1
|
Ruiz-Botella M, Manrique S, Gomez J, Bodí M. Advancing ICU patient care with a Real-Time predictive model for mechanical Power to mitigate VILI. Int J Med Inform 2024; 189:105511. [PMID: 38851133 DOI: 10.1016/j.ijmedinf.2024.105511] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/06/2024] [Revised: 05/24/2024] [Accepted: 05/29/2024] [Indexed: 06/10/2024]
Abstract
BACKGROUND Invasive Mechanical Ventilation (IMV) in Intensive Care Units (ICU) significantly increases the risk of Ventilator-Induced Lung Injury (VILI), necessitating careful management of mechanical power (MP). This study aims to develop a real-time predictive model of MP utilizing Artificial Intelligence to mitigate VILI. METHODOLOGY A retrospective observational study was conducted, extracting patient data from Clinical Information Systems from 2018 to 2022. Patients over 18 years old with more than 6 h of IMV were selected. Continuous data on IMV variables, laboratory data, monitoring, procedures, demographic data, type of admission, reason for admission, and APACHE II at admission were extracted. The variables with the highest correlation to MP were used for prediction and IMV data was grouped in 15-minute intervals using the mean. A mixed neural network model was developed to forecast MP 15 min in advance, using IMV data from 6 h before the prediction and current patient status. The model's ability to predict future MP was analyzed and compared to a baseline model predicting the future value of MP as equal to the current value. RESULTS The cohort consisted of 1967 patients after applying inclusion criteria, with a median age of 63 years and 66.9 % male. The deep learning model achieved a mean squared error of 2.79 in the test set, indicating a 20 % improvement over the baseline model. It demonstrated high accuracy (94 %) in predicting whether MP would exceed a critical threshold of 18 J/min, which correlates with increased mortality. The integration of this model into a web platform allows clinicians real-time access to MP predictions, facilitating timely adjustments to ventilation settings. CONCLUSIONS The study successfully developed and integrated in clinical practice a predictive model for MP. This model will assist clinicians allowing for the adjustment of ventilatory parameters before lung damage occurs.
Collapse
Affiliation(s)
- M Ruiz-Botella
- Departament of Chemical Engineering, Universitat Rovira I Virgili, Tarragona, Spain; Instituto de Investigación Sanitaria Pere i Virgili, Universidad Rovira i Virgili, Tarragona, Spain.
| | - S Manrique
- Instituto de Investigación Sanitaria Pere i Virgili, Universidad Rovira i Virgili, Tarragona, Spain; Critical Care department, Hospital Universitario Joan XXIII, Tarragona, Spain
| | - J Gomez
- Instituto de Investigación Sanitaria Pere i Virgili, Universidad Rovira i Virgili, Tarragona, Spain; Critical Care department, Hospital Universitario Joan XXIII, Tarragona, Spain
| | - M Bodí
- Instituto de Investigación Sanitaria Pere i Virgili, Universidad Rovira i Virgili, Tarragona, Spain; Critical Care department, Hospital Universitario Joan XXIII, Tarragona, Spain; Centro de Investigación Biomédica en Red de Enfermedades Respiratorias (CIBERES). Instituto de Salud Carlos III, Spain
| |
Collapse
|
2
|
Abstract
Our current understanding of protective measures for avoiding ventilator-induced lung injury (VILI) has evolved from targeting low tidal volumes to lowering plateau and driving pressure. Even when pressures across the lung are reliably estimated, however, pressures alone cannot accurately gauge the injury risk; apart from flow rate, inspired oxygen fraction, and currently unmeasurable features of the mechanical microenvironment such as geometry, structural fragility, and vascular perfusion, the frequency with which high-risk tidal cycles are applied must help determine the intensity of potentially damaging energy application. Recognition of a strain threshold for damage by transpulmonary pressure, coupled with considerations of total energy load and strain intensity, has helped shape the unifying concept of VILI generation dependent upon the power transferred from the ventilator to the injured lungs. Currently, under-recognized contributors to the injury process must be addressed to minimize the risk imposed by ventilatory support.
Collapse
Affiliation(s)
- John J Marini
- University of Minnesota, Regions Hospital MS 11203B, 640 Jackson St, St. Paul, MN, 55101, USA.
| |
Collapse
|
3
|
Dissipation of energy during the respiratory cycle: conditional importance of ergotrauma to structural lung damage. Curr Opin Crit Care 2018; 24:16-22. [PMID: 29176330 DOI: 10.1097/mcc.0000000000000470] [Citation(s) in RCA: 17] [Impact Index Per Article: 2.8] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/01/2023]
Abstract
PURPOSE OF REVIEW To describe and put into context recent conceptual advances regarding the relationship of energy load and power to ventilator-induced lung injury (VILI). RECENT FINDINGS Investigative emphasis regarding VILI has almost exclusively centered on the static characteristics of the individual tidal cycle - tidal volume, plateau pressure, positive end-expiratory pressure, and driving pressure. Although those static characteristics of the tidal cycle are undeniably important, the 'dynamic' characteristics of ventilation must not be ignored. To inflict the nonrupturing damage we identify as VILI, work must be performed and energy expended by high stress cycles applied at rates that exceed the capacity of endogenous repair. Machine power, the pace at which the work performing energy load is applied by the ventilator, has received increasing scrutiny as a candidate for the proximate and integrative cause of VILI. SUMMARY Although the unmodified values of machine-delivered energy or power (which are based on airway pressures and tidal volumes) cannot serve unconditionally as a rigid and quantitative guide to ventilator adjustment for lung protection, bedside consideration of the dynamics of ventilation and potential for ergotrauma represents a clear conceptual advance that complements the static parameters of the individual tidal cycle that with few exceptions have held our scientific attention.
Collapse
|
4
|
Engoren M, Rochlen LR, Diehl MV, Sherman SS, Jewell E, Golinski M, Begeman P, Cavanaugh JM. Mechanical strain to maxillary incisors during direct laryngoscopy. BMC Anesthesiol 2017; 17:151. [PMID: 29115945 PMCID: PMC5688811 DOI: 10.1186/s12871-017-0442-z] [Citation(s) in RCA: 5] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/09/2017] [Accepted: 10/31/2017] [Indexed: 01/19/2023] Open
Abstract
Background While most Direct laryngoscopy leads to dental injury in 25–39% of cases. Dental injury occurs when the forces and impacts applied to the teeth exceed the ability of the structures to dissipate energy and stress. The purpose of this study was to measure strain, (which is the change produced in the length of the tooth by a force applied to the tooth) strain rate, and strain-time integral to the maxillary incisors and determine if they varied by experience, type of blade, or use of an alcohol protective pad (APP). Methods A mannequin head designed to teach and test intubation was instrumented with eight single axis strain gauges placed on the four maxillary incisors: four on the facial or front surface of the incisors and four on the lingual or back, near the insertion of the incisor in the gums to measure bending strain as well as compression. Anesthesiology faculty, residents, and certified registered nurse anesthetists intubated with Macintosh and Miller blades with and without APP. Using strain-time curves, the maximum strain, strain rate, and strain time integral were calculated. Results Across the 92 subjects, strain varied 8–12 fold between the 25th and 75th percentiles for all four techniques, but little by experience, while strain rate and strain integral varied 6–13 fold and 15–26 fold, respectively, for the same percentiles. Intubators who had high strain values with one blade tended to have high strains with the other blade with and without the APP (all pairwise correlation rho = 0.42–0.63). Conclusions Strain varies widely by intubator and that the use of the APP reduces strain rate which may decrease the risk of or the severity of dental injury.
Collapse
Affiliation(s)
- Milo Engoren
- Department of Anesthesiology, University of Michigan, Ann Arbor, MI, USA. .,Department of Anesthesiology, Mercy St. Vincent Medical Center, Toledo, OH, USA.
| | - Lauryn R Rochlen
- Department of Anesthesiology, University of Michigan, Ann Arbor, MI, USA
| | - Matthew V Diehl
- Department of Biomedical Engineering, Wayne State University, Detroit, MI, USA
| | - Sarah S Sherman
- Department of Biomedical Engineering, Wayne State University, Detroit, MI, USA
| | - Elizabeth Jewell
- Department of Anesthesiology, University of Michigan, Ann Arbor, MI, USA
| | - Mary Golinski
- Beaumont Graduate Program of Nurse Anesthesia, Oakland University, Rochester, MI, USA
| | - Paul Begeman
- Department of Biomedical Engineering, Wayne State University, Detroit, MI, USA
| | - John M Cavanaugh
- Department of Biomedical Engineering, Wayne State University, Detroit, MI, USA
| |
Collapse
|
5
|
Tonetti T, Vasques F, Rapetti F, Maiolo G, Collino F, Romitti F, Camporota L, Cressoni M, Cadringher P, Quintel M, Gattinoni L. Driving pressure and mechanical power: new targets for VILI prevention. ANNALS OF TRANSLATIONAL MEDICINE 2017; 5:286. [PMID: 28828361 DOI: 10.21037/atm.2017.07.08] [Citation(s) in RCA: 133] [Impact Index Per Article: 19.0] [Reference Citation Analysis] [Abstract] [Key Words] [Subscribe] [Scholar Register] [Indexed: 12/31/2022]
Abstract
Several factors have been recognized as possible triggers of ventilator-induced lung injury (VILI). The first is pressure (thus the 'barotrauma'), then the volume (hence the 'volutrauma'), finally the cyclic opening-closing of the lung units ('atelectrauma'). Less attention has been paid to the respiratory rate and the flow, although both theoretical considerations and experimental evidence attribute them a significant role in the generation of VILI. The initial injury to the lung parenchyma is necessarily mechanical and it could manifest as an unphysiological distortion of the extracellular matrix and/or as micro-fractures in the hyaluronan, likely the most fragile polymer embedded in the matrix. The order of magnitude of the energy required to break a molecular bond between the hyaluronan and the associated protein is 1.12×10-16 Joules (J), 70-90% higher than the average energy delivered by a single breath of 1L assuming a lung elastance of 10 cmH2O/L (0.5 J). With a normal statistical distribution of the bond strength some polymers will be exposed each cycle to an energy large enough to rupture. Both the extracellular matrix distortion and the polymer fractures lead to inflammatory increase of capillary permeability with edema if a pulmonary blood flow is sufficient. The mediation analysis of higher vs. lower tidal volume and PEEP studies suggests that the driving pressure, more than tidal volume, is the best predictor of VILI, as inferred by increased mortality. This is not surprising, as both tidal volume and respiratory system elastance (resulting in driving pressure) may independently contribute to the mortality. For the same elastance driving pressure is a predictor similar to plateau pressure or tidal volume. Driving pressure is one of the components of the mechanical power, which also includes respiratory rate, flow and PEEP. Finding the threshold for mechanical power would greatly simplify assessment and prevention of VILI.
Collapse
Affiliation(s)
- Tommaso Tonetti
- Department of Anesthesiology, Emergency and Intensive Care Medicine, University of Göttingen, Göttingen, Germany
| | - Francesco Vasques
- Department of Anesthesiology, Emergency and Intensive Care Medicine, University of Göttingen, Göttingen, Germany
| | - Francesca Rapetti
- Department of Anesthesiology, Emergency and Intensive Care Medicine, University of Göttingen, Göttingen, Germany
| | - Giorgia Maiolo
- Department of Anesthesiology, Emergency and Intensive Care Medicine, University of Göttingen, Göttingen, Germany
| | - Francesca Collino
- Department of Anesthesiology, Emergency and Intensive Care Medicine, University of Göttingen, Göttingen, Germany
| | - Federica Romitti
- Department of Anesthesiology, Emergency and Intensive Care Medicine, University of Göttingen, Göttingen, Germany
| | - Luigi Camporota
- Department of Adult Critical Care, Guy's and St Thomas' NHS Foundation Trust, King's Health Partners, London, UK.,Division of Asthma, Allergy and Lung Biology, King's College London, London, UK
| | - Massimo Cressoni
- Department of Medicine and Surgery, University of Milano-Bicocca, Milan, Italy
| | - Paolo Cadringher
- Department of Pathophysiology and Transplants, University of Milan, Milan, Italy
| | - Michael Quintel
- Department of Anesthesiology, Emergency and Intensive Care Medicine, University of Göttingen, Göttingen, Germany
| | - Luciano Gattinoni
- Department of Anesthesiology, Emergency and Intensive Care Medicine, University of Göttingen, Göttingen, Germany
| |
Collapse
|