继发性甲状旁腺功能亢进患者甲状旁腺切除术后高钾血症的影响因素及其风险预测模型

Influencing factors of hyperkalemia in patients with secondary hyperparathyroidism after parathyroidectomy and their risk prediction model

目的:探究继发性甲状旁腺功能亢进(SHPT)患者甲状旁腺切除术(PTX)后发生高钾血症的危险因素,并构建其风险预测模型。方法:本研究纳入2015—2021年在广东省人民医院行PTX的透析患者,根据术后血钾情况将患者分为高钾血症组(>5.3 mmol/L)和非高钾血症组(≤5.3 mmol/L)。比较两组患者的临床资料,采用Logistic回归分析探讨SHPT患者PTX后高钾血症的影响因素,并通过绘制受试者工作特征曲线、校准曲线、决策曲线评价模型的区分度、校准度及有效性。结果:共纳入392例患者,高钾血症组98例(25%),非高钾血症组294例(75%)。通过多因素Logistic回归分析,血液透析、术前血钾、术前血清白蛋白(Alb)、血尿素氮(BVN)/血清肌酐(SCr)均为术后高钾血症的预测因素(P<0.05)。纳入经多因素Logistic回归分析筛选的预测因素建立风险预测模型,ROC曲线显示该模型的曲线下面积(AUC)为0.747,通过Bootstrap方法重复抽样1000次对预测模型进行内部验证确认,内部验证AUC为0.733。Hosmer-Lemeshow拟合优度检验结果P=0.723,校正曲线与理想曲线显示出良好的一致性。绘制的临床决策曲线(DCA)结果显示,在阈值概率范围0.06~0.60内模型具有临床效用。结论:血液透析、术前血钾、术前Alb及术前BUN/SCr均为PTX后高钾血症的预测因素,建立的预测模型具有较好的预测能力。

Objective:To investigate the risk factors of hyperkalemia in patients with secondary hyperparathyroidism(SHPT)after parathyroidectomy(PTX)and construct a risk prediction model.Methodology:A single-centre retrospective study that included dialysis patients who underwent parathyroidectomy in Guangdong Provincial People s Hospital from 2015 to 2021,and the patients were divided into hyperkalemia group(>5.3 mmol/L)and non-hyperkalemia group(≤5.3 mmol/L)according to the postoperative serum potassium.The clinical data of the two groups were compared.The influencing factors of hyperkalemia in patients with secondary hyperparathyroidism after parathyroidectomy were analyzed by binary multivariate Logistic regression analysis.The discrimination,calibration,and effectiveness of the model were evaluated by plotting the receiver operating characteristic(ROC)curve,calibration curve,and decision curve.Results:A total of 392 patients were enrolled,98 in the hyperkalemia group and 294 in the non-hyperkalemia group.Through multivariate Logistic regression analysis,preoperative serum potassium,preoperative blood albumin,dialysis mode,and blood urea nitrogen creatinine ratio were predictors for postoperative hyperkalemia(P<0.05).The predictors screened by multivariate Logistic regression analysis were included to establish a risk prediction model.The ROC curve showed that the area under the curve(AUC)of the model was 0.747,and the predictive model was internally validated and confirmed by repeated sampling 1000 times by Bootstrap method,and the AUC of the internal validation was 0.733.The results of Hosmer-Lemeshow goodness-of-fit test was P=0.723,the calibration curve of the model was close to the ideal curve.The decision curve(DCA)indicated that when the probability threshold predicted by the model was 0.06-0.60,the model has clinical utility.Conclusion:Preoperative serum potassium,preoperative blood albumin,dialysis mode,and preoperative blood urea nitrogen creatinine ratio are predictors of hyperkalemia after parathyroidectomy,and the prediction model has a good predictive ability.