Predicting Postthyroidectomy Hypocalcemia: Improving Predictive Ability of Parathyroid Hormone Level

Abstract

We performed a retrospective study of our large patient cohort aiming to examine the ability to predict postthyroidectomy hypocalcemia. A retrospective review of patients who underwent total or near-total thyroidectomy at our institution between 2008 and 2018 was performed. Postoperative hypocalcemia was defined as Ca level <8.0 mg/dL (reference range: 8.9–10.1 mg/dL) within 30 days of operation. Logistic regression was used to develop models for prediction of the occurrence of postoperative hypocalcemia. Inclusion criteria were met by 1463 patients. Hypocalcemia was documented in 223 patients (15%). Models based on parathyroid hormone (PTH) levels alone had an associated receiver operator characteristic with an areas under the curve (AUC) of 0.79. There was an inverse relationship between time of measurement and PTH levels within the first two hours after thyroidectomy ( P < 0.01). When measured two to six hours after closure, the predictive ability of PTH compared favorably (AUC = 0.82) with either earlier (within the first two hours after closure, AUC = 0.79) or later measurement (6–24 hours after closure, AUC = 0.77). When measured between two and six hours postoperatively, PTH < 19 pg/mL had a sensitivity of 90 per cent and negative predictive value of 96 per cent for postoperative hypocalcemia. The model that included the PTH level, concurrently measured total blood calcium level, and time of measurement had an improved predictive ability with an AUC of 0.87. PTH level of 19 pg/mL measured two to six hours after thyroidectomy had a sensitivity of 90 per cent and a negative predictive value of 96 per cent in our cohort. The model including postoperative PTH level, calcium level, and time of measurement may further improve the ability to predict postthyroidectomy hypocalcemia.