Development of a nomogram for predicting severe post-parathyroidectomy hypocalcemia in patients with secondary hyperparathyroidism: a retrospective cohort study.

Background: Severe hypocalcemia affects 25-50% of patients with secondary hyperparathyroidism (sHPT) after total parathyroidectomy. It results from acute postoperative hypoparathyroidism and rapid skeletal mineral uptake. Severe hypocalcaemia leads to significant morbidity, prolonged hospital stays, and increased healthcare costs. Few risk assessment prediction tools exist; they lack external validation and exhibit methodological shortcomings. To enable evidence-based risk stratification and personalised perioperative management, we developed and internally validated a nomogram for predicting severe post-parathyroidectomy hypocalcaemia (SPPH) in patients with sHPT who underwent total parathyroidectomy with autotransplantation. Methods: This retrospective cohort study included 227 patients from two tertiary care institutions with sHPT who underwent total parathyroidectomy with autotransplantation between 2019 and 2024. All subjects had sHPT [parathyroid hormone (PTH) >3× upper limit of normal] in the context of stage 4-5 chronic kidney disease (CKD) or dialysis dependent. Patients with autonomous parathyroid function (calcium-independent PTH elevation) were excluded. Intensive calcium supplementation protocols were administered to all patients. SPPH was defined by: (I) serum calcium <1.9 mmol/L; (II) intravenous calcium requirement beyond postoperative day 3; (III) symptomatic hypocalcemia (tetany, seizures, arrhythmias, or laryngospasm) requiring intravenous calcium and/or urgent cardiac monitoring; or (IV) hospitalization ≥7 days primarily for calcium management. We employed the least absolute shrinkage and selection operator (LASSO) regression for variable selection, followed by multivariable logistic regression. Internal validation utilized 200 bootstrap resamples to derive the optimism-corrected performance estimates. Results: The incidence of SPPH was 49.8% (113/227). Among 157 patients with complete predictor data, LASSO identified four optimal predictors: [adjusted odds ratio (OR) 0.87 per year, 95% confidence interval (CI): 0.76-0.99], preoperative phosphate at one month (adjusted OR 1.29 per mmol/L, 95% CI: 0.96-1.71), preoperative alkaline phosphatase (ALP) at one month (adjusted OR 4.11 per 100 U/L, 95% CI: 1.89-8.92), and preoperative PTH at three months (adjusted OR 0.87 per 10 pmol/L, 95% CI: 0.68-1.13). The model demonstrated good discrimination with an apparent c-statistic of 0.751 (optimism-corrected 0.739), good calibration (Hosmer-Lemeshow P=0.657), and positive net benefit across clinically relevant threshold probabilities of 0.1-0.8. Risk stratification yielded three clinically meaningful groups: low-risk (<30% predicted probability, n=25, 8.0% observed SPPH rate), intermediate-risk (30-70%, n=92, 58.7% rate), and high-risk (>70%, n=40, 72.5% rate). Conclusions: We developed and internally validated a nomogram for preoperative risk stratification of SPPH in patients receiving intensive calcium supplementation protocols using readily available preoperative variables, identifying ALP as the predominant risk factor. The prognostic significance of ALP offers a new perspective beyond the conventional emphasis on PTH levels and insight into the underlying mechanisms of post-operative hypocalcemia. The substantial gradient across risk groups (8.0-72.5%) supports personalized perioperative management. However, multicenter external validation is essential before clinical implementation. Once validated, this tool could enhance patient outcomes and streamline healthcare resources for thousands of patients undergoing parathyroidectomy annually.