Reimagining critical care: Trends and shifts in 21st century medicine

Abstract

Critical care medicine has undergone significant evaluation in the 21^st century, primarily driven by advancements in technology, changes in healthcare delivery, and a deeper understanding of disease processes. Advancements in technology have revolutionized patient monitoring, diagnosis, and treatment in the critical care setting. From minimally invasive procedures to advances imaging techniques, clinicians now have access to a wide array of tools to assess and manage critically ill patients more effectively. In this editorial we comment on the review article published by Padte S et al wherein they concisely describe the latest developments in critical care medicine.

Key Words: Artificintelligenceial; Ventilators; Extracorporeal organ support; Telemedicine; Critical care

Core Tip: Critical care medicine has seen significant advancements across various fronts. Innovations in technology, such as advanced monitoring systems, have enhanced patient care by providing real-time data to clinicians. Pharmacological advancements have led to the development of new drugs and treatment protocols, improving outcomes for critically ill patients. Additionally, there has been a growing emphasis on personalized medicine, tailoring treatments to individual patient needs through genomic and biomarker research. Furthermore, the integration of telemedicine and remote monitoring has expanded access to critical care expertise, particularly in underserved areas. Overall, these advancements have contributed to improved survival rates and quality of life for patients in critical care settings.

INTRODUCTION

Critical care in the 21^st century has undergone a profound transformation primarily driven by technological advancements. One of the most notable trends is the integration of advanced monitoring technologies into everyday practice. Another significant trend is the personalized approach to critical care management. The recognition that each patient responds uniquely to treatment has led to the adoption of precision medicine principles in critical care. Genetic testing, biomarker analysis, and predictive analytics now inform decision-making, allowing clinicians to tailor therapies more precisely to individual patient needs. This personalized approach not only enhances treatment efficacy but also minimizes adverse effects, marking a departure from the one-size-fits-all approach of the past. Artificial intelligence and machine learning algorithms are increasingly used to analyze vast amounts of patient data, predict clinical outcomes, and optimize treatment strategies. AI-powered decision support systems are in development to aid clinicians in making evidence-based decisions quickly and accurately.

REIMAGINING CRITICAL CARE: TRENDS AND SHIFTS IN 21ST CENTURY MEDICINE

Telemedicine, specifically within the realm of critical care known as Tele-ICU, utilizes advanced communication technologies to facilitate real-time, two-way interactions between patients and remote healthcare providers. This innovative approach to medical care allows for the diagnosis and treatment of patients across long distances, addressing the critical need for specialist care in intensive care units (ICUs). The significance of Tele-ICU has been underscored by the varying mortality outcomes associated with different levels of intensivist involvement in patient care. Research has consistently shown that ICUs with higher intensivist engagement tend to report better mortality outcomes compared to those with less intensive oversight[1]. The acute shortage of intensivists in the United States has prompted the exploration of Tele-ICU as a viable solution to bridge the gap in critical care services. This model permits remote specialists to monitor patients continuously, facilitating timely interventions by collaborating with onsite medical teams. The early application of telemedicine in critical care faced challenges; however, subsequent technological advancements have enhanced its feasibility and effectiveness[2]. Notably, a study by Rosenfeld et al[3] in the early 2000s demonstrated the potential of Tele-ICU to improve patient survival rates, marking a pivotal shift in the perception and adoption of telemedicine in ICU settings. Over the years, the adoption of Tele-ICU services has seen substantial growth, with a marked increase in the number of hospitals and ICU beds covered by such services. This expansion is supported by evidence from recent studies and meta-analyses, indicating significant improvements in ICU patient outcomes, including reduced mortality rates and shorter hospital stays.

Despite its promising benefits, the implementation of Tele-ICU faces financial barriers, notably the excessive costs associated with the necessary technology. However, given the ongoing physician shortage and the proven efficacy of Tele-ICU in enhancing patient care, it represents a critical advancement in the delivery of critical care services nationwide, offering a sustainable model for improving patient outcomes in ICUs.

Human intelligence is characterized by its capacity to engage in abstract thinking, employ reasoning to address challenges, grasp intricate concepts, formulate strategies, and glean insights from past encounters. Although adept at recognizing patterns, human intelligence faces constraints due to its finite memory capabilities. In contrast, artificial intelligence (AI) boasts extensive memory storage, excels in tackling multidimensional problems, and possesses the remarkable ability to identify subtle or "fuzzy" connections within extensive datasets[4]. Integrating AI into critical care has ushered in a transformative medical era, marked by enhanced diagnostic accuracy, prognostic precision, and therapeutic decision-making. AI encompasses algorithms designed for tasks requiring complex reasoning, akin to human cognition, such as problem-solving and decision-making. Deep learning (DL), a subset of AI, leverages artificial neural networks and advanced computational methods to develop highly individualized prediction models, significantly surpassing the capabilities of traditional clinical decision-making tools.

Significant advances, particularly with Pirracchio et al[5] super learner models (SL1 and SL2) in 2015, which outperformed traditional mortality prediction scores in ICU settings, marking a significant advancement in patient care. During the COVID-19 pandemic, AI proved invaluable, especially in patient prognosis using chest X-rays and accelerating diagnostics, significantly improving patient management, and showcasing its effectiveness in infectious disease control. In managing sepsis, AI's role in early detection and intervention has shown promising outcomes. By employing predictive analytics, AI models have accurately predicted critical care parameters, like urine output post-fluid administration, enhancing sepsis management and patient prognosis. AI also supports clinical decisions and enhances education and management in ICU settings through language learning models like GPT-4. Yet, its deployment faces challenges such as data quality, ethical concerns, and the need for thorough validation to confirm its efficacy and safety.

Despite these challenges, the outlook for AI in critical care medicine is bright, as continual progress stands ready to elevate the standard of care for critically ill patients. A more formidable yet potentially transformative endeavor lies in creating intelligent machine-learning systems that can constantly evaluate human reactions to critical illness. The ongoing evolution of AI holds the promise of streamlining diagnostic procedures, enhancing prognostic precision, and fine-tuning therapeutic approaches, signaling the dawn of a new era in critical care medicine.

FORGING AHEAD: SHEDDING LIGHT ON OUTDATED PRACTICES IN CRITICAL CARE

Critical care medicine is a rapidly evolving field that aims to provide lifesaving or life-prolonging interventions to patients in acute medical crisis. However, amidst the constant advancements of medical technology and knowledge, certain practices have become outdated. These practices not only failed to optimize patient outcomes but also contribute to inefficiencies and potential harm.

Fluid resuscitation is a cornerstone of septic shock management aimed at restoring adequate tissue perfusion. Previous large randomized studies comparing albumin with crystalloid solutions failed to show differences in clinical outcomes, such as overall mortality[6]. However, when considering the use of crystalloids, clinicians must determine whether saline or a balanced solution should be employed. The traditional approach of administering large volumes of fluids based on static markers such as central venous pressure or pulmonary artery catheter (PAC) readings has come under scrutiny. The Conservative vs Liberal Approach to Fluid Therapy of Septic Shock in Intensive Care (CLASSIC) trial compared a restrictive vs usual care fluid therapy regimen in 1550 adults with septic shock[7]. The median cumulative volume of intravenous fluids administered in the ICU was 2 L lower with a restrictive regimen. Nevertheless, implementing intravenous fluid restriction did not lead to reduced 90-day mortality, fewer serious adverse events, or a decreased need for renal replacement therapy (RRT).

A paradigm shift towards dynamic fluid management strategies, guided by parameters such as stroke volume variation or pulse pressure variation allows for more precise titration of fluid therapy. Additionally, goal directed therapy protocol, which targets specific hemodynamic endpoints, has shown promise in optimizing fluid reduction and improving patient outcomes in septic shock.

PAC has been subject of controversy over the last 35 years. PACs were once used to be a staple for intensivists for monitoring hemodynamics. PACs are now a rare sight in the ICU coinciding with development and use of noninvasive methods such as echocardiogram and ultrasounds.

Passive leg raise is a bedside technique involving the elevation of the patient's legs to 30-45 degrees while supine, causing approximately 300 mL of venous blood to shift from the lower extremities to the central circulation. This action elevates cardiac preload, resulting in a notable increase in mean systemic pressure[8]. In individuals responsive to fluid, this elevation is accompanied by heightened venous return and consequently increased cardiac output. Conversely, in those unresponsive to fluid, the rise in right atrial pressure counterbalances the increase in mean systemic pressure, maintaining the pressure gradient of venous return (and cardiac output) unchanged[8]. Notably, PLR is reversible and can be repeated as needed without fluid administration. Furthermore, PLR has been validated in spontaneously breathing patients and those with cardiac arrhythmias, low tidal volume ventilation, and low lung compliance[9]. An increase in cardiac output of over 10% induced by PLR typically indicates fluid responsiveness with high sensitivity and specificity (85% and 91%, respectively)[9].

Pulse pressure variation is one of the most studied markers of preload responsiveness. It proved to be a remarkable and reliable predictor in patients ventilated at low tidal volumes without cardiac arrhythmias. During mechanical ventilation, positive pressure breaths decrease preload of the right ventricle, which induces a decrease in preload of the left ventricle. If the left ventricular stroke volume changes in response to cyclical positive pressure ventilation, this indicates that the circulatory system is volume dependent.

There were important advances made over the last couple of decades in the field of ventilatory strategies. There is increasing evidence for the use of noninvasive mechanical ventilation in patients with hypercapnic respiratory failure, chronic obstructive pulmonary disease, and cardiogenic pulmonary edema, and for weaning patients off of invasive mechanical support. Invasive mechanical ventilation is a lifesaving intervention for acutely ill patients. Mechanical ventilators are inspiratory assist devices that integrate volumes, pressure, flow, and time (each as dependent or independent variables) to deliver tidal breath under positive pressure. Volume-controlled (VC) ventilation is a popular mode where the tidal volume is set and the pressure that results from delivering the volume is not. Alternatively, if the pressure delivered is set and tidal volume is not, then the patient is receiving pressure-controlled (PC) ventilation. With VC ventilation there are two common strategies of breath sequencing: assist control (AC) and synchronized intermittent mandatory ventilation (SIMV). With PC ventilation, there is AC, pressure-regulated volume controlled, pressure support, and SIMV. Less commonly used advanced modes include airway pressure release ventilation.

The degree of asynchrony can be quantified using the Asynchrony Index, calculated as the ratio of asynchronous breaths to the total breath count (including ventilator cycles and non-triggered breaths), expressed as a percentage. Sinderby et al[10] introduced a standardized, automated approach for measuring asynchronies known as the NeuroSync Index. This method relies on monitoring the Electrical Activity of the Diaphragm (EAdi) and conducting offline analysis of ventilator waveforms to assess asynchrony rates.

There is increasing evidence that suggests the effectiveness of employing neurally adjusted ventilatory assist mode[11], which is guided by the EAdi. This approach optimizes patient-ventilator synchronization and mitigates the risks of both over and under-assistance, conditions that can exacerbate diaphragmatic function, leading to atrophy and fatigue. Another notable advancement in recent years is the integration of ventilators with software capable of detecting and compensating for air leaks. This technological development holds the potential to substantially enhance patient-ventilator synchrony and ultimately improve outcomes.

Contemporary invasive mechanical ventilatory management targets airway plateau pressures as surrogates for alveolar distension pressure. Airway pressure, however, reflects the sum of distending pressures of the lung and chest wall. The use of esophageal pressure manometer (Pes) allows for the estimation of pleural pressures[12]. Pes tracings can be used to individually understand and titrate mechanical ventilatory support in patients with acute respiratory distress syndrome (ARDS).

AI is making its presence known in the world of ventilatory support. AI has the potential to identify ventilation strategies and provide real-time guidance on optimal ventilator settings using patient characteristics, underlying pathology, and imaging data in addition to the waveform data generated by patient monitors and the ventilator. Research indicates that AI algorithms, especially DL algorithms, exhibit impressive performance in the classification of lung diseases. This capability holds significant promise, as AI can assist in identifying various phenotypes of respiratory failure and subsequently develop clinician decision support systems tailored to address them.

More than 50 years ago, synchronous failure of various organ systems, also known as multiorgan failure or multiorgan dysfunction syndrome, was initially characterized as a syndrome encompassing "respiratory failure, hypotension, sepsis, and jaundice"[13]. Over the past century, there has been a proliferation of extracorporeal organ support modalities. The inception of this concept dates back approximately 100 years with the introduction of RRT. Similarly, Gibbon pioneered the use of artificial oxygenation and perfusion support, leading to the success of the first open-heart surgery in 1953[14]. Subsequently, Kolobow et al[15] furthered these advancements by developing an alveolar membrane artificial "heart-lung". Leveraging similar principles, liver-support therapies employing albumin dialysis and carbon dioxide removal devices utilizing membrane oxygenators have since become available.

More than 7% of hospitalized patients experience renal injury[16]. Given the lack of effective pharmacological treatments, acute kidney injury necessitates management primarily through RRT. RRT can be administered using either intermittent or continuous modalities. Intermittent dialysis seeks to replicate kidney function for blood purification over a period of up to 5 hours. Continuous RRT offers advantages such as improved hemodynamic stability, diminished transcellular solute shifts, and enhanced tolerance to fluid removal.

ARDS presents with a mortality rate exceeding 40% and affects approximately 10% of patients in the ICU[17]. Despite the implementation of various effective strategies, such as prone positioning and low tidal volume ventilation, the prognosis for ARDS remains daunting. However, the advent of extracorporeal membrane oxygenation (ECMO)[18] has revolutionized the management of severe ARDS as a last-resort intervention. It holds the potential to reduce mortality in patients who have failed conventional and alternative rescue therapies for ARDS. The prognosis of patients undergoing ECMO therapy is heavily influenced by concurrent non-pulmonary organ failure. Notably, in the EOLIA trial, ECMO decreased mortality from 39% to 22% in patients with a Sequential Organ Failure Assessment (SOFA) score below 11, but its efficacy was diminished in those with a SOFA score of 11 or higher[19,20].

Gattinoni et al[21] introduced the notion of less invasive extracorporeal lung support, focusing solely on carbon dioxide removal (ECCO2R). Due to its reduced blood flow requirements, ECCO2R can be combined with other extracorporeal support devices, such as RRT. Pumpless extracorporeal lung assist (pECLA) has demonstrated efficacy in CO₂ removal while simultaneously enhancing oxygenation.

The liver is one of the three primary organs responsible for detoxification, alongside the kidneys and lungs. Liver failure results in decreased synthesis of beneficial proteins while increasing the levels of circulating protein-bound toxins. The molecular adsorbent recirculating system (MARS)[22] has proven effective in removing bilirubin, ammonia, and creatinine. Gerth et al[22] conducted a subgroup analysis revealing an enhanced 28-day transplant-free survival rate among patients with acute on chronic liver failure (Grade 2 or higher). This indicates a potential benefit of MARS in providing multiorgan support through the elimination of protein-bound toxins.

Finally, within the past decade, a device that combines kidney, lung, and liver support along with blood detoxification has emerged. Known as the Advanced Organ Support (ADVOS) system[23], it facilitates the efficient removal of both water-soluble and protein-bound toxins and metabolites, offering integrated kidney, liver, and lung support for patients experiencing multi-organ failure. This system operates on the principle of albumin dialysis. Among the potential advantages of the ADVOS system is the capability to adjust the pH of the dialysate, tailoring it to the specific needs of the patient during treatment[23].

CONCLUSION

In conclusion critical care medicine has seen significant advancements across various fronts. The advancements in critical care medicine have revolutionized patient outcomes, offering more precise diagnostics, personalized treatments, and improved monitoring techniques. From innovative technologies to enhanced understanding of complex conditions, these developments underscore the commitment of medical professionals to continuously push the boundaries of care, ultimately saving lives and improving the quality of life for patients worldwide.