ICU Acquired Weakness Of Patients With Mechanical Ventilation:Risk Factors Analysis And Prediction

Objective1.To understand the actual occurrence of ICU acquired weakness(ICU-AW)in patients with mechanical ventilation,and to provide reference for the development of corresponding management countermeasures.2.To screen the risk factors of ICU-AW in patients with mechanical ventilation,and to provide reference for the development of targeted intervention measures.3.To establish and validate a risk prediction model for ICU-AW of patients with mechanical ventilation,which improve the awareness of ICU-AW risk warning among medical staff,and provide scientific basis for their assessment of ICU-AW.MethodsThe convenient sampling was performed among 353 adult patients with endotracheal intubation admitted into a hospital in Zhenjiang from November 2018 to May 2020.The general condition,ICU treatment and laboratory data of patients were collected by consulting electronic medical record information and paper case data.General condition include age,gender,diabetes history,hypertension history,APACHE II score,BMI,Whether sepsis or multiple organ failure.ICU treatment include duration of mechanical ventilation,Nutrition type,whether to use continuous renal replacement therapy,immobilize,Whether to use analgesia,sedation,norepinephrine,glucocorticoid,aminoglycoside antibiotics and neuroblockers and other special drugs.Laboratory data include Blood sugar,serum albumin,calcium,and lactic acid.ICU-AW screening was conducted by the Medical Research Council score(MRC-Score).The data of 247 patients from November 2018 to October 2019were used as the modeling column,and the patients were divided into ICU-AW group and non-ICU-AW group according to whether ICU-AW occurred or not.ICU-AW risk factors were preliminarily screened by univariate analysis,variables with statistical significance(P<0.05)from univariate analysis were included in binary Logistic regression analysis,and ICU-AW Logistic regression model was established.At the same time,Hosmer lemeshow(H-L)goodness of fit was used to test the fitting ability of the evaluation model,and the area under curve(AUC)of receiver operating characteristic curve(ROC)was used to evaluate the discrimination ability of the model.Finally,the established ICU-AW risk prediction model is applied to verify the paired data to further verify and evaluate the risk prediction model.In order to facilitate clinical use,We Chat miniprogram was established to predict the occurrence probability of icu-aw in mechanically ventilated patients according to established risk model.Results1.ICU-AWA total of 353 patients with mechanical ventilation who met the inclusion criteria were collected.Among the 247 patients in the modeling group,106 of them had positive results of ICU-AW screening,the incidence of ICU-AW was 42.91%(106/247),and the incidence of ICU-AW was 44.34%(47/106)in the validation group.2.Univariate analysisUnivariate analysis was performed between ICU-AW group and non ICU-AW group There were statistically significant differences in age,mechanical ventilation time APACHE II score,sepsis,multiple organ dysfunction,immobilization,nutrition type,sedatives,analgesics,noradrenaline,aminoglycosides,nerve blockers,blood glucose level,albumin and lactic acid(P<0.05);The two groups had no significant differences in gender,history of hypertension,history of diabetes,continuous renal replacement therapy,glucocorticoid,blood calcium level between the two groups(P>0.05).3.Logistic regressionMultivariate logistic regression analysis showed that six variables entered the final risk prediction model,age(OR=1.043,95%CI 1.018-1.069,P=0.001),mechanical ventilation time(OR=1.140,95%CI:1.069-1.216,P<0.001),APACHE II score(0R=1.081,95%CI:1.028-1.137,P=0.002),blood glucose(OR=1.117,95%CI1.001-1.245,P=0.047),lactic acid(OR=1.459,95%CI:1.232-1.729,P<0.001),Neuromuscular blocking agent(OR=3.499,95%CI:1.561-7.841,P<0.001)were independent predictors of ICU-AW;No correlation was found between sepsis,multiple organ dysfunction,immobilization,nutrition type,sedatives,analgesics,noradrenaline,aminoglycosides,albumin and other factors with ICU-AW.The formula of ICU-AW risk prediction model is as follows:P=1/1+exp(-Z),-Z=8.808-0.042×age-1.252×use of nerve blockers(=0 or 1)-0.078×APACHE-II score-0.110×blood glucose-0.378×blood lactate-0.131×mechanical ventilation time.4.Performance of modelThe results of H-L chi square test showed that there was no significant difference between the predicted value of ICU-AW and the actual situation(x~2=12.085,P=0.147>0.05).The area under the receiver operating characteristic curve(AUC)was 0.869(95%CI:0.824～0.914,P<0.001),which suggestes that the discrimination and consistency of the model are good.When the best cut-off value was 0.503,Youden index reached the maximum value of 0.577,and the combination of sensitivity and specificity is the best,0.754 and 0.823,respectively.5.Clinical application of risk prediction modelVerify the independent data validation of the model for the patients as risk prediction,The cut-off value was obtained by substituting the patient data into the prediction model formula,When the cut-off value is greater than 0.503,ICU-AW will occur.The validation model results show that sensitivity,specificity and accuracy were 70.2%,88.1%and 80.2%,respectively.According to the small program of ICU-AW model constructed by regression equation,medical staff can quickly get the probability of occurrence of ICU-AW by inputting the patient’s age,APACHE II score,mechanical ventilation time,blood glucose,lactic acid,and the value of whether to use neuroblocker.When the probability value exceeds 50.3%,the patient is at higher risk of occurrence of ICU-AW.Conclusions1.The incidence of ICU-AW was higher in patients with mechanical ventilation,and there were many influencing factors.In this study,age,duration of mechanical ventilation,APACHE II score,blood glucose,lactic acid and nerve blockers were selected as independent predictors of ICU-AW.Medical workers should focus on the patients with the above risk factors and take corresponding preventive measures actively to avoid the influence of these factors on the occurrence of ICU-AW.2.In this study,ICU-AW risk prediction model was initially established by Logistic regression modeling.It is verified that the area under ROC curve of this model is 0.869,the sensitivity is 70.2%,and the specificity is 88.1%,indicating that the model has good predictive ability.3.Medical staff can use small procedures developed according to the model to screen mechanically ventilated patients at any time during hospitalization,which can provide reference for risk assessment and clinical prevention and treatment of ICU-AW.