This week in Critical Care

This AudioScholar briefing is generated by artificial intelligence for healthcare professionals and trainees. It is not medical advice.

Welcome to This Week in Critical Care. This week we're covering 10 notable papers spanning respiratory support and ventilator care, resuscitation and post-arrest management, and a pair of papers on acid-base therapy and bedside brain imaging. Let's dive in.

We start with respiratory support, where several papers this week test long-held assumptions about how we ventilate and protect the lungs. The headline is PROSpect, published in The Lancet by Curley and colleagues [1]. This international factorial trial, run across 75 paediatric intensive care units in 18 countries, randomised children with early, moderate-to-severe paediatric ARDS to prone or supine positioning, and to high-frequency oscillatory ventilation or conventional ventilation. The trial stopped early for futility and slow enrolment after 424 of 600 planned patients. Neither intervention changed ventilator-free days at day 28; the medians sat at around 17 to 18 days in every comparison, and day-28 mortality was about ten percent without group differences. The harms, however, were not neutral. Prone positioning was linked to roughly five times the rate of airway plugging or obstruction and about six times the rate of stage 2 or worse pressure injuries, while oscillation was associated with more haemodynamic instability. For children with early moderate-to-severe ARDS, this is the largest randomised evidence to date, and it does not support routine prone positioning or oscillation as a strategy to shorten ventilation, although the early stop leaves some imprecision. Moving to adults, Maillard and colleagues in Intensive Care Medicine revisited the HIGH trial of high flow nasal oxygen in immunocompromised patients with acute hypoxaemic respiratory failure [2]. The original trial found no overall benefit, and this reanalysis confirmed no significant average effect on intubation. But a causal forest machine-learning model found wide heterogeneity, with predicted benefit ranging from about a twenty point absolute reduction in intubation with high flow to a smaller predicted benefit for conventional oxygen. The patients predicted to gain most were older, more often solid organ transplant recipients, less neutropenic, and more hypoxaemic but with less non-respiratory organ failure. The ranking held up in an external observational cohort, though absolute estimates were less well calibrated. This is hypothesis-generating, and the authors themselves call for prospective validation.

Two further papers address the routine care of ventilated adults. In Critical Care, Anushapreethi and colleagues ran a pragmatic cluster-randomised crossover trial in six medical ICUs at a single centre in India, adding 0.12 percent chlorhexidine to a standardised oral-care bundle [3]. Among just over a thousand ventilated patients, chlorhexidine did not reduce ventilator-associated pneumonia, with roughly 17 versus 19 episodes per thousand ventilator-days, did not change 28-day mortality, and was associated with ICU stays nearly a day longer. This adds to existing doubts about chlorhexidine outside cardiac surgery, though it comes from a single institution. On sedation, Wu and colleagues in the Journal of Critical Care analysed propensity-matched Taiwanese patients and found that dexmedetomidine or propofol, compared with benzodiazepines, was associated with ICU stays about two and a half days shorter, extubation about two days earlier, and hospital costs nearly a fifth lower, with no detectable difference in 90-day mortality [4]. It is retrospective and open to residual confounding, but it extends guideline-concordant evidence to an Asian population.

Our second theme is resuscitation, spanning temperature control, extracorporeal support, airway ventilation and traumatic arrest. Interestingly, the same causal forest approach used in the oxygen paper produced the opposite answer when Lileikyte and colleagues applied it to the TTM and TTM2 trials in Critical Care [5]. Across 2,789 out-of-hospital arrest patients, six-month mortality was just under half in both the 33 degree and normothermia groups, and machine-learning methods found no meaningful heterogeneity by ischaemic duration, prehospital factors or admission neurological findings. Posterior probabilities for benefit from hypothermia hovered around a coin flip. Placed side by side, these two papers show that these methods can surface subgroups but do not invent them, and they make the hypothesis of a hidden responder group for deep hypothermia considerably less plausible. Turning to extracorporeal CPR, Alexy and colleagues in Resuscitation followed patients discharged alive after ECPR for refractory shockable out-of-hospital arrest at centres in Minnesota and Prague [6]. Of 728 patients treated, 211 survived to discharge, and among those survivors roughly two thirds were alive at five years, with most deaths clustered in the first 90 days. Good neurological function was present in about four in five survivors at discharge and above ninety percent of those alive beyond a year. Five-year survival fell from over ninety percent in patients under 40 to about half in those over 70. This is retrospective data from two experienced programmes, but it addresses a real gap about the durability of ECPR benefit.

Also in Resuscitation, Aufderheide and colleagues tackled something more basic: whether paramedics actually deliver guideline-concordant ventilation during cardiac arrest [7]. In a before-and-after study of 278 patients in Wisconsin, monitoring with end-tidal carbon dioxide alone left only about seven percent of ventilations on target for both rate and tidal volume, with low tidal volume being the most common error. With a real-time rate and volume feedback device, the proportion on target for both rose more than thirty points, to about 39 percent. The design is non-randomised and reports process measures rather than survival, but it shows how far ventilation quality drifts without feedback. Finally, a third Resuscitation paper from Ramage and colleagues offers a prospective national look at resuscitative thoracotomy across United Kingdom helicopter services and major trauma centres [8]. Of 116 thoracotomies, about seven in ten were performed prehospital. Although tamponade was the most common indication, exsanguinating haemorrhage was the actual finding in about seven in ten patients. Return of circulation occurred in about 43 percent, but only eight patients survived to discharge, survival was much better with tamponade than exsanguination, and there were no survivors beyond ten minutes of traumatic arrest. That ten-minute observation is striking, though it rests on small numbers.

Our final theme covers two papers outside the main clusters. In Intensive Care Medicine, Kitisin and colleagues pooled four randomised trials with 1,544 patients in a Bayesian meta-analysis of sodium bicarbonate for metabolic acidosis in critically ill adults [9]. There was no clear short-term mortality benefit, which was the primary outcome. For the secondary outcome of renal replacement therapy initiation, bicarbonate was associated with about a quarter lower risk, with a high posterior probability of benefit, and no consistent safety signal emerged. The survival question remains open, and the authors call for trials that identify who benefits. And in Minerva Anestesiologica, Caldas and colleagues offer a practical narrative guide to neurological point-of-care ultrasound, covering middle cerebral artery haemodynamics, hydrocephalus, haematoma size, midline shift and intracranial pressure estimation, arguing that most of these are basic or basic-plus skills within reach of intensivists [10].

If you only have time for one paper this week, make it PROSpect in The Lancet [1]. It is the largest randomised test of prone positioning and oscillation in paediatric ARDS, and it reopens the question of whether proning earns its place in children given the absence of benefit and the clear signal of airway and skin harm.

Here is what this week's evidence adds up to in Critical Care. First, in early moderate-to-severe paediatric ARDS, randomised evidence now shows neither prone positioning nor oscillation shortens ventilation, with measurable harms, though an early stop limits precision. Second, machine-learning reanalyses are cutting both ways: they suggest a possible subgroup of immunocompromised patients who benefit from high flow oxygen, which still needs prospective confirmation, while finding no hidden responders to 33 degree temperature control after arrest. Third, chlorhexidine added to an oral-care bundle showed no benefit in a single-centre medical ICU trial, reinforcing existing doubts. Fourth, bicarbonate for metabolic acidosis has not been shown to improve survival, with a signal toward less dialysis that remains a secondary finding. And fifth, observational resuscitation data suggest ECPR survivors largely do well long term, ventilation quality during arrest is poor without feedback, and traumatic thoracotomy outcomes depend heavily on cause and time.

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

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And if you have a paper you have been meaning to read, upload the PDF, or paste any link, at audioscholar dot C C. We will turn it into audio like this one, in any of thirty-one languages.

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