A recent Acute Medicine digest,
as sent to subscribers

Study Type: Longitudinal meta-epidemiological study

Key Question: How prevalent is outcome switching (changing primary outcomes between study registration and publication) in observational cohort studies of interventions?

Key Findings:

• Only 24% (30/124) of cohort studies completely prespecified their primary outcomes in trial registries
• Outcome switching occurred in 48% (60/124) of studies, with only 2 providing explanations for changes
• Among studies with outcome switching, 77% (44/57) favoured statistically significant results by promoting significant outcomes or demoting non-significant ones

Clinical Relevance: This highlights systematic bias in observational research that acute medicine clinicians rely on for evidence-based practice, suggesting many published cohort studies may present misleadingly positive findings.

Limitations: The study only included cohort studies registered within one month of start date, potentially missing studies with longer registration delays and limiting generalisability.

Study Type: Prospective observational cohort study

Key Question: What are the real-world rates of guideline-directed medical therapy implementation at discharge and clinical predictors of successful decongestion in patients with acute heart failure?

Key Findings:

• Among 1,373 Italian patients, 57% of those with HFrEF received all four pillars of guideline-directed therapy at discharge
• Successful decongestion was achieved in 69% of evaluable patients (469/681), meaning 31% had residual congestion at discharge
• Machine learning identified higher eGFR, younger age, lower urea/creatinine ratio, lower CRP, and smaller left atrial volumes as strongest predictors of successful decongestion (AUC 0.80)

Clinical Relevance: This provides benchmarking data for UK acute medicine departments managing heart failure, highlighting that nearly one-third of patients may be discharged with incomplete decongestion despite contemporary care.

Limitations: Single-country study from specialist cardiology units may not reflect broader acute medicine practice or outcomes in general medical wards.

Study Type: Commentary/clinical update

Key Question: How should general physicians recognize and manage Andes hantavirus (ANDV) infections, particularly given its unique human-to-human transmission capacity?

Key Findings:

• ANDV demonstrates sustained human-to-human transmission with reproductive number >2 and 9-40 day incubation period, unlike other hantaviruses
• Hantavirus cardiopulmonary syndrome carries 32-38% case fatality rate, with death occurring mean 6.7 days from symptom onset
• Disease severity correlates with neutrophilia, leukocytosis, lymphopenia, thrombocytopenia, and elevated LDH; no specific antiviral therapy exists

Clinical Relevance: The 2026 cruise ship outbreak affecting 23 nationalities demonstrates how ANDV can rapidly spread internationally, requiring UK acute physicians to maintain diagnostic awareness for severe respiratory presentations with appropriate epidemiological risk factors.

Limitations: This appears to be a clinical update rather than original research, limiting evidence quality assessment.

Study Type: Prospective multicentre cohort study with derivation and validation cohorts

Key Question: Can a clinical prediction score identify VTE patients at risk of cancer diagnosis within 12 months?

Key Findings:

• Cancer diagnosed in 3.29% (95% CI: 3.10-3.48%) of patients within 12 months post-VTE
• Six predictors identified: male sex, age ≥70 years, chronic lung disease, anaemia, platelets >350×10⁹/L, and clinically relevant bleeding during follow-up
• Model showed modest discrimination (AUC 0.60-0.63) but positive predictive value was low (3.3-4.9% in high-risk groups)

Clinical Relevance: This score could inform risk-stratified cancer surveillance strategies in UK acute medicine, potentially avoiding unnecessary investigations in lower-risk VTE patients while current guidelines recommend against routine screening.

Limitations: Modest discriminative performance limits clinical utility for identifying patients who would benefit from enhanced cancer screening.

Study Type: Retrospective cohort study

Key Question: Do anti-Ro/SSA and rheumatoid factor antibodies predict more severe interstitial lung disease in patients with systemic sclerosis?

Key Findings:

• Anti-Ro/SSA antibodies were present in 15.2% of 4,221 systemic sclerosis patients and independently predicted ILD presence (OR 1.24, 95% CI 1.07-1.43)
• Patients with both anti-Ro/SSA and RF antibodies (4.1% of cohort) had significantly worse lung function with FVC 4.12% lower and DLCO 5.4% lower than other groups
• Anti-Ro/SSA positive patients had higher rates of muscular involvement, pulmonary hypertension, and ILD at baseline

Clinical Relevance: These antibodies could improve risk stratification for ILD severity in systemic sclerosis patients presenting to acute medicine, potentially informing early specialist referral and monitoring strategies.

Limitations: Retrospective design from registry data may introduce selection bias and incomplete capture of disease progression patterns.

Study Type: Systematic review and meta-analysis

Key Question: Is hyponatremia associated with increased mortality in adults with traumatic brain injury compared to those with normal sodium levels?

Key Findings:

• Among 6,439 TBI patients across 18 studies, overall mortality in hyponatremic patients was 11% (95% CI: 7-15%)
• Hyponatremia showed no significant association with increased mortality compared to normonatremia (RR: 1.06; 95% CI: 0.80-1.39)
• Subgroup analyses by study design and timing of sodium assessment showed consistent results with no significant differences

Clinical Relevance: This challenges the assumption that hyponatremia independently predicts poor outcomes in TBI patients admitted to UK neurocritical care units, potentially influencing prognostic discussions and treatment priorities.

Limitations: Findings limited by heterogeneity and observational nature of included studies, preventing definitive causal conclusions.

Study Type: Commentary/review

Key Question: How can artificial intelligence tools be effectively implemented in cardiovascular care within internal medicine to improve real-world clinical outcomes?

Key Findings:

• Three near-term applications show promise: AI-augmented ECG interpretation, cardiovascular risk prediction for targeted prevention, and AI-enabled clinical decision support including large language models
• Evaluation should focus on practical metrics beyond accuracy: calibration, positive predictive value at action thresholds, net benefit, alert burden, and override rates
• Successful implementation requires a structured approach (TEIA framework: Train, Explain, Integrate, Accompany) with predefined confirmation pathways and post-deployment monitoring

Clinical Relevance: Provides a pragmatic framework for UK acute medicine departments considering AI implementation, emphasising integration into existing NHS workflows rather than standalone performance metrics.

Limitations: This is a review/commentary rather than original research providing empirical evidence of clinical effectiveness.

Study Type: Retrospective cohort study

Key Question: Does co-management between Neuro-ICU and intensive neurorehabilitation improve outcomes for patients with severe acquired brain injury compared to traditional sequential care?

Key Findings:

• Co-management reduced length of stay from 68 to 53 days and urgent ED transfers from 29.7% to 18.6%
• More patients progressed to rehabilitation units (59.3% vs 23.8%) with fewer requiring long-term care (13.6% vs 28.6%)
• Composite negative outcomes (ED transfers, long-term care, or death) fell from 56% to 25% (p=0.003)

Clinical Relevance: This model could inform development of integrated care pathways for complex neurological patients in NHS acute medicine, potentially reducing readmissions and improving rehabilitation outcomes.

Limitations: Single-centre retrospective design with potential temporal bias limits generalisability.