Patient-ventilator asynchrony during assisted mechanical ventilation

Artificial intelligence in the management of patient-ventilator asynchronies: A scoping review

BACKGROUND

Patient-ventilator asynchronies (PVAs) are frequent complications in mechanically ventilated patients, contributing to adverse outcomes such as ventilator-induced lung injury, prolonged mechanical ventilation, and increased mortality. Artificial intelligence (AI) has emerged as a promising tool for enhancing PVA detection, prediction, and optimization. Despite its growing potential, the full scope of AI applications in this field and persistent gaps in evidence remain inadequately explored.

OBJECTIVE

This scoping review examines the breadth of AI-based approaches for managing PVAs, identifying key methodologies, evaluating research trends, and highlighting limitations in the current literature.

METHODS

A comprehensive search was conducted in PubMed, Embase, Science Direct, IEEE Xplore, and the Cochrane Library without time restrictions. Extracted data included study objectives, AI methodologies, patient populations, performance metrics, and clinical outcomes. The findings were synthesized into thematic categories to map advancements and research gaps.

RESULTS

Twenty-six studies were identified that applied AI techniques to detect, predict, or optimize PVAs. The included studies employed a range of AI methodologies, including convolutional neural networks, long short-term memory networks, and hybrid algorithms. These models demonstrated high predictive performance, with accuracy ranging from 87 % to 99 % and AUROC values exceeding 0.98 for detecting complex asynchronous events. AI systems also showed promise in predicting weaning success and optimizing ventilatory settings through real-time patient-specific adjustments. However, challenges such as reliance on single-center datasets, inconsistencies in calibration, and limited implementation of explainable AI frameworks restrict their clinical applicability.

CONCLUSIONS

AI holds transformative potential in managing PVAs by enabling real-time detection, improved weaning prediction, and personalized ventilatory strategies. However, significant challenges remain, particularly the need for multi-center validation, standardized reporting protocols, and randomized controlled trials to evaluate clinical efficacy. Addressing these gaps is essential for integrating AI into routine critical care practice and transitioning from theoretical models to practical, real-world applications in intensive care units.

Effects of Trigger Algorithms on Trigger Performance and Patient-Ventilator Synchrony

Background: Patient-ventilator synchrony is essential for successful patient-triggered ventilation. This study compared the ability of a trigger algorithm, based on detailed analysis of flow changes (IntelliSync+, Hamilton Medical), to trigger patient breaths with conventional algorithms. Methods: Three models with different lung mechanics (normal, ARDS, and COPD) at 3 severities were simulated with a lung model ventilated in pressure control continuous mandatory ventilation or pressure control continuous spontaneous ventilation (PC-CSV). Inspiratory pressure above PEEP was set at 15 cm H2O and PEEP at 5 cm H2O. Inspiratory trigger was selected from IntelliSync+ (IS+insp), flow trigger (1- 5 L/min), or pressure trigger (-1 to -5 cm H2O). In PC-CSV, expiratory trigger was set at IntelliSync+ (IS+exp) or cycling criteria (5%, 25%, and 40% for ARDS, normal, and COPD, respectively). Measurements were performed with and without leak (50% inspiratory tidal volume). Five breaths per condition were collected to calculate trigger delay time and asynchronous events. Results: For pressure trigger, none of the conditions resulted in 3 successfully triggered consecutive breaths. Overall trigger delay time was significantly longer with flow trigger than with IS+insp in normal (99 vs 81 ms without leak, P < .001; 98 vs 80 ms with leak, P < .001) and ARDS models (334 vs 223 ms without leak, P < .001; 320 vs 236 ms with leak, P = .02). Across all conditions, ineffective efforts occurred more frequently with flow trigger than with IS+insp (7.3% vs 1.5% without leak, P = .01; 10.8% vs 3.0% with leak, P = .01). In PC-CSV, overall cycling delay time with IS+exp was equivalent or longer compared with cycling criteria. Conclusions: In this lung model study, IS+insp demonstrated similar trigger time and fewer ineffective efforts compared with flow trigger even in simulated respiratory conditions, whereas cycling delay time was unaffected by IS+exp because of large variations between conditions.

ARDS Weaning: The Impact of Abnormal Breathing Patterns Detected by Electric Tomography Impedance and Respiratory Mechanics Monitoring

Background: After the improvement of the initial phase of ARDS, when the patients begin spontaneous breathing and weaning from mechanical ventilation, some patients may present abnormal breathing patterns, whose evaluation of the repercussions were poorly studied. This study proposed to evaluate abnormal breathing patterns through the use of electrical impedance tomography (EIT), and clinical, respiratory mechanics, and ventilatory parameters according to the types of weaning from mechanical ventilation. Methods: This was a prospective cohort study of subjects with ARDS who were considered able to be weaned from mechanical ventilation in the clinical-surgical ICU. Weaning types were defined as simple, difficult, and prolonged weaning. EIT, ventilatory, lung mechanics, and clinical data were collected. Data were collected at baseline in a controlled ventilatory mode and, after neuromuscular blocker withdrawal, data were collected after 30 min, 2 h, and 24 h. EIT parameter analysis was performed for ventilation distribution in the lung regions, pendelluft, breath-stacking, reverse-trigger, double-trigger, and asynchrony index. Results: The study included 25 subjects who were divided into 3 groups (9/25 simple, 8/25 difficult, and 8/25 prolonged weaning). The prolonged weaning group showed more delirium, ICU-acquired weakness, stay in ICU, and hospital and ICU mortality. During the change from controlled to spontaneous mode, we observed increased tidal volumes and driving pressures, which were mainly found in the prolonged weaning group when compared with the simple weaning group. The prolonged weaning group showed a higher flow index, more asynchronies during volume-assisted ventilation, a higher incidence of pendelluft, and redistribution of ventilation to posterior regions visualized by EIT. Conclusions: The present study showed abnormal breathing patterns in the prolonged weaning group. The clinical occult findings of abnormal breathing patterns could be monitored, mainly through EIT and with better assessment of pulmonary mechanics.

A Bench Model of Asynchrony in 6 Ventilators Equipped With Waveform-Guided Options

Background: Pressure support ventilation is frequently associated with patient-ventilator asynchrony. Algorithms based on ventilator waveforms have been developed to automatically detect patient respiratory activity and to guide triggering and cycling. The aim of this study was to assess the performance in terms of synchronization of 6 mechanical ventilators, all provided with a waveform-guided software. Methods: This was a bench study to compare standard and new-generation systems simulating different respiratory mechanics, levels of assistance, and respiratory efforts. Six mechanical ventilators were tested: Hamilton G5 (G5) and C6 (C6), IMT bellavista1000 (B1000), Mindray SV300, and Philips RespironicsV200 (V200) and V60 (V60). Apart from V60, the other ventilators were tested twice: with default settings for standard triggering and cycling and with the waveform-guided automation. Results: With the automated settings, breaths with trigger delay ≤ 300 ms increased with B1000, G5, and C6. Ineffective efforts decreased with B1000, G5, C6, and V200. Improvement of triggering was mainly driven by findings obtained in the obstructive profile. With the automated settings, breaths with cycling delay > 300 ms decreased with B1000, G5, C6, and V200 while early cycled breaths increased with B1000. Improvement of cycling was mainly driven by findings obtained in the obstructive profile, whereas worsening of cycling was observed in the restrictive profile with 2 ventilators (B100 and V200). With the automated settings, the asynchrony index (AI) was reduced with G5 and C6 when all the profiles were grouped. In the obstructive profile, the AI decreased with B1000, G5, C6, and V200; in the restrictive profile, the AI increased with B1000. Conclusions: Waveforms-based algorithms have the potential to improve patient-ventilator synchronization. Automation had the most favorable impact when obstructive patients were simulated, while caution should be paid with restrictive ones.

Ineffective respiratory efforts and their potential consequences during mechanical ventilation

The implementation of invasive mechanical ventilation (IMV) in critically ill patients involves two crucial moments: the total control phase, affected among other things by the use of analgesics and sedatives, and the transition phase to spontaneous ventilation, which seeks to shorten IMV times and where optimizing patient-ventilator interaction is one of the main challenges. Ineffective inspiratory efforts (IEE) arise when there is no coordination between patient effort and ventilator support. IIE are common in different ventilatory modes and are associated with worse clinical outcomes: dyspnea, increased sedation requirements, increased IMV days and longer intensive care unit (ICU) and hospital stay. These are manifested graphically as an abrupt decrease in expiratory flow, being more frequent during expiration. However, and taking into consideration that it is still unknown whether this association is causal or rather a marker of disease severity, recognizing the potential physiological consequences, reviewing diagnostic methods and implementing detection and treatment strategies that can limit them, seems reasonable.

An overview of patient-ventilator asynchrony in children

INTRODUCTION

Mechanically ventilated children often have patient-ventilator asynchrony (PVA). When a ventilated patient has spontaneous effort, the ventilator attempts to synchronize with the patient, but PVA represents a mismatch between patient respiratory effort and ventilator delivered breaths.

AREAS COVERED

This review will focus on subtypes of patient ventilator asynchrony, methods to detect or measure PVA, risk factors for and characteristics of patients with PVA subtypes, potential clinical implications, treatment or prevention strategies, and future areas for research. Throughout this review, we will provide pediatric specific considerations.

EXPERT OPINION

PVA in pediatric patients supported by mechanical ventilation occurs frequently and is understudied. Pediatric patients have unique physiologic and pathophysiologic characteristics which affect PVA. While recognition of PVA and its subtypes is important for bedside clinicians, the clinical implications and risks versus benefits of treatment targeted at reducing PVA remain unknown. Future research should focus on harmonizing PVA terminology, refinement of automated detection technologies, determining which forms of PVA are harmful, and development of PVA-specific ventilator interventions.

Impact of Underassisted Ventilation on Diaphragm Function and Structure in a Porcine Model

BACKGROUND

Long-term controlled mechanical ventilation in the intensive care unit induces ventilator-induced diaphragm dysfunction (VIDD). The transition from controlled mechanical ventilation to assisted mechanical ventilation is a challenge that requires clinicians to balance overassistance and underassistance. While the effects of overassistance on the diaphragm are well known, the authors aimed to assess the impact of underassistance on diaphragm function and structure in a piglet model with preexisting VIDD (after long-term controlled mechanical ventilation) or without VIDD (short-term controlled mechanical ventilation).

METHODS

Twenty-two Large White female piglets were anesthetized, ventilated, and separated into two groups: a VIDD group (n = 10) with long-term 72-h controlled mechanical ventilation, and a no-VIDD group (n = 12) with short-term 2-h controlled mechanical ventilation. After sedation reduction at the end of the controlled mechanical ventilation period, each piglet was switched to underassisted ventilation for 2 h. Diaphragm function (supramaximal diaphragm pressure-generating capacity assessed by negative tracheal pressure after transvenous phrenic nerve stimulation) and diaphragm structure (mini-invasive in vivo biopsies) were assessed before and after underassisted ventilation.

RESULTS

In the VIDD group, supramaximal diaphragm pressure-generating capacity decreased by 22% from (mean ± SD) 69.9 ± 12.7 to 54.9 ± 19.7 cm H 2 O ( P = 0.04) after 72 h of controlled mechanical ventilation evidencing VIDD, then dropped by a further 29% from 54.9 ± 19.7 to 38.9 ± 15.5 cm H 2 O ( P < 0.01) after 2 h of underassisted ventilation. Diaphragm pressure-generating capacity remains stable from 55.3 ± 22.7 to 58.2 ± 24 cm H 2 O ( P = 0.24) in the no-VIDD group. Diaphragm structure showed that sarcomeric injuries increase from 13 ± 10% to 24 ± 19% ( P < 0.01) and lipid droplets decrease from 14 ± 8% to 11 ± 6% ( P = 0.03) of the total micrograph area after 2 h of underassisted ventilation in the VIDD group. Sarcomeric injuries and lipid droplets accounted, respectively, for 17 ± 16% and 2 ± 3% of the total micrograph area after underassisted ventilation in the no-VIDD group.

CONCLUSIONS

In this porcine model, a short 2-h exposure of underassisted ventilation induces impairment of diaphragm function with damage to the diaphragm structure in intensive care unit condition with preexisting VIDD.

Monitoring effort and respiratory drive in patients with acute respiratory failure

PURPOSE OF REVIEW

Accurate monitoring of respiratory drive and inspiratory effort is crucial for optimizing ventilatory support during acute respiratory failure. This review evaluates current and emerging bedside methods for assessing respiratory drive and effort.

RECENT FINDINGS

While electrical activity of the diaphragm and esophageal pressure remain the reference standards for assessing respiratory drive and effort, their clinical utility is largely limited to research. At the bedside, airway occlusion maneuvers are the most useful tools: P0.1 is a reliable marker of drive and detects abnormal inspiratory efforts, while occlusion pressure (Pocc) may outperform P0.1 in identifying excessive effort. The Pressure-Muscle-Index (PMI) can help detecting insufficient inspiratory effort, though its accuracy depends on obtaining a stable plateau pressure. Other techniques, such as central venous pressure swings (ΔCVP), are promising but require further investigation. Emerging machine learning and artificial intelligence based algorithms could play a pivotal role in automated respiratory monitoring in the near future.

SUMMARY

Although Pes and EAdi remain reference methods, airway occlusion maneuvers are currently the most practical bedside tools for monitoring respiratory drive and effort. Noninvasive alternatives such as ΔCVP deserve further evaluation. Artificial intelligence and machine learning may soon provide automated solutions for bedside monitoring of respiratory drive and effort.

Therapeutic novelties in acute heart failure and practical perspectives

Acute Heart Failure (AHF) is a leading cause of death and represents the most frequent cause of unplanned hospital admission in patients older than 65 years. Since the past decade, several randomized clinical trials have highlighted the importance and pivotal role of certain therapeutics, including decongestion by the combination of loop diuretics, the need for rapid goal-directed medical therapies implementation before discharge, risk stratification, and early follow-up after discharge therapies. Cardiogenic shock, defined as sustained hypotension with tissue hypoperfusion due to low cardiac output and congestion, is the most severe form of AHF and mainly occurs after acute myocardial infarction, which can progress to multiple organ failure. Although its prevalence is relatively low, cardiogenic shock complicates 12% of acute myocardial infarction. After a brief summary of the epidemiology of AHF and cardiogenic shock, followed by key pathophysiological points, we detailed current treatments in AHF and cardiogenic shock what every anaesthesiologist and intensivist needs to know, based on the latest guidelines and randomized clinical trials published in recent years.

Let's get in sync: current standing and future of AI-based detection of patient-ventilator asynchrony

BACKGROUND

Patient-ventilator asynchrony (PVA) is a mismatch between the patient's respiratory drive/effort and the ventilator breath delivery. It occurs frequently in mechanically ventilated patients and has been associated with adverse events and increased duration of ventilation. Identifying PVA through visual inspection of ventilator waveforms is highly challenging and time-consuming. Automated PVA detection using Artificial Intelligence (AI) has been increasingly studied, potentially offering real-time monitoring at the bedside. In this review, we discuss advances in automatic detection of PVA, focusing on developments of the last 15 years.

RESULTS

Nineteen studies were identified. Multiple forms of AI have been used for the automated detection of PVA, including rule-based algorithms, machine learning and deep learning. Three licensed algorithms are currently reported. Results of algorithms are generally promising (average reported sensitivity, specificity and accuracy of 0.80, 0.93 and 0.92, respectively), but most algorithms are only available offline, can detect a small subset of PVAs (focusing mostly on ineffective effort and double trigger asynchronies), or remain in the development or validation stage (84% (16/19 of the reviewed studies)). Moreover, only in 58% (11/19) of the studies a reference method for monitoring patient's breathing effort was available. To move from bench to bedside implementation, data quality should be improved and algorithms that can detect multiple PVAs should be externally validated, incorporating measures for breathing effort as ground truth. Last, prospective integration and model testing/finetuning in different ICU settings is key.

CONCLUSIONS

AI-based techniques for automated PVA detection are increasingly studied and show potential. For widespread implementation to succeed, several steps, including external validation and (near) real-time employment, should be considered. Then, automated PVA detection could aid in monitoring and mitigating PVAs, to eventually optimize personalized mechanical ventilation, improve clinical outcomes and reduce clinician's workload.

"Evaluation of Different Inspiratory Efforts in a Pediatric Model of Healthy Lung and Pediatric Acute Respiratory Distress Syndrome During Optimized Pressure Support Ventilation and Proportional Assist Ventilation Plus (PAV+): A Bench Study

INTRODUCTION

While proportional ventilator modes have gained popularity in adult patients' ventilatory management, Proportional Assist Ventilation (PAV+) use in pediatric patients with Pediatric Acute Respiratory Distress Syndrome (PARDS) remains unexplored. This study aims to evaluate the effects of optimized PSV and PAV+ on patient-ventilator interaction and respiratory pattern in two pediatric simulated lung models.

METHODS

The study utilized an active lung simulator to replicate two pediatric lung models: one healthy and one with mild PARDS. Each model was ventilated using PAV+ and optimized PSV at four different levels, assessing simulated patient-ventilator interaction and mechanical response to increase inspiratory effort.

RESULTS

In terms of simulated patient-ventilator interaction, in a healthy and mild PARDS lung model and all setting tested, the optimized PSV presented the better patient-ventilator interaction with the shortest values of Inspiratory trigger delay (Delaytrinsp), Pressurization time (Timepress) and Expiratory trigger delay (Delaytrexp) and the highest values of Synchrony time (Timesynch). Only in the lung model with PARDS, during high assistance levels and high Pmus, no significant differences were found in terms of patient ventilation interaction between the two modalities.

CONCLUSIONS

In a healthy lung model, optimized PSV allows optimal simulated patient-ventilator interaction and assistance levels compared to PAV+. On the contrary, in a simulated lung with mild PARDS, PAV+ appears as a valid alternative to PSV, especially under conditions of intense inspiratory effort and high assistance levels.

Visual Patterns of Diaphragmatic Motion in Acute Respiratory Failure: A Prospective Pilot Study

INTRODUCTION

Right diaphragmatic excursion is a reliable and reproducible technique used in intensive care to assess diaphragmatic function. The aim of this study was to investigate the relationship between the appearance of diaphragmatic motion and the etiological diagnosis of patients admitted to the emergency department with acute respiratory failure (ARF).

MATERIALS

A prospective, observational, and multicenter pilot study was conducted. All adult patients admitted in the emergency department with ARF were included. The different visual patterns of diaphragmatic motion were analyzed according to the three main etiologies of ARF encountered in emergency departments.

RESULTS

A total of 39 adult patients were included. We observed a different visual pattern in patients with pneumonia. A sum of plateau times of less than 0.2 s predicted that the main diagnosis was pneumonia, with sensitivity = 89% 95%CI [52%; 100%], specificity = 87% 95%CI [69%; 96%].

CONCLUSION

Our study seems to show that the shape of diaphragmatic motion in patients with ARF secondary to pneumonia is different from that in patients with exacerbation of chronic obstructive pulmonary disease or acute heart failure.

TRIAL REGISTRATION

ClinicalTrials.gov: NCT04591509.

Effect of neurally adjusted ventilator assist versus pressure support ventilation on asynchronies and cardiac function in pediatric liver transplantation

In pediatric liver recipients perioperative factors may affect respiratory and cardiac function, and prolong mechanical ventilation during post-operative period. The use of NAVA can improve the interaction between the patient and the ventilator from both a respiratory and cardiac perspective. The objective of this study is to evaluate the synchronization between the patient and the ventilator, as well as cardiac function, during the application of neurally adjusted ventilatory assist (NAVA) and pressure support ventilation (PSV) in pediatric liver transplant recipients. This is a single-center, prospective, randomized, physiological cross-over controlled trial conducted between 2021 and 2022. Children (1 month-10 years old) who underwent liver transplantation were admitted to the pediatric intensive care unit. Patients were randomised to one of two crossover sequences of ventilation trials of 40 min each (PSV/NAVA/PSV or NAVA/PSV/NAVA). Cardiac function was studied by echocardiogram. Twenty-four patients were enrolled and 21 completed the study. Primary outcomes were variation of asynchrony index (AI) and tricuspid annular plane systolic excursion (TAPSE) during the two ventilation modes. Secondary outcomes were patient-ventilator interaction parameters, gas exchange, left and right ventricular function, and hemodynamic parameters. NAVA compared to PSV: (1) improves patient-ventilator interaction reducing AI (coeff - 6.66 95% CI -11.5 to -1.78, p = 0.008); (2) does not improve TAPSE (coeff 0.62 95% CI -1.49 to 2.74, p < 0.557) No differences in terms of pulmonary gas exchange and hemodynamic parameters were detected. NAVA (when compared to PSV) improves patient-ventilator interaction in terms of asynchronies without affecting cardiac biventricular function.Trial registration: NCT04792788, Registration date: 2021-03-11.

Patient-ventilator synchrony under non-invasive ventilation is improved by an automated real time waveform analysis algorithm: a bench study

BACKGROUND

Because of inherent leaks, obtaining good patient-ventilator synchrony during non-invasive ventilation (NIV) is challenging. The IntelliSync + ® software (Hamilton medical, Bonaduz, CH), that can be used together with the NIV mode, performs real-time automated analysis of airway pressure- and flow-time curves to detect the transition between inspiration and expiration. It then controls the ventilator inspiratory and expiratory valves to improve patient-ventilator synchrony. The main goal of this NIV bench study was to evaluate the impact of IntelliSync + ® on synchrony in the presence of leaks of 9 and 20 L/min in the tested ventilator circuit (no face mask used), with normal, obstructive and restrictive respiratory mechanics and two levels of NIV pressure support (PS 8 and 14 cmH2O). For this, the time needed to trigger the ventilator (Td) and the difference between the end of the simulated breath and the termination of pressurization (Tiex) were measured. The number of classical asynchronies and the ventilator pressurization capacity were also assessed.

RESULTS

Compared to NIV delivered with the classical NIV mode (compensating leaks and limiting inspiratory time to 2 s), activating IntelliSync + ® improved Tiex and, to a lesser extent, Td in clinically relevant setups. IntelliSync + ® also showed a trend towards reducing classical asynchronies, particularly directly after leak flow increase. The impact of the system was most significant with high PS levels and pathological respiratory mechanics. Especially, in the obstructive model, in the presence of large leak (20 L/min) and PS 14 cmH2O, Tiex decreased from 0.61 [0.56-0.64] to 0.16 [0.07-0.18] s and Td from 0.07 [0.06-0.08] to 0.06 [0.06-0.08] s. In less challenging situations, IntelliSync + ® was less beneficial. Overall, ventilator pressurization was improved when IntelliSync + ® was activated.

CONCLUSIONS

In this NIV bench model, IntelliSync + ®, used in addition to NIV-PS, improved both expiratory and inspiratory synchrony. It was particularly efficient in the presence of obstructive and restrictive respiratory mechanics and high-pressure support levels. These pre-clinical results tend to support the ability of IntelliSync + ® to improve patient-ventilator synchrony in the presence of leaks and provide pre-clinical data supporting a clinical evaluation of the automated algorithm during NIV.

Emergency and critical care medicine residents' competency to identify patient ventilator asynchrony using a mechanical ventilator waveform analysis in Addis Ababa, Ethiopia: a multicenter cross-sectional study

BACKGROUND

patient-ventilator asynchrony (PVA) describes a condition in which a suboptimal interaction occurs between a patient and a mechanical ventilator. It is common and often undetected, with a negative impact on patient outcomes if unrecognized and addressed. Mechanical ventilator waveform analysis is a non-invasive and reliable way of identifying PVAs for which advanced methods of identifying PVA are lacking; however, it has not been well studied in residents working in developing setups like Ethiopia.

OBJECTIVES

to assess Emergency and Critical Care Medicine (ECCM) Residents' competency and associated factors to identify PVA using mechanical ventilator (MV) waveform analysis at Saint Paul Hospital Millennium Medical College (SPHMMC) and Tikur Anbesa Specialized Hospital (TASH).

METHODOLOGY

We conducted a cross-sectional study among senior ECCM residents who were on training at TASH and SPHMMC, Addis Ababa. The study enrolled all 91 senior ECCM residents with 80 completing it. A pretested and structured self-administered questionnaire was administered using an internally modified assessment tool. The completed data were collected via web links after being prepared using kobtoolbox. org, coded, manually checked, and exported to version 27 SPSS analysis. Descriptive statistics, the chi-square test, nonparametric tests, and multi-variable logistic regression were used for data analysis.

RESULTS

Eighty senior residents responded out of 91, including 42 from TASH and 38 from SPHMMC. The overall competency of identifying PVA by MV waveforms was 30%. A median of 3 (IQR 1-4) PVAs were correctly identified. Only 1 resident (1.25%) identified all 6 different types of PVAs,;(8.75%) identified 5 PVAs; 20% identified 4 PVAs,22.5% identified 3 PVAs; 17.5% identified 2 PVAs, 13.75% identified 1 PVA Correctly and 16.25% did not identify any PVA. Auto-PEEP was the most frequently identified PVA, and delayed cycling was the least frequently identified PVA. Presenting or attending a seminar on MV waveforms and having lectures on mechanical ventilation increased the probability of identifying ≥ 4 PVAs.

CONCLUSION

The overall competency of identifying PVA by MV waveforms is low among ECCM residents. Presenting or attending seminars on MV waveforms, and having lectures on mechanical ventilation (MV) were associated with increased competency of identifying PVAs by MV waveform analysis.

Evaluation of health care providers' ability to identify patient-ventilator triggering asynchrony in intensive care unit: a translational observational study in China

BACKGROUND

Patient-ventilator asynchrony (PVA) can result in ventilator-induced lung injury (VILI), prolong mechanical ventilation, and ventilator withdrawal failure. The ability of healthcare providers in China to recognize patient-ventilator asynchrony is unknown. The aim of our study was to evaluate the ability and potential influencing factors to correctly identify patient-ventilator triggering asynchrony in tertiary hospitals in China.

METHODS

This was an observational study carried out in 53 tertiary hospitals in China. A total of 191 healthcare providers were asked to finish entry test and evaluation test sequentially. Entry test identified qualified professionals by matching concepts with its corresponding interpretations. Evaluation test assessed the ability in recognizing patient-ventilator asynchrony waveforms by matching asynchrony waveforms with corresponding concepts. A total of 109 qualified professionals were identified. Further analysis based on professional title, role in critical care team, years of experience in managing invasive mechanical ventilation, number of published articles in the field of clinical critical respiratory medicine and training in respiratory waveform/respiratory mechanics was carried out among qualified professionals. A self-innovate Remote-VentlateView platform was used to discriminate the patient-ventilator triggering asynchrony.

RESULTS

Among 109 qualified professionals, the average recognition accuracy was 3.45 out of 8 sets. Inconsistency of concept cognition and waveform recognition of patient-ventilator asynchrony was found among all types of asynchronies. The accuracy of the trained professionals was greater than that of the nontrained professionals for ineffective trigger [76.7% vs. 59.2% (p = 0.009)], auto-trigger [26.7% vs. 12.2% (p = 0.014)] and reverse triggers [30.8% vs. 12.2% (p = 0.002)]. Professionals who published more than 2 articles in the field of critical respiratory performed better on auto-triggers [41.7% vs. 15.9% (p = 0.001)] and reverse triggers [38.9% vs. 19.2% (p = 0.018)]. Neither experience in managing invasive mechanical ventilation nor professional title was associated with the ability of healthcare providers to identify asynchrony.

CONCLUSIONS

Receiving training in mechanical ventilation and conducting critical respiratory clinical research may increase healthcare providers' ability to identify patient-ventilator asynchrony by using waveform analysis. The Remote-VentlateView platform may assist in identifying patient-ventilator asynchronies.

Management of sedation during weaning from mechanical ventilation

PURPOSES OF REVIEW

Critically ill patients frequently require mechanical ventilation and often receive sedation to control pain, reduce anxiety, and facilitate patient-ventilator interactions. Weaning from mechanical ventilation is intertwined with sedation management. In this review, we analyze the current evidence for sedation management during ventilatory weaning, including level of sedation, timing of sedation weaning, analgesic and sedative choices, and sedation management in acute respiratory distress syndrome (ARDS).

RECENT FINDINGS

Despite a large body of evidence from the past 20 years regarding the importance of light sedation and paired spontaneous awakening and spontaneous breathing trials (SATs/SBTs) to promote ventilator weaning, recent studies show that implementation of these strategies lag in practice. The recent WEAN SAFE trial highlights the delay between meeting weaning criteria and first weaning attempt, with level of sedation predicting both delays and weaning failure. Recent studies show that targeted interventions around evidence-based practices for sedation weaning improve outcomes, though long-term sustainability remains a challenge.

SUMMARY

Light or no sedation strategies that prioritize analgesia prior to sedatives along with paired SATs/SBTs promote ventilator liberation. Dexmedetomidine may have a role in weaning for agitated patients. Further investigation is needed into optimal sedation management for patients with ARDS.

Does patient-ventilator asynchrony really matter?

PURPOSE OF REVIEW

Past observational studies have reported the association between patient-ventilator asynchronies and poor clinical outcomes, namely longer duration of mechanical ventilation and higher mortality. But causality has remained undetermined. During the era of lung and diaphragm protective ventilation, should we revolutionize our clinical practice to detect and treat dyssynchrony?

RECENT FINDINGS

Clinicians' ability to recognize asynchronies is typically low. Automatized softwares based on artificial intelligence have been trained to largely outperform human eyesight and are close to be implemented at the bedside. There is growing evidence that in susceptible patients, dyssynchrony may lead to ventilation-induced lung injury (or patient self-inflicted lung injury) and that clusters of such dyssynchronous events have the highest association with poor outcomes. Dyssynchrony may also be associated with harm indirectly when it reflects over-assistance or over-sedation. However, the occurrence of reverse triggering by means of low inspiratory efforts during passive ventilation may prevent diaphragm dysfunction and atrophy and be beneficial.

SUMMARY

Most recent evidence on the topic suggests that synchrony between the patient and the mechanical ventilator is a critical element for protecting lung and diaphragm during the time of invasive mechanical ventilation or may reflect inadequate settings or sedation. Therefore, it is a complex situation, and clinical trials are still needed to test the effectiveness of keeping patient-ventilator interaction synchronous on clinical outcomes.

Spontaneous breathing-induced lung injury in mechanically ventilated patients

PURPOSE OF REVIEW

Recent experimental and clinical studies have suggested that spontaneous effort can potentially injure the lungs. This review summarizes the harmful effects of spontaneous breathing on the lungs during mechanical ventilation in ARDS and suggests potential strategies to minimize spontaneous breathing-induced lung injury.

RECENT FINDINGS

Recent clinical and experimental studies have shown that vigorous spontaneous breathing during mechanical ventilation can potentially injure the lungs due to high transpulmonary pressure, the Pendelluft phenomenon, increased pulmonary perfusion, and patient-ventilator asynchrony. A definitive approach to minimize spontaneous breathing-induced lung injury is the systemic use of neuromuscular blocking agents; however, there is a risk of muscle atrophy. Alternatively, partial paralysis, bilateral phrenic nerve blockade, and sedatives may be useful for decreasing force generation from the diaphragm while maintaining muscle function. A higher positive end-expiratory pressure (PEEP) and prone positioning may reduce force generation from the diaphragm by decreasing neuromechanical efficiency.

SUMMARY

Several potential strategies, including neuromuscular blockade, partial paralysis, phrenic nerve blockade, sedatives, PEEP, and prone positioning, could be useful to minimize spontaneous breathing-induced lung injury.

Double cycling with breath-stacking during partial support ventilation in ARDS: Just a feature of natural variability?

BACKGROUND

Double cycling with breath-stacking (DC/BS) during controlled mechanical ventilation is considered potentially injurious, reflecting a high respiratory drive. During partial ventilatory support, its occurrence might be attributable to physiological variability of breathing patterns, reflecting the response of the mode without carrying specific risks.

METHODS

This secondary analysis of a crossover study evaluated DC/BS events in hypoxemic patients resuming spontaneous breathing in cross-over under neurally adjusted ventilatory assist (NAVA), proportional assist ventilation (PAV +), and pressure support ventilation (PSV). DC/BS was defined as two inspiratory cycles with incomplete exhalation. Measurements included electrical impedance signal, airway pressure, esophageal and gastric pressures, and flow. Breathing variability, dynamic compliance (CLdyn), and end-expiratory lung impedance (EELI) were analyzed.

RESULTS

Twenty patients under assisted breathing, with a median of 9 [5-14] days on mechanical ventilation, were included. DC/BS was attributed to either a single (42%) or two apparent consecutive inspiratory efforts (58%). The median [IQR] incidence of DC/BS was low: 0.6 [0.1-2.6] % in NAVA, 0.0 [0.0-0.4] % in PAV + , and 0.1 [0.0-0.4] % in PSV (p = 0.06). DC/BS events were associated with patient's coefficient of variability for tidal volume (p = 0.014) and respiratory rate (p = 0.011). DC/BS breaths exhibited higher tidal volume, muscular pressure and regional stretch compared to regular breaths. Post-DC/BS cycles frequently exhibited improved EELI and CLdyn, with no evidence of expiratory muscle activation in 63% of cases.

CONCLUSIONS

DC/BS events during partial ventilatory support were infrequent and linked to breathing variability. Their frequency and physiological effects on lung compliance and EELI resemble spontaneous sighs and may not be considered a priori as harmful.

Optimizing Mechanical Ventilation: A Clinical and Practical Bedside Method for the Identification and Management of Patient-Ventilator Asynchronies in Critical Care

The prompt identification and correction of patient-ventilator asynchronies (PVA) remain a cornerstone for ensuring the quality of respiratory failure treatment and the prevention of further injury to critically ill patients. These disruptions, whether due to over- or under-assistance, have a profound clinical impact not only on the respiratory mechanics and the mortality associated with mechanical ventilation but also on the patient's cardiac output and hemodynamic profile. Strong evidence has demonstrated that these frequently occurring and often underdiagnosed events have significant prognostic value for mechanical ventilation outcomes and are strongly associated with prolonged ICU stays and hospital mortality. Halting the consequences of PVA relies on the correct identification and approach of its underlying causes. However, this often requires advanced knowledge of respiratory physiology and the evaluation of complex ventilator waveforms in patient-ventilator interactions, posing a challenge to intensive care practitioners, in particular, those less experienced. This review aims to outline the most frequent types of PVA and propose a clinical algorithm to provide physicians with a structured approach to assess, accurately diagnose, and correct PVA.

Pressure support, patient effort and tidal volume: a conceptual model for a non linear interaction

Pressure support ventilation (PSV) is a form of assisted ventilation which has become frequently used, with the aim of partially unloading the patient's inspiratory muscles. Both under- and over-assistance should be avoided to target a lung- and diaphragm- protective ventilation. Herein, we propose a conceptual model, supported by actual data, to describe how patient and ventilator share the generation of tidal volume (Vt) in PSV and how respiratory system compliance (Crs) affects this interaction. We describe the presence of a patient-specific range of PSV levels, within which the inspiratory effort (Pmus) is modulated, keeping Vt relatively steady on a desired value (Vttarget). This range of assistance may be considered the "adequate PSV assistance" required by the patient, while higher and lower levels may result in over- and under-assistance respectively. As we also show, the determinants of over- and under- assistance borders depend on the combination of Crs and the inspiratory effort which the patient is able to sustain over a period of time. These concepts can be applied at the bedside to understand if the level of assistance is adequate to patient's demand, focusing on the variation of relevant parameters (Vt, Pmus and pressure-muscle-index) as patient reaction to a change in the level of assistance.

A deep learning model for estimating sedation levels using heart rate variability and vital signs: a retrospective cross-sectional study at a center in South Korea

BACKGROUND

Optimal sedation assessment in critically ill children remains challenging due to the subjective nature of behavioral scales and intermittent evaluation schedules. This study aimed to develop a deep learning model based on heart rate variability (HRV) parameters and vital signs to predict effective and safe sedation levels in pediatric patients.

METHODS

This retrospective cross-sectional study was conducted in a pediatric intensive care unit at a tertiary children's hospital. We developed deep learning models incorporating HRV parameters extracted from electrocardiogram waveforms and vital signs to predict Richmond Agitation-Sedation Scale (RASS) scores. Model performance was evaluated using the area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve (AUPRC). The data were split into training, validation, and test sets (6:2:2), and the models were developed using a 1D ResNet architecture.

RESULTS

Analysis of 4,193 feature sets from 324 patients achieved excellent discrimination ability, with AUROC values of 0.867, 0.868, 0.858, 0.851, and 0.811 for whole number RASS thresholds of -5 to -1, respectively. AUPRC values ranged from 0.928 to 0.623, showing superior performance in deeper sedation levels. The HRV metric SDANN2 showed the highest feature importance, followed by systolic blood pressure and heart rate.

CONCLUSIONS

A combination of HRV parameters and vital signs can effectively predict sedation levels in pediatric patients, offering the potential for automated and continuous sedation monitoring in pediatric intensive care settings. Future multi-center validation studies are needed to establish broader applicability.

Ventilator performances for non-invasive ventilation: a bench study

INTRODUCTION

A wide range of recent ventilators, dedicated or not, is available for non-invasive ventilation (NIV) in respiratory or intensive care units (ICU). We conducted a bench study to compare their technical performances.

METHODS

Ventilators, including five ICU ventilators with NIV mode on, two dedicated NIV ventilators and one transport ventilator, were evaluated on a test bench for NIV, consisting of a 3D manikin head connected to an ASL 5000 lung model via a non-vented mask. Ventilators were tested according to three simulated lung profiles (normal, obstructive, restrictive), three levels of simulated air leakage (0, 15, 30 L/min), two levels of pressure support (8, 14 cmH2O) and two respiratory rates (15, 25 cycles/min).

RESULTS

The global median Asynchrony Index (AI) was higher with ICU ventilators than with dedicated NIV ventilators (4% (0; 76) vs 0% (0; 15), respectively; p<0.05) and different between all ventilators (p<0.001). The AI was higher with ICU ventilators for the normal and restrictive profiles (p<0.01) and not different between ventilators for the obstructive profile. Auto-triggering represented 43% of all patient-ventilator asynchrony. Triggering delay, cycling delay, inspiratory pressure-time product, pressure rise time and pressure at mask were different between all ventilators (p<0.01). Dedicated NIV ventilators induced a lower pressure-time product than ICU and transport ventilators (p<0.01). There was no difference between ventilators for minute ventilation and peak flow.

CONCLUSION

Despite the integration of NIV algorithms, most recent ICU ventilators appear to be less efficient than dedicated NIV ventilators. Technical performances could change, however, according to the underlying respiratory disease and the level of air leakage.

Induction of subject-ventilator asynchrony by variation of respiratory parameters in a lung injury model in pigs

BACKGROUND

Subject-ventilator asynchrony (SVA) was shown to be associated with negative clinical outcomes. To elucidate pathophysiology pathways and effects of SVA on lung tissue histology a reproducible animal model of artificially induced asynchrony was developed and evaluated.

METHODS

Alterations in ventilator parameters were used to induce the three main types of asynchrony: ineffective efforts (IE), auto-triggering (AT), and double-triggering (DT). Airway flow and pressure, as well as oesophageal pressure waveforms, were recorded, asynchrony cycles were manually classified and the asynchrony index (AIX) was calculated. Bench tests were conducted on an active lung simulator with ventilator settings altered cycle by cycle. The developed algorithm was evaluated in three pilot experiments and a study in pigs ventilated for twelve hours with AIX = 25%.

RESULTS

IE and AT were induced reliably and fail-safe by end-expiratory hold and adjustment of respiratory rate, respectively. DT was provoked using airway pressure ramp prolongation, however not controlled specifically in the pilots. In the subsequent study, an AIX = 28.8% [24.0%-34.4%] was induced and maintained over twelve hours.

CONCLUSIONS

The method allows to reproducibly induce and maintain three clinically relevant types of SVA observed in ventilated patients and may thus serve as a useful tool for future investigations on cellular and inflammatory effects of asynchrony.

Patient-ventilator asynchrony classification in mechanically ventilated patients: Model-based or machine learning method?

BACKGROUND AND OBJECTIVE

Patient-ventilator asynchrony (PVA) is associated with poor clinical outcomes and remains under-monitored. Automated PVA detection would enable complete monitoring standard observational methods do not allow. While model-based and machine learning PVA approaches exist, they have variable performance and can miss specific PVA events. This study compares a model and rule-based algorithm with a machine learning PVA method by retrospectively validating both methods using an independent patient cohort.

METHODS

Hysteresis loop analysis (HLA) which is a rule-based method (RBM) and a tri-input convolutional neural network (TCNN) machine learning model are used to classify 7 different types of PVA, including: 1) flow asynchrony; 2) reverse triggering; 3) premature cycling; 4) double triggering; 5) delayed cycling; 6) ineffective efforts; and 7) auto triggering. Class activation mapping (CAM) heatmaps visualise sections of respiratory waveforms the TCNN model uses for decision making, improving result interpretability. Both PVA classification methods were used to classify incidence in an independent retrospective clinical cohort of 11 mechanically ventilated patients for validation and performance comparison.

RESULTS

Self-validation with the training dataset shows overall better HLA performance (accuracy, sensitivity, specificity: 97.5 %, 96.6 %, 98.1 %) compared to the TCNN model (accuracy, sensitivity, specificity: 89.5 %, 98.3 %, 83.9 %). In this study, the TCNN model demonstrates higher sensitivity in detecting PVA, but HLA was better at identifying non-PVA breathing cycles due to its rule-based nature. While the overall AI identified by both classification methods are very similar, the intra-patient distribution of each PVA type varies between HLA and TCNN.

CONCLUSION

The collective findings underscore the efficacy of both HLA and TCNN in PVA detection, indicating the potential for real-time continuous monitoring of PVA. While ML methods such as TCNN demonstrate good PVA identification performance, it is essential to ensure optimal model architecture and diversity in training data before widespread uptake as standard care. Moving forward, further validation and adoption of RBM methods, such as HLA, offers an effective approach to PVA detection while providing clear distinction into the underlying patterns of PVA, better aligning with clinical needs for transparency, explicability, adaptability and reliability of these emerging tools for clinical care.

Gastric Pressure Monitoring Unveils Abnormal Patient-Ventilator Interaction Related to Active Expiration: A Retrospective Observational Study

BACKGROUND

Patient-ventilator dyssynchrony is frequently observed during assisted mechanical ventilation. However, the effects of expiratory muscle contraction on patient-ventilator interaction are underexplored. The authors hypothesized that active expiration would affect patient-ventilator interaction and they tested their hypothesis in a mixed cohort of invasively ventilated patients with spontaneous breathing activity.

METHODS

This is a retrospective observational study involving patients on assisted mechanical ventilation who had their esophageal pressure (Peso) and gastric pressure monitored for clinical purposes. Active expiration was defined as gastric pressure rise (ΔPgas) greater than or equal to 1.0 cm H2O during expiratory flow without a corresponding change in diaphragmatic pressure. Waveforms of Peso, gastric pressure, diaphragmatic pressure, flow, and airway pressure (Paw) were analyzed to identify and characterize abnormal patient-ventilator interaction.

RESULTS

76 patients were identified with Peso and gastric pressure recordings, of whom 58 demonstrated active expiration with a median ΔPgas of 3.4 cm H2O (interquartile range = 2.4 to 5.3) observed in this subgroup. Among these 58 patients, 23 presented the following events associated with expiratory muscle activity: (1) distortions in Paw and flow that resembled ineffective efforts, (2) distortions similar to autotriggering, (3) multiple triggering, (4) prolonged ventilatory cycles with biphasic inspiratory flow, with a median percentage (interquartile range) increase in mechanical inflation time and tidal volume of 54% (44 to 70%) and 25% (8 to 35%), respectively and (5) breathing exclusively by expiratory muscle relaxation. Gastric pressure monitoring was required to identify the association of active expiration with these events. Respiratory drive, assessed by the rate of inspiratory Peso decrease, was significantly higher in patients with active expiration (median [interquartile range] dPeso/dt: 12.7 [9.0 to 18.5] vs 9.2 [6.8 to 14.2] cmH2O/sec; P < 0.05).

CONCLUSIONS

Active expiration can impair patient-ventilator interaction in critically ill patients. Without documenting gastric pressure, abnormal patient-ventilator interaction associated with expiratory muscle contraction may be mistakenly attributed to a mismatch between the patient's inspiratory effort and mechanical inflation. This misinterpretation could potentially influence decisions regarding clinical management.

EDITOR’S PERSPECTIVE

Recruitment-Potential-Oriented Mechanical Ventilation Protocol and Narrative Review for Patients with Acute Respiratory Distress Syndrome

Even though much progress has been made to improve clinical outcomes, acute respiratory distress syndrome (ARDS) remains a significant cause of acute respiratory failure. Protective mechanical ventilation is the backbone of supportive care for these patients; however, there are still many unresolved issues in its setting. The primary goal of mechanical ventilation is to improve oxygenation and ventilation. The use of positive pressure, especially positive end-expiratory pressure (PEEP), is mandatory in this approach. However, PEEP is a double-edged sword. How to safely set positive end-inspiratory pressure has long been elusive to clinicians. We hereby propose a pressure-volume curve measurement-based method to assess whether injured lungs are recruitable in order to set an appropriate PEEP. For the most severe form of ARDS, extracorporeal membrane oxygenation (ECMO) is considered as the salvage therapy. However, the high level of medical resources required and associated complications make its use in patients with severe ARDS controversial. Our proposed protocol also attempts to propose how to improve patient outcomes by balancing the possible overuse of resources with minimizing patient harm due to dangerous ventilator settings. A recruitment-potential-oriented evaluation-based protocol can effectively stabilize hypoxemic conditions quickly and screen out truly serious patients.

High Airway Occlusion Pressure Is Associated with Dyspnea and Increased Mortality in Critically Ill Mechanically Ventilated Patients

Rationale: Airway occlusion pressure at 100 ms (P0.1) reflects central respiratory drive. Objectives: We aimed to assess factors associated with P0.1 and whether an abnormally low or high P0.1 value is associated with higher mortality and longer duration of mechanical ventilation (MV). Methods: We performed a secondary analysis of a prospective cohort study conducted in 10 ICUs in France to evaluate dyspnea in communicative MV patients. In patients intubated for more than 24 hours, P0.1 was measured with dyspnea as soon as patients could communicate and the next day. Measurements and Main Results: Among 260 patients assessed after a median time of ventilation of 4 days, P0.1 was 1.9 (1-3.5) cm H2O at enrollment, 24% had P0.1 values >3.5 cm H2O, 37% had P0.1 values between 1.5 and 3.5 cm H2O, and 39% had P0.1 values <1.5 cm H2O. In multivariable linear regression, independent factors associated with P0.1 were the presence of dyspnea (P = 0.037), respiratory rate (P < 0.001), and PaO2 (P = 0.008). Ninety-day mortality was 33% in patients with P0.1 > 3.5 cm H2O versus 19% in those with P0.1 between 1.5 and 3.5 cm H2O and 17% in those with P0.1 < 1.5 cm H2O (P = 0.046). After adjustment for the main risk factors, P0.1 was associated with 90-day mortality (hazard ratio per 1 cm H2O, 1.19 [95% confidence interval, 1.04-1.37]; P = 0.011). P0.1 was also independently associated with a longer duration of MV (hazard ratio per 1 cm H2O, 1.10 [95% confidence interval, 1.02-1.19]; P = 0.016). Conclusions: In patients receiving invasive MV, abnormally high P0.1 values may suggest dyspnea and are associated with higher mortality and prolonged duration of MV.

Application of a cloud platform that identifies patient-ventilator asynchrony and enables continuous monitoring of mechanical ventilation in intensive care unit

Background

Patient-ventilator asynchrony (PVA) frequently occurs in mechanically ventilated patients within the ICU and has the potential for harm. Depending solely on the health care team cannot accurately and promptly identify PVA. To address this issue, our team has developed a cloud-based platform for monitoring mechanical ventilation (MV), comprising the PVA-RemoteMonitor system and the 24-h MV analysis report. We conducted a survey to evaluate physicians' satisfaction and acceptance of the platform in 14 ICUs.

Methods

Data from medical records, clinical information systems, and ventilators were uploaded to the cloud platform and underwent data processing. The data were analyzed to monitor PVA and displayed in the front-end. The 24-h analysis report for MV was generated for clinical reference. Critical care physicians in 14 hospitals' ICUs that involved in the platform participated in a questionnaire survey, among whom 10 physicians were interviewed to investigate physicians' acceptance and opinions of this system.

Results

The PVA-RemoteMonitor system exhibited a high level of specificity in detecting flow insufficiency, premature cycle, delayed cycle, reverse trigger, auto trigger, and overshoot, with sensitivities of 90.31 %, 98.76 %, 99.75 %, 99.97 %, 100 %, and 99.69 %, respectively. The 24-h analysis report supplied essential data about PVA and respiratory mechanics. 86.2 % (75/87) of physicians supported the application of this platform.

Conclusions

The PVA-RemoteMonitor system accurately identified PVA, and the MV analysis report provided guidance in controlling PVA. Our platform can effectively assist ICU physicians in the management of ventilated patients.

Detection and quantitative analysis of patient-ventilator interactions in ventilated infants by deep learning networks

BACKGROUND

The study of patient-ventilator interactions (PVI) in mechanically ventilated neonates is limited by the lack of unified PVI definitions and tools to perform large scale analyses.

METHODS

An observational study was conducted in 23 babies randomly selected from 170 neonates who were ventilated with SIPPV-VG, SIMV-VG or PSV-VG mode for at least 12 h. 500 breaths were randomly selected and manually annotated from each recording to train convolutional neural network (CNN) models for PVI classification.

RESULTS

The average asynchrony index (AI) over all recordings was 52.5%. The most frequently occurring PVIs included expiratory work (median: 28.4%, interquartile range: 23.2-40.2%), late cycling (7.6%, 2.8-10.2%), failed triggering (4.6%, 1.2-6.2%) and late triggering (4.4%, 2.8-7.4%). Approximately 25% of breaths with a PVI had two or more PVIs occurring simultaneously. Binary CNN classifiers were developed for PVIs affecting ≥1% of all breaths (n = 7) and they achieved F1 scores of >0.9 on the test set except for early triggering where it was 0.809.

CONCLUSIONS

PVIs occur frequently in neonates undergoing conventional mechanical ventilation with a significant proportion of breaths containing multiple PVIs. We have developed computational models for seven different PVIs to facilitate automated detection and further evaluation of their clinical significance in neonates.

IMPACT

The study of patient-ventilator interactions (PVI) in mechanically ventilated neonates is limited by the lack of unified PVI definitions and tools to perform large scale analyses. By adapting a recent taxonomy of PVI definitions in adults, we have manually annotated neonatal ventilator waveforms to determine prevalence and co-occurrence of neonatal PVIs. We have also developed binary deep learning classifiers for common PVIs to facilitate their automatic detection and quantification.

Clinical evidence and technical aspects of innovative technology and monitoring of chronic NIV in COPD: a narrative review

INTRODUCTION

Chronic nocturnal noninvasive ventilation (NIV) improves outcomes in COPD patients with chronic hypercapnic respiratory failure. The aim of chronic NIV in COPD is to control chronic hypercapnic respiratory insufficiency and reduce symptoms of nocturnal hypoventilation, thereby improving quality of life. Chronic NIV care is more and more offered exclusively at home, enabling promising outcomes in terms of patient and caregiver satisfaction, hospital care consumption and cost reduction. Yet, to achieve and maintain optimal ventilation, during adaptation and follow-up, effective feasible (home) monitoring poses a significant challenge.

AREAS COVERED

Comprehensive monitoring of COPD patients receiving chronic NIV requires integrating data from ventilators and assessment of the patient's status including gas exchange, sleep quality, and patient-reported outcomes. The present article describes the physiological background of monitoring during NIV and aims to provide an overview of existing methods for monitoring, assessing their reliability and clinical relevance.

EXPERT OPINION

Patients on chronic NIV are 'ideal' candidates for home monitoring; the advantages of transforming hospital to home care are huge for patients and caregivers and for healthcare systems facing increasing patient numbers. Despite the multitude of available monitoring methods, identifying and characterizing the most relevant parameters associated with optimal patient well-being remains unclear.

An adversarial learning approach to generate pressure support ventilation waveforms for asynchrony detection

BACKGROUND AND OBJECTIVE

Mechanical ventilation is a life-saving treatment for critically-ill patients. During treatment, patient-ventilator asynchrony (PVA) can occur, which can lead to pulmonary damage, complications, and higher mortality. While traditional detection methods for PVAs rely on visual inspection by clinicians, in recent years, machine learning models are being developed to detect PVAs automatically. However, training these models requires large labeled datasets, which are difficult to obtain, as labeling is a labour-intensive and time-consuming task, requiring clinical expertise. Simulating the lung-ventilator interactions has been proposed to obtain large labeled datasets to train machine learning classifiers. However, the obtained data lacks the influence of different hardware, of servo-controlled algorithms, and different sources of noise. Here, we propose VentGAN, an adversarial learning approach to improve simulated data by learning the ventilator fingerprints from unlabeled clinical data.

METHODS

In VentGAN, the loss functions are designed to add characteristics of clinical waveforms to the generated results, while preserving the labels of the simulated waveforms. To validate VentGAN, we compare the performance for detection and classification of PVAs when training a previously developed machine learning algorithm with the original simulated data and with the data generated by VentGAN. Testing is performed on independent clinical data labeled by experts. The McNemar test is applied to evaluate statistical differences in the obtained classification accuracy.

RESULTS

VentGAN significantly improves the classification accuracy for late cycling, early cycling and normal breaths (p< 0.01); no significant difference in accuracy was observed for delayed inspirations (p = 0.2), while the accuracy decreased for ineffective efforts (p< 0.01).

CONCLUSIONS

Generation of realistic synthetic data with labels by the proposed framework is feasible and represents a promising avenue for improving training of machine learning models.

Reverse triggering ? a novel or previously missed phenomenon?

BACKGROUND

Reverse triggering (RT) was described in 2013 as a form of patient-ventilator asynchrony, where patient's respiratory effort follows mechanical insufflation. Diagnosis requires esophageal pressure (Pes) or diaphragmatic electrical activity (EAdi), but RT can also be diagnosed using standard ventilator waveforms.

HYPOTHESIS

We wondered (1) how frequently RT would be present but undetected in the figures from literature, especially before 2013; (2) whether it would be more prevalent in the era of small tidal volumes after 2000.

METHODS

We searched PubMed, EMBASE, and the Cochrane Central Register of Controlled Trials, from 1950 to 2017, with key words related to asynchrony to identify papers with figures including ventilator waveforms expected to display RT if present. Experts labelled waveforms. 'Definite' RT was identified when Pes or EAdi were in the tracing, and 'possible' RT when only flow and pressure waveforms were present. Expert assessment was compared to the author's descriptions of waveforms.

RESULTS

We found 65 appropriate papers published from 1977 to now, containing 181 ventilator waveforms. 21 cases of 'possible' RT and 25 cases of 'definite' RT were identified by the experts. 18.8% of waveforms prior to 2013 had evidence of RT. Most cases were published after 2000 (1 before vs. 45 after, p = 0.03). 54% of RT cases were attributed to different phenomena. A few cases of identified RT were already described prior to 2013 using different terminology (earliest in 1997). While RT cases attributed to different phenomena decreased after 2013, 60% of 'possible' RT remained missed.

CONCLUSION

RT has been present in the literature as early as 1997, but most cases were found after the introduction of low tidal volume ventilation in 2000. Following 2013, the number of undetected cases decreased, but RT are still commonly missed. Reverse Triggering, A Missed Phenomenon in the Literature. Critical Care Canada Forum 2019 Abstracts. Can J Anesth/J Can Anesth 67 (Suppl 1), 1-162 (2020). https://doi-org.myaccess.library.utoronto.ca/ https://doi.org/10.1007/s12630-019-01552-z .

Expiratory flow limitation during mechanical ventilation: real-time detection and physiological subtypes

BACKGROUND

Tidal expiratory flow limitation (EFLT) complicates the delivery of mechanical ventilation but is only diagnosed by performing specific manoeuvres. Instantaneous analysis of expiratory resistance (Rex) can be an alternative way to detect EFLT without changing ventilatory settings. This study aimed to determine the agreement of EFLT detection by Rex analysis and the PEEP reduction manoeuvre using contingency table and agreement coefficient. The patterns of Rex were explored.

METHODS

Medical patients ≥ 15-year-old receiving mechanical ventilation underwent a PEEP reduction manoeuvre from 5 cmH2O to zero for EFLT detection. Waveforms were recorded and analyzed off-line. The instantaneous Rex was calculated and was plotted against the volume axis, overlapped by the flow-volume loop for inspection. Lung mechanics, characteristics of the patients, and clinical outcomes were collected. The result of the Rex method was validated using a separate independent dataset.

RESULTS

339 patients initially enrolled and underwent a PEEP reduction. The prevalence of EFLT was 16.5%. EFLT patients had higher adjusted hospital mortality than non-EFLT cases. The Rex method showed 20% prevalence of EFLT and the result was 90.3% in agreement with PEEP reduction manoeuvre. In the validation dataset, the Rex method had resulted in 91.4% agreement. Three patterns of Rex were identified: no EFLT, early EFLT, associated with airway disease, and late EFLT, associated with non-airway diseases, including obesity. In early EFLT, external PEEP was less likely to eliminate EFLT.

CONCLUSIONS

The Rex method shows an excellent agreement with the PEEP reduction manoeuvre and allows real-time detection of EFLT. Two subtypes of EFLT are identified by Rex analysis.

TRIAL REGISTRATION

Clinical trial registered with www.thaiclinicaltrials.org (TCTR20190318003). The registration date was on 18 March 2019, and the first subject enrollment was performed on 26 March 2019.

The effects of real-time waveform analysis software on patient ventilator synchronization during pressure support ventilation: a randomized crossover physiological study

BACKGROUND

Patient-ventilator asynchrony commonly occurs during pressure support ventilation (PSV). IntelliSync + software (Hamilton Medical AG, Bonaduz, Switzerland) is a new ventilation technology that continuously analyzes ventilator waveforms to detect the beginning and end of patient inspiration in real time. This study aimed to evaluate the physiological effect of IntelliSync + software on inspiratory trigger delay time, delta airway (Paw) and esophageal (Pes) pressure drop during the trigger phase, airway occlusion pressure at 0.1 s (P0.1), and hemodynamic variables.

METHODS

A randomized crossover physiologic study was conducted in 14 mechanically ventilated patients under PSV. Patients were randomly assigned to receive conventional flow trigger and cycling, inspiratory trigger synchronization (I-sync), cycle synchronization (C-sync), and inspiratory trigger and cycle synchronization (I/C-sync) for 15 min at each step. Other ventilator settings were kept constant. Paw, Pes, airflow, P0.1, respiratory rate, SpO2, and hemodynamic variables were recorded. The primary outcome was inspiratory trigger and cycle delay time between each intervention. Secondary outcomes were delta Paw and Pes drop during the trigger phase, P0.1, SpO2, and hemodynamic variables.

RESULTS

The time to initiate the trigger was significantly shorter with I-sync compared to baseline (208.9±91.7 vs. 301.4±131.7 msec; P = 0.002) and I/C-sync compared to baseline (222.8±94.0 vs. 301.4±131.7 msec; P = 0.005). The I/C-sync group had significantly lower delta Paw and Pes drop during the trigger phase compared to C-sync group (-0.7±0.4 vs. -1.2±0.8 cmH2O; P = 0.028 and - 1.8±2.2 vs. -2.8±3.2 cmH2O; P = 0.011, respectively). No statistically significant differences were found in cycle delay time, P0.1 and other physiological variables between the groups.

CONCLUSIONS

IntelliSync + software reduced inspiratory trigger delay time compared to the conventional flow trigger system during PSV mode. However, no significant improvements in cycle delay time and other physiological variables were observed with IntelliSync + software.

TRIAL REGISTRATION

This study was registered in the Thai Clinical Trial Registry (TCTR20200528003; date of registration 28/05/2020).

Failing with Cyanosis-Heart Failure in End-Stage Unrepaired or Partially Palliated Congenital Heart Disease

Heart failure in cyanotic congenital heart disease (CHD) is diagnosed clinically rather than relying solely on ventricular function assessments. Patients with cyanosis often present with clinical features indicative of heart failure. Although myocardial injury and dysfunction likely contribute to cyanotic CHD, the primary concern is the reduced delivery of oxygen to tissues. Symptoms such as fatigue, lassitude, dyspnea, headaches, myalgias, and a cold sensation underscore inadequate tissue oxygen delivery, forming the basis for defining heart failure in cyanotic CHD. Thus, it is pertinent to delve into the components of oxygen delivery in this context.

Invasive Mechanical Ventilation

Invasive mechanical ventilation allows clinicians to support gas exchange and work of breathing in patients with respiratory failure. However, there is also potential for iatrogenesis. By understanding the benefits and limitations of different modes of ventilation and goals for gas exchange, clinicians can choose a strategy that provides appropriate support while minimizing harm. The ventilator can also provide crucial diagnostic information in the form of respiratory mechanics. These, and the mechanical ventilation strategy, should be regularly reassessed.

Electrical Impedance Tomography to Monitor Hypoxemic Respiratory Failure

Hypoxemic respiratory failure is one of the leading causes of mortality in intensive care. Frequent assessment of individual physiological characteristics and delivery of personalized mechanical ventilation (MV) settings is a constant challenge for clinicians caring for these patients. Electrical impedance tomography (EIT) is a radiation-free bedside monitoring device that is able to assess regional lung ventilation and changes in aeration. With real-time tomographic functional images of the lungs obtained through a thoracic belt, clinicians can visualize and estimate the distribution of ventilation at different ventilation settings or following procedures such as prone positioning. Several studies have evaluated the performance of EIT to monitor the effects of different MV settings in patients with acute respiratory distress syndrome, allowing more personalized MV. For instance, EIT could help clinicians find the positive end-expiratory pressure that represents a compromise between recruitment and overdistension and assess the effect of prone positioning on ventilation distribution. The clinical impact of the personalization of MV remains to be explored. Despite inherent limitations such as limited spatial resolution, EIT also offers a unique noninvasive bedside assessment of regional ventilation changes in the ICU. This technology offers the possibility of a continuous, operator-free diagnosis and real-time detection of common problems during MV. This review provides an overview of the functioning of EIT, its main indices, and its performance in monitoring patients with acute respiratory failure. Future perspectives for use in intensive care are also addressed.

Flow starvation during square-flow assisted ventilation detected by supervised deep learning techniques

BACKGROUND

Flow starvation is a type of patient-ventilator asynchrony that occurs when gas delivery does not fully meet the patients' ventilatory demand due to an insufficient airflow and/or a high inspiratory effort, and it is usually identified by visual inspection of airway pressure waveform. Clinical diagnosis is cumbersome and prone to underdiagnosis, being an opportunity for artificial intelligence. Our objective is to develop a supervised artificial intelligence algorithm for identifying airway pressure deformation during square-flow assisted ventilation and patient-triggered breaths.

METHODS

Multicenter, observational study. Adult critically ill patients under mechanical ventilation > 24 h on square-flow assisted ventilation were included. As the reference, 5 intensive care experts classified airway pressure deformation severity. Convolutional neural network and recurrent neural network models were trained and evaluated using accuracy, precision, recall and F1 score. In a subgroup of patients with esophageal pressure measurement (ΔPes), we analyzed the association between the intensity of the inspiratory effort and the airway pressure deformation.

RESULTS

6428 breaths from 28 patients were analyzed, 42% were classified as having normal-mild, 23% moderate, and 34% severe airway pressure deformation. The accuracy of recurrent neural network algorithm and convolutional neural network were 87.9% [87.6-88.3], and 86.8% [86.6-87.4], respectively. Double triggering appeared in 8.8% of breaths, always in the presence of severe airway pressure deformation. The subgroup analysis demonstrated that 74.4% of breaths classified as severe airway pressure deformation had a ΔPes > 10 cmH2O and 37.2% a ΔPes > 15 cmH2O.

CONCLUSIONS

Recurrent neural network model appears excellent to identify airway pressure deformation due to flow starvation. It could be used as a real-time, 24-h bedside monitoring tool to minimize unrecognized periods of inappropriate patient-ventilator interaction.

Divergent expiratory braking activity of costal and crural diaphragm

BACKGROUND

There is increasing clinical interest in understanding the contribution of the diaphragm in early expiration, especially during mechanical ventilation. However, current experimental evidence is limited, so essential activity of the diaphragm during expiration and diaphragm segmental differences in expiratory activity, are unknown.

OBJECTIVES

To determine if: 1) the diaphragm is normally active into expiration during spontaneous breathing and hypercapnic ventilation, 2) expiratory diaphragmatic activity is distributed equally among the segments of the diaphragm, costal and crural.

METHODS

In 30 spontaneously breathing male and female canines, awake without confounding anesthetic, we measured directly both inspiratory and expiratory electrical activity (EMG), and corresponding mechanical shortening, of costal and crural diaphragm, during room air and hypercapnia.

RESULTS

During eupnea, costal and crural diaphragm are active into expiration, showing significant and distinct expiratory activity, with crural expiratory activity greater than costal, for both magnitude and duration. This diaphragm segmental difference diverged further during progressive hypercapnic ventilation: crural expiratory activity progressively increased, while costal expiratory activity disappeared.

CONCLUSION

The diaphragm is not passive during expiration. During spontaneous breathing, expiratory activity -"braking"- of the diaphragm is expressed routinely, but is not equally distributed. Crural muscle "braking" is greater than costal muscle in magnitude and duration. With increasing ventilation during hypercapnia, expiratory activity -"braking"- diverges notably. Crural expiratory activity greatly increases, while costal expiratory "braking" decreases in magnitude and duration, and disappears. Thus, diaphragm expiratory "braking" action represents an inherent, physiological function of the diaphragm, distinct for each segment, expressing differing neural activation.

Automated characterization of patient-ventilator interaction using surface electromyography

BACKGROUND

Characterizing patient-ventilator interaction in critically ill patients is time-consuming and requires trained staff to evaluate the behavior of the ventilated patient.

METHODS

In this study, we recorded surface electromyography ([Formula: see text]) signals from the diaphragm and intercostal muscles and esophageal pressure ([Formula: see text]) in mechanically ventilated patients with ARDS. The sEMG recordings were preprocessed, and two different algorithms (triangle algorithm and adaptive thresholding algorithm) were used to automatically detect inspiratory patient effort. Based on the detected inspirations, major asynchronies (ineffective, auto-, and double triggers and double efforts), delayed and synchronous triggers were computationally classified. Reverse triggers were not considered in this study. Subsequently, asynchrony indices were calculated. For the validation of detected efforts, two experts manually annotated inspiratory patient activity in [Formula: see text], blinded toward each other, the [Formula: see text] signals, and the algorithmic results. We also classified patient-ventilator interaction and calculated asynchrony indices with manually detected inspirations in [Formula: see text] as a reference for automated asynchrony classification and asynchrony index calculation.

RESULTS

Spontaneous breathing activity was recognized in 22 out of the 36 patients included in the study. Evaluation of the accuracy of the algorithms using 3057 inspiratory efforts in [Formula: see text] demonstrated reliable detection performance for both methods. Across all datasets, we found a high sensitivity (triangle algorithm/adaptive thresholding algorithm: 0.93/0.97) and a high positive predictive value (0.94/0.89) against expert annotations in [Formula: see text]. The average delay of automatically detected inspiratory onset to the [Formula: see text] reference was [Formula: see text]79 ms/29 ms for the two algorithms. Our findings also indicate that automatic asynchrony index prediction is reliable. For both algorithms, we found the same deviation of [Formula: see text] to the [Formula: see text]-based reference.

CONCLUSIONS

Our study demonstrates the feasibility of automating the quantification of patient-ventilator asynchrony in critically ill patients using noninvasive sEMG. This may facilitate more frequent diagnosis of asynchrony and support improving patient-ventilator interaction.

Exploring the Impact of Artificial Intelligence on Global Health and Enhancing Healthcare in Developing Nations

BACKGROUND

Artificial intelligence (AI), which combines computer science with extensive datasets, seeks to mimic human-like intelligence. Subsets of AI are being applied in almost all fields of medicine and surgery.

AIM

This review focuses on the applications of AI in healthcare settings in developing countries, designed to underscore its significance by comprehensively outlining the advancements made thus far, the shortcomings encountered in AI applications, the present status of AI integration, persistent challenges, and innovative strategies to surmount them.

METHODOLOGY

Articles from PubMed, Google Scholar, and Cochrane were searched from 2000 to 2023 with keywords including AI and healthcare, focusing on multiple medical specialties.

RESULTS

The increasing role of AI in diagnosis, prognosis prediction, and patient management, as well as hospital management and community healthcare, has made the overall healthcare system more efficient, especially in the high patient load setups and resource-limited areas of developing countries where patient care is often compromised. However, challenges, including low adoption rates and the absence of standardized guidelines, high installation and maintenance costs of equipment, poor transportation and connectivvity issues hinder AI's full use in healthcare.

CONCLUSION

Despite these challenges, AI holds a promising future in healthcare. Adequate knowledge and expertise of healthcare professionals for the use of AI technology in healthcare is imperative in developing nations.

Understanding the mechanisms of ventilator-induced lung injury using animal models

Mechanical ventilation is a life-saving therapy in several clinical situations, promoting gas exchange and providing rest to the respiratory muscles. However, mechanical ventilation may cause hemodynamic instability and pulmonary structural damage, which is known as ventilator-induced lung injury (VILI). The four main injury mechanisms associated with VILI are as follows: barotrauma/volutrauma caused by overstretching the lung tissues; atelectrauma, caused by repeated opening and closing of the alveoli resulting in shear stress; and biotrauma, the resulting biological response to tissue damage, which leads to lung and multi-organ failure. This narrative review elucidates the mechanisms underlying the pathogenesis, progression, and resolution of VILI and discusses the strategies that can mitigate VILI. Different static variables (peak, plateau, and driving pressures, positive end-expiratory pressure, and tidal volume) and dynamic variables (respiratory rate, airflow amplitude, and inspiratory time fraction) can contribute to VILI. Moreover, the potential for lung injury depends on tissue vulnerability, mechanical power (energy applied per unit of time), and the duration of that exposure. According to the current evidence based on models of acute respiratory distress syndrome and VILI, the following strategies are proposed to provide lung protection: keep the lungs partially collapsed (SaO2 > 88%), avoid opening and closing of collapsed alveoli, and gently ventilate aerated regions while keeping collapsed and consolidated areas at rest. Additional mechanisms, such as subject-ventilator asynchrony, cumulative power, and intensity, as well as the damaging threshold (stress-strain level at which tidal damage is initiated), are under experimental investigation and may enhance the understanding of VILI.

Asynchrony Injures Lung and Diaphragm in Acute Respiratory Distress Syndrome

OBJECTIVES

Patient-ventilator asynchrony is often observed during mechanical ventilation and is associated with higher mortality. We hypothesized that patient-ventilator asynchrony causes lung and diaphragm injury and dysfunction.

DESIGN

Prospective randomized animal study.

SETTING

University research laboratory.

SUBJECTS

Eighteen New Zealand White rabbits.

INTERVENTIONS

Acute respiratory distress syndrome (ARDS) model was established by depleting surfactants. Each group (assist control, breath stacking, and reverse triggering) was simulated by phrenic nerve stimulation. The effects of each group on lung function, lung injury (wet-to-dry lung weight ratio, total protein, and interleukin-6 in bronchoalveolar lavage), diaphragm function (diaphragm force generation curve), and diaphragm injury (cross-sectional area of diaphragm muscle fibers, histology) were measured. Diaphragm RNA sequencing was performed using breath stacking and assist control ( n = 2 each).

MEASUREMENTS AND MAIN RESULTS

Inspiratory effort generated by phrenic nerve stimulation was small and similar among groups (esophageal pressure swing ≈ -2.5 cm H 2 O). Breath stacking resulted in the largest tidal volume (>10 mL/kg) and highest inspiratory transpulmonary pressure, leading to worse oxygenation, worse lung compliance, and lung injury. Reverse triggering did not cause lung injury. No asynchrony events were observed in assist control, whereas eccentric contractions occurred in breath stacking and reverse triggering, but more frequently in breath stacking. Breath stacking and reverse triggering significantly reduced diaphragm force generation. Diaphragmatic histology revealed that the area fraction of abnormal muscle was ×2.5 higher in breath stacking (vs assist control) and ×2.1 higher in reverse triggering (vs assist control). Diaphragm RNA sequencing analysis revealed that genes associated with muscle differentiation and contraction were suppressed, whereas cytokine- and chemokine-mediated proinflammatory responses were activated in breath stacking versus assist control.

CONCLUSIONS

Breath stacking caused lung and diaphragm injury, whereas reverse triggering caused diaphragm injury. Thus, careful monitoring and management of patient-ventilator asynchrony may be important to minimize lung and diaphragm injury from spontaneous breathing in ARDS.

Automated evaluation of typical patient-ventilator asynchronies based on lung hysteretic responses

BACKGROUND

Patient-ventilator asynchrony is common during mechanical ventilation (MV) in intensive care unit (ICU), leading to worse MV care outcome. Identification of asynchrony is critical for optimizing MV settings to reduce or eliminate asynchrony, whilst current clinical visual inspection of all typical types of asynchronous breaths is difficult and inefficient. Patient asynchronies create a unique pattern of distortions in hysteresis respiratory behaviours presented in pressure-volume (PV) loop.

METHODS

Identification method based on hysteretic lung mechanics and hysteresis loop analysis is proposed to delineate the resulted changes of lung mechanics in PV loop during asynchronous breathing, offering detection of both its incidence and 7 major types. Performance is tested against clinical patient data with comparison to visual inspection conducted by clinical doctors.

RESULTS

The identification sensitivity and specificity of 11 patients with 500 breaths for each patient are above 89.5% and 96.8% for all 7 types, respectively. The average sensitivity and specificity across all cases are 94.6% and 99.3%, indicating a very good accuracy. The comparison of statistical analysis between identification and human inspection yields the essential same clinical judgement on patient asynchrony status for each patient, potentially leading to the same clinical decision for setting adjustment.

CONCLUSIONS

The overall results validate the accuracy and robustness of the identification method for a bedside monitoring, as well as its ability to provide a quantified metric for clinical decision of ventilator setting. Hence, the method shows its potential to assist a more consistent and objective assessment of asynchrony without undermining the efficacy of the current clinical practice.

A Novel Application of Spectrograms with Machine Learning Can Detect Patient Ventilator Dyssynchrony

Patients in intensive care units are frequently supported by mechanical ventilation. There is increasing awareness of patient-ventilator dyssynchrony (PVD), a mismatch between patient respiratory effort and assistance provided by the ventilator, as a risk factor for infection, narcotic exposure, lung injury, and adverse neurocognitive effects. One of the most injurious consequences of PVD are double cycled (DC) breaths when two breaths are delivered by the ventilator instead of one. Prior efforts to identify PVD have limited efficacy. An automated method to identify PVD, independent of clinician expertise, acumen, or time, would potentially permit early, targeted treatment to avoid further harm. We performed secondary analyses of data from a clinical trial of children with acute respiratory distress syndrome. Waveforms of ventilator flow, airway pressure and esophageal manometry were annotated to identify DC breaths and underlying PVD subtypes. Spectrograms were generated from those waveforms to train Convolutional Neural Network (CNN) models in detecting DC and underlying PVD subtypes: Reverse Trigger (RT) and Inadequate Support (IS). The DC breath detection model yielded AUROC of 0.980, while the multi-target detection model for underlying dyssynchrony yielded AUROC of 0.980 (RT) and 0.976 (IS). When operating at 75% sensitivity, DC breath detection had a number needed to alert (NNA) 1.3 (99% specificity), while underlying PVD had a NNA 1.6 (98.5% specificity) for RT and NNA 4.0 (98.2% specificity) for IS. CNNs using spectrograms of ventilator waveforms can identify DC breaths and detect the underlying PVD for targeted clinical interventions.