Concordance between GH determination and IGF-I in acromegaly using two IGF-I methods
S Hepburn, SM Chambers, JA Gilbert, AM McGregor, JP Miell & SJB Aylwin
Introduction and aims: Following treatment for acromegaly, both growth hormone (GH) and insulin-like growth factor-I (IGF-I) levels are predictive of mortality. These data are derived from studies of either a single GH or the mean circadian GH, with a threshold of 2 mcg/l. However, consensus target criteria (Giustina et al. 2000) require: a nadir GH of <1 mcg/l on OGTT and IGF-I within the age/sex-adjusted normal range. We aimed to determine the degree to which normalized IGF-I was concordant with basal GH (<2 mcg/l), OGTT mean GH (<2 mcg/l) or nadir GH (<1 mcg/l). In addition, we have examined the performance of two IGF-I assays in their degree of concordance with GH parameters.
Methods: Basal, mean and nadir GH values were determined in 75 consecutive GH/OGTT evaluations amongst 55 acromegalic subjects. Serum IGF-I (n=37) was measured using a DSL immunoradiometric assay before Nov 2004 and subsequently (n=38) using a Nichols immunoassay. Comparison of IGF-I (expressed as % of upper limit of the age/sex-adjusted reference range) and GH determinations were made by log-linear regression. Contingency tables were derived to establish the concordance between IGF-I and GH remission.
Results: Strong correlations were found between IGF-I and basal, mean and nadir GH with the strongest for nadir GH (r=0.65, P<0.0001). The regression coefficients were equivalent for IGF-I values determined by both assays. However, when a contingency was performed there was a discrepancy between the two assays. The Nichols assay had improved concordance between IGF-I and nadir GH (3/38 had high IGF-I but nadir GH <1 mcg/l). In contrast, the DSL assay produced 9/37 similarly discordant results.
Conclusions: A nadir GH <1 mcg/l on OGTT showed the greatest correlation with IGF-I. However, although IGF-I results from two assays were concordant over the assay range, at the threshold between active disease and remission there was a magnified discordance between assays.