Litigation related to central venous access by anaesthetists: an analysis of claims against the NHS in England 1995–2009


  • A response to a previously published article or letter can be submitted to the Online Correspondence section at Please note that a selection of this correspondence will be reproduced (possibly in modified form) in the journal. All correspondence intended for publication in Anaesthesia should be addressed to Dr Steve Yentis, Editor-in-Chief, and submitted as an email attachment to Copy should be prepared in the usual style of the Correspondence section. Authors must follow the advice about references and other matters contained in the Guidance for Authors at Correspondence presented in any other style or format will be returned to the author for revision. All correspondence submissions should be accompanied by a completed Author Declaration Form which can be accessed via a link under ‘Covering letter’ in the Guidance for Authors (as above). The completed Author Declaration Form should be sent either by e-mail with the submission or by fax to (0)207 681 1008.

  • Visit the Anaesthesia Correspondence website at and comment on any article or letter in this issue of the Journal.

This letter presents data from the NHS Litigation Authority (NHSLA) relating to anaesthetists and central venous access between 1995 and 2009. The data have important limitations regarding completeness and validity that affect their interpretation, and readers are strongly urged to read the full methodological descriptions in previous related publications [1, 2]. This study is an analysis of litigation and not complications, lack of denominators meaning that the incidence of litigation cannot be determined. ‘Cost’ describes the value attributed to a closed claim by the NHSLA and includes all legal costs and any payment made to the claimant, but not the cost of the NHSLA itself. Cost is not ‘damages’ [2].

There were 26 claims (2.5% of all assessable anaesthesia claims): 19 were closed and 13 led to cost. Twenty (77%) claims described a severe or fatal outcome [3] with 10 deaths, increasing to 11 if the death of an unborn child (stage of pregnancy not stated) is included. The total cost of closed claims was £1.68 m; median (IQR [range]) £27 000 (£0–£93 000 [£0–£790 000]).

Two claims’ details were very limited and prevented complete analysis: of the remaining 24 claims, two were obstetric and none paediatric. Routes of access are shown in Table 1: where the nature of the injury indicated use of the internal jugular or subclavian route these claims are described as ‘upper body approach’. There were 22 descriptions of immediate harm, two of delayed harm and two in which there was inadequate detail to judge. Trauma to 23 nearby structures was described in 20 claims (Table 1).

Table 1.   Central venous access claims: routes of access and types of traumatic injury. Values are number.
 Internal jugular (= 10)* Subclavian (= 1) Femoral (= 2)†Upper limb and Hickman (= 11)‡ Not stated (= 2)
  1. *One claim described haemothorax and haemopericardium.

  2. †One claim described arterial and nerve injury.

  3. ‡One claim described bilateral pneumothorax.

Arterial puncture81140
Other vascular injury00010
Nerve injury00100

There were 14 claims relating to arterial punctures; 10 were closed, five with cost: total cost £1.27 m; median (IQR [range]) £16 000 (£0–£113 000 [£0–£790 000]). Sites of arterial puncture were eight carotid, one internal mammary, one femoral and four unspecified during upper limb access. All eight descriptions of carotid artery injury involved the internal jugular approach; five described arterial puncture and three catheter placement in the carotid. There were five deaths (four strokes and one pharyngeal bleed), two non-fatal strokes and one ‘brain damage’. Seven were closed, and three led to cost (total £0.95 m; median (IQR [range]) £0 (£0–£79 000 [£0–790 000]). The six claims of other arterial punctures described one death, one stroke, two thoracotomies and one possible brain damage: three were closed (total £324 000; median (IQR [range]) £33 000 (£16 000–162 000 [£0–292 000]).

Four claims described haemopericardium (two internal jugular and two ‘upper body’ approaches), all of which were fatal: three during insertion and one delayed catheter erosion through the right atrial wall. One claim was closed, with a cost of £72 000. Other claims of trauma described unspecified fatal ‘vascular injury’, bilateral pneumothoraces, empyema following intended intravenous fluid administration into a misplaced pleural catheter and femoral nerve injury. There were two claims of a retained wire (without permanent harm) and one of stroke due to air embolism (failure to place a cap on a central line port). There were no claims relating to local, distant or generalised infective complications. No claims described the use of ultrasound techniques, or their omission.

Claims relating to central venous access, although representing a small proportion of claims against anaesthetists, are marked by high severity, having the highest proportion of severe and fatal outcomes of all 13 clinical subgroups we examined [1]. They account for < 3% of all anaesthesia-related claims but > 14% of deaths. Overall cost was relatively low. Many claims described misplaced internal jugular catheters leading to puncture or cannulation, of the carotid artery. These misplaced catheters resulted most commonly in stroke, approximately half of which were fatal. Carotid puncture may occur in up to 10% of attempts at internal jugular cannulation [4] while stroke is rare, often delayed and perhaps not recognised as a complication of venous cannulation [5]. The claims in this dataset are sentinel events and probably the ‘tip of the iceberg’ in terms of near misses. The data do not enable an analysis of whether the use of ultrasound to guide cannulation, as recommended since September 2002 [6], reduces litigation. In 2006, Bodenham wrote ‘in the past it was possible to defend clinicians who did not use ultrasound on the basis that it was not yet routine or of proven benefit, but I believe that this position will become increasingly untenable in the future’ [7]. It is possible that injury to adjacent vascular structures during central venous cannulation without ultrasound guidance will, in the future, be judged more severely.

Several claims relating to central venous access cited failures of basic technique: a central line that was too long was left in place and eroded the right atrium leading to fatal haemopericardium; a cap left off a central line led to air embolism and stroke; and two wires were lost during line insertion. All these injuries are preventable by standard good practice and the descriptions hint at inadequate technique and therefore perhaps inadequate training.

Claims related to central venous access in the American Society of Anesthesiologists Closed Claims Project have marked similarities in both quantitative aspects and case descriptions [8]. Comparison with the Australian Incident Monitoring Study (AIMS) suggests that central venous access events leading to clinical incident reports have similar mechanisms to those leading to litigation, but differ in outcome severity [9].

Of interest, no claim was related to infective complications. Perhaps, in the future, claims related to infective complications will lead to litigation.

Claims relating to central venous cannulation are a numerically small but clinically and financially important group. These claims describe events (and by implication a larger group of cases of patient harm) that appear to be largely preventable by good quality care. A reduction in the number of central lines placed is one solution. The majority of claims relate to mechanical injury to adjacent structures and are likely to be preventable with use of optimal ultrasound techniques and high standards of care. Complications of central venous access are likely to represent a suitable area of focus for risk reduction in all trusts. Ensuring high levels of compliance with NICE recommendations on ultrasound guidance for internal jugular line insertion, and routine good care, would apparently prevent most claims.

No external funding and no competing interests declared.


We thank Dr Shaun Scott and Dr Ruxandra Mihai for work classifying the NHS Litigation Authority dataset; and Ruth Symons, Risk Manager, NHS Litigation Authority for help with data interpretation.