日间经皮肾镜碎石取石术患者延迟出院的危险因素分析和列线图构建

Delayed discharge following day-surgery percutaneous nephrolithotomy:risk factor evaluation and nomogram construction

目的探讨日间经皮肾镜碎石取石术(PCNL)患者延迟出院的危险因素,并构建风险预测列线图。方法回顾性审查370例日间PCNL患者的电子病历,根据患者是否在24 h内出入院,将其分为延迟出院组(61例)和正常出院组(309例)。采用单因素分析比较两组的一般资料、术后1 h生命体征和血液学指标。通过Logistic回归分析确定延迟出院的危险因素,绘制列线图以预测延迟出院的风险,应用校准曲线评估列线图的内部校准性能。结果370例患者中有61例(16.5%)出现延迟出院,18例(4.9%)住院时间超过48 h。Logistic回归分析显示年龄>65岁(OR:3.992,95%CI:1.792~8.892,P=0.001)、通道数量(OR:3.139,95%CI:1.548~6.367,P=0.002)、术中严重出血(OR:12.207,95%CI:2.761~53.973,P=0.001)、术后1 h体温>38℃(OR:4.623,95%CI:1.445~14.789,P=0.010)、中性粒细胞比率>75%(OR:1.951,95%CI:1.008~3.777,P=0.047)和钾<3.5 mmol/L(OR:5.469,95%CI:1.641~18.226,P=0.006)是日间PCNL患者延迟出院的独立危险因素。延迟出院预测模型的ROC曲线下面积为0.784,通过危险因素构建的列线图具有良好的区分度和准确性。结论基于年龄、通道数量、术中严重出血、术后感染和低钾血症构建的延迟出院风险预测列线图预测性能良好,可帮助医护人员快速识别术后延迟出院高风险患者,以便尽早实施医疗干预。

Objective To explore the risk factors of delayed discharge following day-surgery percutaneous nephrolithotomy(PCNL)and construct a risk prediction nomogram.Methods 370 patients underwent day-surgery PCNL were retrospectively reviewed.According to whether the patients were admitted and discharged within 24 h,they were divided into delayed discharge group(n=61)and normal discharge group(n=309).Univariate analysis was used to compare the general data,vital signs and hematological indexes of the two groups.Furtherly,risk factors of delayed discharge was evaluated in multivariate logistic regression analysis.Nomogram was plotted,and calibration curve was applied to assess the internal calibration of the nomogram.Results 61 of 370 patients(16.5%)suffered delayed discharge,18 patients(4.9%)were hospitalized for more than 48 h.Age over 65 years(OR:3.992,95%CI:1.792-8.892,P=0.001),number of tracts(OR:3.139,95%CI:1.548-6.367,P=0.002),intraoperaive severe bleeding(OR:12.207,95%CI:2.761-53.973,P=0.001),body temperature over 38℃(OR:4.623,95%CI:1.445-14.789,P=0.010),neutrophil ratio greater than 75%(OR:1.951,95%CI:1.008-3.777,P=0.047)and potassium less than 3.5 mmol/L(OR:5.469,95%CI:1.641-18.226,P=0.006)were identified as independent predictors of delayed discharge following day-surgery PCNL.The area under the ROC curve of delayed discharge prediction model is 0.784,and the nomogram constructed by risk factors has good discrimination and accuracy.Conclusion The nomogram constructed based on age,number of channels,intraoperative severe bleeding,postoperative body temperature,neutrophil ratio,and hypokalemia has good performance in predicting the risk of delayed discharge.It can help medical staff quickly identify high-risk patients with delayed discharge,so that medical intervention can be implemented as soon as possible.