|Home | About | Journals | Submit | Contact Us | Français|
We acknowledge the suggestions provided by Jorge and Arnhold for improvement of the score for identifying subjects eligible for SHOX testing. These authors refer to two publications, those of Binder et al1 and Jorge et al.2 Both publications are based on a small number of subjects, less than a tenth of our sample size, investigating the value of anthropometric measurements for identifying subjects who should be tested for SHOX haploinsufficiency. In principle, results are consistent between the three publications.
Binder et al suggest using the limb:trunk ratio ((leg length + arm span)/sitting height) as a criterion. Leg length was calculated as height minus sitting height. In this approach, the deficits in limb length are additive, which may add to the diagnostic accuracy and the positive predictive value. In contrast, we used the arm span:height and sitting height:height ratios, staying with the raw data.
Jorge et al suggest using the sitting height:height ratio expressed as standard deviation scores (SDS) referring to a Dutch reference population. However, it is unclear how well the Dutch population, which is known to be among the tallest worldwide, reflects other ethnicities.
We agree with Jorge and Arnhold that the sitting height:height ratio and also the arm span:height ratio are age‐dependent. This age dependence is especially obvious in both boys and girls <3 years of age. For the age range of 3–10 years, as in our study, there is some further decline for sitting height:height (from 57% to 53%; 50th centile) and an increase of arm span:height (from 96% to 100%). The scoring system was developed for a population with a mean age of about 7.5 years, which means that very young children may be less likely to receive score points for these two criteria; this however, does not a priori exclude them from SHOX testing.
The relatively small number of subjects with SHOX deficiency (n=68) did not allow us to divide this population further according to age. This would have made sensible statistics impossible. However, we accept the suggestion by Jorge and Arnhold to include both the age‐dependence of sitting height:height and arm span:height and also the extremity:trunk ratio suggested by Binder et al in future analyses of sufficiently large numbers of subjects, which will hopefully be available from Lilly's observational study GeNeSIS (Genetics and Neuroendocrinology of Short Stature International Study).
We applied the variables as listed in our suggested score system for two main reasons: (1) use of raw data (eg arm span, sitting height and height, all in cm) makes the use of the score system easy in daily practice; and (2) converting absolute values to SDS always introduces an additional source of error, introducing a measure (SD) that changes non‐linearly with age and depends on the reference population, which may not be available to the physician.
Competing interests: None declared.