Quantification of Error in the Calculation of ZScores in Neonates

We agree with the conclusions in the article by den Dekker et al . regarding the caution with which Z scores should be interpreted, particularly in neonates. It is well established that widely differing results are generated by different Z -score algorithms, but it is also important to quantify the errors created by body surface area (BSA) estimation and observer error.

The generic use of the Haycock algorithm for the calculation of BSA risks misinterpretation of the population data, and this is of particular importance when allometric calculations are based on a single marker of body size. As the authors point out, the algorithm of Du Bois and Du Bois will underestimate BSA in neonates (see Figure 1 ) compared with the Haycock formula. Pettersen et al . used the formula of Du Bois and Du Bois, and use of the Haycock formula will produce a higher estimate of BSA and consequently a disproportionately low Z score. The error is exaggerated for “stockier” neonates, with higher weight-for-length centiles, and a difference of up to 10% in estimated BSA is common ( Figure 1 ). A commonly used “weight only” estimation of BSA (BSA = 0.1023 × weight 0.68 ) results in a similar magnitude error in the other direction. In real terms, this will result in variations of a mitral valve Z score of up to 0.6 simply by choice of BSA algorithm.

Figure 1

Percentage error of the BSA formulas commonly used in the neonatal population, plotted against body length. The percentage error is the percentage by which the calculated BSA would be altered from the value calculated by the Haycock algorithm (at the same weight centile) if one of the alternative formulas were used (Du Bois and Du Bois, Boyd, Mosteller, or weight only). A positive value indicates that the alternative formula computes a higher BSA than the Haycock formula. Weight centiles are calculated using the weight-for-length centiles produced by the World Health Organization data. Errors are proportionately greater for all formulas in the patients with higher weight for length ( dashed red lines ).

Only gold members can continue reading. Log In or Register to continue

May 31, 2018 | Posted by in CARDIOLOGY | Comments Off on Quantification of Error in the Calculation of ZScores in Neonates

Full access? Get Clinical Tree

Get Clinical Tree app for offline access