常规术中神经监测对甲状腺切除术并发症的影响：一项倾向性评分匹配的回顾性队列研究。

文献信息

中文摘要

背景： 术中神经监测（IONM）常用于甲状腺手术中，以帮助识别喉返神经（RLN），并可能减少相关术后并发症。然而，常规IONM在降低并发症发生率方面的疗效仍存在争议。在本研究中，我们探讨了IONM对甲状腺切除术后并发症的影响。方法：这项回顾性单中心研究纳入了2021年至2024年间进行的214例甲状腺切除术。患者被分为2组：第1组（n=107）接受常规IONM手术，第2组（n=107）未接受IONM，仅采用传统直接可视化。比较两组术后并发症（RLN麻痹、低钙血症、甲状旁腺功能减退症和意外甲状旁腺切除术）。主要结局为任何RLN麻痹，在神经水平进行分析。还进行了预先设定的倾向性评分匹配和敏感性分析。结果：第1组术前钙水平高于第2组（P=.002）。两组术后钙和甲状旁腺激素水平以及低钙血症和甲状旁腺功能减退症发生率无差异；这些结局中最小的P值在校正多重性前为.04，校正后为.40。第1组有198条风险神经，第2组有204条。第1组中3/198条神经（1.5%）发生任何RLN麻痹，第2组中2/204条（1.0%）发生（P=.68）。短暂性RLN麻痹分别发生于3/198条（1.5%）和1/204条（0.5%）（P=.37），而永久性RLN麻痹分别发生于0/198条（0%）和1/204条（0.5%）（P>.99）。在倾向性评分匹配队列中未观察到任何RLN麻痹的差异（P>.99），在预先设定的亚队列和敏感性分析中也观察到类似结果。结论：与常规直接可视化相比，常规IONM与RLN麻痹、低钙血症或甲状旁腺功能减退症的显著减少无关。极少的RLN事件数量和宽置信区间表明统计精度有限，且该研究把握度不足以排除RLN麻痹风险中具有临床意义的差异。

英文摘要

Intraoperative neuromonitoring (IONM) is commonly used during thyroid surgery to facilitate recurrent laryngeal nerve (RLN) identification and potentially reduce related postoperative complications. However, the efficacy of routine IONM in decreasing complication rates remains debated. In this study, we investigated the impact of IONM on postoperative complications after thyroidectomy. This retrospective single-center study included 214 thyroidectomy operations performed between 2021 and 2024. Patients were divided into 2 groups: Group 1 (n = 107) underwent surgery with routine IONM, and Group 2 (n = 107) without IONM, using only conventional direct visualization. Postoperative complications (RLN palsy, hypocalcemia, hypoparathyroidism, and incidental parathyroidectomy) were compared between groups. The primary outcome was any RLN palsy, analyzed at the nerve level. Prespecified propensity score matching and sensitivity analyses were also performed. Preoperative calcium levels were higher in Group 1 than in Group 2 (P = .002). Postoperative calcium and parathyroid hormone levels and the rates of hypocalcemia and hypoparathyroidism did not differ between the groups; the smallest P value across these outcomes was.04 before and.40 after adjustment for multiplicity. There were 198 nerves at risk in Group 1 and 204 in Group 2. Any RLN palsy occurred in 3/198 nerves (1.5%) in Group 1 and 2/204 (1.0%) in Group 2 (P = .68). Transient RLN palsy occurred in 3/198 (1.5%) and 1/204 (0.5%), respectively (P = .37), whereas permanent RLN palsy occurred in 0/198 (0%) and 1/204 (0.5%), respectively (P > .99). No difference in any RLN palsy was observed in the propensity score-matched cohort (P > .99), and comparable findings were observed in the pre specified sub cohort and sensitivity analyses. Routine IONM was not associated with a significant reduction in RLN palsy, hypocalcemia, or hypoparathyroidism compared with conventional direct visualization. The very small number of RLN events and the wide confidence intervals indicate limited statistical precision, and the study was not sufficiently powered to exclude clinically meaningful differences in the risk of RLN palsy.