Objective: To assess the predictive value of postoperative thyrotrophin releasing hormone (TRH)-induced GH responsiveness in relation to (late) postoperative outcome in patients in remission after surgery for acromegaly.
Patients: and methods One hundred and twenty-nine patients underwent surgery for acromegaly in our institution between 1977 and 1996. TRH tests and oral glucose tolerance tests (GTT) were performed and serum IGF-I concentrations were measured pre- and postoperatively and during follow-up. Criteria for postoperative remission were a mean serum GH concentration < 5 mU/l and/or serum GH after an oral glucose tolerance test < 1 mU/l immunofluorometric assay (IFMA) or < 2.5 mU/l (radioimmunosassay), together with a normal serum IGF-I concentration.
Results: Preoperatively, the TRH-induced GH response was highly variable, with gradual overlap between 'nonresponders' and 'responders'. Arbitrarily defined as a doubling of serum GH concentration, 45.6% of patients were 'responders' to TRH. GH response after TRH injection was significantly correlated to the TRH-induced prolactin response but not to preoperative GH concentration or adenoma size. After surgery, remission was achieved in 83 of the 129 patients. Postoperative remission was significantly correlated to mean preoperative serum GH concentration and preoperative glucose-suppressed serum GH but not to tumour class. Seventy-one patients with early postoperative remission were followed without adjuvant treatment for a mean of 9.4 +/- 0.7 years (range 0-23 years). Forty-one of these patients were TRH responsive as defined by at least doubling of the serum GH concentration preoperatively. Of the 71 patients, 12 developed recurrence of disease, as defined by insufficient GH suppression during oral GTT, and elevated IGF-I and mean serum GH concentration. Irrespective of the preoperative response to TRH, the initial postoperative TRH test was predictive of developing disease recurrence with a sensitivity of 75% and a specificity of 100% when an absolute GH increase of 3.75 mU/l was chosen to define paradoxical responsiveness. A stimulated GH 1.6 times basal was predictive of recurrence with a sensitivity of 83% and a specificity of 73%, and 2.1 times basal was predictive with a sensitivity of 75% and a specificity of 80%. None of the 32 patients with postoperative normalization of the preoperatively present TRH-induced GH response, defined as a postoperative GH increase < 3.75 mU/l, developed recurrence of disease, while all nine patients with a GH increase above this level developed recurrence of acromegaly.
Conclusion: To our knowledge this is the first report which addresses the value of early postoperative TRH-induced GH responsiveness in predicting late surgical outcome using receiver-operating characteristic (ROC) curves to redefine a postoperative paradoxical response instead of arbitrarily chosen criteria. An absolute postoperative GH response of more than 3.75 mU/l was associated with recurrence in all nine patients, while all 32 patients with normalization of previously paradoxical response are still in remission. Our findings from this study demonstrate that the TRH test is a valuable tool in the early identification of patients at risk of developing postoperative recurrence of acromegaly.