ObjectiveTo discover the indicators and develop a model for predicting protracted hypoparathyroidism (HPT) after thyroid cancer surgery in order to guide the early therapy for patients with HPT.MethodsThe clinical and postoperative pathological data of patients with thyroid cancer who received surgical treatment in the Xuanwu Hospital and Beijing Pinggu Hospital from January 2019 to December 2020 were retrospectively analyzed. The potential indicators of postoperative HPT and protracted HPT were analyzed by logistic and LASSO regression analysis. A nomogram for predicting protracted HPT was constructed in the training set, and the discrimination and consistency of the nomogram were verified in the training set and the validation set respectively.ResultsAccording to the inclusion and exclusion criteria, a total of 464 patients diagnosed with thyroid cancer were finally included in the study. Among the 100 patients with postoperative HPT (except 1 case of incomplete data), 62 patients showed short-term HPT and 37 patients developed protracted HPT. Multivariate logistic regression analysis showed that the preoperative intact parathyroid hormone (iPTH) level [OR=0.953, 95%CI (0.931, 0.976), P<0.001], lobectomy with contralateral partial lobectomy [OR=3.247, 95%CI (1.112, 9.485), P=0.031], and total thyroidectomy [OR=11.096, 95%CI (5.432, 22.664), P<0.001] were related to postoperative HPT. The multivariant logistic regression analysis revealed that postoperative iPTH level was a predictive factor for protracted HPT [OR=0.719, 95%CI (0.588, 0.879), P=0.001]. The area under receiver operating characteristic curve (AUC) value of postoperative iPTH level in predicting protracted HPT was 0.848 [95%CI (0.755, 0.942)]; The cut-off value was 9.405 ng/L, and its specificity and sensitivity were 0.659 and 0.944, respectively. Moreover, the AUC value of the nomogram model including postoperative iPTH level and other clinicopathologic features (extraglandular invasion, cumulative maximum tumor diameter, and central lymph node dissection) for predicting protracted HPT was 0.900 [95%CI (0.817, 0.982)]; The cut-off score was 118.891, and its specificity and sensitivity were 0.772 and 0.944, respectively; The Hosmer-Lemeshow goodness of fit test indicated good fit of nomogram (χ2=8.605, P=0.377). The AUC value of the nomogram was 0.640 [95%CI (0.455, 0.826)] in the validation set (Pinggu Hospital data). The Hosmer-Lemeshow goodness of fit test also indicated good fit of nomogram (χ2=12.266, P=0.140).ConclusionsThe postoperative iPTH level is an important influencing factor of protracted HPT. The nomogram prediction model based on postoperative iPTH level and other clinicopathologic features has a favorable predictive value for protracted HPT.