Physiology Is Vital to Precision Medicine in Acute Respiratory Distress Syndrome and Sepsis

Predicting mortality and risk factors of sepsis related ARDS using machine learning models

Sepsis related acute respiratory distress syndrome (ARDS) is a common and serious disease in clinic. Accurate prediction of in-hospital mortality of patients is crucial to optimize treatment and improve prognosis under the new global definition of ARDS. Our study aimed to use machine learning models to develop models that can effectively predict the in-hospital mortality of patients with sepsis related ARDS, calculate the mortality, and to identify related risk factors under the new global definition of ARDS. Based on MIMIC database, our study included 3470 first-time admission records of patients with sepsis related ARDS. After excluding 4 patients under the age of 18, 75 patients with less than 24 h stay in ICU, and 5 cases with missing indicators > 30%, finally 3386 cases were retained. The variance inflation factor (VIF) analysis was used to test the collinearity of the explanatory variables. The data were divided into the training set and the test set according to the ratio of 7:3. Six models, extreme gradient boosting (XGBoost), light gradient boosting (LightGBM), random forest (RF), classification and regression tree (CART), naive bayes (NB) and logistic regression (LR), were designed for training and testing. In the training set, XGBoost (AUROC = 0.951, 95% CI 0.942-0.961), LR (AUROC = 0.835, 95% CI 0.817-0.854), RF (AUROC = 1.0, 95% CI 1.0-1.0), LightGBM (AUROC = 1.0, 95% CI 1.0-1.0), CART (AUROC = 0.831, 95% CI 0.811-0.852), NB (AUROC = 0.793, 95% CI 0.772-0.814). In the test set, XGBoost (AUROC = 0.833, 95% CI 0.804-0.861), LR (AUROC = 0.82695% CI 0.796-0.856), RF (AUROC = 0.846, 95% CI 0.818-0.874), LightGBM (AUROC = 0.827, 95% CI 0.798-0.856), CART (AUROC = 0.753, 95% CI 0.718-0.787), NB (AUROC = 0.799, 95% CI 0.768-0.831). The RF model has the best performance on the test set. Further analyze the feature importance ranking and partial dependence plots of random forest model. Acute physiology and chronic health evaluation III (APACHE III), bicarbonate, anion gap and non-invasive blood pressure systolic were identified as the four most important risk characteristics. In this study, a variety of machine learning models have been successfully constructed to predict the in-hospital mortality of patients with sepsis related ARDS, among which the RF model performs well. Key risk factors identified include APACHE III, bicarbonate, anion gap and non-invasive blood pressure systolic. The identification of these factors helps clinicians to assess patients' conditions more accurately and develop personalized treatment plans, thereby improving the survival rate and prognosis quality of patients under the new global definition of ARDS.

Subphenotyping prone position responders with machine learning

BACKGROUND

Acute respiratory distress syndrome (ARDS) is a heterogeneous condition with varying response to prone positioning. We aimed to identify subphenotypes of ARDS patients undergoing prone positioning using machine learning and assess their association with mortality and response to prone positioning.

METHODS

In this retrospective observational study, we enrolled 353 mechanically ventilated ARDS patients who underwent at least one prone positioning cycle. Unsupervised machine learning was used to identify subphenotypes based on respiratory mechanics, oxygenation parameters, and demographic variables collected in supine position. The primary outcome was 28-day mortality. Secondary outcomes included response to prone positioning in terms of respiratory system compliance, driving pressure, PaO2/FiO2 ratio, ventilatory ratio, and mechanical power.

RESULTS

Three distinct subphenotypes were identified. Cluster 1 (22.9% of whole cohort) had a higher PaO2/FiO2 ratio and lower Positive End-Expiratory Pressure (PEEP). Cluster 2 (51.3%) had a higher proportion of COVID-19 patients, lower driving pressure, higher PEEP, and higher respiratory system compliance. Cluster 3 (25.8%) had a lower pH, higher PaCO2, and higher ventilatory ratio. Mortality differed significantly across clusters (p = 0.03), with Cluster 3 having the highest mortality (56%). There were no significant differences in the proportions of responders to prone positioning for any of the studied parameters. Transpulmonary pressure measurements in a subcohort did not improve subphenotype characterization.

CONCLUSIONS

Distinct ARDS subphenotypes with varying mortality were identified in patients undergoing prone positioning; however, predicting which patients benefited from this intervention based on available data was not possible. These findings underscore the need for continued efforts in phenotyping ARDS through multimodal data to better understand the heterogeneity of this population.

Pathophysiological Markers of Acute Respiratory Distress Syndrome Severity Are Correlated With Ventilation-Perfusion Mismatch Measured by Electrical Impedance Tomography

OBJECTIVES

Pulmonary ventilation/perfusion (V/Q) mismatch measured by electrical impedance tomography (EIT) is associated with the outcome of patients with the acute respiratory distress syndrome (ARDS), but the underlying pathophysiological mechanisms have not been fully elucidated. The present study aimed to verify the correlation between relevant pathophysiological markers of ARDS severity and V/Q mismatch.

DESIGN

Prospective observational study.

SETTING

General ICU of a university-affiliated hospital.

PATIENTS

Deeply sedated intubated adult patients with ARDS under controlled mechanical ventilation.

INTERVENTIONS

Measures of V/Q mismatch by EIT, respiratory mechanics, gas exchange, lung imaging, and plasma biomarkers.

MEASUREMENTS AND MAIN RESULTS

Unmatched V/Q units were assessed by EIT as the fraction of ventilated nonperfused plus perfused nonventilated lung units. At the same time, plasma biomarkers with proven prognostic and mechanistic significance for ARDS (carbonic anhydrase 9 [CA9], hypoxia-inducible factor 1 [HIF1], receptor for advanced glycation endproducts [RAGE], angiopoietin 2 [ANG2], gas exchange, respiratory mechanics, and quantitative chest CT scans were measured. Twenty-five intubated ARDS patients were included with median unmatched V/Q units of 37.1% (29.2-49.2%). Unmatched V/Q units were correlated with plasma levels of CA9 (rho = 0.47; p = 0.01), HIF1 (rho = 0.40; p = 0.05), RAGE (rho = 0.46; p = 0.02), and ANG2 (rho = 0.42; p = 0.03). Additionally, unmatched V/Q units correlated with plateau pressure ( r = 0.38; p = 0.05) and with the number of quadrants involved on chest radiograph ( r = 0.73; p < 0.01). Regional unmatched V/Q units were correlated with the corresponding fraction of poorly aerated lung tissue ( r = 0.62; p = 0.01) and of lung tissue weight (rho: 0.51; p = 0.04) measured by CT scan.

CONCLUSIONS

In ARDS patients, unmatched V/Q units are correlated with pathophysiological markers of lung epithelial and endothelial dysfunction, increased lung stress, and lung edema. Unmatched V/Q units could represent a comprehensive marker of ARDS severity, reflecting the complex organ pathophysiology and reinforcing their prognostic significance.

Association Between Baseline Driving Pressure and Mortality in Very Old Patients with Acute Respiratory Distress Syndrome

Rationale: Because of the effects of aging on the respiratory system, it is conceivable that the association between driving pressure and mortality depends on age. Objectives: We endeavored to evaluate whether the association between driving pressure and mortality of patients with acute respiratory distress syndrome (ARDS) varies across the adult lifespan, hypothesizing that it is stronger in older, including very old (⩾80 yr), patients. Methods: We performed a secondary analysis of individual patient-level data from seven ARDS Network and PETAL Network randomized controlled trials ("ARDSNet cohort"). We tested our hypothesis in a second, independent, national cohort ("Hellenic cohort"). We performed both binary logistic and Cox regression analyses including the interaction term between age (as a continuous variable) and driving pressure at baseline (i.e., the day of trial enrollment) as the predictor and 90-day mortality as the dependent variable. Measurements and Main Results: On the basis of data from 4,567 patients with ARDS included in the ARDSNet cohort, we found that the effect of driving pressure on mortality depended on age (P = 0.01 for the interaction between age as a continuous variable and driving pressure). The difference in driving pressure between survivors and nonsurvivors significantly changed across the adult lifespan (P < 0.01). In both cohorts, a driving pressure threshold of 11 cm H2O was associated with mortality in very old patients. Conclusions: Data from randomized controlled trials with strict inclusion criteria suggest that the effect of driving pressure on the mortality of patients with ARDS may depend on age. These results may advocate for a personalized age-dependent mechanical ventilation approach.

Lung ultrasound is associated with distinct clinical phenotypes in COVID-19 ARDS: A retrospective observational study

BACKGROUND

ARDS is a heterogeneous syndrome with distinct clinical phenotypes. Here we investigate whether the presence or absence of large pulmonary ultrasonographic consolidations can categorize COVID-19 ARDS patients requiring mechanical ventilation into distinct clinical phenotypes.

METHODS

This is a retrospective study performed in a tertiary-level intensive care unit in Israel between April and September 2020. Data collected included lung ultrasound (LUS) findings, respiratory parameters, and treatment interventions. The primary outcome was a composite of three ARDS interventions: prone positioning, high PEEP, or a high dose of inhaled nitric oxide.

RESULTS

A total of 128 LUS scans were conducted among 23 patients. The mean age was 65 and about two-thirds were males. 81 scans identified large consolidation and were classified as "C-type", and 47 scans showed multiple B-lines with no or small consolidation and were classified as "B-type". The presence of a "C-type" study had 2.5 times increased chance of receiving the composite primary outcome of advanced ARDS interventions despite similar SOFA scores, Pao2/FiO2 ratio, and markers of disease severity (OR = 2.49, %95CI 1.40-4.44).

CONCLUSION

The presence of a "C-type" profile with LUS consolidation potentially represents a distinct COVID-19 ARDS subphenotype that is more likely to require aggressive ARDS interventions. Further studies are required to validate this phenotype in a larger cohort and determine causality, diagnostic, and treatment responses.

Mechanical Ventilation, Past, Present, and Future

Mechanical ventilation (MV) has played a crucial role in the medical field, particularly in anesthesia and in critical care medicine (CCM) settings. MV has evolved significantly since its inception over 70 years ago and the future promises even more advanced technology. In the past, ventilation was provided manually, intermittently, and it was primarily used for resuscitation or as a last resort for patients with severe respiratory or cardiovascular failure. The earliest MV machines for prolonged ventilatory support and oxygenation were large and cumbersome. They required a significant amount of skills and expertise to operate. These early devices had limited capabilities, battery, power, safety features, alarms, and therefore these often caused harm to patients. Moreover, the physiology of MV was modified when mechanical ventilators moved from negative pressure to positive pressure mechanisms. Monitoring systems were also very limited and therefore the risks related to MV support were difficult to quantify, predict and timely detect for individual patients who were necessarily young with few comorbidities. Technology and devices designed to use tracheostomies versus endotracheal intubation evolved in the last century too and these are currently much more reliable. In the present, positive pressure MV is more sophisticated and widely used for extensive period of time. Modern ventilators use mostly positive pressure systems and are much smaller, more portable than their predecessors, and they are much easier to operate. They can also be programmed to provide different levels of support based on evolving physiological concepts allowing lung-protective ventilation. Monitoring systems are more sophisticated and knowledge related to the physiology of MV is improved. Patients are also more complex and elderly compared to the past. MV experts are informed about risks related to prolonged or aggressive ventilation modalities and settings. One of the most significant advances in MV has been protective lung ventilation, diaphragm protective ventilation including noninvasive ventilation (NIV). Health care professionals are familiar with the use of MV and in many countries, respiratory therapists have been trained for the exclusive purpose of providing safe and professional respiratory support to critically ill patients. Analgo-sedation drugs and techniques are improved, and more sedative drugs are available and this has an impact on recovery, weaning, and overall patients' outcome. Looking toward the future, MV is likely to continue to evolve and improve alongside monitoring techniques and sedatives. There is increasing precision in monitoring global "patient-ventilator" interactions: structure and analysis (asynchrony, desynchrony, etc). One area of development is the use of artificial intelligence (AI) in ventilator technology. AI can be used to monitor patients in real-time, and it can predict when a patient is likely to experience respiratory distress. This allows medical professionals to intervene before a crisis occurs, improving patient outcomes and reducing the need for emergency intervention. This specific area of development is intended as "personalized ventilation." It involves tailoring the ventilator settings to the individual patient, based on their physiology and the specific condition they are being treated for. This approach has the potential to improve patient outcomes by optimizing ventilation and reducing the risk of harm. In conclusion, MV has come a long way since its inception, and it continues to play a critical role in anesthesia and in CCM settings. Advances in technology have made MV safer, more effective, affordable, and more widely available. As technology continues to improve, more advanced and personalized MV will become available, leading to better patients' outcomes and quality of life for those in need.

Hemoglobin signal network mapping reveals novel indicators for precision medicine

Precision medicine currently relies on a mix of deep phenotyping strategies to guide more individualized healthcare. Despite being widely available and information-rich, physiological time-series measures are often overlooked as a resource to extend insights gained from such measures. Here we have explored resting-state hemoglobin measures applied to intact whole breasts for two subject groups - women with confirmed breast cancer and control subjects - with the goal of achieving a more detailed assessment of the cancer phenotype from a non-invasive measure. Invoked is a novel ordinal partition network method applied to multivariate measures that generates a Markov chain, thereby providing access to quantitative descriptions of short-term dynamics in the form of several classes of adjacency matrices. Exploration of these and their associated co-dependent behaviors unexpectedly reveals features of structured dynamics, some of which are shown to exhibit enzyme-like behaviors and sensitivity to recognized molecular markers of disease. Thus, findings obtained strongly indicate that despite the use of a macroscale sensing method, features more typical of molecular-cellular processes can be identified. Discussed are factors unique to our approach that favor a deeper depiction of tissue phenotypes, its extension to other forms of physiological time-series measures, and its expected utility to advance goals of precision medicine.

Leveraging Data Science and Novel Technologies to Develop and Implement Precision Medicine Strategies in Critical Care

Precision medicine aims to identify treatments that are most likely to result in favorable outcomes for subgroups of patients with similar clinical and biological characteristics. The gaps for the development and implementation of precision medicine strategies in the critical care setting are many, but the advent of data science and multi-omics approaches, combined with the rich data ecosystem in the intensive care unit, offer unprecedented opportunities to realize the promise of precision critical care. In this article, the authors review the data-driven and technology-based approaches being leveraged to discover and implement precision medicine strategies in the critical care setting.

Laying the Groundwork for Physiology-Guided Precision Medicine in the Critically Ill

Canonical critical care syndromes such as sepsis and acute respiratory distress syndrome (ARDS) include patients with markedly heterogeneous biology.1 This, paired with decades of randomized controlled trials (RCTs) that were traditionally viewed as "negative," has stalled progress in improving patient outcomes.2 However, emerging awareness of sub-phenotypes based on differences in biomarker profiles and resulting heterogeneous treatment effects have led to calls for precision medicine in which therapies are targeted to those most likely to benefit.3.

Septic shock in obstetrics: guidelines of the All-Russian public organization “Federation of Anesthesiologists and Reanimatologists”

The article reflects the main provisions of the clinical guidelines on septic shock in obstetrics, approved by the All-Russian public organization “Federation of Anesthesiologists-Resuscitators” in 2022. The relevance of the problem is associated with high mortality and morbidity rates from sepsis and septic shock in obstetrics. The main issues of etiology, pathogenesis, clinical picture, methods of laboratory and instrumental diagnostics, features of using the qSOFA, SOFA, MOEWS, SOS, MEWC, IMEWS scales for sepsis verification are consistently presented. The article presents the starting intensive therapy (the first 6–12 hours) of the treatment of septic shock in obstetrics, taking into account the characteristics of the pregnant woman's body. The strategy of prescribing vasopressors (norepinephrine, phenylephrine, epinephrine), inotropic drugs (dobutamine) is described, antibiotics and optimal antibiotic therapy regimens, features of infusion and adjuvant therapy are presented. The issues of surgical treatment of the focus of infection and indications for hysterectomy, as well as the organization of medical care and rehabilitation of patients with sepsis and septic shock were discussed. The basic principles of prevention of sepsis and septic shock in obstetrics are described. The criteria for the quality of medical care for patients with septic shock and the algorithms of doctor's actions in the diagnosis and intensive care of patients with septic shock in obstetrics are presented.