To quantify in-patient mortality from thyroid surgery, explore potential causes and suggest an accountability framework.

A retrospective analysis of the UK registry in Endocrine and thyroid surgery (UKRETS) database. Non-parametric statistical tests.

Between February 1999 and March 2022, 83,410 thyroid operations were registered into UKRETS; in-patient deaths were recorded in 56. Eleven were excluded because of subsequent healthcare interactions; in-patient mortality was 0.054% (45 patients). Mortality in malignant cases was 6-fold higher than for benign disease (29/17404[0.17%] vs. 15/51758[0.029%]; p<0.001). Histology was unknown in one case. In benign disease (27% bilateral surgery, 53% lobectomy, 20% other/unknown) death was associated with thyroid surgery to permit airway control (4;27%), cardiac complications (4;27%), sepsis, mesenteric vein thrombosis, and traumatic brain injury. One death followed re-operation for haemorrhage. In thyroid cancer, tumour was irresectable in 17(59%), resectable in 3(10%) and unknown in the remainder. Bilateral surgery was attempted/performed in 14(48%), lobectomy in 7(24%), biopsy in 2(7%) and unknown/other in 6(20%). Mortality according to histology was anaplastic 10/157(6.37%), lymphoma 1/100(1%), MTC 2/677(0.30%), FTC 4/3071(0.13%), PTC 5/12446(0.04%).

In-hospital mortality following thyroidectomy is 1 in 2000 and 6-fold higher in malignant disease; largely driven by ATC (220-fold). Most malignant deaths occur with irresectable disease. Establishment of regional MDTs might identify such cases and prevent hopeless surgery. Mortality in benign disease is often unrelated to thyroid surgery itself. We propose 'airway control' is introduced as a novel variable for indication-to-operate in endocrine registries and that a suitable forum is developed where the learning from such deaths is discussed.