A recent Intensive Care digest,
as sent to subscribers

Study Type: Retrospective cohort study (two ICUs, single tertiary centre, 2016–2022; n=139)

Key Question: What are the outcomes of patients with ALS admitted to ICU with acute respiratory failure, and which admission factors predict mortality?

Key Findings:

• Mortality was 48.2% at 3 months, rising to 61.2% at 6 months and 74.1% at 12 months. Most patients (79.9%) received non-invasive ventilation; 13.7% were intubated and 2.2% underwent tracheostomy.
• Hypoxaemia at admission (PaO₂/FiO₂ <300) independently predicted 3-month mortality (HR 2.27, 95% CI 1.28–4.01); older age was also associated (HR 1.03 per year).
• Among 3-month survivors, 79.2% were living at home.

Clinical Relevance: Admission hypoxaemia may help inform early prognostic discussions and shared decision-making when ALS patients present to UK critical care, though ICU admission is not futile for all.

Limitations: Single-centre retrospective design with no ALS-specific disease severity or functional baseline data reported in the abstract, limiting generalisability and confounder adjustment.

Study Type: Prospective cohort study (two tertiary hospitals, Athens, Greece; n=130)

Key Question: How do ex vivo innate immune responses differ between COVID-19 pneumonia and non-COVID-19 CAP, and do they vary across pandemic periods or relate to severe respiratory failure (SRF) and mortality?

Key Findings:

• PBMCs from COVID-19 patients (n=102) produced more TNFα and IL-6, but less IL-1RA, than those from non-COVID-19 CAP (n=28) on Day 1.
• Omicron-infected patients showed higher TNFα and lower IL-6 production than earlier variants.
• On multivariate analysis, a fall in PBMC IL-6 production over time was associated with progression to SRF and 28-day mortality (effect sizes not reported in abstract). Serum microRNA findings were exploratory and from a small subset.

Clinical Relevance: Findings suggest declining functional IL-6 responsiveness may flag deterioration, but this is a research observation with no bedside application or UK treatment implications yet.

Limitations: Small non-COVID comparator group, single-country setting and ex vivo assays limit generalisability, and the abstract lacks effect sizes and confidence intervals.

Study Type: Multicentre observational cohort study (Japan), with overlap propensity score weighting.

Key Question: Is sitting at the bedside within 72 hours of ICU admission associated with death or functional decline at 3 months post-discharge in mechanically ventilated survivors receiving rehabilitation?

Key Findings:

• Of 423 patients, 173 sat early. No association with the primary outcome of death or functional decline (aRR 0.94, 95% CI 0.80–1.09).
• Exploratory outcomes favoured early sitting for readmission or death (aRR 0.53, 95% CI 0.32–0.88; BH-adjusted p = 0.027), but this attenuated after adjusting for Day-3 clinical status. The cognitive impairment signal did not survive multiplicity adjustment or a stricter definition.

Clinical Relevance: A 72-hour sitting milestone should not be treated as a validated target for UK early mobilisation pathways, although sitting appears feasible in clinically stable patients.

Limitations: Residual confounding by clinical stability is likely, and the exploratory findings probably reflect patient selection rather than any effect of sitting.

Study Type: Retrospective multicentre cohort study (ancillary unsupervised machine-learning analysis of the national PLEXHIA cohort, France)

Key Question: Can distinct clinical phenotypes be identified among ICU patients with severe ANCA-associated vasculitis (AAV) and diffuse alveolar haemorrhage (DAH), and do they differ in mortality?

Key Findings:

• Among 197 patients (2012-2024), Kohonen self-organising maps identified three clusters using admission variables only: MPO pulmonary-renal (MPO-PR), mild-respiratory (Mild-R), and PR3 multi-organ (PR3-MO; younger, severe respiratory failure, multiorgan failure).
• Organ support differed: RRT was most frequent in MPO-PR (73%); invasive ventilation in MPO-PR (71%) and PR3-MO (68%).
• 30-day mortality differed significantly: 23% (MPO-PR), 14% (PR3-MO), 4% (Mild-R).

Clinical Relevance: Early phenotyping at ICU admission may help UK intensivists and rheumatology/nephrology teams anticipate organ support needs and prognosis, and could inform future phenotype-targeted immunosuppression trials.

Limitations: Retrospective design with unsupervised clustering and no external validation, so phenotypes are hypothesis-generating and cannot guide treatment (90-day outcomes and cluster sizes not reported in the abstract).

Study Type: Prospective observational cohort study (single-cohort, n=84).

Key Question: Do dynamic changes in left ventriculo-arterial coupling (LVAC) correlate with changes in capillary refill time (CRT) during the first 6 hours of septic shock resuscitation?

Key Findings:

• CRT change differed across LVAC trajectory groups (p=0.013): improved with uncoupled-to-coupled transition (−33.8%) and worsened with coupled-to-uncoupled transition (+17.5%); persistently coupled/uncoupled patients showed heterogeneous CRT responses.
• In uncoupled patients, CRT fell after dobutamine (−26.2%, p=0.007) but not after fluids (−9.5%, p=0.656), although both raised stroke volume by ~14%. CRT was unchanged in already-coupled patients.
• The authors conclude the LVAC-CRT relationship is weak, with peripheral perfusion often evolving independently of cardiac-arterial energy transmission.

Clinical Relevance: Supports a multimodal approach to resuscitation: improving stroke volume or LVAC alone does not reliably restore peripheral perfusion, so CRT-guided assessment (already familiar from ANDROMEDA-SHOCK) should not be replaced by macrohaemodynamic targets.

Limitations: Small observational sample with few intervention events (31 fluid, 32 inotrope), limiting causal inference and subgroup power.

Study Type: Retrospective single-centre cohort study (landmark Cox and competing-risk analyses)

Key Question: Is decannulating from V-V ECMO despite persistently impaired respiratory mechanics on sweep gas-off trial ("permissive weaning") associated with worse outcomes than conventional weaning in ARDS?

Key Findings:

• Of 280 successfully decannulated patients, 117 (42%) had permissive weaning. This group had longer ECMO duration (median 19 vs 12 days).
• 90-day post-decannulation mortality was higher with permissive weaning (adjusted HR 4.45, 95% CI 1.98-9.97).
• No difference in liberation from mechanical ventilation by day 90 (adjusted subdistribution HR 0.90, 95% CI 0.70-1.17).

Clinical Relevance: For UK ECMO centres, the findings caution against decannulating on oxygenation and CO₂ clearance alone while plateau/driving pressures, PEEP or tidal volumes remain abnormal, though they cannot establish that delaying decannulation improves outcomes.

Limitations: Retrospective, single-centre design with likely confounding by indication (permissive group had higher BMI, more COVID-19 ARDS and longer support), and only patients successfully decannulated were analysed.

Study Type: Retrospective observational cohort study (pooled data from three ICU repositories; 808 patients, 1614 echocardiograms).

Key Question: Is time-varying left ventricular systolic function (LVEF) independently associated with ICU mortality in critically ill COVID-19 patients?

Key Findings:

• At baseline, 27% of patients had LVEF ≤50% and 17% had hyperdynamic LV function (LVEF >70%). Abnormal LVEF was most common at the first examination (27%) and least common at the fifth (17%), while hyperdynamic function peaked at the fourth examination.
• After adjustment (including acute cor pulmonale, LVOT VTI and clinical covariates), neither abnormal LVEF (aHR 1.17, 95% CI 0.90–1.53) nor hyperdynamic LVEF (aHR 1.04, 95% CI 0.79–1.38) was independently associated with ICU mortality.

Clinical Relevance: LVEF phenotype alone appears to add little prognostic information in severe COVID-19 once RV function and forward flow are considered, suggesting echocardiographic assessment should focus on RV function and flow rather than LVEF.

Limitations: Retrospective design with echocardiograms performed at clinician discretion, and LVEF may be too insensitive to detect subtle myocardial injury.

Study Type: Multicentre retrospective cohort study (10 hospitals, Taiwan; March 2019–January 2020).

Key Question: Do patients with new psychoactive substance (NPS) poisoning have greater clinical severity than those with traditional illicit drug (TID) poisoning?

Key Findings:

• Of 1,378 suspected intoxications, 362 (26.3%) were LC-MS/MS-positive: 147 NPS, 130 TID, 85 mixed. Synthetic cathinones and ketamine analogues predominated among NPS.
• Overall critical outcome (ICU admission or ED death) rate was 18.3%. Adjusted OR for critical illness, NPS vs TID: 0.94 (95% CI 0.51–1.74).
• No difference in short-term survival across NPS, TID and mixed groups (log-rank p = 0.665).

Clinical Relevance: NPS poisoning carries comparable risk of critical illness to conventional drugs, supporting equivalent vigilance and escalation thresholds in UK EDs and ICUs, where NPS exposure is increasingly encountered.

Limitations: Retrospective design with a wide confidence interval (possibly underpowered), and a composite outcome that does not capture ICU-specific interventions or longer-term mortality; Taiwanese drug patterns may not generalise to the UK.