Central Line Air Emboli: one death every day?

Central lines are essential for the care of many patients.

Unfortunately worldwide data indicate around one patient is dying every day from avoidable central line related air emboli.

We would appreciate your feedback on the numerous case series and reports presented below. (Please click on any links in red).

Many cases occur because of accidental line disconnections (this problem can be engineered out – see here). Many other cases occur when central lines are removed with patients sitting upright – staff should ensure patients are lying flat or slightly head down to central line removal.

Rarely a patient may not be able to lie flat for a sufficient period for line removal. In this case we recommend the following:

Patients who can’t lie flat or have other risk factors have an increased propensity for air embolism on central line removal. In these circumstances there needs to be an escalation of procedure:

Assess patient for increased risk of air embolism before removal of the CVC. Risks include:
– Respiratory compromise (can generate a large negative intrathoracic pressure increasing risk and rate of air entrainment)
– Intravascular depletion (which can lead to a greater negative intravascular pressure, increasing the risk and rate or air entrainment)
– Inability to lie flat for an extended period (if unable to lie flat may create a pressure gradient that favours movement of air into the circulation)
– Low body mass index (smaller tract between the atmosphere and the vessel)

If the patient is unable to tolerate lying in the supine position or are considered high risk, the following should occur:
– Do not remove the CVAD in the first instance
– Contact an experienced critical care medical officer to review and manage the patient
– Delay removal until the risks can be minimised (if possible)
– Vascular access nurse (or experienced nurse) remove central line when appropriate
– Critical care medical officer in attendance
– Ensure alternative venous access already obtained
– Remove line in an adequate environment – monitored, with arrest trolley at hand
– Patient to remain monitored until satisfied risks of air embolism have abated

This link provides some very useful information regarding central line management.

These are more than just statistics.


(Image above from CNN report)

Sufficient air can enter blood vessels rapidly, over 1 or 2 seconds, causing a cardiac arrest or cerebral air embolus which often leads to death.

The magnitude of this problem is not well known – adverse event being dispersed in place and time.  Also the usual healthcare safety approach of education, alerts and policy writing doesn’t appear to have impacted on preventing this issue. In fact reports demonstrate the frequency of this adverse event is only increasing (see here).

We discuss how a human factors approach may significantly help in reducing adverse event frequency (see here).

Your feedback is greatly appreciated.

Thank you.

How frequently do central line related air emboli deaths occur? (see here):

Even the most conservative estimates (5million CVCs per year in US, air embolus 0.1% of CVCs, and 23% mortality) indicate 1,150 deaths per year from this avoidable complication in the US alone.

Does education and experience help reduce the incidence? (see here)

Unfortunately studies suggest that those with greater experience may be less likely to follow appropriate procedures to minimise the risk of CVC related air embolus. Also the impact of education tends to wane rapidly with time.

If CVC related air embolus is so prevalent where are all the reports?

Below we’ve collated numerous case series and an overwhelming amount of case reports which highlight the prevalence of this problem. Please note we have no access to error report databases (see here) – these cases represent only some of those available via the internet – we have little doubt there are multiple more.

As is the nature of adverse events in healthcare they tend to be dispersed in time and place. This often starves us of the impetus required to put effective system measures in place.

We’ve only just started to collate this data, however one may already start to appreciate the magnitude of this issue.

Please help us put a stop to these avoidable adverse events: (read here)

Thank you.


Case Series

158 cases of ‘cerebral’ air emboli related to CVC. 21% mortality.

52 of 86 cases of air embolus presenting for HBOT at one institution

15 cases one institution radiology dept in 5 years

11 reports NPSA UK (2009 to 2011) all on removal of CVC 9 in sitting position


8 reported deaths 51 near miss events NSW CEC since April 2012

6 CVC related air emboli AIMS database

10 cases (4 deaths) air embolus on removal of dialysis catheter

26 patients suffering cerebral air embolus

6 cases related to CVC removal

19 cases of CVC air embolus, 7 fatal

8 cases of air embolism from central line removal

By 1987 total of 79 cases in world literature, 25 (32%) had fatal outcome

2 fatal CVC related cerebral air emboli & literature review

5 cases of CVC related air emboli in lung transplant recipients

Several cases of air embolus related to central lines in NPSA

9 reports NPSA patient deterioration from CVC removed upright

3 reports of harm from CVC being left open to air

2 cases of homonymous hemianopia from CVC related cerbral air emboli

2 cases (one death) post removal of central line

2 paediatric patients air embolus on insertion of CVC

6 patients suffering neurological injury related to CVC management

Median claim for VAE US$ 325,000, 100% payment. 2007 range $25,800 to $4,120,00

 Case Reports – Insertion

Case Reports – Maintenance

Disconnection / ports left open

Death of 29yr old living renal donor transplant recipient

Death – AAGBI alert

Severe cerebral air embolus

33yr female, 17 weeks pregnant. Child died at 4 months of age.

77yr male cerebral air embolus

62yr female death

32yr male disconnected CVC

66yr female collapsed from CVC ports being left open

62 yr male air embolus from disconnection

Death from sheath disconnection

81yr female death from CVC port being left open

Disconnection of CVC lead to stroke & later death. Over US$18million payment.

65yr male died post CVC disconnection

Pulmonary air embolism

Death post misconnection

38yr male air embolus from misconnection of CVC

Fatal air embolism from CVC misconnection

2yr old air embolism from misconnection

Cerebral air embolus from CVC misconnection

56yr female misconnection

81yr male death – air in portacath

72yr male – severe permanent neurological injury post catheter misconnection

Other e.g. damage to central line

79yr male dies from cerebral air embolus after lacerating CVC while shaving

71 yr male – severe neurological impairment

52yr male lacerated CVC while shaving

72yr male lacerated CVC while shaving

62yr female air embolus post rupture of CVC

Air embolus from accidental laceration of CVC

Case Reports – Removal

1. Patient not supine


16yr female – CVC removal in preparation for discharge

50yr male – CVC removed in preparation for discharge

53yr male – Dialysis catheter removed, US$10million settlement

Father of 3 – death on catheter removal, US$8.2million settlement

23yr female – US$6million settlement

18yr male death post CVC removal

Death on removal of CVC, patient sat up

65yr male, severe cerebral air embolus ultimately causing death

Lethal pulmonary air embolus on removal of dialysis catheter

Death 8 days post op from CVC removal

Lethal case of air embolus from dialysis catheter removal upright

43yr male death post CVC removal

Male – stroke and death over 2 months later from CVC removal upright

Female (mother) death from CVC removal in upright position

Death from removal of double lumen catheter in upright position

36yr male died post CVC removal

73yr male died post removal of double lumen dialysis catheter

44yr male, Liver Transplant, death post CVC removal

18yr male, Liver Transplant, death post CVC removal

Female, death post removal of CVC

68yr female death post CVC removal

83yr male death post CVC removal in upright position

20yr male death from CVC removal upright

Neurological Impairment

Acute decompensation, CVC removed by resident – review by Dr Peter Pronovost

57yr female minimal long term injury

Female – severe neurological injury, US$3.85million settlement

35yr – central line removed by 1st yr resident, severe injury, US$5.75 million

54yr female – severe neurological injury, US$3.8million settlement

Collapse and acute hemiplegia

95yr female, permanent neurological impairment

54yr male air embolus and neurological deficit post PA catheter removal

50yr female massive cerebral air embolus and severe neurological injury

Young male air embolus post CVC removal

20yr male – treated with hyperbaric oxygen therapy

43yr male cerebral air embolus from CVC removal

Air bubbles on echocardiography post CVC removal

89yr female post CVC removal

76yr male stroke from air embolus post CVC removal

23yr male coronary air embolism and MI

57yr female cerebral air embolism post removal of CVC

71yr male CVC removed ’45 degrees head up’

No residual deficit

52yr male 8 days post op

60yr male

27yr male 30 mins post removal

73yr male air embolus post CVC removal by House Officer

2. Central line insertion site not sealed sufficiently

Death of 84yr old

51yr old, cerebral air embolism, site not sealed appropriately

49 female stroke post CVC removal

65yr male air embolus post air embolus removal

Near fatal embolism, fibrin sheath as portal of air entry

76yr male air embolus via residual tract

Please note we have only just started to collate this date. We have little doubt there are hundreds of other individual case reports in journals which are not included above. These reports only represent a fraction of the true number of cases. We will continue to update this list.

Please help stop these avoidable deaths: click here.

Estimates of Frequency:

1. How many central lines are inserted per year?

In US an estimated 6 million lines inserted per year and increasing.

5 million per year in US (NEJM 2003)

More than 5million per year in US. Approximately 8% of hospitalised patients.

As many as 5million CVCs per yr in US. 15% experience complications.

48% of ICU patients. 15million CVC days per year in US.

Estimated 200,000 CVCs per year in NHS (BMJ 2013)

2. What proportion of the central lines are complicated by air emboli?

0.2 to 1% of CVCs

0.2 to 1% of CVCs

Incidence may be as high as 0.8% (AAGBI)

1 in 47 to 1 in 3000 cases

1 in 772 

1 in 800

0.1 to 2% of CVC insertions

0.13 to 0.5% of insertions

0.13% of interventional radiology CVC manipulations

15 of 11,583 radiographically inserted central lines

(Gas embolism complicates 2.65 per 100,000 hospitalizations)

3. What proportion of the air emboli result in death?

Mortality rate 23 to 50%

Mortality rate 23%

Review of 24 cases – 50% died, of remainder 42% had permanent neurological injury

Review of 26 cases – 23% died

4 out of 10 cases resulted in death

Mortality rate up to 30%

Awareness, Experience & Education

Survey: Only 31% of nurses (whose job description included removal of central lines) reported using all the recommended procedures. (see here)

Nurses more aware than doctos of risks of air embolism on CVC removal (see here)

Air embolism entirely preventable complication, but not widely known among practitioners (see here)

Are we missing too many cases? (see here)

Nurse Survey. In overall group comparison, few differences were found between nurses and physicians in terms of patient positioning at CVC insertion or removal. Nurses were more likely than physicians to request air-occlusive dressings after CVC removal (19 of 53 [36%] vs. 12 of 140 [9%]; p < .001), but there was no difference between nurses and physicians in awareness of VAE as the reason for choosing one patient position or dressing over another (29% vs. 39%, respectively). Critical care nurses with <=2 years of experience more often placed the patient in the supine or the Trendelenburg position for CVC removal than nurses with >2 years of experience (71% vs. 26%; p = .03).

Although most physicians (127, 91%) chose the Trendelenburg position for CVC insertion, only 42 physicians (30%) reported concern for VAE. On CVC removal, only 36 physicians (26%) cited concern for VAE. Some physicians (13, 9%) reported elevating the head of the bed during CVC removal, possibly increasing the risk of VAE. Awareness of VAE or its prevention did not correlate with the level of physician training, experience, or specialty. After the educational intervention, concern for and awareness of proper methods of prevention of VAE improved (p < .001). At 6-month follow-up, reported use of the Trendelenburg position continued, but concern cited for VAE had returned to baseline. (see here)

8 thoughts on “Central Line Air Emboli: one death every day?

  1. The exact incidence is hard to estimate. We must extrapolate data from a number of sources. Furthermore, Australian data is more sparse than it is in the US, but what is known is that the event is both common and associated with significant harm- including death. ‘One death every day’ seems a reasonable conclusion based on the numbers available. What is certain is that gas embolism is entirely preventable and unnecessary central venous cannulation is a common problem.


  2. Another great post Rob.

    I think it’s a bit like the conjecture surrounding Marty Makary’s assertion that iatrogenic death is the third ranked cause of death in the US – the real number is as not important as the realisation that any number is too big. It clearly occurs at a scale that must be curbed.

    The only benefit in exploring the incidence further is to be able to put it in context for the average punter to make it real -“a death per day” is quite memorable and provocative.

    Of course the other reason to pin down incidence is to monitor effectiveness of interventions – but our current monitoring of adverse events is so poor that accurate numbers are elusive. We need to evolve a better culture of tracking outcomes and managing them. Until then, nothing will change.


  3. This makes tough for pretty sombre reading – to think all these deaths are totally preventable. From my local experience and hearing of CVC air-embolism deaths in nearby hospitals I have absolutely no doubt that to say CVCs contribute to “one death a day around the world” is almost certainly an understatement. We must do something about it. Great work and good luck Patient Safe.


  4. Quite astonishing (and depressing) reading. Your research clearly shows the following regarding CVC related VAE…
    – its a major safety issue
    – its much more common than we think
    – when it occurs the consequences can be dire
    – it falls into a category of preventable morbidity and mortality
    – medico-legal costs are high if it occurs
    Calculating exact numbers of CVC related VAE is hard (?impossible) however, estimates of ‘one death a day’ would seem to be entirely justifiable. Action needs to be taken as a matter of urgency to reduce the risk to patients.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s