Auxology-based cut-off values for biochemical testing of GH secretion in childhood.

OBJECTIVE

The diagnosis of GH deficiency (GHD) in childhood requires GH tests with arbitrary cut-offs. We aimed to define GH cut-offs based on auxology.

DESIGN

From a total of 349 children diagnosed with GHD between 1985 and 2005 at our hospital, we excluded all children who had additional characteristics likely to interfere with growth velocity. Age at start of therapy was defined as 4 to 8/9 years (girls/boys). Auxological inclusion criteria were pathological growth velocity, height at start of therapy >1.5 SD below the target, and efficient catch-up growth during GH therapy. Basal IGF-I/IGFBP-3, GH response to arginine and spontaneous GH secretion at night had been measured by the same polyclonal RIA. The reference was a group of 108 normally growing age-matched children with Turner syndrome or born small for gestational age tested during the same time period.

RESULTS

We identified 52 children with GHD who fulfilled the inclusion criteria. ROC analysis showed the best diagnostic accuracy at a peak GH cut-off for arginine of 6.6 μg/L (sensitivity, 84.3%; specificity, 75.5%; AUC=0.83) and at a peak GH cut-off during spontaneous secretion at night of 7.3 μg/L (sensitivity, 96.8%; specificity, 82.4%; AUC=0.93). Our arbitrarily defined GH cut-offs had been higher. Children diagnosed with GHD in the past with GH test values above the new cut-offs were less responsive to GH therapy (P=0.007).

CONCLUSIONS

Here we provide a new rational approach which allows the substitution of arbitrarily defined GH cut-offs by those based on auxology.