This Week in Critical Care — Jun 6, 2026

Welcome to This Week in Critical Care. This week we're covering 10 notable papers spanning the management of sepsis and shock, advances in respiratory failure, and the integration of new technologies into ICU operations. Let's dive in.

We begin with a deep dive into sepsis and shock, a perennial challenge in the ICU. A major paper in Critical Care Medicine helps us define the sickest of these patients by operationalizing new consensus criteria for refractory septic shock [7]. In a large, multicohort retrospective study of over 15,000 patients in the United States, researchers defined refractory septic shock as the concurrent need for a norepinephrine equivalent dose over 0.5 micrograms per kilogram per minute and a lactate greater than 2 millimoles per liter. They found this devastating condition occurred in about one in five patients with septic shock. The clinical implications are stark: hospital mortality was 64.4% in this group, and after risk adjustment, the odds of dying were nearly five times higher compared to septic shock patients who did not meet these criteria. This work gives us a concrete definition to identify a high-risk population for enrollment in future clinical trials of rescue therapies.

While we define the problem, other research explores potential treatments. A retrospective cohort study in Shock evaluates adjunctive tocilizumab, an IL-6 receptor antibody, in pediatric septic shock [4]. Among 58 children, those who received tocilizumab had a dramatically shorter median duration of shock—84 hours versus 240 hours—a shorter ICU stay, and significantly lower mortality, at about 19% versus 46% in historical controls. The benefit was particularly pronounced in patients with Gram-negative sepsis, who had much higher initial IL-6 levels. Despite this more severe initial inflammation, their mortality after treatment was not significantly different from the Gram-positive cohort, suggesting the IL-6 blockade effectively blunted the hyperinflammatory surge. This provides compelling, albeit retrospective, evidence for targeted immunotherapy in this population.

Ventilator management in sepsis is another key topic, and a multi-cohort observational analysis in the Journal of Critical Care investigated the impact of PEEP levels [3]. Analyzing data from 844 mechanically ventilated septic patients, the study stratified them into higher or lower PEEP groups based on the ARDSnet table. In two of the three cohorts, a higher PEEP strategy was associated with significantly lower mortality—for example, a 30-day mortality of 22% versus nearly 42% in one prospective cohort. Importantly, this was achieved without evidence of hemodynamic deterioration; vasopressor requirements and lactate levels were similar or even improved in the high PEEP groups. The authors appropriately caution that these are observational, hypothesis-generating findings that require confirmation in a randomized trial, but they challenge the common fear that higher PEEP is necessarily detrimental to hemodynamics in septic patients.

Rounding out our sepsis theme, a narrative review in Critical Care urges us to rethink a cornerstone of shock monitoring: lactate [8]. The authors argue that hyperlactatemia, especially in sepsis, is often not a simple sign of tissue hypoxia. Instead, it should be viewed as an integrated metabolic signal reflecting both stress-induced production and, critically, impaired organ clearance. The kidneys, in particular, play an underappreciated role in lactate utilization. When sepsis-induced acute kidney injury occurs, this clearance mechanism is compromised, directly contributing to elevated lactate levels. This perspective recasts lactate as a marker of systemic stress and organ crosstalk, rather than just anaerobic metabolism, with important implications for how we interpret it at the bedside, particularly in patients with AKI. This complements a review in Critical Care Medicine on multimodal perfusion assessment, which advocates for an integrated approach using tools like capillary refill time as a clinical anchor, complemented by other measures to assess physiologic coherence rather than chasing a single number [5].

Next, we turn to respiratory failure and liberation from the ventilator. A fascinating study in Critical Care asks whether all ARDS is created equal, comparing postoperative to medical ARDS [1]. In a large retrospective analysis of over 1,000 patients, postoperative ARDS, which accounted for 42% of cases, had a more favorable trajectory and significantly lower 90-day mortality—36% versus 49% for medical ARDS. After adjustment, the risk of death remained about 30% lower. The prognostic factors also differed markedly. For postoperative ARDS, mortality was linked to extrapulmonary organ dysfunction and the type of surgery, with no independent association with respiratory mechanics. In contrast, for medical ARDS, mortality was strongly associated with the severity of respiratory failure, including the PaO2/FiO2 ratio and driving pressure. This suggests postoperative ARDS is a distinct clinical subtype, driven more by systemic and surgical factors than the lung injury itself, arguing for a management focus on perioperative prevention and early detection of surgical complications.

Once a patient is recovering, weaning from mechanical ventilation presents its own hurdles. The WEAN-US study, also in Critical Care, investigated a comprehensive, ultrasound-based assessment for predicting weaning success in difficult-to-wean patients [2]. This single-center prospective study found that weaning failure, which occurred in 27% of patients, is a complex process. An integrated predictive model that combined ultrasound assessment of diaphragmatic function, lung aeration, and cardiac function, along with measures of peripheral muscle strength and dyspnea, had excellent predictive performance, with an AUROC of 0.88. This multimodal approach significantly outperformed any single predictor. Interestingly, the study teased out different mechanisms of failure: cardiac dysfunction was more strongly associated with failing a spontaneous breathing trial, whereas neuromuscular impairment, like low handgrip strength, was more linked to post-extubation failure. This supports a more individualized, physiology-guided approach to ventilator liberation.

Finally, we look at ICU operations and future technologies. A study in Critical Care Medicine reports on the implementation of a remote respiratory therapy service in a donor center ICU [6]. Over a 12-month period, remote respiratory therapists using a telecritical care platform performed nearly 4,000 procedures, totaling over 1,700 hours of care, including full remote ventilator management. On-site RT support was needed only about 6% of the time, mainly for transport. Critically, there were no airway losses, emergency activations, or delays in care. This model resulted in an estimated savings of 2.2 full-time-equivalent staff and over $300,000 in labor costs, all while maintaining excellent organ procurement outcomes. This demonstrates a safe and effective model for extending specialized expertise and addressing staffing shortages.

Looking further into the future, another paper in Critical Care Medicine describes the temporal validation of an EMR-integrated machine learning system called BEST-AI [10]. Deployed in a Japanese ICU, the system generates hourly predictions for six key outcomes, including mortality and extubation. In a forward-in-time validation cohort, the model showed strong discrimination, with AUROCs ranging from 0.85 to 0.96, and generally good calibration. The system was successfully integrated into clinical workflow with automated updates and visualizations, providing clinicians with real-time probabilistic forecasting without generating prescriptive alerts. This work represents a crucial step in moving predictive analytics from theoretical models to validated, operational tools at the bedside. And on the basic science front, a paper in Science Translational Medicine identifies a potential new therapeutic pathway for idiopathic pulmonary fibrosis [9]. Researchers found that deficiency in Toll-like receptor 5, or TLR5, is associated with IPF in humans and worse fibrosis in mice. Activating TLR5 in mice protected them from fibrosis, a mechanism that appears to be mediated by improving lung dysbiosis after injury. This work links innate immunity, the lung microbiome, and fibrosis, opening a new avenue for potential therapies.

If you only have time for one paper this week, make it the study in Critical Care comparing postoperative and medical ARDS [1]. These findings strongly suggest that postoperative ARDS is a distinct clinical entity with a different prognosis and different drivers of mortality. This has immediate implications for how we counsel families, prognosticate, and design future clinical trials, reminding us to look beyond the lungs in our surgical patients with ARDS.

Here are the key takeaways from this week in Critical Care: First, postoperative ARDS appears to be a separate phenotype from medical ARDS, with lower mortality that is driven more by extrapulmonary and surgical factors than by the severity of the lung injury itself. Your management focus should broaden accordingly [1]. Second, for your difficult-to-wean patients, a multimodal assessment incorporating ultrasound of the heart, lungs, and diaphragm, plus measures of peripheral muscle strength, provides a much better prediction of weaning outcomes than any single parameter alone [2]. Third, a new operational definition for refractory septic shock—a norepinephrine dose over 0.5 mics per kilo per minute plus a lactate over 2—identifies a cohort comprising one in five septic shock patients who face a grim mortality rate of over 60%. This gives a clear target for future rescue therapy trials [7]. Fourth, in pediatric septic shock, retrospective data suggests adjunctive tocilizumab is associated with markedly improved survival and faster shock resolution, especially in severe Gram-negative infections, supporting the concept of targeted immunomodulation [4]. Finally, tele-respiratory therapy, including full remote ventilator management, can be safely and effectively implemented, offering a viable model to improve specialized care delivery and address staffing shortages [6].

That's your roundup for This Week in Critical Care. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.