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