Objective To explore the risk factors of postoperative infection after external ventricular drainage(EVD)and to construct a nomogram prediction model.Methods The clinical data of 100 patients treated with EVD from June 2020 to July 2022 were retrospectively analysed.They were divided into the infected(n=18)and the non-infected(n=82)groups according to whether an associated infection occurred after EVD.Logistic regression was used to analyse the independent risk factors for post-EVD infections,and a nomogram prediction model was constructed for the occurrence of associated infections after EVD,and the predictive efficacy of the model was verified using the subject's work characteristics(ROC)curve.Results The percentages of patients with intracranial haemorrhage,combined other systemic infections,combined craniotomy,preoperative presence of an artificial airway,bilateral placement of tubes,cerebrospinal fluid sampling frequency≥3 times/week,catheter retention time≥7 d,and postoperative albumin≤35 g/L,as well as the length of stay in the ICU were compared between the two groups,and the differences were statistically significant(P＜0.05).Logistic regression analysis of the percentage of patients with combined other systemic infections,combined craniotomy,cerebrospinal fluid sampling frequency＞3 times/week,catheter reten-tion time＞7 d,and long ICU hospital stay were independent risk factors for the occurrence of associated infections after EVD.The AUC of the column-line graph prediction model for the occurrence of associated infections after EVD was 0.894(95％CI:0.817～0.947),with a sensitivity of 72.00％and a specificity of 98.8％.Conclusion Combined with other systemic infections,combined with craniotomy,cerebrospinal fluid sampling frequency≥3 times/week,catheter retention time＞7 d and long ICU stay are independent risk factors for the occurrence of associated infections after EVD,and the column chart model constructed based on the above factors has a good predictive efficacy.
External ventricular drainageInfectionRisk factorsNomogram model
万方数据
维普
