Prediction of bone mass gain by bone turnover parameters after parathyroidectomy for primary hyperparathyroidism: neural network software statistical analysis.

Abstract

Background: Primary hyperparathyroidism (pHPT) is the most frequent endocrine hypersecretion disease, and parathyroidectomy is the only curative option, since pharmacologic therapy reduces hypercalcemia but does not impede parathyroid hormone hypersecretion. According to guidelines from the National Institutes of Health, parathyroidectomy is associated with bone mass increase in some asymptomatic patients, while in others bone mass is not changed after surgery. Therefore, we performed the present study in an attempt to elucidate whether a preoperative biochemical bone parameter can be predictive of a significant vertebral bone mass increase in patients with pHPT. Methods: For each patient we analyzed the following preoperative parameters: parathyroid hormone, urinary calcium excretion, urinary type I collagen cross-linked N-telopeptide (NTX), osteocalcin, and vertebral computerized bone mineralography. All patients underwent vertebral computerized bone mineralography 12 months after the operation. Statistical analysis was carried out by a neural network program, an event-predicting software modeled on human brain neuronal connections, which is able to examine independent statistical parameters. Results: The patients presenting with high preoperative bone turnover (especially high NTX levels) will have a 5% vertebral bone mass gain in 83.33% of cases after surgery, independently of the National Institutes of Health guidelines. Conclusions: A high preoperative NTX level seems to be the best predictor parameter for postoperative vertebral bone mass gain in patients with pHPT. Our study also illustrates that neural network software may be a valuable method to help elucidate which pHPT patients should undergo surgical treatment