丘脑胶质瘤患者术后血钾水平变化规律及对预后的影响研究

Postoperative trajectory of serum potassium levels and its impact on prognosis in patients with thalamic glioma

摘要:

目的:探究丘脑胶质瘤患者术后血钾水平动态变化轨迹,及其与患者预后的相关性。方法:选取2020年2月至2025年5月于本院进行手术切除病灶且术后病理为丘脑胶质瘤的患者116例。分别于患者术后12 h、术后2 d、术后7 d和术后30 d,检测患者血钾水平,并记录患者一般资料、手术参数、术后并发症及预后情况。各时间点血钾水平相关性分析使用Pearson相关性分析；血钾变化轨迹构建使用潜变量增长曲线模型(LGCM)进行；血钾变化轨迹分析使用增长混合模型(GMM)进行；使用Logistic回归模型分析不同血钾变化轨迹亚组与患者预后不良发生的关联性；使用Logistic回归模型分析影响患者血钾变化轨迹的相关因素；使用分层Logistic回归分析和森林图分析血钾变化轨迹与不同年龄、性别、WHO分级、BMI、术中补液量、WBC、CRP和Cor亚组中预后不良的相关性。使用受试者工作特征曲线(ROC)分析血钾变化轨迹对预后不良发生的预测价值,并计算曲线下面积(AUC)。结果:丘脑胶质瘤患者术后血钾水平总体呈先降后升的变化趋势,术后早期血钾水平较术前基线有所下降,术后2 d左右降至最低水平,此后逐渐回升,至术后30 d时基本恢复至正常范围。各时间点血钾水平之间存在明显正相关性(P＜0.05)。LGCM分析显示,术后血钾水平的初始状态存在明显个体差异,血钾变化速率亦存在明显个体间变异(P＜0.05)。GMM分析识别出三种不同的血钾变化轨迹亚组:持续稳定组、先降后升组和持续低水平组,持续稳定组、先降后升组和持续低水平组的术后并发症发生率分别为23.40%、31.82%和76.00%,预后不良发生率分别为4.26%、34.09%和52.00%,三组间差异均具有统计学意义(P＜0.05)。持续低水平组患者中肿瘤WHO分级Ⅳ级比例(64.00%)明显高于持续稳定组(14.89%)和先降后升组(22.73%),持续低水平组患者术中补液量明显高于持续稳定和先降后升组,差异具有统计学意义(P＜0.05)。多因素Logistic回归分析显示,在持续稳定组与持续低水平组、先降后升组与持续低水平组的对比中,肿瘤WHO分级Ⅳ级、术中补液量较大均为患者血钾呈持续低水平轨迹变化的独立危险因素(P＜0.05)。分层分析结果表明,血钾变化轨迹与预后不良的关联在不同年龄、性别、WHO分级、BMI、术中补液量、WBC、CRP和Cor亚组间均保持一致,各亚组交互作用检验均无统计学意义(Pinteraction均＞0.05)。ROC曲线分析表明,血钾变化轨迹对术后预后不良具有较好的预测价值,其预测效能数值上高于多数单一时间点血钾水平,但DeLong检验显示仅与术后12 h血钾的AUC差异具有统计学意义；联合预测的AUC最高(0.785)。结论:丘脑胶质瘤患者术后血钾水平存在不同的变化轨迹,其中持续低水平轨迹与术后预后不良密切相关。血钾动态变化轨迹分析较传统单一时间点检测具有更优的预后预测效能。

Abstract:

Objective: To investigate the dynamic trajectory of postoperative serum potassium levels in patients with thalamic glioma and to evaluate its association with patient prognosis.Methods: A total of 116 patients who underwent surgical resection in our hospital between February 2020 and May 2025 and were pathologically diagnosed with thalamic glioma were included. Serum potassium levels were measured at 12 h, 2 d, 7 d, and 30 d postoperatively. Clinical data, surgical parameters, postoperative complications, and prognostic outcomes were collected. Pearson correlation analysis was used to assess correlations of potassium levels across different time points. The trajectory of serum potassium changes was constructed using latent growth curve modeling (LGCM); the analysis of serum potassium change trajectories was performed using growth mixture modeling (GMM). Logistic regression models were used to examine the association between different subgroups of serum potassium change trajectories and the occurrence of poor patient outcomes, as well as to identify factors influencing serum potassium change trajectories. Stratified logistic regression analysis and forest plots were employed to assess the relationship between serum potassium change trajectories and subgroups defined by age, sex, WHO grade, BMI, intraoperative fluid volume, WBC, CRP, and cortisol (Cor). The predictive value of serum potassium change trajectories for poor outcomes was evaluated using receiver operating characteristic (ROC) curves, and the area under the curve (AUC) was calculated. Results: Postoperative serum potassium levels exhibited an overall trend of initial decline followed by gradual recovery. Compared with preoperative baseline levels, potassium decreased in the early postoperative period, reached the lowest level around postoperative day 2, and then progressively increased, returning to the normal range by postoperative day 30. Significant positive correlations were observed among potassium levels at different time points (P < 0.05). LGCM analysis revealed significant inter-individual variability in both initial potassium levels and rates of change (P < 0.05).