 11 years of age had significantly higher HR (p < 0.001) during sleep than Caucasian and Hispanic children when body mass indices (BMIs) were equal (Archbold, Johnson, Goodwin, Rosen, & Quan, 2010). This study also found that girls had sleeping HRs near 3.5 beats/minute faster than boys.
Studies on Blood Pressure in Children. As noted earlier, the most cited information on BP norms is from the Fourth Report. Current BP nomograms are presented to account for age, gender, and height percentiles simultaneously (NHBPEP Working Group on High Blood Pressure in Children and Adolescents, 2004). Before 1993, BP tables were presented accounting for age and gender only, but prior research suggested physical maturity (vs. chronologic age) is the primary indicator of BP (Gillum, Prineas, & Horibe, 1982). In 1993, Rosner, Prineas, Loggie, and Daniels proposed that weight was not appropriate for accounting for maturity because the inclusion of obese children in the sample might lead to "normal" values that were unhealthy.
Researchers in the United Kingdom found weight was a stronger predictor of normative BP than height when adjusting for age and gender in a sample of children and young adults ranging in age from 4 to 24 years (N = 22,901) (Jackson, Thalange, & Cole, 2007). They found that systolic blood pressure (SBP) and diastolic blood pressure (DBP) rose with age, but there was a marked increase at puberty with boys, which may have been linked with increase of weight at male puberty. Jackson et al. (2007) also redefined high BP at the 98th percentiles and high-normal between the 91st and 98th percentiles.
The issue of weight or BMI to predict higher BP, both systolic and diastolic, has been examined in the United States and China. Falkner and associates (2006) did a retrospective chart review of 18,618 children 2 to 19 years of age from pediatric primary care sites. BMI was a strong predictor of BP, even in children 2 to 5 years of age (n = 6331). Other predictors of BP were age, height, insurance type, and gender. In a sample of 208,513 young Chinese children approximately one month to 7 years of age, re searchers compared obese with non-obese children (He, Ding, Fong, & Karlberg, 2000). Differences in SBP and DBP between groups became significant at 3 years of age for both girls and boys.
Normative values for VS have been based on percentiles in the distribution of data from epidemiologic surveys. With evidence that weight and BMI factor into an increase in BP, Rosner, Cook, Portman, Daniels, and Falkner (2008) cautioned against adjusting normative values based on sampling that may include obese children.
Other Factors in Determining BP Limits. Several studies gathering ata on national normative BP values have been reported from countries as diverse as India and Saudi Arabia (Al Salloum, El Mouzan, Al Herbish, Al Omar, & Qurashi, 2009; Krishna et al., 2006). Krishna and colleagues (2006) eliminated both undernourished and obese children from the sample. Although there was no attempt to statistically compare values for Indian children to their American counterparts, Indian values for children 3 to 5 years of age were higher than those in the Fourth Report. The study from Saudi Arabia compared Saudi 90th percentile values to those of Turkish and American children and found a variance among the three samples (Al Salloum et al., 2009). A study in the United States compared rates of elevated BP between ethnic groups (White, Black, or Hispanic, by selfreport) in U.S. children (Rosner et al., 2009). This secondary analysis of the data set from the Fourth Report (n = 58,698) found that adjusting for BMI, Hispanic boys were more likely to have hypertension than were White boys (OR 1.21, p = 0.002), while Black boys were more likely to be prehypertensive than White boys (OR 1.32, p < 0.001). Girls did not have any differences between groups in hypertension after adjusting for BMI; however, Black girls were more likely to have prehypertension than White girls (OR 1.23, p < 0.001) and Hispanic girls were less likely to be prehypertensive compared to White girls (OR = 0.80, p = 0.01).
Studies have raised questions regarding diurnal changes and differences in method of obtaining BP. Lurbe and associates (1996) obtained 24 hours of ambulatory and conventional (oscillatory) BP reading on 228 normotensive children 6 to 16 years of age. The children were instructed to avoid vigorous physical activity while being monitored. Comparing average daytime BP between 0800 and 2000 to nighttime average measures taken between 2400 and 0600, Lurbe et al. (1996) found an average drop in SBP of 12.6 ± 6.7 in boys and 11.4 ± 5.7 in girls at night. DBP also dropped an average of 14.2 ± 5.9 (boys and girls combined). These significant nocturnal drops in BP were recorded in more than 80% of the children in the first phase of the study. However, the diurnal curve of BP parameters was not reproducible in a subsample of 31 children (Lurbe et al., 1996).
A difference between methods of BP measurement was also found significant. Several studies have pointed to BP measured by oscillatory were higher than those measured with ausculatory means (Lurbe et al., 1996; Midgley, Wardhaugh, Macfarlane, Magowan, & Kelnar, 2009; Park, Menard, & Schoolfield, 2005). Park et al. (2005) caution about false diagnoses of hypertension if the Fourth Report charts are used to evaluate BPs obtained with an oscillatory method. These studies point to the lack of agreement on what are considered "normal" values for BP in children and the variations due to method of measurement. A weakness in many studies is the possible variance and measurement error due to protocol around the data collection and factors, such as cuff size, position of the child, and techniques of taking a measure.
Defining Lower Limits of BP. Haque and Zaritsky (2007) further explored data from the NHBPEP to develop tables for hypotension as defined by values that fall below the 5th percentile. These authors used the given values for the 50th and 95th percentiles, with the assumption of normal distribution, to calculate the values for 5th percentile based on gender, age, and height percentile. This work was not intended for use as normative data.
The literature on early warning tools was reviewed to determine what was known about significant change in VS parameters. Pediatric early warning tools have been developed in an effort to better predict deterioration and provide care that is timely. There were multiple tools identified in the literature (Duncan, Hutchison, & Parshuram, 2006; Edwards, Powell, Mason, & Oliver, 2009; Egdell, Finlay, & Pedley, 2008; Haines, Perrott, & Weir, 2006; Monaghan, 2005; Parshuram, Hutchison, & Middaugh, 2009). These tools include varying parameters to assess deterioration of a child. The Brighton Paediatric Early Warning Score measures deterioration on three items of behavior, and cardiovascular and respiratory changes (Monaghan, 2005). For example, the respiratory item gives one point for more than 10 breaths per minute over the normal rate, two points for more than 20 above normal, and three points for five below the normal respiratory rate. Although HR and RR are included in the items for this tool, there is little information on how these cutoff points were determined or what "normal" RR and HR were used. Subsequent studies (Akre et al., 2010; Tucker, Brewer, Baker, Demeritt, & Vossmeyer, 2009) using this tool used the parameters listed in widely used pediatric nursing textbooks (T. Brewer, personal communication, May, 2009).
In contrast, the Paediatric Early Warning System Score (PEWS) and the Bristol Paediatric Early Warning Tool (PEW) have additional parameters, including oxygen therapy, demographics, medications, potassium levels, and seizure activities as indicators of deterioration (Duncan et al., 2006; Haines et al., 2006). These tools offer age-specific parameters for VS; however, only Duncan et al. (2006) report on the development of parameters, which was by modified Delphi to garner expert opinion. No research studies were cited for VS parameters in the early warning literature.
The objectives of the early warning research were directed at validating these tools rather than establishing parameters for specific VS. These studies offer some guidance on VS as indicators of deterioration. SBP was included in several tools, but only Parshuram et al. (2009) evaluated the SBP score for ability to discriminate between controls and ICU admissions. Although the SBP did not statistically predict ICU admission, the authors evaluated it to be clinically important and retained the SBP item in the tool (Parshuram et al., 2009). The Cardiff and Vale pediatric early warning system and the PEWS tools include a SBP item, but the reported analyses include only the ability of the tool as a whole to discriminate and do not address the items separately (Duncan et al., 2006; Edwards et al., 2009). The authors of the Brighton tool noted that while adult early warning tools use BP as a predictor, change in BP is considered a late sign of shock in children and was not included in the pediatric tool (Monaghan, 2005).
HR and RR are included in all early warning tools reviewed. When these parameters were evaluated for ability to discriminate deterioration, Haines and associates (2006) found that bradycardia alone was not a predictor, and tachycardia was a useful discriminator only when tachycardia persisted following a fluid bolus of 2 to 20 ml/kg. However, tachypnea was useful discriminator (Haines et al., 2006). Tume (2007) reviewed charts from children with unplanned admission to ICU or high dependency units (HDU) from the wards over a fourmonth period. This study using the Bristol tool found that tachypnea alone would have triggered the tool in 25% of those children admitted to ICU. Tume (2007) noted that respiratory distress was the main reason for admission in 55% cases of ICU admission and 54% HDU admissions.
A review of the early warning literature reveals there is lack of agreement on both the normative values for VS and what is considered a critical change in those parameters. No references to research studies were found in determining VS criteria within the early warning literature reviewed.