A novel algorithm to assess the quality of 12-lead ECG recordings: validation in a real telecardiology application.

Objective. Automatic detection of Electrocardiograms (ECG) quality is fundamental to minimize costs and risks related to delayed diagnosis due to low ECG quality. Most algorithms to assess ECG quality include non-intuitive parameters. Also, they were developed using data non-representative of a real-world scenario, in terms of pathological ECGs and overrepresentation of low-quality ECG. Therefore, we introduce an algorithm to assess 12-lead ECG quality, Noise Automatic Classification Algorithm (NACA) developed in Telehealth Network of Minas Gerais (TNMG).Approach. NACA estimates a signal-to-noise ratio (SNR) for each ECG lead, where 'signal' is an estimated heartbeat template, and 'noise' is the discrepancy between the template and the ECG heartbeat. Then, clinically-inspired rules based on SNR are used to classify the ECG as acceptable or unacceptable. NACA was compared with Quality Measurement Algorithm (QMA), the winner of Computing in Cardiology Challenge 2011 (ChallengeCinC) by using five metrics: sensitivity (Se), specificity (Sp), positive predictive value (PPV),F2, and cost reduction resulting from adoption of the algorithm. Two datasets were used for validation: TestTNMG, consisting of 34 310 ECGs received by TNMG (1% unacceptable and 50% pathological); ChallengeCinC, consisting of 1000 ECGs (23% unacceptable, higher than real-world scenario).Main results. Both algorithms reached a similar performance on ChallengeCinC, although NACA performed considerably better than QMA in TestTNMG (Se = 0.89 versus 0.21; Sp = 0.99 versus 0.98; PPV = 0.59 versus 0.08;F2= 0.76 versus 0.16 and cost reduction 2.3 ± 1.8% versus 0.3 ± 0.3%, respectively).Significance. Implementing of NACA in a telecardiology service results in evident health and financial benefits for the patients and the healthcare system.

Physiological mechanism and spatial distribution of increased alveolar dead-space in early ARDS: An experimental study.

BACKGROUND: We aimed to investigate the physiological mechanism and spatial distribution of increased physiological dead-space, an early marker of ARDS mortality, in the initial stages of ARDS. We hypothesized that: increased dead-space results from the spatial redistribution of pulmonary perfusion, not ventilation; such redistribution is not related to thromboembolism (ie, areas with perfusion = 0 and infinite ventilation-perfusion ratio, V ˙ / Q ˙ ), but rather to moderate shifts of perfusion increasing V ˙ / Q ˙ in non-dependent regions. METHODS: Five healthy anesthetized sheep received protective ventilation for 20 hours, while endotoxin was continuously infused. Maps of voxel-level lung ventilation, perfusion, V ˙ / Q ˙ , CO2 partial pressures, and alveolar dead-space fraction were estimated from positron emission tomography at baseline and 20 hours. RESULTS: Alveolar dead-space fraction increased during the 20 hours (+0.05, P = .031), mainly in non-dependent regions (+0.03, P = .031). This was mediated by perfusion redistribution away from non-dependent regions (-5.9%, P = .031), while the spatial distribution of ventilation did not change, resulting in increased V ˙ / Q ˙ in non-dependent regions. The increased alveolar dead-space derived mostly from areas with intermediate V ˙ / Q ˙ (0.5≤ V ˙ / Q ˙ ≤10), not areas of nearly "complete" dead-space ( V ˙ / Q ˙ >10). CONCLUSIONS: In this early ARDS model, increases in alveolar dead-space occur within 20 hours due to the regional redistribution of perfusion and not ventilation. This moderate redistribution suggests changes in the interplay between active and passive perfusion redistribution mechanisms (including hypoxic vasoconstriction and gravitational effects), not the appearance of thromboembolism. Hence, the association between mortality and increased dead-space possibly arises from the former, reflecting gas-exchange inefficiency due to perfusion heterogeneity. Such heterogeneity results from the injury and exhaustion of compensatory mechanisms for perfusion redistribution.

Recurrence plots for the assessment of patient-ventilator interactions quality during invasive mechanical ventilation.

Inappropriate patient-ventilator interactions' (PVI) quality is associated with adverse clinical consequences, such as patient anxiety/fear and increased need of sedative and paralytic agents. Thus, technological devices/tools to support the recognition and monitoring of different PVI quality are of great interest. In the present study, we investigate two tools based on a recent landmark study which applied recurrence plots (RPs) and recurrence quantification analysis (RQA) techniques in non-invasive mechanical ventilation. Our interest is in how this approach could be a daily part of critical care professionals' routine (which are not familiar with dynamical systems theory methods and concepts). Two representative time series of three typical PVI "scenarios" were selected from 6 critically ill patients subjected to invasive mechanical ventilation. First, both the (i) main signatures in RPs and the (ii) respective signals that provide the most (visually) discriminant RPs were identified. This allows one to propose a visual identification protocol for PVIs' quality through the RPs' overall aspect. Support for the effectiveness of this visual based assessment tool is given by a RQA-based assessment tool. A statistical analysis shows that both the recurrence rate and the Shannon entropy are able to identify the selected PVI scenarios. It is then expected that the development of an objective method can reliably identify PVI quality, where the results corroborate the potential of RPs/RQA in the field of respiratory pattern analysis.

Estimation of confidence limits for descriptive indexes derived from autoregressive analysis of time series: Methods and application to heart rate variability.

The growing interest in personalized medicine requires making inferences from descriptive indexes estimated from individual recordings of physiological signals, with statistical analyses focused on individual differences between/within subjects, rather than comparing supposedly homogeneous cohorts. To this end, methods to compute confidence limits of individual estimates of descriptive indexes are needed. This study introduces numerical methods to compute such confidence limits and perform statistical comparisons between indexes derived from autoregressive (AR) modeling of individual time series. Analytical approaches are generally not viable, because the indexes are usually nonlinear functions of the AR parameters. We exploit Monte Carlo (MC) and Bootstrap (BS) methods to reproduce the sampling distribution of the AR parameters and indexes computed from them. Here, these methods are implemented for spectral and information-theoretic indexes of heart-rate variability (HRV) estimated from AR models of heart-period time series. First, the MS and BC methods are tested in a wide range of synthetic HRV time series, showing good agreement with a gold-standard approach (i.e. multiple realizations of the "true" process driving the simulation). Then, real HRV time series measured from volunteers performing cognitive tasks are considered, documenting (i) the strong variability of confidence limits' width across recordings, (ii) the diversity of individual responses to the same task, and (iii) frequent disagreement between the cohort-average response and that of many individuals. We conclude that MC and BS methods are robust in estimating confidence limits of these AR-based indexes and thus recommended for short-term HRV analysis. Moreover, the strong inter-individual differences in the response to tasks shown by AR-based indexes evidence the need of individual-by-individual assessments of HRV features. Given their generality, MC and BS methods are promising for applications in biomedical signal processing and beyond, providing a powerful new tool for assessing the confidence limits of indexes estimated from individual recordings.

Effects of different levels of pressure support on intra-individual breath-to-breath variability.

BACKGROUND: Evidence exists that during pressure support ventilation (PSV), the addition of an extrinsic (ie, ventilator-generated) breath-to-breath variability (BBV) of breathing pattern improves respiratory function. If BBV is beneficial per se, choosing the PS level that maximizes it could be considered a valid strategy for conventional PSV. In this study, we evaluated the effect of different PS levels on intrinsic BBV in acutely ill, mechanically ventilated subjects to determine whether a significant relationship exists between PS level and BBV magnitude. METHODS: Fourteen invasively mechanically ventilated subjects were prospectively studied. PS was adjusted at 20 cm H2O and sequentially reduced to 15, 10, and 5 cm H2O. Arterial blood gas analysis and pressure at 0.1 s after the onset of inspiration (P0.1) were measured at each PS level. Airway and esophageal pressure and air flow were continuously recorded. Peak inspiratory flow, tidal volume (VT), breathing frequency, and pressure-time product (PTP) were calculated on a breath-by-breath basis. The breathing pattern variability was assessed by the coefficient of variation of the time series of VT, peak inspiratory flow, and breathing frequency from ∼ 60 consecutive breath cycles at each PS level. A general linear model for repeated measures was applied, with PS as an independent factor. A significance level of .05 was considered. RESULTS: Despite a large inter-individual difference in all measured variables (P < .001), the coefficient of variation was as low as 30%, and no significant differences in the coefficient of variation of peak inspiratory flow, breathing frequency, and VT between PS levels were observed (P > .15). Additionally, a significant increase in P0.1, PTP, and breathing frequency (P < .01) and a reduction in VT (P < .001) were observed with PS reduction. CONCLUSIONS: Despite a significant increase in spontaneous activity with PS reduction, BBV was not influenced by the PS level and was as low as 30% for all evaluated parameters.

Low-frequency heart rate variability is related to the breath-to-breath variability in the respiratory pattern.

Changes in heart rate variability (HRV) at "respiratory" frequencies (0.15-0.5 Hz) may result from changes in respiration rather than autonomic control. We now investigate if the differences in HRV power in the low-frequency (LF) band (0.05-0.15 Hz, HRV(LF)) can also be predicted by respiration variability, quantified by the fraction of tidal volume power in the LF (V(LF,n)). Three experimental protocols were considered: paced breathing, mental effort tasks, and a repeated attentional task. Significant intra- and interindividual correlations were found between changes in HRV(LF) and V(LF,n) despite all subjects having a respiratory frequency above the LF band. Respiratory parameters (respiratory period, tidal volume, and V(LF,n)) could predict up to 79% of HRV(LF) differences in some cases. This suggests that respiratory variability is another mechanism of HRV(LF) generation, which should be always monitored, assessed, and considered in the interpretation of HRV changes.

A numerical model of the respiratory modulation of pulmonary shunt and PaO2 oscillations for acute lung injury.

It is an accepted hypothesis that the amplitude of the respiratory-related oscillations of arterial partial pressure of oxygen (DeltaPaO2) is primarily modulated by fluctuations of pulmonary shunt (Deltas), the latter generated mainly by cyclic alveolar collapse/reopening, when present. A better understanding of the relationship between DeltaPaO2, Deltas, and cyclic alveolar collapse/reopening can have clinical relevance for minimizing the severe lung damage that the latter can cause, for example during mechanical ventilation (MV) of patients with acute lung injury (ALI). To this aim, we numerically simulated the effect of such a relationship on an animal model of ALI under MV, using a combination of a model of lung gas exchange during tidal ventilation with a model of time dependence of shunt on alveolar collapse/opening. The results showed that: (a) the model could adequately replicate published experimental results regarding the complex dependence of DeltaPaO2 on respiratory frequency, driving pressure (DeltaP), and positive end-expiratory pressure (PEEP), while simpler models could not; (b) such a replication strongly depends on the value of the model parameters, especially of the speed of alveolar collapse/reopening; (c) the relationship between DeltaPaO2 and Deltas was overall markedly nonlinear, but approximately linear for PEEP>or=6 cmH2O, with very large DeltaPaO2 associated with relatively small Deltas.

Heart-rate and blood-pressure variability during psychophysiological tasks involving speech: influence of respiration.

Changes in heart-rate and systolic arterial pressure variability (HRV and SAPV) indexes have been used in psychophysiology to assess autonomic activation, including during tasks involving speech. The current article clearly demonstrates in a sample of 25 adult subjects that the erratic and broadband respiratory patterns during such tasks violate the usual assumption that respiration is limited to the high-frequency band (0.15-0.4 Hz). For these tasks, interindividual differences and rest-task changes in HRV and SAPV in the low-frequency band (0.04-0.15 Hz) can be explained, to a large extent, by variations in the respiratory volume signal. This makes the use of HRV and SAPV as markers of autonomic function during these tasks highly questionable. Furthermore, a number of subjects with long respiratory period at rest were identified, whose presence in the sample can bias the estimation of baseline rest values.