| Background and objectives:Chronic obstructive pulmonary disease(COPD)is a preventable and treatable disease.Patients with structural abnormalities but moderately or normally functioning lungs are considered to have pre-COPD.Due to dynamic hyperinflation,patients with COPD often experience exercise limitation.Expiratory positive airway pressure(EPAP)devices can provide positive pressure during expiration to reduce dynamic compression.However,the long-term efficacy and safety of domestic EPAP device usage have not been thoroughly studied.Moreover,patients with COPD or pre-COPD suffers from persistent airflow limitation,and dynamic hyperinflation(DH)occurs when ventilation demand is imposed such as during exercise.The 6MWT is a standard procedure used to evaluate exercise capacity in patients with COPD,and there is still a lack of understanding of the collected nasal airflow respiratory signals.Deterministic learning algorithm is a new method of machine learning for dynamic environments proposed in recent years for nonlinear dynamic systems that generate periodic or regressive trajectories,enabling dynamic local accurate modeling of their unknown systems,such as the respiratory system.Therefore,the first part of this study aimed to evaluate the efficacy and safety of EPAP devices in patients with COPD.The second part of this study aimed to develop an inexpensive and convenient method for diagnosing COPD,including pre-COPD.Methods:Part 1:Efficacy and safety of the expiratory positive airway pressure(EPAP)devices for respiratory rehabilitation in patients with COPD.From July 2021 to December 2022,a total of 52 patients with COPD or pre-COPD were included based on inclusion and exclusion criteria and were randomly divided into the control group(sham EPAP device,EPAP=0,n=26)and the intervention group(EPAP device,EPAP=5 cmH2O,n=26).Patients were required to use the device for a total of 4 hours per day.Follow-up was conducted at 2,4,and 6 months to collect data on the six-minute walk test(6MWT),pulmonary function test,partial pressure of end-tidal carbon dioxide(PETCO2),Modified British Medical Research Council(mMRC)score,Baseline Dyspnea Index and Transition Dyspnea Index(BDI and TDI scores),and Likert scale score.Mixed-effects model analysis was used to compare the differences between groups to evaluate the efficacy and safety.Part 2:Characterization of respiratory airflow signals in COPD based on deterministic learning and respiratory dynamical diagram.From July 2021 to December 2022,a total of 25 healthy volunteers and 52 patients with COPD or pre-COPD were enrollment.6MWT was performed for all subjects according to the standard ATS/ERS procedure,and nasal respiratory airflow signals were collected during the test using Nox T3? recorder.Finally,64 valid nasal respiratory airflow signals were collected,including 25 from healthy volunteers and 39 from patients with COPD or pre-COPD.Further,one-dimensional respiratory signal was converted into a two-dimensional signal and the two-dimensional dynamic system was modeled by deterministic learning to obtain the dynamics features.Then,dynamics information related to COPD was extracted,and the modeling results was visualized to obtain a respiratory dynamic diagram(RDG)and RDGs were classified and analyzed according to the pulmonary function and chest CT scan of the patients.Results:Part 1:Efficacy and safety of the expiratory positive airway pressure(EPAP)devices for respiratory rehabilitation in patients with COPD.(1)Patients in the intervention group showed significant improvement in exercise tolerance.①Patients in the intervention group showed an increase in walking distance(6MWD)immediately after 2 months(P=0.003),with further increases after 4 and 6 months(P<0.001).Comparison between groups revealed that the improvement in 6MWD after 4 months reached the minimal clinically important difference(130.44 m,95%CI:65.33-195.55 m,P<0.001).②Patients in the intervention group showed significant improvement in dyspnea after the 6MWT after 2 months(P<0.001),and the improvement increased further after 4 months(P=0.012).Between-group comparisons revealed varying degrees of improvement in dyspnea after 2 months(P=0.029),4 months(P=0.002),and 6 months(P=0.002)in the intervention group,but none reached the minimal clinically important difference(MCID).③ The greatest improvement in leg fatigue was achieved in the intervention group after 2 months(P=0.002).A comparison between groups revealed that leg fatigue in the intervention group showed improvement after 6 months(P=0.013),but none reached the MCID.④Patients’ oxygen saturation at the beginning of the 6MWT(Pre-SpO2)improved,and between-group comparisons revealed varying degrees of improvement in Pre-SpO2 after 2 months(P=0.002),4 months(P=0.038),and 6 months(P=0.003)in the intervention group.⑤Patients’ minimal oxygen saturation during the 6MWT(Min-SpO2)improved,and a comparison between two groups revealed varying degrees of improvement in Min-SpO2 after 2 months(P=0.002),4 months(P=0.019),and 6 months(P=0.002)in the intervention group.⑥ Patients’ oxygen saturation at the end of the 6MWT(End-SpO2)improved,and comparisons between groups revealed improvements in End-SpO2 after 2 months(P=0.049),4 months(P=0.033),and 6 months(P<0.001)in the intervention group.⑦ Compared to the control group,patients in the intervention group showed a significant improvement in the Δ oxygen saturation/walk distance ratio after 6 months of intervention(P=0.043).(2)Patients in the intervention group showed improvement in lung function.① Inspiratory capacity(IC)increased in the intervention group after 6 months compared to the control group(P=0.039).②Forced expiratory volume in one second(FEV1)showed different degrees of improvement after 2,4,and 6 months in the intervention group(P<0.05).③Forced expiratory volume in one second/forced vital capacity(FEV1/FVC)showed different degrees of improvement after 2,4,and 6 months in the intervention group(P<0.05).④ Maximum voluntary ventilation(MVV)showed different degrees of improvement after 2,4,and 6 months in the intervention group(P<0.05).(3)Patients in the intervention group showed significant improvement in dyspnea in daily life.①Compared with the control group,the mMRC scores showed different degrees of improvement after 2,4,and 6 months in the intervention group(P<0.05).② Compared with the control group,the Transition Dyspnea Index(TDI)scores showed different degrees of improvement after 2,4,and 6 months in the intervention group(P<0.05).(4)The safety of the long-term use of EPAP device was confirmed,and the EPAP device did not cause carbon dioxide retention.PETCO2 showed some decrease in the intervention group after 4 and 6 months of intervention compared to the control group(P<0.05).Part 2:Characterization of respiratory airflow signals in COPD based on deterministic learning and respiratory dynamical diagram.(1)A respiratory dynamics diagram was generated for each subject in this study,which is also known as a butterfly diagram because of its resemblance to the left and right wings of a butterfly.The left wing represents the expiratory phase and the right wing represents the inspiratory phase.(2)In this study,it was found that the left and right wings were symmetrical in the diagrams of healthy volunteers.In contrast,most of the patients with COPD or pre-COPD involving different disease severity showed a significant left-wing reduction in their diagrams.Moreover,the left-wing reduction persisted during exercise.(3)In this study,it was also found that some of the patients with left-wing reduction in the diagrams were those whose with pre-COPD.This indicates that the deterministic learning algorithm used in this study may provide more sensitive information than conventional pulmonary function tests and may identify the presence of airflow limitation in patients at the early stage of COPD.Conclusions:(1)Long-term use of EPAP devices at home is safe and effective in improving exercise capacity,lung function and the degree of dyspnea in daily life of patients with COPD(including pre-COPD).(2)EPAP devices are efficient,safe and portable,and may be used in respiratory rehabilitation on a large scale in the future,which has important clinical application value.(3)The respiratory dynamics diagrams generated by deterministic learning have good interpretability and may demonstrate the presence of airflow limitation and dynamic hyperinflation in patients with COPD or pre-COPD,making it possible to diagnose COPD or pre-COPD with waveforms generated by calm breathing,which is expected to be a useful tool for future clinical aid in the diagnosis of COPD or pre-COPD.(4)This method has the advantage of non-invasive,convenient and inexpensive.It facilitates the diagnosis of patients with pre-COPD,thus enabling the diagnosis and treatment of COPD to be moved forward and facilitating future breakthroughs in COPD prevention and control. |
PDF Full Text Request