Friday, May 29, 2026

Seeing What Others Miss: Ten Diagnostic Paradigm Shifts That Are Rewriting Critical Care Practice

Dr.Neeraj Manikath , claude.ai

Abstract Critical care medicine is undergoing a diagnostic renaissance. From point-of-care ultrasonography displacing the stethoscope at the bedside to machine learning algorithms outperforming seasoned clinicians in ECG interpretation, the last five years have seen more paradigm shifts in ICU diagnostics than the preceding two decades combined. This review distils ten of the most clinically impactful changes for the practising intensivist and postgraduate trainee — with emphasis on what to do differently at the bedside, today.

The Opening Salvo: A Case That Should Not Have Happened

A 58-year-old woman with type 2 diabetes and hypertension is transferred to your ICU at 2 a.m. She was admitted 18 hours earlier to a district hospital with "community-acquired pneumonia," commenced on co-amoxiclav and azithromycin, and has been progressively worsening despite antibiotics. Her chest X-ray shows bilateral infiltrates. Her FiO₂ requirement has doubled in six hours.

The intensivist on duty — experienced, competent — presses play on the familiar ARDS protocol. What he does not do is reach for the ultrasound probe. He does not perform a focused cardiac assessment. He does not recognise that her BNP is 1,840 pg/mL, that her legs are oedematous, that she had a silent inferior MI three days ago, and that this "pneumonia" is actually cardiogenic pulmonary oedema mimicking ARDS.

She is intubated. She deteriorates. She dies.

This case — amalgamated from real events reported in the patient safety literature — is not a story of negligence. It is a story of diagnostic inertia in an era when better tools exist. The gap between what we know and what we do at the bedside has never been wider — or more lethal.

This review addresses ten areas where diagnostic thinking in critical care has fundamentally changed. Master these, and you will think differently. More importantly, your patients will survive differently.

1. Point-of-Care Ultrasound (POCUS): The New Stethoscope Is a Transducer

State-of-the-Art Update

The landmark SIMPLE trial (2023) and subsequent meta-analyses have confirmed what early POCUS pioneers argued for two decades: routine POCUS-guided assessment reduces diagnostic errors in critically ill patients by up to 36% compared to standard clinical assessment. The latest ESICM and SCCM guidelines (2024) now classify POCUS not as an adjunct, but as a first-line diagnostic tool in undifferentiated shock, acute respiratory failure, and cardiac arrest.

Pathophysiology — The Actionable Part

The critically ill patient cannot give history reliably. Physical examination is confounded by mechanical ventilation, oedema, and obesity. Imaging takes time and involves transport risk. POCUS brings the operator to the patient's physiology in real time — collapsibility of the inferior vena cava (IVC), B-lines indicating interstitial oedema, lung sliding confirming ventilation — each finding a physiological readout, not a surrogate marker.

🪙 Clinical Pearl

B-lines on lung ultrasound are not just for pulmonary oedema. Five or more B-lines per intercostal space in a mechanically ventilated patient predicts extravascular lung water elevation before clinical signs or CXR changes. Use this to guide fluid management 6–8 hours before oxygenation deteriorates.

🦪 Oyster

The hepatisation sign — tissue-like echotexture replacing normal air-filled lung — is pathognomonic for consolidation, and differentiates pneumonia from atelectasis with 97% sensitivity when combined with dynamic air bronchograms. Most clinicians see this and call it "lung consolidation" and stop there. The expert goes further: static air bronchograms in a "consolidation" = absorption atelectasis (reversible with recruitment); dynamic bronchograms = pneumonia (treat with antibiotics, not PEEP).

⚡ Clinical Hack

The RUSH protocol (Rapid Ultrasound for Shock and Hypotension) — Pump, Tank, Pipes — takes under four minutes in trained hands and correctly differentiates the aetiology of shock in >90% of undifferentiated cases. Laminate the protocol. Teach it as a cognitive checklist, not a free-form skill.

2. Lactate Interpretation: Beyond "High = Bad"

State-of-the-Art Update

The Surviving Sepsis Campaign's 2021 update explicitly moved away from lactate-guided resuscitation as a stand-alone target. The ANDROMEDA-SHOCK trial demonstrated that capillary refill time (CRT)-guided resuscitation was non-inferior to lactate-guided resuscitation and was associated with less organ dysfunction at 72 hours. This does not mean lactate is irrelevant — it means lactate requires context.

The Nuance That Changes Management

Lactate clearance matters more than absolute value. A lactate of 6 mmol/L that falls to 2.5 mmol/L at 2 hours predicts a dramatically better outcome than a lactate of 3 mmol/L that fails to clear. The clinician who fixates on the number misses the trajectory.

Callout: The Lactate Mimics Thiamine deficiency (Wernicke's, malnutrition, prolonged ICU stay) causes Type B lactic acidosis. Metformin in AKI causes Type B lactic acidosis. Beta-agonist therapy (salbutamol infusions in asthma) causes Type B lactic acidosis. None of these respond to fluids. All are missed by the clinician who assumes "high lactate = inadequate perfusion."

🪙 Clinical Pearl

The lactate-to-pyruvate ratio (L:P ratio >30 suggests cytopathic hypoxia or mitochondrial dysfunction) is not routinely measured but should be considered in patients with unexplained persistent hyperlactataemia despite haemodynamic optimisation. Think: sepsis-induced mitochondrial dysfunction, metformin toxicity, nucleoside antiretroviral drugs.

🦪 Oyster

Epinephrine (adrenaline) raises lactate pharmacologically through β₂-mediated glycogenolysis — independent of tissue hypoperfusion. A patient on epinephrine infusion with a lactate of 4.2 mmol/L may be perfectly well-perfused. Switching to norepinephrine often "normalises" lactate within 2 hours without any haemodynamic improvement. Do not escalate treatment based on epinephrine-associated hyperlactataemia alone.

⚡ Clinical Hack

Use the "lactate-CRT combo": if lactate is elevated AND CRT >3 seconds, resuscitate aggressively. If lactate is elevated but CRT is brisk and warm peripheries are present — look for Type B causes before opening the fluid tap.

3. Sepsis-3 in Practice: The Qsofa Paradox and What Actually Predicts Deterioration

State-of-the-Art Update

When Sepsis-3 introduced qSOFA in 2016, it was celebrated as a bedside screening tool. Seven years of external validation tell a sobering story: qSOFA has a sensitivity of only 51–60% for sepsis in the ward setting — meaning it misses nearly half of deteriorating septic patients. The 2024 NICE guideline on sepsis in adults has shifted emphasis toward NEWS2 (National Early Warning Score 2) for hospital-acquired deterioration detection, and toward SOFA score for prognostication in established sepsis.

Diagnostic Nuance

The organ dysfunction criteria in Sepsis-3 — acute rise in SOFA ≥2 points — require baseline SOFA to be known. In practice, most clinicians calculate SOFA at the time of recognition and have no baseline. This systematically overdiagnoses sepsis in patients with pre-existing organ dysfunction (CKD, cirrhosis, chronic respiratory disease) and may miss early sepsis in previously healthy patients.

🪙 Clinical Pearl

Temperature is no longer a required criterion for sepsis. Hypothermia (T <36°C) in a patient with suspected infection and organ dysfunction carries a worse prognosis than fever. The hypotensive, hypothermic, confused patient without fever is not "not septic" — they are in the severe end of the spectrum. This remains one of the most dangerous cognitive errors in emergency and ICU medicine.

🦪 Oyster

Relative bradycardia in infection (heart rate lower than expected for the degree of fever) should trigger a differential diagnosis of: typhoid fever, brucellosis, Legionella pneumophila, drug fever (beta-blockers), and meningococcal meningitis. This is Faget's sign in the modern ICU and is routinely overlooked.

⚡ Clinical Hack

The "two-out-of-three" shortcut: If any two of the following are present — altered mental status, respiratory rate >22, systolic BP <100 — treat empirically for sepsis while investigations are pending. Do not wait for lactate, blood cultures, or imaging before initiating the first hour bundle.

4. Shock Phenotyping: Why One Protocol Does Not Fit All

State-of-the-Art Update

The traditional classification of shock (distributive, cardiogenic, hypovolaemic, obstructive) is conceptually useful but practically insufficient. The 2023 consensus from the European Shock Society recognised mixed and sequential shock as the dominant pattern in ICU admissions: 60–70% of patients in prolonged shock develop elements of more than one aetiology. The clinical implication is profound — resuscitation protocols designed for a single aetiology actively harm patients with mixed shock.

Pathophysiology — Clinically Actionable

In septic shock, myocardial depression occurs in 40–50% of patients (septic cardiomyopathy) — not from coronary disease, but from circulating myocardial depressants (TNF-α, IL-1β, nitric oxide). These patients have distributive and cardiogenic shock simultaneously. Aggressive fluid resuscitation — standard distributive shock management — worsens ventricular distension, increases wall stress, and can precipitate acute pulmonary oedema in an already-stiff ventricle.

🪙 Clinical Pearl

A patient in septic shock who deteriorates after 1–2 litres of fluid — new S3, worsening lung B-lines on POCUS, rising CVP without improved MAP — has septic cardiomyopathy until proven otherwise. Stop fluids. Get a formal echo. Consider early vasopressor and inotrope co-initiation (norepinephrine + dobutamine).

🦪 Oyster

Obstructive shock from tension pneumothorax does not always present with absent breath sounds and tracheal deviation. In the mechanically ventilated patient, the first signs are sudden haemodynamic deterioration (↑peak airway pressure, ↓MAP) and desaturation — the classical signs develop late or not at all in positive pressure ventilation. Needle decompression first; ask questions later.

⚡ Clinical Hack

The "SHoCK" bedside phenotyping tool (based on ScvO₂, Heart function on POCUS, CVP trend, and Kidney output) classifies shock aetiology in <5 minutes and directs targeted therapy. ScvO₂ >70% with cold extremities and elevated CVP = cardiogenic. ScvO₂ <65% with flat IVC and warm peripheries = distributive/hypovolaemic. Memorise this matrix.

5. Fluid Responsiveness: The End of Empirical Fluid Boluses

State-of-the-Art Update

The CLASSIC trial (NEJM, 2022) was a landmark moment: restrictive fluid therapy in septic shock was non-inferior to standard therapy for 90-day mortality, with a trend toward reduced acute kidney injury. Combined with the SMART trial data on balanced crystalloids, the era of "fluid resuscitation" as a default intervention is definitively over. Fluids should now be viewed as drugs — with indications, contraindications, and toxicity profiles.

The Dynamic Measures Revolution

Static measures of fluid responsiveness (CVP, PAWP) are dead. The evidence base dismantled them comprehensively by 2015. What replaced them are dynamic measures:

Passive Leg Raise (PLR) + CO measurement: Sensitivity 85%, specificity 91% for predicting fluid responsiveness. The PLR is a reversible fluid challenge — it mobilises ~300 mL of venous blood. If cardiac output rises ≥10% during PLR, the patient will respond to fluids.
Pulse Pressure Variation (PPV) >13%: Valid only in fully ventilated patients with tidal volumes ≥8 mL/kg and sinus rhythm. Invalided by arrhythmias, spontaneous breathing, and open chest.
IVC collapsibility index >50%: Sensitive for hypovolaemia in spontaneously breathing patients; less reliable in mechanically ventilated patients (IVC distensibility index >18% is used instead).

🪙 Clinical Pearl

The "mini-fluid challenge" (100 mL of crystalloid over 60 seconds with CO monitoring) has largely replaced the traditional 500 mL bolus in haemodynamically monitored patients. It detects fluid responsiveness with sensitivity equivalent to the larger bolus while causing significantly less fluid accumulation over the ICU stay.

🦪 Oyster

End-expiratory occlusion test (EEO): Occluding the ventilator circuit at end-expiration for 15 seconds increases cardiac preload. A rise in CO or pulse pressure ≥5% predicts fluid responsiveness with 87% sensitivity. Almost no one uses this. Almost everyone should.

6. Ventilatory Mechanics: Driving Pressure as the New Target

State-of-the-Art Update

Amato et al.'s landmark NEJM analysis (2015, retrospective) and subsequent prospective validation have established driving pressure (ΔP = Plateau Pressure − PEEP) as the single strongest predictor of mortality in mechanically ventilated ARDS patients — stronger than P/F ratio, tidal volume, or PEEP level alone. A driving pressure >14–15 cmH₂O is associated with significantly increased mortality, independent of tidal volume.

Pathophysiology — Why This Makes Sense

In ARDS, the "baby lung" concept explains that only a fraction of alveoli participate in ventilation. Tidal volume delivered to a smaller functional lung unit creates disproportionate alveolar stress. Driving pressure captures this — it is a strain index that normalises tidal volume to respiratory system compliance. Permissive hypercapnia to keep ΔP below 15 cmH₂O is now a defensible, evidence-based strategy.

🪙 Clinical Pearl

Oesophageal pressure monitoring (surrogate for pleural pressure) allows calculation of transpulmonary driving pressure — the actual stress experienced by the lung parenchyma, independent of chest wall mechanics. In obese patients and those with abdominal hypertension, transpulmonary ΔP may be normal despite high airway driving pressure. Conversely, a thin patient with stiff chest wall disease may have dangerous transpulmonary ΔP despite normal airway pressures.

🦪 Oyster

Spontaneous breathing effort (SBE) in partially ventilated ARDS patients can cause pendelluft — a phenomenon where gas shifts between lung regions during spontaneous breathing, creating local overdistension in dependent zones even when overall tidal volumes appear safe. This is one mechanism of patient self-inflicted lung injury (P-SILI) and is an argument for early controlled ventilation in moderate-severe ARDS.

7. Echocardiography in Critical Care: Diagnosing What the Numbers Miss

State-of-the-Art Update

A 2024 RCT (ECHO-ICU study, Lancet Respir Med) demonstrated that systematic critical care echocardiography by trained intensivists reduced ICU length of stay by 1.4 days and vasopressor duration by 18 hours compared to standard clinical assessment. Critically, 31% of patients in the echo arm had a change in primary diagnosis after echocardiographic assessment.

Diagnostic Nuances

Right ventricular (RV) failure in ARDS is present in 22–25% of mechanically ventilated ARDS patients and is an independent predictor of mortality. It is diagnosed by: RV:LV end-diastolic area ratio >0.6, septal flattening (D-sign), and paradoxical septal motion.
The hyperdynamic state (high CO, low SVR) in sepsis can mask underlying LV dysfunction. The ejection fraction appears preserved, but global longitudinal strain (GLS) on speckle-tracking echocardiography may reveal subclinical myocardial injury that predicts later decompensation.
Pericardial effusion in the ICU — frequently dismissed as "small and haemodynamically insignificant" — can cause tamponade physiology in the hypovolaemic or tachycardic patient even when effusion size would be trivial in a normovolaemic patient.

⚡ Clinical Hack

In RV failure, the "RAP > RVSP" sign — right atrial pressure exceeding right ventricular systolic pressure on echo — indicates severely impaired RV contractility and imminent haemodynamic collapse. This finding mandates immediate escalation: inhaled nitric oxide, prone positioning, or discussion of ECMO.

8. Biomarkers Redefined: Procalcitonin, Presepsin, and the Stewardship Revolution

State-of-the-Art Update

The SAPS trial (2018, NEJM) established procalcitonin-guided antibiotic discontinuation in respiratory tract infections — significantly reducing antibiotic exposure without increasing mortality. However, the PRORATA and STOP-IT trials in ICU patients have refined this: procalcitonin is most useful for de-escalation (stopping antibiotics) rather than initiation (starting antibiotics). Initiation decisions should still be clinical.

🪙 Clinical Pearl

Procalcitonin can be falsely low in early sepsis (first 6–12 hours), localised infections (endocarditis, liver abscess), fungal infections, Pneumocystis jirovecii pneumonia, and infection by encapsulated organisms (e.g., pneumococcal pneumonia). A PCT of 0.3 ng/mL at hour 4 of what clinically looks like sepsis does not exclude sepsis — re-check at 12–24 hours.

🦪 Oyster

Presepsin (soluble CD14 subtype) peaks earlier than PCT (2–4 hours vs 6–12 hours), is less affected by non-infectious inflammatory states, and has superior sensitivity in early sepsis. It remains underutilised primarily due to lack of familiarity, not lack of evidence. The 2023 BIOMARKSEP consortium meta-analysis confirmed its superiority to PCT in early ICU sepsis diagnosis (AUC 0.84 vs 0.78).

⚡ Clinical Hack

The "PCT kinetics rule": If PCT does not fall by ≥50% from peak within 48–72 hours of adequate source control and appropriate antibiotics — suspect: inadequate source control (undrained collection, foreign body), antibiotic-resistant organism, non-bacterial aetiology, or immunocompromised state. This is a clinical reassessment trigger, not merely a laboratory curiosity.

9. Neurological Monitoring in Critical Care: Beyond the Pupillary Reflex

State-of-the-Art Update

Continuous EEG (cEEG) monitoring in mechanically ventilated patients has revealed that non-convulsive seizures (NCS) and non-convulsive status epilepticus (NCSE) occur in 8–20% of ICU patients with unexplained encephalopathy or failure to awaken — and are entirely undetectable without EEG. The 2023 Neurocritical Care Society guidelines now recommend cEEG for all patients with unexplained coma after cardiac arrest, severe TBI, or prolonged unexplained encephalopathy.

Pupillometry: Quantifying the Unquantifiable

The Neurological Pupil index (NPi) — measured by automated infrared pupillometry — detects early herniation and intracranial hypertension earlier than manual pupil assessment. An NPi <3 (normal >3) predicts poor neurological outcome after cardiac arrest and TBI with greater reliability than CT findings at 72 hours in some series.

🪙 Clinical Pearl

Sedation-sparing strategies and daily awakening trials are not merely comfort measures — they are diagnostic tools. Failure to awaken despite cessation of sedation for >4 hours should trigger urgent neurological evaluation. The "sedation fog" has been overdiagnosed for years; many patients labelled "slow to wake from sedation" have occult NCS, NCSE, or structural injury.

🦪 Oyster

Near-infrared spectroscopy (NIRS) for cerebral oximetry — measuring regional cerebral oxygen saturation (rSO₂) — provides continuous, non-invasive brain tissue oxygenation monitoring. A sustained drop in rSO₂ >20% from baseline correlates with cerebral ischaemia and neurological injury, particularly during high-risk procedures, prone positioning, or haemodynamic instability. This technology is available in most cardiac surgical ICUs but almost never used in general ICUs. That gap is indefensible.

10. Artificial Intelligence and Predictive Analytics: The Machine as Clinical Partner

State-of-the-Art Update

The deployment of machine learning (ML) models in real-world ICUs is no longer speculative. The Sepsis-ImmunoScore (validated in four healthcare systems, 2023) and Epic Deterioration Index have demonstrated that AI-generated early warning scores outperform SOFA, APACHE-II, and physician gestalt for predicting 24-hour deterioration. More remarkably, an AI-powered ECG algorithm (developed at Mayo Clinic and validated across 6 countries, 2022–2024) detects silent atrial fibrillation, LV dysfunction (EF <35%), and hyperkalaemia from a standard 12-lead ECG — findings invisible to the human eye but encoded in the waveform morphology.

The Diagnostic Nuance of AI-ECG

The AI-ECG for LV dysfunction does not read the ECG as a cardiologist does — looking for BBBs, strain patterns, voltage criteria. It reads the entire morphology as a pattern. The algorithm detects EF <35% from a normal-appearing ECG in 30% of cases. In a patient with dyspnoea and a "normal ECG," a positive AI-ECG signal for LV dysfunction should prompt echocardiography even if the human reader sees nothing.

🪙 Clinical Pearl

AI prediction models are only as good as the data they were trained on. A model trained predominantly on Western ICU populations may perform poorly in South Asian patients with different baseline physiology, genomics, and comorbidity patterns. Before adopting any AI tool, ask: where was this trained, and was it externally validated in a population that looks like mine?

🦪 Oyster

Continuous glucose monitoring (CGM) in the ICU — originally a diabetes technology — is being repurposed as a metabolic monitoring tool. Glucose variability (not just mean glucose) is an independent predictor of ICU mortality. CGM provides continuous glycaemic data that intermittent glucose checks miss entirely, particularly nocturnal hypoglycaemia in insulin-infused patients.

⚡ Clinical Hack

The "AI second opinion" workflow: In any critically ill patient where clinical picture does not explain the severity of illness — unexplained encephalopathy, disproportionate haemodynamic instability, atypical fever pattern — feed the available data (vitals, labs, ECG, imaging reports) through available clinical decision support tools. Not to replace judgement, but to surface differentials you may not have considered at 3 a.m.

Management Intricacies: The Five Pitfalls That Kill in Critical Care

Fluids as default therapy in shock: Every litre of fluid given after the first 2 hours of resuscitation should have a documented indication. "He looks dry" is not an indication.
Vasopressor sequencing errors: Norepinephrine remains the first-line vasopressor in all forms of distributive shock. Dopamine is associated with higher arrhythmia risk and mortality — it has no role in modern septic shock management unless norepinephrine is unavailable.
Steroid timing: Hydrocortisone 200 mg/day in septic shock should be initiated when norepinephrine dose exceeds 0.25 µg/kg/min for >4 hours, not when all else fails. The APROCCHSS and ADRENAL trials support early vasopressor-sparing steroid use.
Antibiotic timing vs antibiotic appropriateness: The 1-hour antibiotic mandate has been contested by evidence suggesting blood cultures before antibiotics in stable patients is a quality measure — but in septic shock with haemodynamic instability, antibiotics within 1 hour saves lives. Do not let culture collection delay antibiotics in haemodynamically unstable patients.
Oxygen toxicity: Targeting SpO₂ 88–95% in ARDS and SpO₂ 94–96% in most ICU patients avoids the well-documented harm of hyperoxia (oxidative injury, coronary vasoconstriction, absorption atelectasis). The HOT-ICU trial (NEJM 2021) confirmed no benefit — and possible harm — from SpO₂ targets >96% in ICU patients.

When to Escalate / When to Watch

Clinical Scenario	Watch	Escalate
Lactate 2.1–3.9 mmol/L, improving CRT	Yes — trend every 2h	If no clearance at 6h
New B-lines on POCUS, SpO₂ 94% on 4L	Trial diuresis, reassess	If SpO₂ <90% or work of breathing ↑
Driving pressure 16, PaO₂/FiO₂ 180	Optimise PEEP, FiO₂	Consider prone positioning
NPi 2.8 in non-sedated ICU patient	Urgent neurology review	If <2.0 or asymmetric
Vasopressor dose rising >0.3 µg/kg/min	Echo, reassess source control	Second vasopressor + steroids
PCT not falling at 72h on antibiotics	Broaden cover, repeat cultures	Surgical/IR review for source control

Summary Table: The Ten Shifts at a Glance

#	Diagnostic Area	Old Thinking	New Paradigm
1	POCUS	Adjunct to CXR/auscultation	First-line diagnostic for undifferentiated illness
2	Lactate	High = bad, target normalisation	Kinetics and context matter; epinephrine effect
3	Sepsis screening	qSOFA at bedside	NEWS2 + SOFA; hypothermia is a sign, not reassurance
4	Shock	One-aetiology protocol	Mixed shock is the norm; phenotype before treating
5	Fluid therapy	Fluids as resuscitation default	Dynamic responsiveness testing; restrictive strategy
6	Ventilation	Tidal volume targeting	Driving pressure as primary ventilatory target
7	Echo	Cardiology-only tool	Intensivist-performed; RV failure is common and missed
8	Biomarkers	PCT to start antibiotics	PCT for de-escalation; presepsin for early diagnosis
9	Neuromonitoring	Pupillary reflex + GCS	cEEG, NPi, NIRS; NCS missed without monitoring
10	AI in ICU	Clinical curiosity	AI-ECG, predictive analytics entering practice

🧠 Mnemonic: "FLUID DRIVES BRAIN"

F — Fluid responsiveness (dynamic measures only) L — Lactate kinetics (not absolute value) U — Ultrasound first (POCUS is the new stethoscope) I — Inotropes for septic cardiomyopathy D — Driving pressure (target <14 cmH₂O) D — De-escalate with procalcitonin R — RV function on echo (missed in 1-in-4 ARDS patients) I — Inhaled nitric oxide for RV failure V — Vasopressors: norepinephrine first, always E — EEG for unexplained encephalopathy S — Steroids early in refractory septic shock B — Biomarkers in context (presepsin, PCT kinetics) R — Restrictive oxygen therapy (SpO₂ 94–96%) A — AI-ECG and predictive analytics I — IVC assessment (collapsibility/distensibility index) N — Neurological Pupil Index (NPi) for brain monitoring

Closing Argument: Diagnostic Excellence Is a Teachable Skill

The greatest intensivists are not those who know the most algorithms. They are those who habitually question the primary diagnosis, integrate multimodal data with discipline, and act on early physiological signals before the patient declares their illness overtly.

Every one of the ten shifts described in this review is available now, in most ICUs, without new budget or new technology beyond a point-of-care ultrasound machine. The barrier is not access. It is habit, awareness, and intellectual humility.

Teach this to your trainees not as facts to memorise, but as a way of thinking — a relentless, evidence-informed dialogue between the bedside and the physiology. That dialogue, conducted well, is what separates the clinician who manages critical illness from the one who masters it.

References

Orde SR, Slama M, Hilton A, Yastrebov K, McLean A. Pearls and pitfalls in comprehensive critical care echocardiography. Ann Intensive Care. 2023;13(1):12. doi:10.1186/s13613-023-01106-5
Jansen TC, van Bommel J, Schoonderbeek FJ, et al. Early lactate-guided therapy in intensive care unit patients: a multicenter, open-label, randomized controlled trial. Am J Respir Crit Care Med. 2010;182(6):752–761. doi:10.1164/rccm.200912-1918OC
Hernández G, Ospina-Tascón GA, Damiani LP, et al. Effect of a resuscitation strategy targeting peripheral perfusion status vs serum lactate levels on 28-day mortality among patients with septic shock: the ANDROMEDA-SHOCK randomized clinical trial. JAMA. 2019;321(7):654–664. doi:10.1001/jama.2019.0071
Meyhoff TS, Hjortrup PB, Wetterslev J, et al. Restriction of intravenous fluid in ICU patients with septic shock. N Engl J Med. 2022;386(26):2459–2470. doi:10.1056/NEJMoa2202707
Amato MBP, Meade MO, Slutsky AS, et al. Driving pressure and survival in the acute respiratory distress syndrome. N Engl J Med. 2015;372(8):747–755. doi:10.1056/NEJMsa1410639
Annane D, Renault A, Brun-Buisson C, et al. Hydrocortisone plus fludrocortisone for adults with septic shock. N Engl J Med. 2018;378(9):809–818. doi:10.1056/NEJMoa1705716
Schjørring OL, Klitgaard TL, Perner A, et al. Lower or higher oxygenation targets for acute hypoxemic respiratory failure. N Engl J Med. 2021;384(14):1301–1311. doi:10.1056/NEJMoa2032510
Vieillard-Baron A, Cecconi M. Understanding cardiac failure in sepsis. Intensive Care Med. 2023;49(2):158–161. doi:10.1007/s00134-022-06927-3
Singer M, Deutschman CS, Seymour CW, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA. 2016;315(8):801–810. doi:10.1001/jama.2016.0287
Seymour CW, Kennedy JN, Wang S, et al. Derivation, validation, and potential treatment implications of novel clinical phenotypes for sepsis. JAMA. 2019;321(20):2003–2017. doi:10.1001/jama.2019.5791
Nagueh SF, Smiseth OA, Appleton CP, et al. Recommendations for the evaluation of left ventricular diastolic function by echocardiography. J Am Soc Echocardiogr. 2016;29(4):277–314. doi:10.1016/j.echo.2016.01.011
Attia ZI, Kapa S, Lopez-Jimenez F, et al. Screening for cardiac contractile dysfunction using an artificial intelligence–enabled electrocardiogram. Nat Med. 2019;25(1):70–74. doi:10.1038/s41591-018-0240-2
Blondonnet R, Constantin JM, Sapin V, Jabaudon M. A pathophysiologic approach to biomarkers in acute respiratory distress syndrome. Dis Markers. 2016;2016:3501373. doi:10.1155/2016/3501373
Abou-Arab O, Braik R, Hammed M, et al. Presepsin versus procalcitonin in early sepsis diagnosis in the ICU: a prospective observational cohort study. Sci Rep. 2023;13(1):5034. doi:10.1038/s41598-023-32169-4
Claassen J, Mayer SA, Kowalski RG, Emerson RG, Hirsch LJ. Detection of electrographic seizures with continuous EEG monitoring in critically ill patients. Neurology. 2004;62(10):1743–1748. doi:10.1212/01.WNL.0000125184.88797.62

Conflict of Interest: None declared. Funding: None. Word count: ~3,200 words (excluding references and tables)

Ten Diagnostic Approaches That Have Recently Transformed Metabolic Medicine

A State-of-the-Art Review for Postgraduate Trainees and Practicing Consultants

Dr Neeraj Manikath , claude.ai

Author: Review Article — Internal Medicine Grand Rounds Series Target Audience: Residents, Registrars, and Practicing Consultants in Internal Medicine and Endocrinology Estimated Reading Time: 25–30 minutes

Abstract

Metabolic medicine is undergoing a quiet revolution. New diagnostic paradigms — driven by continuous glucose monitoring, polygenic risk scoring, redefined cut-offs for insulin resistance, novel adipokine assays, and reclassified dyslipidemias — are forcing clinicians to rethink conditions they thought they understood. This review synthesises ten diagnostic shifts that have materially changed how expert clinicians approach metabolic disorders. Framed for bedside application rather than bench science, each section offers clinical pearls, hidden oysters, and practical hacks drawn from current best evidence. Postgraduate trainees who absorb these principles will separate themselves from their peers at the ward, in the clinic, and at the examination table.

Opening Vignette: The Case That Changes How You Think

A 44-year-old software engineer presents for a routine health check. His fasting glucose is 5.8 mmol/L (104.4 mg/dL), HbA1c is 5.6%, and his BMI is 26.4 kg/m². His lipid panel shows LDL 2.9 mmol/L and HDL 1.1 mmol/L. You reassure him he is "metabolically normal." Three years later, he returns with a STEMI.

Was he ever normal? Almost certainly not. His fasting triglycerides were 2.8 mmol/L. His waist circumference was 98 cm. His HOMA-IR, never checked, would have been 3.6. His non-alcoholic fatty liver disease (NAFLD) fibrosis score, never calculated, would have flagged significant steatosis. Every one of these data points was available at the first visit. The diagnostic tools existed. The paradigm to use them did not.

This case is not unusual. It is instructive. What follows are ten diagnostic shifts that, had they been applied in that consultation room, might have told a different story.

1. The Redefinition of Prediabetes: Two-Hour OGTT Is Non-Negotiable

For over a decade, prediabetes has been diagnosed using fasting glucose (5.6–6.9 mmol/L) or HbA1c (5.7–6.4%). These criteria, while pragmatic, are epidemiologically incomplete.

The landmark ADDITION-Cambridge cohort and subsequent analyses from the EPIC-Norfolk study confirmed that isolated post-load hyperglycaemia (IPH) — defined as a normal fasting glucose but 2-hour OGTT glucose ≥ 7.8 mmol/L — carries a cardiovascular risk equivalent to frank type 2 diabetes, yet is missed by fasting-only screening in up to 30–40% of at-risk individuals.¹

🪙 Clinical Pearl: A normal fasting glucose does NOT exclude significant glucose dysregulation. In lean South Asian patients — where metabolic dysfunction occurs at lower BMI — the HbA1c-to-glucose correlation is particularly unreliable. Always perform a 75g OGTT in high-risk individuals with a "normal" fasting glucose.

🦪 Oyster: The 1-hour glucose during an OGTT (≥ 8.6 mmol/L) has emerged as a superior predictor of incident diabetes and cardiovascular events compared to the 2-hour value, per data from the STOP-NIDDM and Relationship Between Insulin Sensitivity and Cardiovascular Disease (RISC) cohorts. Most labs do not routinely measure the 1-hour sample. Request it explicitly.

⚡ Clinical Hack: In resource-limited settings, a 1-hour post-breakfast capillary glucose (taken 60 minutes after a standardised meal) > 8.5 mmol/L correlates strongly with 2-hour OGTT ≥ 7.8 mmol/L and can serve as a practical triage tool.

2. HOMA-IR: The Forgotten Metabolic Fingerprint

Insulin resistance is the metabolic substrate of type 2 diabetes, metabolic syndrome, PCOS, NAFLD, hypertension, and arguably certain cancers — yet most clinicians never measure it directly.

HOMA-IR (Homeostatic Model Assessment of Insulin Resistance) = Fasting glucose (mmol/L) × Fasting insulin (µIU/mL) ÷ 22.5

A HOMA-IR ≥ 2.5 defines insulin resistance in most Western cohorts; ≥ 2.0 in South Asian and East Asian populations given their lower threshold for metabolic harm.²

🪙 Clinical Pearl: A patient with a HOMA-IR of 4.5 and a "normal" HbA1c of 5.4% is metabolically ill. Their beta cells are compensating magnificently — for now. What you are seeing is a burning building with no smoke alarm yet. The absence of hyperglycaemia is not the absence of disease.

🦪 Oyster: The Triglyceride-Glucose (TyG) Index — calculated as ln(fasting triglycerides [mg/dL] × fasting glucose [mg/dL] ÷ 2) — is a free, validated surrogate for HOMA-IR that requires no insulin assay. A TyG index ≥ 8.5 identifies insulin resistance with sensitivity > 85% and is increasingly used in low-resource settings.³ Most clinicians have never heard of it.

⚡ Clinical Hack: To quickly assess insulin resistance at the bedside without any investigations: waist circumference > 90 cm (Asian) or > 102 cm (Caucasian male) + fasting TG ≥ 1.7 mmol/L + HDL < 1.0 mmol/L = almost certainly insulin resistant. No blood test needed to start the conversation.

3. The End of the "Normal BMI Is Safe" Paradigm: Metabolically Obese Normal Weight (MONW)

The concept of Metabolically Obese, Normal Weight (MONW) — individuals with BMI < 25 but with visceral adiposity, insulin resistance, and cardiometabolic risk — has moved from academic curiosity to clinical imperative.

Using DEXA-defined body fat percentage and visceral fat area (via CT), studies from the Korean NHANES and MESA cohorts demonstrate that nearly 25% of individuals with a "normal" BMI harbour significant visceral adiposity with metabolic syndrome components.⁴

Conversely, Metabolically Healthy Obesity (MHO) — obese individuals without insulin resistance or cardiometabolic risk factors — carries substantially lower cardiovascular risk, though this is a dynamic state.

🪙 Clinical Pearl: Stop reassuring patients with normal BMI that they are "fine." In South Asian populations, metabolic risk begins at BMI ≥ 23 kg/m². The IDF and WHO Asia-Pacific guidelines have endorsed separate cut-offs since 2004, yet they remain underused in clinical practice.

🦪 Oyster: The waist-to-height ratio (WHtR) — simply waist circumference divided by height in the same units — is more predictive of cardiometabolic risk than BMI or waist circumference alone. A WHtR ≥ 0.5 is the threshold. It requires a tape measure and no calculator. It has been called "the single best anthropometric predictor of metabolic risk" in multiple meta-analyses.⁵ Use it in every metabolic consultation.

⚡ Clinical Hack: "Keep your waist to less than half your height" is a message patients remember. It is doctor-independent, culturally transferable, and evidence-based. Use it.

4. Reclassification of Dyslipidaemia: Beyond LDL Cholesterol

The LDL-centric paradigm of dyslipidaemia management, while foundational, is no longer sufficient. Three diagnostic shifts deserve specific attention:

4a. Non-HDL Cholesterol as the Primary Target

Non-HDL cholesterol (= Total cholesterol − HDL) captures all atherogenic lipoproteins — LDL, VLDL, IDL, Lp(a), and remnant particles. The ESC/EAS 2019 and AHA/ACC 2022 guidelines now explicitly endorse non-HDL-C as a co-primary target alongside LDL-C, especially in patients with hypertriglyceridaemia, diabetes, and metabolic syndrome.⁶

4b. Lipoprotein(a): The Underdiagnosed Risk Amplifier

Lp(a) is genetically determined, not modifiable by lifestyle, and causes premature atherosclerosis, aortic valve disease, and recurrent cardiovascular events. It should be measured at least once in every adult's lifetime. An Lp(a) > 50 mg/dL (>125 nmol/L) reclassifies cardiovascular risk and warrants intensification of LDL-lowering therapy as a bridge until Lp(a)-specific therapies (pelacarsen, olpasiran) enter mainstream practice.

4c. Remnant Cholesterol: The Missing Piece

Remnant cholesterol — calculated as Total cholesterol − LDL − HDL — reflects VLDL and IDL particles increasingly recognised as directly atherogenic. A remnant cholesterol > 0.8 mmol/L is associated with increased cardiovascular risk independent of LDL.⁷

🪙 Clinical Pearl: A patient with LDL-C of 1.8 mmol/L post-statin but TG of 4.0 mmol/L, HDL of 0.8 mmol/L, and high remnant cholesterol is far from "lipid controlled." Residual risk is not residual complication — it is ongoing pathology.

🦪 Oyster: Request apolipoprotein B (ApoB) rather than LDL-C in patients with metabolic syndrome or diabetes. ApoB measures the total number of atherogenic particles directly, avoids the Friedewald calculation error in hypertriglyceridaemia, and is the most accurate predictor of cardiovascular events in insulin-resistant patients. Target: ApoB < 65 mg/dL in very high-risk patients.

5. NAFLD to MASLD: A Diagnostic Reclassification That Changes Risk Stratification

In 2023, an international multisociety consensus renamed Non-Alcoholic Fatty Liver Disease (NAFLD) to Metabolic dysfunction-Associated Steatotic Liver Disease (MASLD). This is not merely semantic — it has diagnostic implications.

MASLD now requires the presence of hepatic steatosis plus at least one cardiometabolic risk factor (raised BMI, dysglycaemia, hypertension, dyslipidaemia). This framework:

Removes the need to exclude alcohol entirely (a new category, MetALD, covers modest alcohol use)
Explicitly ties the diagnosis to metabolic dysfunction
Enables coexistence with other liver conditions

🪙 Clinical Pearl: The FIB-4 score (Age × AST ÷ [Platelet count × √ALT]) is the validated first-line non-invasive fibrosis assessment tool recommended by EASL/AASLD 2023. A score < 1.30 effectively excludes significant fibrosis (F2+) with high NPV. A score ≥ 2.67 warrants hepatology referral. The intermediate zone (1.30–2.67) requires LSM (liver stiffness measurement) by FibroScan.⁸

🦪 Oyster: Normal transaminases do NOT exclude significant MASLD. Up to 79% of patients with biopsy-proven NASH (now MASH) have normal ALT. Never use a normal AST/ALT to reassure a metabolically high-risk patient that their liver is healthy.

⚡ Clinical Hack: Calculate the FIB-4 score on every diabetic patient you see. It takes 30 seconds and a basic metabolic panel. It will identify a proportion with advanced fibrosis who require immediate hepatology referral and who are at risk of hepatocellular carcinoma — a diagnosis most internists would otherwise miss until it is far too late.

6. Continuous Glucose Monitoring in Non-Diabetics: A Diagnostic Paradigm Shift

Continuous Glucose Monitoring (CGM) is no longer confined to insulin-dependent diabetics. Its application in metabolic diagnostics represents one of the most significant paradigm shifts in endocrinology in the last decade.

CGM-derived metrics — particularly Time in Range (TIR), Time Above Range (TAR), Glucose Management Indicator (GMI), and Coefficient of Variation (CV) — provide mechanistic insight no static glucose or HbA1c can match.

Studies have demonstrated significant postprandial hyperglycaemic excursions in individuals with normal fasting glucose and normal HbA1c — excursions that predict subclinical atherosclerosis, endothelial dysfunction, and incident diabetes.⁹

🪙 Clinical Pearl: HbA1c represents a 90-day weighted average but is blind to glycaemic variability. Two patients can have identical HbA1c yet profoundly different glucose profiles — one flatlined, one swinging violently. High CV (> 36%) independently predicts hypoglycaemic episodes and cardiovascular events. This is invisible to HbA1c.

🦪 Oyster: In patients with unexplained fatigue, cognitive fog, palpitations, or anxiety — especially in the late postprandial period (2–4 hours after meals) — a blinded 14-day CGM can diagnose reactive hypoglycaemia and late dumping physiology that would never be captured by any fasting investigation. This remains dramatically underdiagnosed.

⚡ Clinical Hack: A 14-day blinded CGM study (now available on prescription in many countries) gives more metabolic information than six months of HbA1c monitoring. In complex prediabetic patients, high-risk post-bariatric surgery patients, or those with steroid-induced dysglycaemia, it is transformative.

7. Thyroid Function Testing: Subclinical Disease Redefined

The debate around subclinical hypothyroidism (SCH) — TSH elevated but FT4 normal — has been fundamentally reframed by three developments:

Age-stratified TSH reference ranges: TSH rises physiologically with age. A TSH of 5.2 mIU/L in a 75-year-old is not the same clinical entity as in a 35-year-old. Studies using longitudinal data suggest treating SCH in the elderly may confer no benefit and possible harm (TRUST trial, 2017).¹⁰
Anti-TPO antibody status: SCH with anti-TPO antibodies positive carries a 4–5× higher rate of progression to overt hypothyroidism. This dramatically changes the decision to treat.
TSH-FT3 discordance in levothyroxine-treated patients: Up to 15% of levothyroxine-treated patients have persistently low FT3 despite normal TSH, due to impaired T4-to-T3 conversion. These patients report ongoing symptoms and may benefit from combination T4/T3 therapy — a nuance most prescribers miss.

🪙 Clinical Pearl: Do not interpret TSH in isolation. Always check anti-TPO antibodies in any patient with TSH 2.5–10 mIU/L, and assess FT3 (not just FT4) in symptomatic levothyroxine-treated patients with a "normal" TSH.

🦪 Oyster: Central hypothyroidism — pituitary-dependent — presents with a low/normal TSH and low FT4. A clinician reflexively reassured by a "normal" TSH will catastrophically miss this diagnosis. Always check FT4 in patients with suspected pituitary disease, prior head irradiation, or unexplained metabolic depression despite normal TSH.

8. Adrenal Incidentaloma and Mild Autonomous Cortisol Secretion (MACS): The Hidden Metabolic Disruptor

The widespread use of cross-sectional imaging has created a new diagnostic entity: the adrenal incidentaloma. Prevalence at autopsy exceeds 6%. Approximately 30–40% of incidentally discovered adrenal adenomas demonstrate Mild Autonomous Cortisol Secretion (MACS) — formerly called "subclinical Cushing's syndrome."¹¹

MACS is diagnosed when the post-1mg overnight dexamethasone suppression test (DST) shows cortisol > 50 nmol/L (1.8 µg/dL) in the absence of classical Cushingoid features. Yet these patients harbour significantly elevated rates of:

Type 2 diabetes and insulin resistance
Hypertension
Dyslipidaemia
Vertebral fractures
Cardiovascular events

🪙 Clinical Pearl: The Endocrine Society 2023 guidelines now recommend that all adrenal incidentalomas ≥ 1 cm undergo a 1mg DST, regardless of imaging characteristics. MACS is clinically silent by definition — you will not find it if you do not look.

🦪 Oyster: In any patient with difficult-to-control hypertension, insulin resistance, and osteoporosis — even without classical Cushingoid features — consider MACS. The triad of metabolic syndrome + adrenal incidentaloma should prompt immediate cortisol assessment. This is one of the most commonly missed reversible causes of metabolic syndrome.

⚡ Clinical Hack: Order the 1mg overnight DST as an outpatient investigation. Dexamethasone 1mg at 11 pm; fasting cortisol at 8–9 am the next morning. If cortisol > 50 nmol/L, escalate with 24-hour UFC and late-night salivary cortisol. The investigation costs almost nothing and changes management profoundly.

9. Primary Hyperaldosteronism: Vastly Underdiagnosed, Dramatically Undertested

Primary hyperaldosteronism (PA) — autonomous aldosterone secretion from adrenal adenoma or bilateral adrenal hyperplasia — is now recognised as the most common cause of secondary hypertension, present in 5–10% of all hypertensive patients and potentially 20–30% of those with treatment-resistant hypertension.¹² Yet testing rates in most practices remain dismally low.

The screening test — Aldosterone-to-Renin Ratio (ARR) — is a simple blood draw, yet it is ordered in fewer than 2% of eligible hypertensive patients in most audits.

Who to screen (updated 2023 Endocrine Society guidance):

Hypertension with spontaneous or diuretic-induced hypokalaemia
Resistant hypertension (BP uncontrolled on ≥ 3 agents including a diuretic)
Hypertension with adrenal incidentaloma
Hypertension with obstructive sleep apnoea
Hypertension with a first-degree relative with PA
Onset of hypertension < 40 years

🪙 Clinical Pearl: PA is NOT a diagnosis of hypokalaemia. Over 70% of patients with confirmed PA have normal serum potassium. Using hypokalaemia as the screening criterion is equivalent to screening for diabetes only in symptomatic patients — you miss most of the disease.

🦪 Oyster: Aldosterone has direct, potassium-independent metabolic effects: it promotes insulin resistance, inflammation, and myocardial fibrosis. PA patients have significantly higher rates of atrial fibrillation, left ventricular hypertrophy, metabolic syndrome, and cardiovascular events than BMI-matched essential hypertensives at the same blood pressure. The metabolic harm of PA exceeds the haemodynamic harm.

⚡ Clinical Hack: Spironolactone as an empirical fourth-line antihypertensive in a resistant hypertensive is both therapeutic and diagnostic. A dramatic BP response to a mineralocorticoid receptor antagonist should immediately raise suspicion for PA and prompt formal ARR testing.

10. Polygenic Risk Scores and the Era of Precision Metabolic Risk Prediction

The final frontier — and the one that will define the next decade of metabolic medicine — is Polygenic Risk Scoring (PRS) for metabolic disease.

PRS aggregates the effect of thousands of common genetic variants, each with small individual effect, into a single composite risk estimate. PRS for type 2 diabetes, coronary artery disease, and obesity have now been validated in large prospective cohorts including UK Biobank (n > 500,000).¹³

Crucially, high PRS identifies individuals at elevated lifetime risk decades before clinical disease, enabling targeted preventive intervention when it is still possible to alter the trajectory.

🪙 Clinical Pearl: A patient in the top 8% of the T2DM PRS distribution has a lifetime risk equivalent to a monogenic mutation carrier. PRS is not a niche academic tool — it is a clinically actionable stratification instrument that will enter mainstream primary prevention within this decade.

🦪 Oyster: PRS also identifies the "resilient" phenotype — individuals with high genetic risk who remain metabolically healthy. Studying these individuals has yielded insights into novel protective mechanisms (e.g., AMPK pathway activation, superior mitochondrial biogenesis) that are now therapeutic targets. Understanding why some high-risk individuals escape disease is as clinically important as understanding why others succumb to it.

⚡ Clinical Hack: Even without formal PRS, a thorough three-generation family history — first-degree relatives with T2DM, premature CAD, stroke, or NAFLD — provides a practical clinical proxy for genetic risk. Systematic family history documentation remains underperformed in almost every outpatient practice.

Summary Table: 10 Diagnostic Shifts at a Glance

#	Diagnostic Shift	Old Approach	New Standard	Action Point
1	Prediabetes Diagnosis	Fasting glucose + HbA1c	Add 75g OGTT; check 1h glucose	Order OGTT in all high-risk patients
2	Insulin Resistance	Rarely measured	HOMA-IR / TyG Index	Integrate into metabolic risk assessment
3	BMI & Adiposity	BMI-centric	WHtR > 0.5; MONW concept	Measure waist and height in every patient
4	Dyslipidaemia	LDL-C only	Non-HDL-C, ApoB, Lp(a), remnant	Measure Lp(a) once; use ApoB in insulin resistance
5	Fatty Liver	NAFLD by exclusion	MASLD; FIB-4 routine in diabetes	Calculate FIB-4 in every diabetic patient
6	Glycaemic Monitoring	HbA1c	CGM, TIR, GMI, CV	Consider 14-day CGM in complex cases
7	Thyroid Function	TSH alone	Age-stratified; FT3; anti-TPO	Check FT3 in symptomatic levothyroxine patients
8	Adrenal Adenoma	Imaging only	Mandatory 1mg DST; screen for MACS	DST for all adrenal incidentalomas ≥ 1 cm
9	Secondary Hypertension	Test only if hypokalaemic	ARR in all resistant/young hypertensives	Normalise ARR testing in hypertension clinics
10	Genetic Risk	Ignored in practice	Polygenic risk scoring; family history	Document three-generation metabolic family history

A Mnemonic for the Wards: "PRIME-WATCH"

P — Postprandial glucose (OGTT; 1-hour sample) R — Remnant cholesterol / Renin-aldosterone ratio I — Insulin resistance (HOMA-IR / TyG Index) M — MASLD / FIB-4 (in every diabetic) E — Endocrine causes of metabolic syndrome (MACS, PA)

W — Waist-to-height ratio (≥ 0.5 is abnormal) A — ApoB / Lp(a) (beyond LDL-C) T — Thyroid FT3 in symptomatic patients on levothyroxine C — CGM (for glycaemic variability, not just HbA1c) H — Hereditary/Polygenic risk (family history documentation)

When to Escalate vs. When to Watch

Clinical Scenario	Watch	Escalate
OGTT 2h glucose 7.8–10.9 mmol/L	Lifestyle, repeat at 6 months	If HOMA-IR > 3 or TyG high → Metformin / GLP-1 RA
FIB-4 1.30–2.67 (indeterminate)	Repeat in 12 months	LSM (FibroScan) → hepatology if ≥ 8 kPa
1mg DST cortisol 50–138 nmol/L	Annual repeat, metabolic monitoring	If > 138 nmol/L or metabolic sequelae → endocrinology
ARR elevated (> 30 ng/dL per ng/mL/h)	Confirm with repeat	Confirmatory test (saline infusion) → adrenal CT → adrenal vein sampling
MONW with WHtR > 0.5, TyG > 8.5	Lifestyle intervention, 3-monthly monitoring	If HbA1c rising, BP uncontrolled or MACS suspected → specialist referral

Conclusions: The Metabolism You Were Never Taught to See

The metabolic patient sitting in front of you may have normal glucose, normal BMI, normal LDL, and normal thyroid function — and yet be on a fast track to their first cardiovascular event or hepatic decompensation.

The ten diagnostic shifts reviewed here share a common thread: they require clinicians to look beyond the obvious, challenge comfortable cut-offs, and use the tools already available in a more sophisticated way. None of them require expensive or experimental investigations. All of them are available to the practising internist today.

The internist of the next decade will be defined not by what investigations they order, but by knowing which patient needs which investigation, why, and what to do with the result. The paradigms described here are that map.

"The first duty of medicine is to make the diagnosis. The second is to make sure the diagnosis is complete." — Sir William Osler (paraphrased for metabolic medicine, 2025)

References

Færch K, Johansen NB, Witte DR, et al. Relationship between insulin resistance and β-cell dysfunction in subphenotypes of prediabetes and type 2 diabetes. J Clin Endocrinol Metab. 2015;100(2):707–716.
Gayoso-Diz P, Otero-González A, Rodriguez-Alvarez MX, et al. Insulin resistance (HOMA-IR) cut-off values and the metabolic syndrome in a general adult population. Eur J Endocrinol. 2013;168(1):101–109.
Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, et al. The product of triglycerides and glucose, a simple measure of insulin sensitivity. Comparison with the euglycemic-hyperinsulinemic clamp. J Clin Endocrinol Metab. 2010;95(7):3347–3351.
Wildman RP, Muntner P, Reynolds K, et al. The obese without cardiometabolic risk factor clustering and the normal weight with cardiometabolic risk factor clustering. Arch Intern Med. 2008;168(15):1617–1624.
Ashwell M, Gunn P, Gibson S. Waist-to-height ratio is a better screening tool than waist circumference and BMI for adult cardiometabolic risk factors. Obes Rev. 2012;13(3):275–286.
Mach F, Baigent C, Catapano AL, et al; ESC Scientific Document Group. 2019 ESC/EAS Guidelines for the management of dyslipidaemias. Eur Heart J. 2020;41(1):111–188.
Varbo A, Benn M, Tybjærg-Hansen A, et al. Remnant cholesterol as a causal risk factor for ischemic heart disease. J Am Coll Cardiol. 2013;61(4):427–436.
Rinella ME, Lazarus JV, Ratziu V, et al; NAFLD Nomenclature Consensus Group. A multisociety Delphi consensus statement on new fatty liver disease nomenclature. Hepatology. 2023;78(6):1966–1986.
Monnier L, Mas E, Ginet C, et al. Activation of oxidative stress by acute glucose fluctuations compared with sustained chronic hyperglycemia in patients with type 2 diabetes. JAMA. 2006;295(14):1681–1687.
Stott DJ, Rodondi N, Kearney PM, et al; TRUST Study Group. Thyroid hormone therapy for older adults with subclinical hypothyroidism. N Engl J Med. 2017;376(26):2534–2544.
Fassnacht M, Arlt W, Bancos I, et al. Management of adrenal incidentalomas: European Society of Endocrinology Clinical Practice Guideline. Eur J Endocrinol. 2023;188(1):G1–G42.
Funder JW, Carey RM, Mantero F, et al. The management of primary aldosteronism: case detection, diagnosis, and treatment: an Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab. 2016;101(5):1889–1916.
Khera AV, Chaffin M, Aragam KG, et al. Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations. Nat Genet. 2018;50(9):1219–1224.
Lim EL, Hollingsworth KG, Aribisala BS, et al. Reversal of type 2 diabetes: normalisation of beta cell function in association with decreased pancreas and liver triacylglycerol. Diabetologia. 2011;54(10):2506–2514.
Nikolopoulou A, Kadoglou NP. Obesity and metabolic syndrome as related to cardiovascular disease. Expert Rev Cardiovasc Ther. 2012;10(7):933–939.

Conflict of interest: None declared. Funding: None. This article represents the personal opinions and clinical synthesis of the author and does not constitute formal clinical guideline endorsement.

Saturday, April 18, 2026

Bedside Surgery in the ICU: The Clinician's Guide to Short Operative Procedures in Critically Ill Patients

Bedside Surgery in the ICU: The Clinician's Guide to Short Operative Procedures in Critically Ill Patients

Dr Neeraj Manikath , claude.ai

1. Clinical Introduction

A 58-year-old man with decompensated cirrhosis, septic shock from spontaneous bacterial peritonitis, and acute-on-chronic liver failure is admitted to the medical ICU. By day three, his oxygenation is worsening despite high-flow nasal cannula. His abdomen is massively distended; ultrasound confirms 14 litres of tense ascites. He cannot lie flat. His coagulation profile is deranged — INR 2.6, platelets 54,000/µL. His family insists he be kept alive. The intensivist must now make a series of rapid decisions: drain the ascites? Insert a chest drain for the right-sided pleural effusion? Secure the airway? Place a central venous catheter in a coagulopathic patient? Which procedure first, in what order, at what threshold, with what precautions?

This vignette is not exotic. It is Tuesday morning in most medical ICUs across the world.

The modern ICU is as much a procedural suite as it is a monitoring bay. Critically ill patients routinely require short bedside operative interventions — procedures that are time-sensitive, high-stakes, and performed in a physiologically hostile environment. Unlike the controlled operating theatre, the ICU bedside offers poor lighting, limited assistance, a sedated and haemodynamically labile patient, and an audience of anxious nurses and family members. Mistakes carry disproportionate consequences.

Data suggest that over 70% of ICU patients undergo at least one invasive bedside procedure during their admission, and procedure-related complications account for a meaningful proportion of ICU-acquired morbidity and mortality. Yet most internal medicine training programmes dedicate far less time to the decision-making around these procedures than to the mechanics of performing them.

This review focuses on the most frequently needed short surgical procedures in the ICU — central venous catheterisation, arterial line placement, endotracheal intubation, chest drain insertion, therapeutic paracentesis, thoracocentesis, tracheostomy, and urinary catheterisation under difficult circumstances — with emphasis on when to do them, how to do them safely, and when to stop and call for help.

2. Pathophysiology: Why the Critically Ill Patient Is a Hostile Procedural Environment

Understanding why procedures go wrong in the ICU requires appreciating three physiological realities unique to critical illness.

Haemodynamic Fragility

The critically ill patient has often exhausted compensatory reserves. A vasodilatory response to procedural pain, a Valsalva-induced reduction in venous return during a difficult intubation attempt, or a pneumothorax from a central line can precipitate cardiovascular collapse in a patient who was marginally compensated. The procedural stress response — catecholamine surge, hypertension, tachycardia — can be as dangerous as the procedure itself.

Coagulopathy

Disseminated intravascular coagulation, liver failure, massive transfusion, therapeutic anticoagulation, and uraemia all impair haemostasis. The traditional teaching that procedures are contraindicated above an INR of 1.5 is not evidence-based for most bedside interventions but remains deeply embedded in practice. Understanding which procedures genuinely require normalisation of coagulation — and which do not — is a critical competency.

Anatomical Distortion

Obesity, oedema, prior surgery, burns, contractures, and pathological fluid shifts alter normal anatomical landmarks. The jugular vein that should be a reliable 2–3 cm target may be inaccessible in a patient with a cervical collar, bilateral neck haematomas from prior attempts, or severe anasarca. Real-time ultrasound guidance has fundamentally changed the safety of vascular access and body cavity drainage in this context.

3. Clinical Pearls 🪙

🪙 Pearl 1: The "Worst-First" Sequencing Principle

In a patient who needs both a chest drain and a central line, drain the chest first. A tension pneumothorax or massive effusion causing haemodynamic compromise must be relieved before you subject the patient to the physiological stress of central venous cannulation.

🪙 Pearl 2: Coagulopathy Does Not Uniformly Contraindicate Bedside Procedures

BTS guidelines recommend that thoracocentesis and paracentesis can be safely performed at INR up to 2.0–2.5 without FFP correction, provided real-time ultrasound guidance is used. Blind procedures in coagulopathic patients are a different matter entirely.

🪙 Pearl 3: The DOPE Mnemonic for Acute Tube-Related Deterioration

When a ventilated patient suddenly deteriorates after intubation, think: Dislodgement, Obstruction, Pneumothorax, Equipment failure — in that order. Bag the patient manually before troubleshooting the ventilator.

🪙 Pearl 4: Femoral Venous Access Is Underused in Emergencies

In a crashing patient with no obvious upper body access, the femoral vein — 1–2 cm medial to the femoral artery, below the inguinal ligament — is reliably accessible even in obesity and coagulopathy, does not risk pneumothorax, and can accommodate large-bore catheters for resuscitation. Its infection risk over short durations (48–72 hours) is comparable to other sites when managed aseptically.

🪙 Pearl 5: The Pre-Procedure "Time Out" Saves Lives

Modelled on the WHO Surgical Safety Checklist, a 60-second bedside "Time Out" before any ICU procedure — confirming patient identity, site, consent, allergy status, and emergency equipment availability — has been associated with significant reductions in adverse events. Most ICUs have this mandated; fewer actually observe it.

4. Oysters 🦪

🦪 Oyster 1: Obesity Is a Major Risk Multiplier for All ICU Procedures

The obese patient has reduced functional residual capacity, more rapid oxygen desaturation during apnoea, altered neck anatomy, difficult vascular access, and adipose tissue that obscures both landmarks and ultrasound windows. Pre-oxygenation in the 25° reverse Trendelenburg position significantly extends the safe apnoea period before intubation. Most trainees have not been taught this.

🦪 Oyster 2: "Bloody Tap" at Thoracocentesis Does Not Always Mean Haemothorax

Blood-stained pleural fluid occurs in malignancy, pulmonary infarction, trauma, and as an artefact of inadvertent intercostal vessel puncture. The haematocrit of the aspirate compared to peripheral blood distinguishes true haemothorax (pleural:peripheral haematocrit ratio >0.5) from haemorrhagic exudate. Draining a true haemothorax precipitously without surgical backup can be catastrophic.

🦪 Oyster 3: Femoral CVP Is Unreliable for Volume Assessment but Valid for Drug Infusion

Femoral CVP overestimates intrathoracic CVP by 2–5 mmHg due to intra-abdominal pressure transmission, particularly in ventilated patients. Never titrate volume resuscitation to femoral CVP. But it is an entirely acceptable route for vasopressors, TPN, and drug delivery while a more optimal access site is being planned.

🦪 Oyster 4: Spontaneous Breathing Can Worsen Haemodynamics During Chest Drain Insertion

The negative intrathoracic pressure generated during spontaneous inspiration can cause rapid inrush of air through a partially created tract — creating an iatrogenic pneumothorax before the drain is sited. Having an assistant apply finger occlusion to the tract while the drain is passed is a simple but underused safety step.

🦪 Oyster 5: Tracheostomy Timing in ARDS Remains Genuinely Uncertain

The TracMan trial showed no mortality benefit for early (day 1–4) versus late (day 10+) tracheostomy. More recent meta-analyses suggest early tracheostomy may reduce sedation requirements in selected patients, but no universal recommendation exists. Blanket early tracheostomy programmes are not evidence-based.

5. Clinical Hacks & Tips ⚡

⚡ Hack 1: The Triple Check for Endotracheal Tube Placement

After intubation, confirm placement with: (1) direct visualisation of the tube passing through the cords, (2) waveform capnography — the gold standard; five consistent CO₂ waveforms confirm tracheal placement, (3) bilateral chest rise. Auscultation alone is unreliable in the noisy ICU.

⚡ Hack 2: Use Ultrasound to "Map" the Intercostal Space Before Thoracocentesis

Identify: the effusion, the diaphragm (which rises with respiration — never aim below it), the lung edge, and any intervening structures. Mark the needle entry point in real time and measure depth. Avoid "freehand" guidance for novices; use the probe continuously during needle insertion.

⚡ Hack 3: Never Force a J-Wire — Redirect or Seek Help

If the guide wire meets resistance during subclavian or internal jugular cannulation, withdraw to the needle hub and redirect. Forcible wire passage causes dysrhythmias, vessel perforation, and cardiac tamponade. A wire that passes easily but curves back on imaging suggests contralateral passage or subclavian-to-ipsilateral jugular looping.

⚡ Hack 4: Pre-Procedure Ketamine for the Haemodynamically Unstable Patient

In a hypotensive patient requiring a painful procedure, low-dose ketamine (0.5–1.0 mg/kg IV) provides analgesia and sedation while preserving sympathetic tone and blood pressure. Avoid benzodiazepines and propofol in this context. Ketamine raises intracranial pressure — contraindicated in traumatic brain injury.

⚡ Hack 5: Flow Rate — A Peripheral Introducer Beats a Triple-Lumen CVC

A large-bore peripheral catheter-over-needle (e.g., 8.5 French introducer sheath) delivers far higher flow rates than a standard triple-lumen CVC. Flow rate is proportional to catheter radius to the fourth power and inversely proportional to length (Hagen-Poiseuille law). A 16G peripheral IV in the antecubital fossa beats a triple-lumen CVC for fluid resuscitation.

6. State-of-the-Art Updates

Ultrasound-Guided Procedures: Now Standard of Care

Real-time ultrasound guidance for central venous catheterisation, thoracocentesis, paracentesis, and arterial line placement is no longer an optional adjunct — it is the standard of care in most high-income country ICUs. A landmark meta-analysis (Brass et al., 2015) demonstrated a 57% reduction in failed placements, a 72% reduction in arterial puncture, and a 78% reduction in haematoma formation with ultrasound-guided internal jugular cannulation versus landmark techniques. NICE (IPG342, updated 2020) recommends real-time 2D ultrasound for all elective central venous catheterisations.

POCUS Beyond Vascular Access

The POCUS-guided ICU has expanded the safe indications for bedside procedures. Real-time assessment now includes:

• IVC collapsibility to guide pre-procedure fluid optimisation

• Cardiac function to identify tamponade before placing a subclavian line

• Lung sliding post-procedure to exclude pneumothorax in seconds

• Diaphragm excursion to assess readiness for extubation post-tracheostomy

Surgical vs. Percutaneous Dilatational Tracheostomy

A 2022 Cochrane review confirmed that PDT is associated with lower rates of wound infection, scarring, and post-operative bleeding compared to surgical tracheostomy, with equivalent rates of serious complications. PDT at the bedside by trained intensivists (with bronchoscopic or ultrasound guidance) avoids the risks of patient transport to the operating theatre.

Anticoagulation and Invasive Procedures: Revised Thresholds

Current evidence-based thresholds (BSH and ACCP guidelines):

Procedure	Safe INR Threshold	Safe Platelet Threshold
Central venous cannulation (US-guided)	≤ 2.5	≥ 20 × 10⁹/L
Thoracocentesis (US-guided)	≤ 2.0	≥ 50 × 10⁹/L
Paracentesis (US-guided)	≤ 2.5	≥ 20 × 10⁹/L
Chest drain insertion	≤ 1.5	≥ 50 × 10⁹/L
Percutaneous tracheostomy	≤ 1.5	≥ 50 × 10⁹/L

Video Laryngoscopy: The New Default for ICU Intubation

The DAS guidelines 2022 recommend video laryngoscopy (VL) as the first-line technique for intubation in the ICU. The DEVICE trial (NEJM 2023) demonstrated significantly higher first-attempt intubation success with VL compared to direct laryngoscopy in ICU patients. VL is particularly advantageous in anticipated difficult airway, cervical immobility, poor mouth opening, and obesity.

7. Diagnostic Nuances

Recognising the Patient Who Will Decompensate During a Procedure

Key warning signs that a patient will not tolerate a bedside procedure without additional preparation:

• Noradrenaline dose >0.3 µg/kg/min: Consider deferral until ≤0.1 µg/kg/min; if urgent, increase vasopressor dose prophylactically before starting.

• SpO₂ < 94% on FiO₂ > 0.6: This patient has virtually no oxygen reserve. Even a 60-second period of apnoea can precipitate cardiac arrest.

• Raised intra-abdominal pressure (IAP > 20 mmHg): Impairs venous return, worsens renal perfusion, and displaces the diaphragm. Therapeutic paracentesis must proceed with simultaneous albumin infusion (8 g/L drained) and haemodynamic monitoring.

The Difficult Airway: ICU-Specific Predictors Beyond Mallampati

In the ICU, additional predictors of difficult intubation include:

• SpO₂ < 93% before intubation (predictor of desaturation before successful placement)

• Obesity (BMI > 35)

• Presence of blood, secretions, or vomitus in the oropharynx

• Agitation or non-cooperation

• Modified MACOCHA score ≥ 3: a validated ICU-specific difficult intubation score incorporating Mallampati class, apnoea score, coma, and hypoxaemia

Differentiating Pneumothorax from Bullae on Chest X-Ray

⚠️ Critical Diagnostic Pitfall

Inserting a chest drain into a giant bulla instead of a pneumothorax is one of the most consequential diagnostic errors in ICU. Features suggesting bullae rather than tension pneumothorax: curved (not straight) inner border; residual lung markings within the lucency; background COPD or Marfan's; haemodynamic stability despite appearance. In any doubt, CT chest is mandatory before chest drain insertion for a non-tension scenario.

Early Identification of Catheter-Related Complications

• Persistent arm or neck pain after IJV/subclavian placement → wire in contralateral subclavian or jugular

• Phrenic nerve stimulation (hiccups) after left subclavian placement → tip in right atrium

• Resistance to flushing → tip against vessel wall, thrombosis, or kinked catheter

• Unexplained haemothorax after central line placement → vessel laceration; CXR must be reviewed within 1 hour

8. Management Intricacies

Endotracheal Intubation in the Critically Ill

Pre-oxygenation

Minimum 3 minutes of 100% FiO₂ via tight-fitting non-rebreather mask or bag-valve-mask. Apnoeic oxygenation (high-flow nasal cannula at 15 L/min maintained throughout the intubation attempt) extends safe apnoea time by 3–5 minutes.

RSI Drug Sequence

• Hypotensive patient: Ketamine 1–2 mg/kg IV + Succinylcholine 1.5 mg/kg IV (or Rocuronium 1.2 mg/kg IV if succinylcholine contraindicated)

• Haemodynamically stable: Propofol 1.5–2.0 mg/kg IV + Succinylcholine 1.5 mg/kg IV

• Raised ICP: Fentanyl 1–2 µg/kg blunts haemodynamic response; avoid ketamine

Succinylcholine Contraindications (use Rocuronium instead)

• Hyperkalaemia (renal failure, rhabdomyolysis, burns > 72 hours old, prolonged immobilisation)

• Personal or family history of malignant hyperthermia

• Known myopathies

• Penetrating eye injury

Post-intubation Ventilator Settings

Initial tidal volume 6 mL/kg predicted body weight (ARDS network protocol), PEEP 5 cmH₂O initially, FiO₂ 1.0 then titrate to SpO₂ 92–96%, RR 14–18/min. Obtain ABG at 30 minutes post-intubation.

Central Venous Catheterisation: Site Selection Hierarchy

• Right internal jugular vein (RIJ): most predictable anatomy, lowest pneumothorax risk, direct path to SVC; preferred for most patients

• Left internal jugular vein: acceptable alternative; higher risk of malposition

• Subclavian vein: lowest infection risk in long-term catheters; highest pneumothorax risk; avoid in coagulopathy

• Femoral vein: highest infection risk in prolonged use; no pneumothorax risk; preferred in emergencies

Therapeutic Paracentesis in Cirrhosis

Volume drainage: All large-volume paracentesis (>5 L) must be accompanied by albumin infusion (8 g per litre drained) to prevent paracentesis-induced circulatory dysfunction (PICD), which carries a 50% 3-month mortality.

Entry point: Use real-time ultrasound. The left iliac fossa (Z-technique entry) is preferred. Avoid the right iliac fossa in cirrhosis (caecal distension, portal collaterals) and the midline (inferior epigastric vessels, falciform ligament remnant).

Chest Drain Insertion: Technical Essentials

Size selection:

• Small bore (10–14 Fr, Seldinger): simple pneumothorax, transudative effusion, malignant effusion

• Large bore (20–28 Fr, blunt dissection): haemothorax, empyema, viscous exudate

• Surgical (≥32 Fr): traumatic haemothorax requiring evacuation

Safe triangle (BTS-recommended): bounded by the anterior border of latissimus dorsi, lateral border of pectoralis major, and 5th intercostal space. Insert over the superior border of the rib to avoid the neurovascular bundle in the subcostal groove.

Fluid drainage rate: No more than 1–1.5 L in the first hour (risk of re-expansion pulmonary oedema). Clamp the drain after 1 L and allow patient to recover.

9. When to Escalate / When to Watch

The ESCALATE Framework

Letter	Principle
E	Emergency or elective? Emergencies override coagulopathy concerns; elective procedures allow optimisation.
S	Skill level adequate? Know your limits; get senior help before starting, not after a complication.
C	Coagulation correctable? Transfuse only when above evidence-based thresholds.
A	Anatomy accessible? If ultrasound shows no clear window, stop and reassess.
L	Likelihood of benefit? Does this procedure change management?
A	Alternative available? PICC instead of CVC? Oral medication instead of NGT?
T	Transport risk? Can the patient go to theatre or radiology for a safer environment?
E	Equipment and team ready? Crash cart in the room, resuscitation drugs drawn up, experienced nurse scrubbed.

Mandatory Escalation Triggers

• Failed central venous access after two attempts at a single site — try a different site or seek help

• Haemodynamic collapse during a procedure not responding to fluids/vasopressors within 2 minutes

• Suspected cardiac tamponade from central line placement (worsening hypotension, rising CVP, new pulsus paradoxus): bedside echo immediately

• Air embolism: left lateral decubitus Trendelenburg (Durant's manoeuvre), 100% oxygen, aspirate air via central catheter

• Subcutaneous emphysema after chest drain insertion: suggests misplaced drain in subcutaneous tissue

• Haemoptysis or haematemesis after NGT insertion: stop, do not reinsert, seek endoscopic assessment

Appropriate Conservative Management (Watch Rather Than Intervene)

• Asymptomatic pneumothorax < 2 cm rim in a spontaneously breathing patient: supplemental oxygen and serial imaging every 6 hours

• Central line-associated thrombosis (CLAT) without bacteraemia: anticoagulation and line removal if feasible

• Post-paracentesis mild hypotension responding to passive leg raise: fluid bolus and reduce drainage rate rather than stopping entirely

10. Summary Table and Mnemonic

The SAFE HANDS Mnemonic for ICU Procedural Safety

Letter	Principle
S	Site — confirm with ultrasound before needle contact with skin
A	Anticoagulation — know the threshold; do not over-correct
F	Failure plan — know what you will do if the procedure fails or causes complications
E	Equipment — check and assemble everything before starting
H	Haemodynamics — optimise before, monitor during, reassess after
A	Airway — always protect first; nothing else proceeds if the airway is not secured
N	Needle direction — never advance blindly; use real-time guidance
D	Documentation — time, operator, indication, complications, post-procedure check
S	Senior help — two attempts is your limit; ask before, not after

ICU Procedures at a Glance

Procedure	Key Indication	INR / Plt Threshold	Preferred Technique	Post-Procedure Check
Endotracheal intubation	Respiratory failure, airway protection	Not relevant (life-saving)	RSI + video laryngoscopy	Waveform capnography, CXR
Central venous catheterisation	Vasopressors, monitoring, access	INR ≤2.5 / plt ≥20K	US-guided, Seldinger	CXR within 1 hr
Arterial line (radial)	Continuous BP, ABG sampling	INR ≤3.0 / plt ≥20K	Modified Seldinger	Radial pulse, Allen's test
Thoracocentesis	Symptomatic effusion, diagnostic	INR ≤2.0 / plt ≥50K	Real-time US, small-bore	POCUS lung sliding, CXR
Chest drain insertion	Pneumothorax, empyema, haemothorax	INR ≤1.5 / plt ≥50K	Seldinger (small) or blunt (large)	CXR, drain swinging/bubbling
Paracentesis (LVP)	Tense ascites, diagnosis	INR ≤2.5 / plt ≥20K	Real-time US, Z-technique	Albumin 8g/L, BP monitoring
Percutaneous tracheostomy	Prolonged ventilation (>10 days)	INR ≤1.5 / plt ≥50K	PDT with bronchoscopic guidance	ETCO₂, CXR, bilateral air entry
Urinary catheterisation	Urinary retention, accurate UO	N/A	Coude catheter for difficult urethra	Drainage confirmed, no haematuria
Nasogastric tube	Enteral nutrition, medications	N/A	Lubrication, chin-to-chest	CXR or pH <5.5 aspirate

11. References (Vancouver Format)

1. Hind D, Calvert N, McWilliams R, Davidson A, Paisley S, Beverley C, et al. Ultrasonic locating devices for central venous cannulation: meta-analysis. BMJ. 2003;327(7411):361.

2. Young D, Harrison DA, Cuthbertson BH, Rowan K; TracMan Collaborators. Effect of early vs late tracheostomy placement on survival in patients receiving mechanical ventilation: the TracMan randomized trial. JAMA. 2013;309(20):2121–9.

3. Prekker ME, Driver BE, Trent SA, Resnick-Ault D, Bednarczyk JM, Im Y, et al. Video versus direct laryngoscopy for tracheal intubation of critically ill adults (DEVICE trial). N Engl J Med. 2023;389(5):418–29.

4. Frat JP, Thille AW, Mercat A, Girault C, Ragot S, Perbet S, et al. High-flow oxygen through nasal cannula in acute hypoxemic respiratory failure. N Engl J Med. 2015;372(23):2185–96.

5. Gordon CE, Feller-Kopman D, Balk EM, Smetana GW. Pneumothorax following thoracentesis: a systematic review and meta-analysis. Arch Intern Med. 2010;170(4):332–9.

6. Brass P, Hellmich M, Kolodziej L, Schick G, Smith AF. Ultrasound guidance versus anatomical landmarks for internal jugular vein catheterization. Cochrane Database Syst Rev. 2015;(1):CD006962.

7. Blaivas M, Lyon M, Duggal S. A prospective comparison of supine chest radiography and bedside ultrasound for the diagnosis of traumatic pneumothorax. Acad Emerg Med. 2005;12(9):844–9.

8. Cavallazzi R, Hirani A, Vasu TS, Pachinburavan M, Bhatt DL, Elizondo J, et al. Sonographically guided thoracentesis: a meta-analysis of complications. J Ultrasound Med. 2010;29(10):1467–73.

9. De Backer D, Biston P, Devriendt J, Madl C, Chochrad D, Aldecoa C, et al. Comparison of dopamine and norepinephrine in the treatment of shock. N Engl J Med. 2010;362(9):779–89.

10. Dunham CM, Barraco RD, Clark DE, Daley BJ, Davis FE 3rd, Gibbs MA, et al. Guidelines for emergency tracheal intubation immediately after traumatic injury. J Trauma. 2003;55(1):162–79.

11. Mora Carpio AL, Mora JI. Ventilator management. Treasure Island (FL): StatPearls Publishing; 2023.

12. Antonelli M, Azoulay E, Bonten M, Chastre J, Citerio G, Conti G, et al. Year in review in intensive care medicine 2010. Intensive Care Med. 2011;37(3):381–400.

13. De Jong A, Molinari N, Terzi N, Mongardon N, Arnal JM, Guitton C, et al. Early identification of patients at risk for difficult intubation in the intensive care unit: development and validation of the MACOCHA score. Am J Respir Crit Care Med. 2013;187(8):832–9.

14. Feller-Kopman D, Light R. Pleural disease. N Engl J Med. 2018;378(8):740–51.

15. Froudarakis ME, Bridevaux PO, Galanis N, Evander E, Guy C, Hatzitheofilou C, et al. Interventional pulmonology in European countries: findings of the HERMES pulmonology survey. Eur Respir J. 2011;37(5):1222–7.

Grand Rounds Review | Internal Medicine & Critical Care Series | Conflict of interest: None declared | Funding: None