Reducing major procedural complications from central venous catheterisation


In this month’s issue of the Journal, Cook provides important insights into UK anaesthetic legal cases, related to central venous catheters (CVCs) [1]. Key issues in reducing such risks are prevention based on the mechanisms of injury, plus early recognition and safe management of complications [2]. Successful defence of such cases is often hampered by the latter. Cook’s report mirrors other international series, with small numbers producing disproportionately high mortality and costs. It is likely that actual numbers, including those with limited injury, are much higher, as the report only relates to claims from the NHS Litigation Authority (NHSLA). Less dramatic complications will be even higher in numbers, and overall will probably cost even more in clinical terms, but are not discussed here.

The numbers and outcomes put these cases on a scale of risk similar to major anaesthetic airway events [3]. There is currently a national UK focus on catheter-related infection (National Patient Safety Agency, Matching Michigan [4]) in which the number of infections per 1000 catheter-days is recorded; poor or complicated insertions are intuitively important preventable factors. Similar numerical data are lacking for procedural complications, but could be audited recorded locally and nationally (e.g. a future Royal College of Anaesthetists National Audit Project).

Anaesthetists should contrast their interest and expertise in airway problems (e.g. specialist societies, lectures, national algorithms, difficult intubation trolleys), to that for CVCs; it may well be lacking for the latter. Nationally, there is little formalised training or accreditation of such practice, which is assumed to be learnt in general anaesthetic experience. The number of such procedures performed by anaesthetists is increasing, but training opportunities are missed locally [5].

Complications are likely to be more frequent in critical care [6], but may go unnoticed or unreported due to other problems. Procedures are more difficult due to restricted sites, thrombosed veins, coagulopathy, restless patients and hypovolaemia. Moreover, complications are likely to be more serious, for example, a pneumothorax or haemothorax in a critically ill patient receiving controlled ventilation.

Cook correctly cites ultrasound guidance as important in the prevention of complications, but the later elements of procedures are equally relevant. There is little good evidence for many clinical techniques that have developed over time in radiology, cardiac catheter labs and operating theatres. Currently, it is very rare for anaesthetists to visit other areas for training in such techniques, despite the obvious potential benefits. Anaesthetists are increasingly present in such areas to provide anaesthesia for interventional procedures, so training opportunities could be made available.

Mechanisms of injury

Many different sites of injury are described adjacent or connected to the vascular tree [2]. Certain patterns emerge from such reports on high value legal claims, relating to local pressure effects from arterial haematoma, massive bleeding into the chest/abdomen, and strokes. Below, I attempt to provide some clinical background to the cases in Cook’s report, explaining how and why such serious complications occur and why they are sometimes missed, and highlighting some principles of safe early management.

Collateral damage from needle placement

This includes damage to adjacent arteries, lung, pleura and nerves. Even if the original needle trajectory is correct, empty veins often have to be transfixed for cannulation. Structures behind the vein are liable to injury, e.g. the carotid artery or thyrocervical trunk. Arteries may also overlie the vein, e.g. the superficial femoral artery in the groin [7].

Needle damage is largely avoidable by the routine use of ultrasound to ensure accurate real-time placement of the needle tip. This skill limits many practitioners, who have had insufficient instruction or experience. Today, 8 years after national UK guidance [8], there are still clinicians unwilling/unable to use it to its full potential. For example, cardiologists in my own hospital justify not using it routinely for pacing, on the premise that introducer sheaths are not CVCs! Others cite a lack of evidence base for sites other than the internal jugular vein, but it is intuitive that similar advantages apply at all sites. I believe it is unethical to perform trials with landmark techniques taken to the point of complication, just to prove benefit for all access sites. Training issues related to ultrasound have been covered in a recent editorial [9].

Insertion of devices

Guidewires generally follow a path of least resistance centrally, but may go astray from all access sites, including the right internal jugular vein. Without real-time screening or other aids, there is no certainty that they have not passed over the midline, into vein branches, down an arm, or out of the vein. Excessive force must not be used as guidewires can exit veins, to pass into the pleura, mediastinum or other structures.

Dilators and catheters passed along a guidewire will enlarge the tract to their diameter. If the guidewire is kinked or angulated, and further force is applied to the dilator/catheter, it will tear the vein wall and exit into adjacent tissues. The guidewire should be repeatedly checked to ensure that it moves freely through the dilator/catheter, to ensure no distortion. Care must be taken and if any resistance is felt, the procedure must be stopped, or further imaging obtained to guide alternative approaches. Such judgements are very much an acquired skill, and there are no useful manikins to aid training. Mobile image intensifiers are available in most theatre suites, and provide the optimal tool, but are very rarely used by anaesthetists outside pain procedures; why is this the case?

Positioning of the catheter tip

An adequate length of the distal catheter, lying in the long axis of the superior vena cava with its tip above the pericardial reflection, is ideal (approximately level with the carina on chest X-ray). This is not always achievable, particularly with left-sided catheters, contrary to textbook guidance [10]. The majority of short-term CVCs are inserted without real-time imaging, with a subsequent chest X-ray to confirm central placement. Other aids include ECG guidance (although this does not distinguish between arterial, venous and mediastinal placement) and plain, on-table X-ray imaging. The use of these varies widely.

Many units do not stock suitable ranges of catheter lengths for each route of insertion. A minimal range of adult lengths would be 15, 20 and 24 cm. Failure to achieve good catheter positions will increase the risk of catheter-related thrombosis, arrhythmias, perforation of the vein’s wall, hydrothorax, cardiac tamponade, catheter failure and pain during injections.

Catheters may be misplaced within the venous system, by following an abnormal path to the neck, the arm or contralateral side. Such catheters need re-positioning if they are to stay in situ for anything more than a few hours. Alternatively, the patient may have a normal variant of anatomy, e.g. left-sided superior vena cava or acquired stenosis of the great veins [11]. Specialist advice before using catheters in such situations is required. More seriously, catheters on a plain chest X-ray may appear to follow the approximate normal path, but are not correctly sited in the superior vena cava. Analysis of cross-sectional anatomy shows that catheters in the superior vena cava, right pleural space, right internal mammary vessels, azygous system, ascending aorta or mediastinum cannot be reliably differentiated on a single plane chest X-ray. Many clinicians still rely on chest X-ray appearances alone; however, it can only confirm that the catheter has passed centrally and is not kinked.

If a catheter is in an unusual position or malfunctions, time should be taken to ascertain its position, and it should not just be pulled out. Bedside tests include transducing the pressures of all its lumens (via a fluid column or transducer) and aspiration of blood from the lumens for estimation of oxygen partial pressure or saturation. None of these bedside tests is entirely reliable. Definitive localisation requires injection of contrast down the catheter with an image intensifier or other imaging.

Specific complications


Tears of great veins are probably more common during dilator/catheter insertion than generally realised, due to thin vein walls. As a result of low venous pressure, connective tissue, muscle or other structures will usually tamponade local bleeding without further problems. Major bleeding occurs when a tear connects directly to the low pressure pleural space [12]. Chest veins at particular risk, immediately adjacent to the pleura, include the superior vena cava (right border), azygous system, hemi-azygous system and internal mammary. It is revealing to observe these vessels on CT scans, or at thoracotomy/thoracoscopy.

Damage to the arterial tree can cause similar problems, when the hole made by a needle alone is sufficient. The subclavian arteries and their internal mammary branches protrude and can bleed directly into the apex of the pleural spaces. Alternatively, a more distant enlarging arterial haematoma may burst into the pleura. Similar mechanisms apply to the peritoneal space, with cases of fatal concealed haemorrhage from iliac vessels. Management relies on drainage, leaving dilators/catheters in place to reduce bleeding, and urgent repair by surgery or interventional radiology.

Cardiac tamponade

Case series suggest that this occurs from two distinct mechanisms. First, the pericardial space (reflection) extends up on to the arch of the aorta and the proximal part of its branches and hence a needle puncture at this site may lead to bleeding into the pericardium. Surprising as it might appear, this is a recognised problem, usually after attempted right subclavian approaches [13]. Second, erosion of catheters through the lower superior vena cava or right atrium can occur, allowing infusion of fluid into the pericardium [14]. Reported cases suggest that it is typically pressurised fluid infusion, rather than low pressure venous bleeding, that is the problem. Cardiac perforation is probably common with pacing wires, but rarely causes tamponade.

A high index of suspicion is required to prompt, and then confirm, the diagnosis by echocardiography. Regrettably, it is often a post mortem finding. Management requires attempted aspiration of infused fluid through the catheter if still in situ, urgent pericardiocentesis, and stenting or surgical repair if required.

Arterial damage in the limbs/neck

Insertion of needles, guidewires, dilators and catheters may cause damage to arteries at the puncture site or more centrally. A local haematoma or false aneurysm may cause skin and tissue loss, nerve damage from compression [15] and upper airway compression. Surgery will be required to decompress large haematomas urgently, irrespective of how the leak is eventually closed (surgery or radiology). Arterial dissection, thrombosis, embolus and unintentional cannulation may cause distal ischaemic damage, with particular relevance to the carotid artery.

Case histories suggest that accidental arterial catheter placement may not be as obvious clinically as one might expect, as attending staff seek to find alternative erroneous explanations for malfunctioning catheters, back-bleeding, infusion pump alarms, and signs of thrombosis including a stroke or transient ischaemic attack. Bleeding or haematoma at the puncture site is not necessarily seen until removal of the catheter. A high index of suspicion is required.

Removal of large bore catheters situated in the carotid artery needs careful consideration, to avoid emboli to the brain, and to seal the arterial defect. Systemic heparinisation and removal of the device, with either a surgical procedure similar to a carotid endarterectomy (venting clot and suturing defect), or a radiological stenting procedure, are the two preferred options. Removing such devices and pressing for 20 min to prevent carotid haemorrhage will further risk brain ischaemia from haematoma, dislodged emboli and a lack of blood flow [16]. Similar considerations apply to larger arterial cannulae elsewhere. In the short-term, it is generally safe to leave dilators, catheters and guidewires in position, whilst the situation is evaluated. It is generally recommended not to remove catheters larger than 9 Fr from arteries without using a percutaneous occlusion device or surgical closure.


Complications of CVC insertion constitute a major burden due to serious patient harm and suffering, and medicolegal costs. The serious cases in Cook’s report are likely to represent only a small fraction of overall numbers. Many procedural complications are avoidable by increased education and experience of operators. Early recognition of problems and timely referral to appropriate specialties can often prevent permanent damage.

Competing interests

No external funding and no competing interests declared.