ETM Radiology: That’s not a knife!

In true Aussie trauma style, we pay homage to Crocodile Dundee, with a case of penetrating chest injury. Can you troubleshoot the chest in this tricky case?

A 49 year old man presents after being stabbed multiple times in the left anterior chest.

Paramedics state that en-route he was tachycardic to 125 bpm, his oxygen saturations were dropping to 88% despite supplemental oxygen, he had absent breath sounds on the left, and they therefore decompressed his left chest with a pneumocath/pleurocath, placed in the left anterior 2nd intercostal space attached to a heimlich valve.

On arrival the patient was alert, GCS 15, with a persistent tachycardia and persistently low sats around 90-92% despite supplemental oxygen. He had 3 x 2cm stab wounds to his left anterior precordium, and the pneumocath was in situ.

The Police are in attendance and produce the weapon for inspection by the treating team. (iPhone used for size comparison)

This is a knife!

This is a knife!

[DDET What structures may be injured in stab wounds to the left anterior chest? Click for answer]

Stab wounds to the left anterior chest may cause:

  • Pneumothorax
  • Haemothorax: with bleeding from chest wall tissue or intercostal muscle, intercostal or internal mammary vessels, lung laceration, major vessel injury or cardiac injury
  • Cardiac injury with or without tamponade
  • Tracheo-bronchial injury

Depending on the length of the blade and angle of insertion, other structures that may be injured include:

  • Neck – airway, vessels & nerves
  • Thyroid (thymus in babies/kids)
  • Diaphragm (and sub-diaphragmatic organs)
  • Spine – it would have to be a very long knife!
  • Thoracic duct

[DDET What rapid bedside investigations would you perform?(Click for answer)]
Chest x-ray
There is debate about the utility of chest x-ray in patients who will undergo a CT chest regardless of the x-ray findings, however it’s still recommended as a quick, low to no risk, non-invasive procedure, especially in penetrating chest trauma.

This is the patient’s first chest x-ray:

Chest X-ray 1 on arrival

Chest X-ray 1 on arrival

It shows an apparently inflated left lung (small residual apical pneumothorax) with a pneumocath in situ, and no haemothorax. The cardiac contour looks normal.
Do you notice anything else about this image?

Bedside ultrasound
E-FAST can detect pneumothorax, haemothorax and pericardial effusion/tamponade with greater sensitivity than chest x-ray.
However this is operator dependent and other factors such as obesity, subcutaneous emphysema and chest wall haemaotoma may hinder the views. In this case, there was poor views of the heart on bedside ultrasound. No massive pericardial effusion but we couldn’t exclude it.


What happened next?
Given the obvious penetrating injuries, and pneumocath insertion, an intercostal catheter was inserted. Only 50 ml of blood drained on insertion.

[DDET Click to see the first post-chest drain x-ray]

Chest x-ray 2 - post ICC insertion

Chest x-ray 2 – post ICC insertion

It shows an apparently inflated lung, with a fairly well positioned chest tube (1 or 2 rib spaces higher is ideal)


So he should get better now, right?

The patient remained tachycardic, the sats hovered around 92-94% on oxygen, and his blood pressure remained around 100 systolic. Not the fantastic response to decompression/ICC insertion we were hoping for!

[DDET What could be going on in this patient’s chest to explain the lack of improvement post chest tube insertion?]

1) Occult injury
Lung or treacheo-bronchial injury +/- ongoing air leak/tension could explain the ongoing hypoxia and haemodynamic compromise, but the lung looks well inflated! Doesn’t it?

Haemorrhage could explain the haemodynamic compromise but not the low sats. There’s no gross haemothorax on the film and only 50ml of blood was drained from the chest drain.

Tamponade? Could be causing the signs we’re seeing, but he’s remarkably “stably unstable”. i.e. He’s not deteriorating. One would expect a patient who is tamponading from a hole in the heart to deteriorate in front of you, or at least be more unstable. This patient’s vitals are relatively static.

2) Chest tube:
The chest tube may not be in the pleural space, but the drain was seen to swing & bubble on insertion.

The drain was checked and it is filled/setup correctly.
Could it be kinked or blocked with clot? Possibly, but this is very rare.
Could the connection between tube and drain be loose? This was checked and found to be well connected with no leak/entrainment of air.


What happened next? 

The patient was taken to CT, and this is his scan

[radiologystack cats=18]

It shows a well positioned intra-pleural chest tube, a small haemothorax, and a persistent pneumothorax. Look closely at the pneumocath. It’s intra-ulmonary (i.e. it’s in the lung), possibly with the tip in a bronchus. This has created a fistula between the patient’s airway and the outside of the chest, which has impaired lung re-expansion. This also raises the possibility of an air leak from lacerated lung, that is allowing air to enter the pleural space faster than the chest tube can withdraw it.  Look closely at the base of the heart as well. There is no large pericardial effusion, but there is some blood sitting near the apex.

[DDET The the lab rings. Can you guess what the troponin was?]

28,000! Yep, and that’s the new high sensitivity troponin (with a bad STEMI usually peaking around 2,000-3,000). Not only does this guy have a likely broncho-atmospheric fistula from the pneumocath, and/or lacerated lung causing a persistent air leak, but his myocardium is injured as well.


On return to the resus cubicle, the pneumocath was promptly removed, the first chest drain was placed on suction with no effect, and a second chest tube was inserted. He went to theatre shortly thereafter for a median sternotomy.

[DDET Here’s the chest x-ray post 2nd chest tube insertion]

Post 2nd chest tube insertion

Post 2nd chest tube insertion

After the 2nd tube was inserted, the sats started coming up, the pulse slowed a bit, and the Thoracics Team arrived to whisk him off to theatre.


Injuries found at operation:

Lacerated lung from knife wounds – repaired

Hole in lung from pneumocath – repaired

Lacerated myocardium – heart chambers not penetrated – repaired

Lessons from this case:

1) Chest X-ray has limitations in trauma
Go back to the initial chest x-ray. What’s wrong with it? It’s supine. In fact all of them are. In pentrating chest trauma with a systolic BP > 90, don’t be afraid to sit the patient up for the chest x-ray. It may have shown the non-inflated lung, and moved the team to the 2nd chest tube more quickly. Supine chest x-rays can easily miss a pneumo or haemothorax. Ultrasound is better, however in this case, it didn’t help due to the overlying tissue damage, and subcutaneous emphysema.

For a great discussion on the utility of chest x-rays in trauma – head over to St Emlyn’s where Simon Carley does a great podcast, and engages in some lively debate in the comments section on ditching superfluous, low sensitivity, time-wasting procedures and processes in resus.

If you’re going to bother to get a chest x-ray in penetrating chest trauma, and the patient can tolerate sitting up without fainting from hypotension – sit them up!

2) A non-inflating lung with an adequately positioned chest tube means there’s an air leak

A broncho-pleural connection will allow air to continue to enter the pleural space, sometimes faster than a chest tube can get it out. The management for this is to apply wall suction to the drain (NOT the tube!), and if this doesn’t work, insert a 2nd chest drain. A broncho-atmospheric connection (eg from a pneumocath) means the inhaled air will bypass the pleural space, and the trapped pleural air won’t escape, rendering the pneumocath useless. If there is a hole in the chest wall that has not closed over you can get a broncho-atmospheric connection, however the pressures in the airway and atmosphere should be roughly the same, so you shouldn’t get tension.  Without a broncho-pleural connection, a non-sealed chest wall hole will act as an open pneumothorax, and draw air from outside inwards during inspiration, (the “sucking chest wound”), and you may see bubbling from the wound on exhalation as some of the air in the pleural space is expelled. The treatment for this is either a 3-sided dressing (which are notoriously hard to get to work), or insert a chest tube elsewhere in the hemithorax (not in the hole) and repair/close/seal the sucking wound.

3) Knives are dangerous

As the London HEMS know all to well, with their case series of roadside thoracotomies for penetrating chest injury, knives vs chests all too often ends in tragedy.  As demonstrated by the lack of utility of chest x-rays in this case, your threshold for getting a CT scan for anyone stabbed in the chest should be very low. Wound exploration of thick chest walls with intercostal muscles that close over the tract can be falsely reassuring. Patients can bleed and deteriorate slowly, so initially normal vital signs can be falsely reassuring. And as in this case, chest x-ray (even with an adequately positioned ICC) can fail to detect significant pneumothoraces.

4) Pneumocaths/pleurocaths don’t always work, and may cause problems

There are various commercially available pleural drainage kits used by ambulance and pre-hospital services. Most involve a seldinger type kit, usually fitted to a heimlich valve (a one way valve with a rubber sleeve inside a clear plastic chamber). Complications of pneumocaths/pleurocaths include:

  • Too shallow – failure to reach the pleural space – subcutaneous air may be aspirated giving a false positive result without treating the pneumothorax
  • Too deep – can penetrate the lung/other organs
  • They can get blocked with blood or kink – leads to re-accumulation of pneumothorax
  • Intercostal vessel injury – haemorrhange

In this case the intrapulmonary insertion rendered the pleurocath ineffective.

pleurocath heimlich


2 Comments on “ETM Radiology: That’s not a knife!”

  1. bevan R

    Hi Andy,

    A few other signs on the CXR not to miss- even with the rotated film the trachea and mediastinum are still shifted right. A clue there must be either collapsed lung on the right (which cannot be seen in the CXR) or air/blood expanding in the left side (you can still see a big pneumothorax even in the supine film).

    Another potential edit, HS troponin often peaks in the 100,000s in bad STEMIs.

  2. David Berger

    You should have seen absence of lung sliding anteriorly on ultrasound — did you?

    Plus, there was, as the previous poster said, significant mediastinal and tracheal shift, worse after the first chest tube was inserted.

    Good case.

Leave a Reply