Evaluation of surgery for acromegaly: role of intraoperative growth hormone measurement?

Objective: Intraoperative growth hormone (GH) measurement has earlier been tried to improve surgery for acromegaly. We calculated GH half-life after a denomectomy and evaluated the possible role of this variable in predicting the final outcome of pituitary surgery in 28 consecutive patients with acro megaly. The sensitivity, specificity and predictive values were determined in relation to the results from GH suppression during an oral glucose load and IGF-l 3 months postoperatively. The GH half-life data were also compare d to the corresponding results obtained from GH measurements between 60 min and 180 min after adenomectomy, and early, within I week, postoperatively. Results: GH half-life less than or equal to 31 min was recorded in 8/13 cu red patients but also in 2/15 unsuccessful cases. A mean GH concentration l ess than or equal to4.4 mU/L between 60 min and 120 min after adenomectomy was found in 11/13 cured subjects but also in 3/15 not cured patients. A me an GH less than or equal to4.0 mU/L between 90 min and 180 min was found in 11/13 cured and in 4/15 not cured patients. A mean early postoperative GH concentration less than or equal to2.6 mU/L was noted in all 13 cured patie nts, but also in 2/13 unsuccessful cases. The specificity of early postoper ative GH less than or equal to2.6 mU/L was 100% compared to 62% for a GH ha lf-life less than or equal to 31 min (p < 0.05) and 85% for the GH mean val ues between 60 min and 120 min and 90 min and 180 min, respectively. The se nsitivity for persistent disease of values above the four cut-off limits us ed was between 73% and 87%. The positive predictive value for a mean early postoperative GH value >2.6 mU/L was 100%, and 72% for a GH half-life > 31 min (n.s.). Conclusion: Although intraoperative GH half-life might be usefu l in some cases, it was not a reliable tool for predicting outcome of pitui tary surgery in acromegaly. In cases with a 50% decrease of a basal GH conc entration >5.5 mU/L, mean GH values less than or equal to4 to less than or equal to4.4 mU/L late intraoperatively were more informative but not as goo d as those obtained from the mean of a series of GH values drawn on one occ asion within 1 week postoperatively, offering a 100% specificity for cure i f less than or equal to2.6 mU/L. Intraoperative GH half-life measurements s hould therefore be used with caution. The predictive values of the cut-off limits used in this study should be further evaluated before general applic ation.