A Blood-Draw Timing- and Season-Adjusted PTH-Calcium Model for Predicting Post-Endoscopic/Robotic Thyroidectomy...

Area of Science:

• Endocrinology
• Surgical Complications
• Predictive Modeling

Background:

• Post-thyroidectomy hypocalcemia is a frequent complication, potentially delaying patient discharge.
• Current prediction methods using parathyroid hormone (PTH) and calcium (Ca) levels often overlook circadian and seasonal variations.
• Optimizing prediction requires accounting for temporal factors in blood sample collection.

Purpose of the Study:

• To develop and validate a predictive model for early post-thyroidectomy hypocalcemia.
• To assess the impact of incorporating blood-draw timing and season into hypocalcemia prediction.
• To improve the efficiency of postoperative management by refining prediction accuracy.

Main Methods:

• A retrospective and prospective cohort study of adults undergoing endoscopic or robotic thyroidectomy.
• Development and validation of predictive models (Model A: PTH only; Model B: PTH + Ca; Model C: Model B + timing/season/covariates).
• Primary outcome: biochemical hypocalcemia within 48 hours post-surgery, defined by corrected Ca <2.10 mmol/L or ionized Ca <1.10 mmol/L.

Main Results:

• Model C, incorporating timing and season, demonstrated superior predictive performance (AUROC 0.87 in derivation, 0.84 in validation) compared to models using only PTH or PTH+Ca.
• The model showed excellent calibration and maintained strong performance in prospective validation.
• Decision-curve analysis indicated Model C's potential to reduce unnecessary prophylactic supplementation by approximately 11% per 100 patients.

Conclusions:

• Incorporating blood-draw timing and seasonal variations significantly enhances the early prediction of post-thyroidectomy hypocalcemia.
• The refined predictive model offers a more accurate assessment than standard PTH/Ca measurements alone.
• This approach can lead to more efficient and targeted postoperative management strategies.