Timing of defibrillation (DF) to interrupt ventricular fibrillation (VF) is of utmost importance as it implies interruption of chest compression and,in case of failure, the risk of adding further damage to the already critical myocardial condition. In such scenario, a real-time indicator of the probability of success is needed . This study was aimed to the capability of “Amplitude Spectrum Area” (AMSA) to predict DF outcome in a large database of out-of-hospital cardiac arrest. Electrocardiographic (ECG) data recorded by automated external defibrillators were obtained from 8.419 cardiac arrest events occurring in 7 provinces in Lombardia Region, Italy, between 2008 and 2009. Among these events, only VF/VT cardiac arrests receiving DFs were selected (n=1055). A 2 sec ECG window ending at 0.5 sec before DF was analyzed and AMSA calculated, after fast Fourier transformation. DF was defined as successful in the presence of spontaneous rhythm 40 bpm starting within 60 secs from the DF. Threshold values of AMSA able to discriminate DF outcome were individuated and sensitivity, specificity, accuracy, positive and negative predictive values (PPV, NPV) were calculated. The area under the receiver operating characteristic (ROC) curve was measured. A total of 2.442 quality DF events, including 1055 first attempts and 1.387 subsequent ones were included in the analyses. DF success rate was of 26%, 27%, and 25.2% for all, first, and subsequent DFs, respectively. AMSA was significantly greater prior to successful DFs, compared to that preceding unsuccessful ones (13.8 vs. 6.9 mV-Hz, and 13.9 vs. 6.8 mV-Hz, and 13.7 vs. 7 mV-Hz, for all, first, and subsequent DFs respectively). Intersection of sensitivity, specificity and accuracy curves identified a threshold value of AMSA of approximately 9.5 mV-Hz, able to predict DF outcome, with a balanced sensitivity, specificity and accuracy of 80%, for all, first, and subsequent DFs . Moreover, intersection of PPV and accuracy curves identified a threshold value of AMSA of approximately 15 mV-Hz able to predict a successful DF with a PPV and accuracy of 80%, for all, first, and subsequent DF attempts. AMSA values greater than 27 mV-Hz correctly predicted the success of DF with a PPV value of 100%. AMSA below 8 mV-Hz correctly predicted the DF failure with a NPV of > 95%, for all, first, and subsequent DFs. Further decreases in AMSA values below 4 mV-Hz achieved a NPV of 100%. Area under ROC curves was 0.872, 0.869, and 0.875 for all, first, and subsequent DFs, respectively In this large patient population, an AMSA algorithm was capable to predict DF outcome with high accuracy. A specific AMSA threshold in order to predict DF outcome, i.e. success or failure, may be identified during CPR. An AMSA-based DF decision therefore would be an useful approach to guide the best CPR intervention.
La scelta del momento di procedure alla defibrillazione (DF) di un paziente in fibrillazione ventricolare (FV) comporta l’interruzione del massaggio cardiaco e, in caso di inefficacia, il rischio di aggiungere un ulteriore danno miocardico. Perciò è assai importante individuare un indicatore affidabile della probabilità di successo utilizzabile in tempo reale. In questo studio è stata valutata la capacità della Amplitude Spectrum Area (AMSA) nel predire l’esito della defibrillazione in una vasta popolazione di pazienti con arresto cardiaco extraospedaliero. Sono stati acquisiti gli elettrocardiogrammi (ECG) dalla memoria dei defibrillatori esterni utilizzati dagli operatori del 118 di 7 province lombarde nel periodo 2008-2009 in 8419 eventi di arresto cardiaco. Di essi sono stati analizzati i 1055 nei quali è stata effettuata una DF. L’AMSA è stato calcolato analizzando una finestra di 2 sec di ECG che terminava 0.5 sec prima della DF utilizzando una trasformata di Fourier per l’ampiezza e la frequenza. Il successo della DF è stato definito come la presenza di un ritmo stabile a frequenza di 40 bpm 60 sec dopo la DF. Sono stati individuati i valori soglia dell’AMSA per discriminare il risultato della DF, sono state calcolate sensibilità, specificità, predittività positiva e negativa ed è stata misurata l’area sotto la curva ROC. Sono stati inclusi 2442 eventi di DF, fra cui 1055 prime DF e 1387 successive. Sono risultate efficaci rispettivamente il 26% di tutte le DF, il 27% delle prime e il 25.2% delle successive. L’AMSA è risultato significativamente maggiore prima delle DF efficaci rispetto a quello ottenuto prima delle DF inefficaci (13.8 vs 6.9 mV-Hz, 13.9 vs. 6.8 mV-Hz, e 13.7 vs. 7 mV-Hz rispettivamente per tutte le DF ,le prime e le successive). L’intersezione delle curve di sensibilità, specifictà e predittività ha identificato un valore soglia di AMSA di 9.5 mV-Hz capace di predire il risultato della DF con un valore bilanciato dell’80% per tutte, le prime e le successive DF. Valori di AMSA > 27 mV-Hz hanno correttamente previsto un successo del 100%, mentre un valore di 8,5 mV-Hz ha previsto l’insuccesso con predittività negativa > 95% ed un valore di 4 mV-Hz indicava il 100% di probabilità di insuccesso. L’area sotto la curva ROC è risultata 0.872, 0.869 e 0.875 rispettivamente per tutte le DF, le prime e le successive. In questa vasta popolazione di pazienti con arresto cardiaco un algoritmo AMSA è stato capace di predire l’esito della DF con notevole accuratezza. Una soglia specifica di AMSA che identifichi le probabilità di successo o insuccesso è ottenibile anche durante massaggio cardiaco. Pertanto la decisione di attuare la DF basata sull’AMSA può essere utile per migliorare l’esito della rianimazione cardiopolmonare.
AMPLITUDE SPECTRUM AREA AS A PREDICTOR OF SUCCESSFUL DEFIBRILLATION: THRESHOLD VALUES ANALYSIS IN A LARGE DATABASE OF OUT-OF-HOSPITAL CARDIAC ARREST TREATED BY DC-SHOCK
FINZI, ANDREA ALFONSO
2013
Abstract
Timing of defibrillation (DF) to interrupt ventricular fibrillation (VF) is of utmost importance as it implies interruption of chest compression and,in case of failure, the risk of adding further damage to the already critical myocardial condition. In such scenario, a real-time indicator of the probability of success is needed . This study was aimed to the capability of “Amplitude Spectrum Area” (AMSA) to predict DF outcome in a large database of out-of-hospital cardiac arrest. Electrocardiographic (ECG) data recorded by automated external defibrillators were obtained from 8.419 cardiac arrest events occurring in 7 provinces in Lombardia Region, Italy, between 2008 and 2009. Among these events, only VF/VT cardiac arrests receiving DFs were selected (n=1055). A 2 sec ECG window ending at 0.5 sec before DF was analyzed and AMSA calculated, after fast Fourier transformation. DF was defined as successful in the presence of spontaneous rhythm 40 bpm starting within 60 secs from the DF. Threshold values of AMSA able to discriminate DF outcome were individuated and sensitivity, specificity, accuracy, positive and negative predictive values (PPV, NPV) were calculated. The area under the receiver operating characteristic (ROC) curve was measured. A total of 2.442 quality DF events, including 1055 first attempts and 1.387 subsequent ones were included in the analyses. DF success rate was of 26%, 27%, and 25.2% for all, first, and subsequent DFs, respectively. AMSA was significantly greater prior to successful DFs, compared to that preceding unsuccessful ones (13.8 vs. 6.9 mV-Hz, and 13.9 vs. 6.8 mV-Hz, and 13.7 vs. 7 mV-Hz, for all, first, and subsequent DFs respectively). Intersection of sensitivity, specificity and accuracy curves identified a threshold value of AMSA of approximately 9.5 mV-Hz, able to predict DF outcome, with a balanced sensitivity, specificity and accuracy of 80%, for all, first, and subsequent DFs . Moreover, intersection of PPV and accuracy curves identified a threshold value of AMSA of approximately 15 mV-Hz able to predict a successful DF with a PPV and accuracy of 80%, for all, first, and subsequent DF attempts. AMSA values greater than 27 mV-Hz correctly predicted the success of DF with a PPV value of 100%. AMSA below 8 mV-Hz correctly predicted the DF failure with a NPV of > 95%, for all, first, and subsequent DFs. Further decreases in AMSA values below 4 mV-Hz achieved a NPV of 100%. Area under ROC curves was 0.872, 0.869, and 0.875 for all, first, and subsequent DFs, respectively In this large patient population, an AMSA algorithm was capable to predict DF outcome with high accuracy. A specific AMSA threshold in order to predict DF outcome, i.e. success or failure, may be identified during CPR. An AMSA-based DF decision therefore would be an useful approach to guide the best CPR intervention.File | Dimensione | Formato | |
---|---|---|---|
phd_unimi_R08598.pdf
accesso aperto
Dimensione
1.01 MB
Formato
Adobe PDF
|
1.01 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/85521
URN:NBN:IT:UNIMI-85521