Heart failure (HF) occurs when the heart is not able to eject enough blood from its chambers in order to satisfy the needs of the body. Patients suffering from chronic HF often experience periods of deterioration with characteristics signs of systemic and pulmonary congestion, which consists in fluid accumulation in the lungs and peripheries. Trans-thoracic impedance has been proposed as a promising solution for tracking fluid accumulation in the lungs, anticipating heart failure worsening, and assessing the onset of congestion. A wearable textile electrode patient monitoring device (BIM) was used to measure bioimpedance, ECG and accelerations. The aim of this thesis was to assess the onset of congestion in HF patients, starting from trans-thoracic impedance signals. The thesis is composed by 2 parts: Assessment of congestion: algorithms based on trans thoracic impedance measurements for the discrimination of congested from stable dry patients were validated. Signal processing for predicting cardiac output: starting from trans-thoracic signals, a variation synchronous with the heart activity was extracted in order to predict cardiac output. An important and significant decrease in impedance was present in HF congested patients, but BIM measurements did not reach sensitive and specific discrimination of congestion cases. A new EMD-based technique is presented in this thesis which is able to remove respiratory and motion artifacts from trans-thoracic bioimpedance signals. This method showed better performance (*p<0.05) as compared to band pass filtering, typical and reliable technique in impedance cardiography. Averaging techniques were used to extract a waveform with respect to the R-peaks. A common shape was found and four features were extracted. Although none of the features were reliable enough to find a significant correlation to cardiac output in HF patients, the typical waveform was found to be related to the physiological blood changes of the heart cycle, systole, and diastole.
L’insufficienza cardiaca è definite come l’incapacità del cuore di fornire il sangue in quantità adeguata rispetto all’effettiva richiesta dell’organismo. Pazienti che soffrono di questa patologia sono spesso soggetti ad un peggiorameno delle condizioni di salute che portano a congestione sistemica o polmonare. Questa situazione consiste nell’accumulo di fluidi nei polmoni e nelle periferie, e porta a sintomi tra cui dispnea, fatica ed intolleranza all’esercizio fisico. L’impedenza trans-toracica è stata proposta come una promettente soluzione per monitorare l’accumulo di liquidi nei polmoni, anticipare il peggioramento di insufficienza cardiaca e valutare l’insorgenza di congestione anche in assenza di sintomi clinici. In questa tesi è stato utilizzato un dispositivo indossabile (Bioimpedance Monitor, BIM) capace di misurare l’impedenza trans-toracica, l’elettrocardiogramma e le accelerazioni lungo i tre assi. Lo scopo della tesi è di valutare l’insorgenza di congestione in pazienti con insufficienza cardiaca, a partire da segnali di impedenza registrati con il BIM. Il lavoro di divide in due parti: Assessment of congestion: sono stati validati algoritmi basati sull’impedenza trans-toracica al fine di discriminare pazienti congestionati da quelli non congestionati. Signal processing for predicting cardiac output: è stata estratta una variazione sincrona con i battiti cardiaci, al fine di predire il valore di gittata cardiaca nel tempo a partire dal segnale di impedenza. E’ stato riscontrato un decremento significativo del valore di impedenza nel caso di pazienti congestionati, ma le registrazioni del BIM non hanno raggiunto alti valori di sensitività e specificità nella discriminazione di casi di congestione. Una nuova tecnica basata sull’EMD è presentata in questa tesi. Questa è capace di rimuovere artifatti respiratori e di movimento per estrarre esclusivamente la componente cardiaca. Questo metodo ha mostrato performance migliori (*p<0.05) rispetto al filtraggio passa banda, considerata una tecnica affidabile nell’impedance cardiography. Sono state usate tecniche di averaging al fine di estrarre la forma d’onda rispetto ai picchi R. E’ stata trovata una forma comune in tutti i segnali e sono state estratte quattro features a partire da ogni forma d’onda. Sebbene nessuna feature è risultata essere abbastanza affidabile da mostrare una correlazione significativa con la gittata cardiaca, la forma d’onda riflette le variazioni di sangue fisiologiche relative al ciclo cardiaco, sistole e diastole.
Signal processing of trans-thoracic bioimpedance from a wearable device to predict cardiac output and congestion in chronic heart failure patients
BRUNO, PASQUALE CLAUDIO
2013/2014
Abstract
Heart failure (HF) occurs when the heart is not able to eject enough blood from its chambers in order to satisfy the needs of the body. Patients suffering from chronic HF often experience periods of deterioration with characteristics signs of systemic and pulmonary congestion, which consists in fluid accumulation in the lungs and peripheries. Trans-thoracic impedance has been proposed as a promising solution for tracking fluid accumulation in the lungs, anticipating heart failure worsening, and assessing the onset of congestion. A wearable textile electrode patient monitoring device (BIM) was used to measure bioimpedance, ECG and accelerations. The aim of this thesis was to assess the onset of congestion in HF patients, starting from trans-thoracic impedance signals. The thesis is composed by 2 parts: Assessment of congestion: algorithms based on trans thoracic impedance measurements for the discrimination of congested from stable dry patients were validated. Signal processing for predicting cardiac output: starting from trans-thoracic signals, a variation synchronous with the heart activity was extracted in order to predict cardiac output. An important and significant decrease in impedance was present in HF congested patients, but BIM measurements did not reach sensitive and specific discrimination of congestion cases. A new EMD-based technique is presented in this thesis which is able to remove respiratory and motion artifacts from trans-thoracic bioimpedance signals. This method showed better performance (*p<0.05) as compared to band pass filtering, typical and reliable technique in impedance cardiography. Averaging techniques were used to extract a waveform with respect to the R-peaks. A common shape was found and four features were extracted. Although none of the features were reliable enough to find a significant correlation to cardiac output in HF patients, the typical waveform was found to be related to the physiological blood changes of the heart cycle, systole, and diastole.File | Dimensione | Formato | |
---|---|---|---|
thesis.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Testo della tesi
Dimensione
3.05 MB
Formato
Adobe PDF
|
3.05 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/88584