| Background:China is entering the rapid development stage of population aging.The incidence of hip fracture in the elderly is high.Pulmonary infection is one of the most common complications after hip fracture surgery in the elderly,which affects the prognosis of patients,brings a heavy economic burden to patients and their families,and can even lead to death.Early identification and intervention can help reduce the incidence of lung infections and improve patient outcomes.However,there is still a lack of clinically feasible tools for assessing the risk of pulmonary infection after hip fracture in the elderly.Objective:The main purpose of this study is to explore the risk factors of pulmonary infection in elderly patients after hip fracture surgery,and to find the objective indicators that may be related to the occurrence of postoperative pulmonary infection,especially laboratory indicators.We also build a risk prediction model according to objective indicators,so as to help clinical doctors and nurses find high-risk patients with postoperative pulmonary infection in the early stage of admission and provide the basis for their early prevention of postoperative pulmonary infection.Methods:The clinical data of elderly patients with hip fracture in Shenzhen Second People’s Hospital from January 2012 to December 2020 were retrospectively analyzed,including general clinical data,24-hour laboratory indicators after admission,operation-related variables,and whether postoperative pulmonary infection occurred.The diagnostic criteria of postoperative pulmonary infection referred to the Chinese guidelines for the diagnosis and treatment of hospital-acquired pneumonia and ventilator-associated pneumonia in adults（2018 Edition）.1.The patients were divided into three groups according to their age: the young elderly（60 ～ 69 years old）,the middle-aged elderly（70 ～ 79 years old）,and the elderly（≥ 80 years old）.The relationship between age,the number of complications,and postoperative pulmonary infection was analyzed.2.The occurrence of postoperative pulmonary infection is regarded as the dependent variable,and then the factor with statistical significance in univariate factor analysis was screened by LASSO.Finally,the independent risk factors of postoperative pulmonary infection were obtained by multivariate logistic regression analysis with the backward LR method.The original data of 1341 cases are divided into a training sets and test sets according to the ratio of 7:3,and the ROSE algorithm is used to process the a training set to form a new balanced training set.Logistic regression is used to build the prediction model,and the area under the ROC curve and DCA were used to evaluate the performance and application value of the model.Based on the original data and artificial data,the risk prediction model is established respectively,the prediction performance of the two is compared,and the optimal risk prediction model is selected to establish the nomogram.Results:1.Base situationThere were 1341 elderly patients after hip fracture surgery in this study,including79 patients with postoperative pulmonary infection,accounting for 5.89%.The average age of the patients was（77.59 ± 8.64）years old.Women accounted for 72.93%,femoral neck fracture accounted for 61.74%,the incidence of in-hospital death was 0.29%,and the length of hospital stay was 10（8,15）days.2.Relationship between age,number of complications,and postoperative pulmonary infectionMultivariate analysis showed that the risk of postoperative pulmonary infection in the younger aged group with the number of complications ≥ 3,the middle-aged elderly group with the number of complications ≥3,the oldest elderly group with the number of complications <3 and oldest elderly group with a number of complications ≥3 was5.882 times（95CI%:1.126～30.727,P=0.036）,5.345 times（95CI%:1.316～21.715,P=0.019）,4.165 times（95CI%:1.181～14.688,P=0.026）and 12.265times（95CI%:3.223～46.678,P<0.001）higher than that in the younger aged group with the number of complications < 3.3.Multivariate analysis of postoperative pulmonary infectionAge（OR=1.054,95%CI: 1.022～1.088,P=0.001）,number of complications ≥ 3（OR=2.813,95%CI: 1.667～4.748,P<0.001）,admission to ICU（OR=10.213,95%CI:4.302～24.243,P<0.001）,high WBC level（OR=1.121,95%CI: 1.039～1.208,P=0.003）and low AGR level（OR=0.199,95%CI: 0.071～0.555,P=0.002）were independent risk factors for postoperative pulmonary infection.4.Construction and validation of postoperative pulmonary infection risk modelThere were 939 cases in the original training set,only including 58 patients with postoperative pulmonary infection,which is an unbalanced dataset.The new training set generated by the ROSE algorithm,contains 1762 cases.In the new training set,the ratio of positive cases（845 cases）to negative cases（917 cases）was close to 1.The formula of ROSE-WAGH model is Z = 0.034 × Age-0.416 × Gender assignment + 0.094 ×WBC-1.240 × AGR-0.014 × HGB-0.103.The AUC of the model is 0.781,the sensitivity is 92.4%,the specificity is 57.1%,the accuracy is 90.5%,the threshold probability area of the maximum net benefit value is 3% ～ 43%,and the Youden index is 0.652,the value on the corresponding nomogram is 238.Conclusions1.The overall incidence of pulmonary infection after hip fracture surgery in the elderly was 5.89%.2.When the number of complications was ≥3,the risk of postoperative pulmonary infection in the younger,middle and older groups was higher than the sum of the individual effects of age and number of complications ≥ 3,and the risk increased significantly in the older group.3.It was found that when the patient’s age was less than 80 years old,the risk of postoperative pulmonary infection was mainly related to the number of complications,while when the age was ≥80 years old,the risk of postoperative pulmonary infection was related to both age and the number of complications.4.Age,number of complications ≥3,admission to ICU,high WBC level,and low AGR level were independent risk factors for pulmonary infection after hip fracture surgery in the elderly.5.Patients with nomogram scores of more than 238 are the key observation objects of medical staff.The prediction model of pulmonary infection after hip fracture surgery in the elderly based on objective indicators has certain application value and is easy to popularize in the clinic.The model based on the artificial data processed by the ROSE algorithm has higher sensitivity and accuracy than that of the model based on the original training data. |
PDF Full Text Request