Article Text
Abstract
Objectives To gather data on the ages and weights of children aged between 1 and 16 years in order to assess the validity of the current weight estimation formula ‘Weight(kg)=2(age+4)’ and the newly derived formula ‘Weight=3(age)+7’.
Design Retrospective study using data collected from paediatric attendances at an emergency department (ED).
Setting A large paediatric ED in a major UK city.
Patients 93 827 children aged 1–16 years attending the ED between June 2003 and September 2008.
Main outcome measures Percentage weight difference between the child's actual weight and the expected weight, the latter determined by ‘Weight(kg)=2(age+4)’ and by ‘Weight(kg)=3(age)+7’, in order to compare these two formulae.
Results The weights of seriously ill children were recorded in only 20.5% of cases, necessitating a weight estimate in the remainder. The formula ‘Weight=2(age+4)’ underestimated children's weights by a mean of 33.4% (95% CI 33.2% to 33.6%) over the age range 1–16 years whereas the formula ‘Weight=3(age)+7’ provided a mean underestimate of 6.9% (95% CI 6.8% to 7.1%). The formula ‘Weight=3(age)+7’ remains applicable from 1 to 13 years inclusive.
Conclusions Weight estimation is of paramount importance in paediatric resuscitation. This study shows that the current estimation formula provides a significant underestimate of children's weights. When used to calculate drug and fluid dosages, this may lead to the under-resuscitation of a critically ill child. The formula ‘Weight=3(age)+7’ can be used over a larger age range (from 1 year to puberty) and allows a safe and more accurate estimate of the weight of children today.
- Child
- weight
- estimation
- formula
- resuscitation
- anaesthesia
- intensive care
- paediatrics
- paediatrics
- paediatric emergency med
- paediatrics
- paediatric resuscitation
Statistics from Altmetric.com
- Child
- weight
- estimation
- formula
- resuscitation
- anaesthesia
- intensive care
- paediatrics
- paediatrics
- paediatric emergency med
- paediatrics
- paediatric resuscitation
Introduction
There are many reasons why weighing an unwell child is not possible. It may be that the child cannot be moved due to trauma or pain. Active resuscitation of the child might preclude attempts at measuring weight. However, the weight of children provides the cornerstone of their management—determining, as it does, the fluid quantities, drug dosages, DC shock energy and ventilator tidal volumes. Currently, the formula used by the advanced paediatric life support (APLS) course estimates weight of a child between 1 and 10 years and is in use in Europe, South Africa, Australia and New Zealand.1 This APLS formula is given by ‘Weight (in kg)=2×(age in years +4)’. The age is the child's age in whole years (ie, the child's age at last birthday).
Several recent studies have questioned the validity of this formula, showing that it significantly underestimates children's weight.2 A revised formula from the Nottingham Paediatric Weight Study (NPWS) has been proposed,3 ‘Weight (in kg)=3×(age)+7’, and the age is ascertained as before. Given the previous study findings, the questions asked in this paper are:
Does the new formula still provide a better estimate than the current formula in a new study population?
Is the new formula applicable to a larger age range?
Does the new formula provide accuracy in the most urgent cases?
Methods
With the assistance of a medical statistician, a sample size calculation was carried out to establish the minimum number of children needed to provide sufficient data to enable individual age group analysis. Thus, using a 5% level of statistical significance with power of 80%, this was determined to be 400; 258 953 sets of data were collected. Details of weight, age at presentation, sex and emergency department (ED) streaming category (table 1) were collected retrospectively from June 2003 to September 2008 from the database at the ED, The Sheffield Children's Hospital, Sheffield UK. These data were recorded mainly by the triage nurse and occasionally by other nursing staff or doctors. The children were weighed on a Seca 861 floor-mounted scale (Seca, Hamburg, Germany) or a Marsden MPDC 250 chair scale (Marsden, Henley on Thames, UK). A Weymed BMI 30 (H Fereday and Sons, Harlow, UK) was used for infants and smaller children. All of the scales were regularly calibrated. Weights were recorded with children wearing a minimum of clothing and without shoes. Removal of data outliers occurred for points lying outside 4SD and subjectively if it was clear that data had been entered incorrectly, were missing or corrupted. The raw data included children aged <1 year and these were not used in the final analysis. The ages 1–16 years inclusive were selected to look at a wider age range than the 1–10 years covered by the APLS formula. Male and female subgroups were examined individually to assess whether any formula would be applicable to both.
Results
A total of 164 724 sets of data included children aged <1 year, missing or corrupted data and were not used in the final analysis (see Appendix 1 for breakdown). The remaining 94 229 complete sets of data covered the ages 1–16 years inclusive. The number of data outliers removed totalled 402 and this paper used only the weight and age data of the remaining 93 827 children.
In triage categories A and B (‘critically ill’ children) there were 345 patients. Seventy-one of these had weights recorded. The percentage of weights recorded by ED triage category is shown in table 1.
When considered as subgroups, the weights of boys and girls were statistically different (table 2). Boys had a 1.8% higher mean weight and their mean age was older than the mean age for girls.
For the individual age group analysis (see table 3) the new formula was the same as the current formula for children aged 1 year and more accurate at all other ages, although it overestimated weights in those aged 4–8 years (maximum 5.31% at 5 years). The new formula was more accurate in age groups outside the original 1–10 years.
For the age range 1–10 years inclusive, a mean underestimate of 21.9% (95% CI 21.7% to 22.1%) was found when comparing actual weights with those given by the APLS formula. Using ‘Weight=3(age)+7’, there was a mean underestimate of 2.0% (95% CI 1.9% to 2.2%) for the same age range. For the age range 1–13 years inclusive the results were 29.0% (95% CI 28.8% to 29.2%) and 4.8% (95% CI 4.6% to 5.0%), respectively (table 4). Also shown are the results for the age group 1–16 years. These older age groups are shown to check accuracy outside the previous 1–10 years age range. The 11–13 years age group is highlighted to compare the ability of the two formulae to predict weights throughout puberty. The accuracy of both formulae for critically ill children (triage categories A and B) is also shown in table 4.
Figure 1 shows a graph of these results. The inflexion point at age 16 reflects the paucity of data for 16-year-old children attending a children's ED. The average for age groups 1–15 years was 6248 children whereas there were just 102 children aged 16 years.
Discussion
The formula ‘Weight=2(age+4)’
A literature search found no evidence supporting the formula ‘Weight=2(age+4)’. When comparing weights of post-war children given in the Textbook of Paediatrics published in 1952,4 this formula works well. However, a formula applicable to post-war children is unlikely to apply today, and both this study and the NPWS confirm this. The APLS formula was assessed by several other papers, all of which found it to be a significant underestimation of the true weight, the error becoming greater as the age and weight increased. Carrol et al showed the APLS formula as having a mean difference of 24% in 169 clinic patients, but stated no age ranges or whether this was an over or underestimate. Argall et al found the APLS formula to have a mean underestimate of 3.52 kg. However, this was not expressed as a percentage difference so is difficult to interpret.5 Black et al found the APLS formula to have a mean underestimate of 4.7% in the 10–25 kg group, 20% in the 25–40 kg and 42% in those >40 kg.2 Luscombe6 studied anaesthetic cases in New Zealand and the UK and found the APLS formula had a mean underestimate of 24.7% and 18.5%, respectively. The NPWS looked at 17 000 ED children retrospectively (UK) and found the APLS formula to have a mean underestimate of 18.8% (95% CI 18.42% to 19.48%). Breakdown by years showed it to underestimate weights at all ages with a steady rise from 11.47% aged 1 year to 38.26% at 10 years. It is only recommended for use up to the age of 10 years, further limiting its usefulness.
Concerning obesity
Children's weight is increasing.7 It is easy to confuse increasing weight due to better nutrition with obesity, especially as the popular press focuses on the extremes. The formula ‘Weight=3(age)+7’ was derived using a linear regression analysis on just two standard deviations to exclude the effect of any overweight or underweight children. In the NPWS data, ‘cleaning’ involved removing weights more than four standard deviations from the mean. In this study, ‘cleaning’ was also used for each age group. Hence, misleading extremes were excluded from the derivation of the formula and subsequent analysis.
The APLS formula has clearly become a victim of better nourished children. With a mean underestimate of more than 20% (nearly 40% at age 10 years), its place as a weight estimation tool is questionable.
How best to estimate?
A formula-based method is one option for weight estimation; however, several others have been suggested. The Broselow tape has stood the test of time and is particularly useful if the age is not known as a length (height)-based estimate has established accuracy.8 It requires a tape or similar measuring system to be present at the time of the resuscitation which may not always be available (eg, in the prehospital environment). Accurately measuring the length of a child may pose problems in a resuscitation scenario. Simply estimating by looking has been shown to be poor,9–12 although a recent article found parental estimation to be more accurate than either the APLS formula or the Broselow tape. However, it may be difficult and unfair to expect parents to safely state their child's weight during resuscitation, and the study only used clinically stable children, excluding emergency cases.13 The use of growth charts to estimate weights by using the 50th centile has shown better accuracy than the current formula14 but, once again, requires the information to be readily available. Other formulae such as the mid-arm circumference or shoe size-based systems are available but not straightforward.15 Any new formula should be more accurate, be safe from large overestimates in weight and be simple to calculate. These three criteria were pivotal in the derivation of the formula ‘Weight=3(age)+7’.
Is a calculated formula useful?
A calculated formula remains helpful. It can be used when no equipment is available as, for example, in the prehospital setting. Also, it allows calculations to be made before the child's arrival at the ED, enabling advance drug/fluid preparation. Clearly the age must be known, and UK prehospital practitioners telephone the age to ED staff when transferring critically ill children. However, there are concerns about calculation errors when using formulae.16 While human error cannot be avoided whatever system is used, great care was taken by the authors of the NPWS to produce a simple new formula. Paediatric resuscitation is stressful and complex calculations are best avoided. Indeed, a team from Brisbane derived three separate formulae (known as the ‘Best Guess’ formulae) depending on the age of the child.17 The authors feel that replacing one formula with several only increases the potential for error. In another small study a ‘devised weight estimation method’ was proved comparable to the Broselow tape.18 However, it involved an estimate of body habitus combined with a height measurement to produce a weight, adding an extra level of complexity for similar accuracy.
Gender and the applicable age range
The age range of 1–10 years limits the usefulness of the APLS formula. Its accuracy is poor in older age groups within its age range and effectively unusable outside this. The current resuscitation guidelines state that ‘the onset of puberty, which is the physiological end of childhood, is the most logical landmark for the upper age limit for use of paediatric guidelines.’19 The validity of the formula ‘Weight=3(age)+7’ was therefore checked for ages up to (and beyond) late puberty. Pubertal ages are subject to wide variation but may be thought of as between 8 and 13 years in girls and between 9 years 6 months and 13 years 6 months in boys.20 The new formula performs well over these age ranges and can be considered as applicable from 1 year of age until puberty.
There was a statistical difference between the weights of boys and girls. However, the mean weight difference is not clinically significant at just 1.8% and may simply reflect the slightly higher mean age of the boys in the sample (see table 2).
The critically ill child
Thankfully, the number of critically ill children presenting to the ED is small, so there were a very small number to study. As stated, weighing these children is difficult. It is of paramount importance that a weight estimation tool is applicable to this group. The new formula predicted their weights well whereas the APLS formula showed a large weight underestimate (28.7%).
Limitations of study
It is not known if these data apply to all ethnic groups in the UK as ethnicity was not recorded. This data sample is from one UK city, although the NPWS showed similar results. This study may not apply to the developing world where disease and poor nutrition are prevalent. The study did not compare the formulae with other methods of weight estimation. Both formulae give the same result at age 1 year, with an underestimate of 19.6%. The re-attendance rate is 5.26% but, owing to the anonymity of the data, we were unable to identify and remove repeat attendees. It is not clear (and may be difficult to show) whether a more accurate formula will alter clinical outcome.
Conclusions
Paediatric weight estimation has been scrutinised in ever more detail. The ‘gold standard’ is to weigh the child, but this is difficult for critically ill children. Accurate weight estimation is therefore paramount. From this, drug doses and other parameters are derived which are vital for the optimal resuscitation of a critically ill child. While any method will only provide an estimate, the following points need emphasising:
A simple formula remains a useful method to estimate children's weights.
The APLS formula provides a poor weight estimate. To continue with an inaccurate formula with no evidence base cannot be considered good medical practice.
The formula ‘Weight=3(age)+7’ works simply, safely and accurately from ages 1 to puberty, and is a validated evidence-based system to estimate the weight of children today.
Acknowledgments
The authors acknowledge Dr Tracey Young (Trent RDSU) for providing statistical advice, Phillip Richardson (Sheffield Children's Hospital) for collecting the data, and Nicola Luscombe (emergency nurse practitioner) and Suzanne Owens (emergency nurse practitioner) for proof reading and advice.
Appendix
References
Footnotes
Funding MDL received a research grant from Doncaster and Bassetlaw NHS Foundation Trust.
Competing interests None.
Ethics approval Ethical permission was obtained from the North Sheffield Research Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.