The APLS formula to calculate weight in children is a commonly used method, especially for critically ill children in whom it is impractical or unsafe to acquire weight on a scale. During resuscitation of children, weight is used to guide drug dosages, intravenous (IV) fluid boluses, equipment size, defibrillation and cardioversion dosages. As described in the introduction, there are a number of methods used to estimate weight in children [3–9]. However, At the Eric Williams Medical Sciences Complex, the most commonly used method is the APLS formula.
It has been shown by several studies that the original APLS formula underestimates weight, however all of these studies have been performed on non-Caribbean populations [3–7]. In 2007, Luscombe and Owens examined data from over 17000 children and found that the APLS formula was found to have underestimated weight by a mean of 18.8% [3]. Several subsequent studies in Australia and the United Kingdom also demonstrated the tendency for the APLS formula to underestimate weight in children in developed countries [4–6, 9]. This included a review of 93827 children over a 5 year period from 2003 to 2008 by Luscombe et al. [10]. In light of this, the most recent edition of the APLS manual recommends the use of the Luscombe and Owens formula in children aged 6 – 12 years old, with retention of the original APLS formula for those aged 1 – 5 years.
In India, however, Varghese et al. examined 500 outpatient children and found the APLS formula to overestimate weight in their population by a mean of 2–3 kg [7]. In addition, a 2010 study of the accuracy of various weight estimation methods in South African children concluded that the APLS formula and Broselow tape were more accurate than the Luscombe and Owens formula over all age groups in this population [11].
The Broselow tape was designed for use in children from 45 cm to 145 cm in length [5]. Although it has been validated in several studies as a reliable tool for estimating weight [5, 7, 11–13], it is not commonly used in Trinidad. One of the largest studies of the Broselow tape was performed by Lubitz et al. in the United States of America [1]; out of 937 patients it was found that 79% of patients had estimated weights using the Broselow tape which were within 15% of their actual weights. Krieser et al. compared the Broselow tape to the APLS formula and found it to be more accurate, with 61% of patients having Broselow weights within 10% of actual weights compared to 34% of patients having APLS weights within 10% of actual weights [5]. Interestingly, Cattermole et al. found mid-arm circumference to be more accurate and precise than age-based rules for predicting weight in school aged children, and as accurate as the Broselow tape [14].
Krieser et al. [5] showed that parental estimate of children’s weight was a reliable method of weight estimation; 78% of the 410 children studied had an estimated weight within 10% of their actual weight and the mean difference between estimated and measured weight was −0.6 kg [5]. A previous study performed by Harris et al. had demonstrated that out of 100 children from 0–8 years of age, 84 had estimated weights within 15% of actual weights when parental estimate was used [15]. Leffler et al. also demonstrated that parental estimate was within 10% of actual weights in 80% of cases [16] and Goldman et al. demonstrated that parental estimate was within 10% of actual weights in 73% of cases [17]. This method would have to be tested further in a Trinidadian population to determine whether it is as accurate as in other countries.
It is clear from the previously cited studies from India and South Africa that age-based formulae for estimating weight may not be applicable to developing countries, and that studies on these formulae will yield differing results in different settings, given the variation in body habitus between children from developed and developing countries. In light of this, formulae used in the first world for weight estimation should be tested before they are adopted in developing countries, such as Trinidad.
This study showed that the APLS formula did not significantly underestimate weight in the 1–5 year age group compared to other formulae. This is in contrast to the evidence that has been emerging worldwide, where there has been a tendency for the APLS formula to significantly underestimate weight. This may be in keeping with the UNICEF progress for children report, which found a larger proportion of underweight Trinidadian children than in first world countries [18]. This and other evidence suggests that Trinidadian children in this age group weigh less than their first world counterparts [19].
The Luscombe and Owens formula was no more accurate at estimating weight than the APLS formula in our population. The APLS formula was also found to be marginally more precise than the Luscombe and Owens formula, with 45.6% having estimated weight within 10% of measured weight using the APLS formula as opposed to 42.3% using the Luscombe and Owens formula. The new derived formula ([2.5 × age] + 8) was more accurate than either the APLS or Luscombe and Owens formula. However, the overall accuracy and precision of all three formulae were not found to be significantly different.
These findings suggest that the APLS formula is acceptable for use in the 1–5 year old age group. Although the new formula was slightly better than the APLS formula, it is the recommendation of the authors that the APLS formula be retained for weight estimation. This is because the APLS formula is already familiar to medical staff in Trinidad and Tobago, and the ease of recalling a familiar formula would make it a more practical choice. This is also in light of the fact that adopting the new formula would not produce a significant improvement in weight estimation.
The study has several limitations. Firstly, ethnicity, gender and socioeconomic status were not taken into account. While there may have been some variation in weights based on these factors, it was thought unnecessary to analyse these subgroups separately, as it is unlikely that separate formulae for each of these categories would be practical for use in the emergency situation. In addition, the one study of the relationship between ethnicity and weight in children in Trinidad did not show any significant difference [20].
Secondly, the study did not include children aged 6 years or older, so it is not known whether the formula would be applicable to older children. The younger (pre-school) age group was specifically investigated in this study, as these children make up the majority of patients presenting to the Paediatric Emergency Department. In addition, the authors felt that these smaller, younger patients were more likely to suffer ill effects of miscalculation of dosage of medication than the older age group. However, it is well accepted that children’s weights do not bear a linear relation to age, and it would be necessary to perform a similar study on older children (6 – 12 years) to test the accuracy and precision of the various weight estimation formulae on this age group.
Finally, the study was restricted to the island of Trinidad. While the results of this study may be generalisable to the rest of the Caribbean region, the authors intend to do a more extensive study of weight estimation in children across the Caribbean.