Objective To study the predictive values of lung ultrasound(LUS)score and Downes score in selecting respiratory support strategies for newborns with dyspnea.Methods From September 2021 to July 2022,newborns admitted to our hospital with dyspnea were selected and assigned into the non-invasive respiratory support(N)group,invasive respiratory support(I)group and control(C)group based on the respiratory support strategies on admission.LUS scores and Downes scores at 6,24,and 48 h after birth were recorded.ROC curves were drawn to determine the predictive values of LUS and Downes scores for respiratory support strategies.Results A total of 263 cases were enrolled,including 105 cases in N group,56 cases in I group and 102 cases in C group.The differences of LUS and Downes scores between the three groups at the same timepoint were statistically significant with I group had the highest scores,N group second and C group lowest(P＜0.05).LUS and Downes scores within each group at different timepoints were significantly different(P＜0.05).In all three groups,LUS and Downes scores were decreased with longer duration of treatment.LUS score,Downes score and PaO2/FiO2 were positively correlated with each other(P＜0.05).The area under the curve(AUC)of LUS score and Downes score predicting non-invasive respiratory support within 6 h after birth were 0.900(95％CI 0.861-0.940,P＜0.05)and 0.889(95％CI 0.847-0.931,P＜0.05),respectively,with the same cutoff of 2.5.The AUC of the combination of LUS and Downes scores predicting non-invasive respiratory support was 0.944(95％CI 0.915-0.973,P＜0.05).The AUC of LUS score and Downes score predicting invasive respiratory support were 0.979(95％CI 0.963-0.995,P＜0.05)and 0.831(95％CI 0.760-0.902,P＜0.05),respectively,with the same cutoff of 5.5.The AUC of the combination of LUS and Downes scores predicting invasive respiratory support was 0.985(95％CI 0.972-0.998,P＜0.05).Conclusions Both LUS score and Downes score have certain predictive values for respiratory support strategies in newborns with dyspnea.
维普
万方数据