Valor del gradiente de descenso de parathormona y del número de paratiroides identificadas en la predicción precoz de hipocalcemia tras tiroidectomía total

Zusammenfassung

INTRODUCTION Total thyroidectomy is the mainstay of therapy for bilateral benign and malignant thyroid disease. Hypoparathyroidism is the most frequent complication of total thyroidectomy, and it contributes to increase the morbidity and economics cost associated with this procedure. The main risk factor for its development is hypoparathyroidism, secondary to transitory or permanent parathyroid gland dysfunction after surgery. Usually, this gland dysfunction generates mild and transitory hipocalcaemia. However, a severe deficiency with high risk for the patient can occur in a considerable percentage of cases, including intact parathormone (iPTH) permanent deficiency. The possibility to perform an early, reliable and simple detection method for this complication becomes particularly important to prevent it. A large number of predictors of hypoparathyroidism have been identified. The more relevant factors are related to the type of surgery and some early postoperative biochemical parameters. The objective of this study was to identify relevant factors for this complication, develop a hypocalcaemia predictive model, and validate it in an independent cohort of patients. METHODS An observational, single-center study was performed in a cohort of consecutive patients undergoing total thyroidectomy for any indication. The operations were carried out between May 2012 and September 2015 (study group) and between November 2015 and October 2016 (validation group). Exclusion criteria included age lesser than 14 years old, completion thyroidectomy, concomitant parathyroid disease, patients on calcium or bisphosphonate supplementation therapy, and the need of emergent reoperation for hemorrhage. Patients with missed significant data or lost in the follow-up were also excluded. Demographic, clinical, analytical, surgical and postoperative variables were collected. The incidence of postoperative hypocalcaemia and permanent hypoparathyroidism was calculated. Bivariate and multivariate analysis were performed to identify variables related to the development of laboratory-confirmed hypocalcaemia, and to detect independent predictors. Based on independent predictors of hypocalcaemia, we generated a predictive model using a multivariate logistic regression formula to calculate the hypocalcaemia risk of each patient. The model was subsequently applied to a validation cohort of patients. To assess the predictive power of our model, the area under the ROC curve was calculated. RESULTS During the study period, 352 and 118 patients were included in the study and validation group, respectively. A total of 73 patients (20.7%) developed postoperative laboratory-confirmed hypocalcaemia in study group, which was symptomatic in 43 subjects (12.2% of the total). During long-term follow-up, 11 patients (3.1%) developed permanent hypoparathyroidism. Bivariate analysis results showed that in patients who developed hypocalcaemia, lymph node dissections were more frequent (p=0.036), the length of surgery was longer (p=0.038), the number of identified parathyroid glands was higher(p=0.001), and the gradient of PTH decline was higher (p<0.001). The percentage of patients who had suffered an incidental parathyroidectomy was also higher (p=0.002). Finally, absolute posthyroidectomy iPTH levels at 4 hours after surgery were lower in patients who developed hypocalcaemia (p<0.001). A close and lineal relationship between the number of parathyroid glands identified intraoperatively and the development of postoperative hypocalcaemia was proven: a greater number of parathyroid glands identified was correlated with higher risk of this complication. This risk with one, two, three and four identified glands were respectively 9.4, 15, 23.4 and 34.8 percent, respectively. The variables included in multivariate analysis were performance or not of lymph node dissection, the number of parathyroid glands identified during surgery, the gradient of PTH decline, and the incidental parathyroidectomy. This model demonstrated that the number of parathyroid glands identified and the gradient of PTH decline were independent predictors of postoperative hypocalcaemia. Based on both variables, we named our scale for prediction of laboratory-confirmed hypocalcaemia as NuGra (Number of parathyroid glands identified- Gradient of decline) scale. Its formula (exportable to spreadsheets and computer programs) predicts the individual risk of each patient for developing hypocalcaemia. The predictive power of the model was high (AUROC 0.902, CI 0.857–0.947). When the NuGra scale was applied to the validation group, we obtained an AUROC of 0.956 (CI 0.919–0.993). CONCLUSIONS A greater number of parathyroid glands identified during total thyroidectomy, and a higher gradient of iPTH decline at 4h after surgery, are independent predictors of postoperative hypocalcaemia. The combination of both parameters has allowed us the development of a predictive scale. This model predicts the individual risk of each patient for developing hypocalcaemia, with a high predictive power. NuGra scale was applied to a validation cohort of patients and it obtained very high accuracy level.