Page 2 of Standardization of Body Surface Area Calculations

Back to BSA report page 1. of A report by Thanh Vu, B.Sc. (pharm),

( click here to go to the Body Surface Area BSA calculator )

Currently several institutions use the DuBois formula.  There are several arguments for and against the use of the Du Bois formula for calculating body surface area.  In favor of the Du Bois formula is the fact that the nine subjects used to derive it were of diverse body shapes.  However, many investigators have since questioned the accuracy of the formula.  Results from a study by Jones et al⁵ indicate that the equation is not the most accurate of those available.  Jones et al noted that for their formula DuBois and DuBois only measured on leg and one arm, assuming the body to be symmetrical.  This assumption has since been found to be invalid⁶. Jones et al also reported that the formula grossly overestimated the surface area of obese people.  Mitchell et al⁷ compared the "measured" surface area with predicted surface area from the Du Bois formula and showed that although the two methods were highly correlated, the predicted area was systematicall lower than the measured area. This underestimate was greater in small individuals, thus the authors suggested that the Du Bois formula may be invalid when the predicted surface area is less than 1.3 m² (i.e. in children).  Haycock et al⁸ reported that the Du Bois formula underestimated surface area up to 8% in infants, especially as values fell below 0.7 m².

The above results were supported by Wand and his colleagues in 1992.  Since the Du Bois formula was derived from a small number of subjects, several of whom had skeletal deformities, and included only one child,  Wang et al questioned the validity and accuracy of the formula for infants and for individuals with extremes of stature, such as those who are pregnant, obese, or very tall.  Wang et al assessed the Du Bois formula for accuracy using the Root Mean Squared Error (RMSE) method of prediction.  The RMSE measures concordance between measured and predicted data.  They found that the Du Bois formula systematically underestimates BSA by almost 5% and that the tendency to underestimate BSA is slightly greater in infants than in others, but this difference is not likely to be clinically relevant⁹.

The results from the RMSE method of preduction by Wang et al showed that of the 15 different formula for predicting BSA, 8 have an RMSE less than 8%. The Du Bois formula ranks fifth for the lowest RMSE (Table 1).   Despite this, the Du Bois formula continues to be used, probably in regard more to tradition than to accuracy, and a number of major drug producers still provide their clients with nomograms based on this formula.  In addition, as Turcotte observed, nomograms printed in several standard texts were reproduced incorrectly from the original, leading to systematic under-estimation of surface area.  The compounded risks of using such nomograms may be quite large.  To put this in perspective, a change in BSA from 1.87 to 1.60 is equivalent to a weight loss of 22kg in a woman who weight 80.5kg and is 158cm tall.  This would correspond to a change in a doxorubicin dose (50mg/m²) from 93.5mg to 80mg (Table 3).

It is particularly important to standardize BSA calculation at an institution which conducts clinical trials.  There are several tools available to aid in the calculation of BSA including slide rules, nomograms, purpose built calculators and software.  Several types of slide rules and calculators are currently in use at the Cross Cancer Institute (CCI).  The way BSA is determined at the CCI is not very consistent.  Some health care professionals use a BSA calculator while others use a slide rule or nomogram.   In order to maintain consistency, the way BSA is determined at CCI would need to be standardized, and pharmacy provided with a patients height and weight so that the dose may be confirmed.  In the pharmacy, BSA should be calculated to three significant figures.  Slide rules and nomograms are incapable of calculating with this degree of accuracy.  In addition, they suffer from error associated with their analog nature and the formula they are based on.  For example, after studying a three-scale nomographic chart designed by Boothby and Sandiford, which was based on the Du Bois formula, Turcotte determined that BSA values obtained from it were smaller than Du Bois values by 8%¹⁰.  After being shifted upward, leading to an 8% underestimate of surface area for all values of height and wieght.  Recalling that the Du Bois formula underestimated actual BSA by 5%, according to Wang et al, it is clear that this tool underestimates BSA by 13%.  Purpose built calculators are more accurate in their calculations, but are still hampered by reliance on inaccurate formulae.

The solution for an institution that enters all of its orders on a central computer system available to pharmacists, nurses, technicians, and physicians alike is simple.   Software is available or can be developed that uses not only a reliable formula, but calculates BSA to three significant figures with unwavering certainty and accuracy.

This BSA report continues on page 3

Back to the Index of calculators of ideal weight, body mass index and breast cancer risk.