Editorial II

It’s NICE to see in the dark

David H. T. Scott1

1 Department of Cardiothoracic Surgery, The Royal Infirmary of Edinburgh, Edinburgh EH3 9YW1, UK, Email: david.scott@ed.ac.uk

The National Institute for Clinical Excellence (NICE) has published its recommendations for the use of ultrasound locating devices for placing central venous cannulae.1 The benefits of ultrasound guidance have been emphasized in this journal before.2 NICE concluded:

‘Two-dimensional (2-D) imaging ultrasound guidance is recommended as the preferred method for insertion of central venous catheters (CVCs) into the internal jugular vein (IJV) in adults and children in elective situations.’

‘The use of two-dimensional (2-D) imaging ultrasound guidance should be considered in most clinical circumstances where CVC insertion is necessary either electively or in an emergency situation.’

‘It is recommended that all those involved in placing CVCs using two-dimensional (2-D) imaging ultrasound guidance should undertake appropriate training to achieve competence.’

‘Audio-guided Doppler ultrasound guidance is not recommended for CVC insertion.’

The provisional report hoped for full implementation of its recommendations by the end of the financial year 2004–5, but the implementation date has been removed from the latest version.1

While the NICE Appraisal Committee considered evidence from manufacturers, an expert in intensive care, and professional and specialist groups, the basis for their recommendations was an assessment report prepared by the School of Health and Related Research, University of Sheffield.3 The report considered the available techniques for central venous access: the traditional landmark method, using surface anatomy and palpation of the artery beside which the vein usually, but not always, lies;4 two-dimensional ultrasound (2-D), which gives an ultrasound picture of the underlying structures, enabling their identification, and visualization of the path and effect of the operator’s needle; audio-guided Doppler ultrasound (AGDU) using a special needle or sterile attachment that produces an audio signal derived from flow in the vessel under attack; and surgical cut down.

The English language literature was searched for randomized controlled trials, and 20 met the selection criteria. These studies showed that ‘2-D ultrasound guidance was significantly better than the landmark method for all five outcome variables’. The outcome measures were: failed catheter placement, catheter placement complications, failure on the first catheter placement, the number of attempts to achieve successful catheterization, and speed of insertion. Firm conclusions related only to insertion at the internal jugular venous site. Data on speed of insertion were less impressive. Nevertheless, the evidence for the benefits of 2-D ultrasound of the internal jugular vein site is overwhelming. Efficacy at other sites, and in children, was also found but not supported by many randomized clinical trials.

Audio-guided Doppler ultrasound is not as successful a technology as 2-D. For the internal jugular vein, the pooled results of the trials considered suitable by the assessment group showed that it can improve catheter placement and reduce complications, but only failed catheter placements (61% reduction in risks, 95% confidence interval (CI) 8–83%, P=0.03), and risk of failure on the first catheter placement attempt (43% reduction in risk, 95% CI 12–63%, P=0.01) reached statistical significance.3 Successful insertion took longer using Doppler guidance than with the landmark method (35 s longer, 95% CI –54 to 124 s, P=0.4). Audio-guided Doppler ultrasound significantly increased the risk of failure at the subclavian vein, compared with the landmark technique. It was not endorsed by NICE who concluded, ‘Audio-guided Doppler ultrasound guidance is not recommended for CVC insertion’.1

There were no randomized control trials comparing surgical cut down with any other method. In my working place, the radiologists have taken over the insertion of long-term tunnelled lines, which used to be the province of the surgeons (although this was not the case in the USA before 1999).5

The introduction of new technology can be justified in many ways. Enthusiasts seem to act on the principle ‘We have the technology so let us use it’. Whereas it is important that everything is evaluated by somebody, the availability of a technology is not a very good reason for everybody adopting it. Right ventricular ejection fraction measurement and the Penaz technique of arterial pressure analysis using the Finapres are examples of technologies that work but are not very useful in anaesthetic practice. Many use the banner of best practice to justify innovation, although cynics may brand them as dedicated followers of fashion. Health service managers wonder ‘Will it save money this year?’. Health economists claim to look at the global, long-term picture and its effects on the whole nation. While I have believed the arguments of many health economists, I have rarely been able to convince my NHS managers to adopt their recommendations if it meant spending money from this year’s budget, to save money in the future. Despite its name, the National Institute for Clinical Excellence looks at economic considerations as well as clinical excellence. No economic evaluations of central line insertion were identified in the literature, and the only submission made to NICE that included economic evaluation was the assessment report prepared by the School of Health and Related Research, who devised their own model.3

Economic evaluation includes looking at the capital cost of equipment, its maintenance, running costs, and depreciation. The additional cost of training has to be identified and the opportunity costs, both lost and gained, must be identified. Opportunities lost may include the time taken to set up equipment, and opportunities gained include the reduction in time that theatre staff spend waiting for the central line to be inserted. There are also costs involved in reduction of complications in patients, and the reduction in litigation and other costs when serious complications occur. NICE expects that the introduction of ‘ultrasound for all’, will cost £15 000 000 in capital costs for the scanners, and £14 000 000 for additional training costs. It will add £10 to the cost of inserting a line but, by reducing the number of complications and the time taken to eventually get the line in place, it will reduce the total cost of the procedure by £2 per patient.

Many of the assumptions made by the School of Health and Related Research3 are open to debate. They used a capital cost of £11 000 for a scanner and assumed that a machine purchased today will be replaced with no scrap value in 3 yr time, giving a capital cost per annum of £3882. I am still using my first SiteRite, purchased for me by the British Heart Foundation in 1991.6 It currently retails for £7000, giving an annual capital cost, so far, of less than one-fifth of the figure used by NICE. As it should have many more years of useful life, the final value will be even lower.

The economic model also includes an annual maintenance charge of £1350. This may be the charge that an external organization would make, but we have four machines and no maintenance contract. In 25 machine years of use, we have had to send two probes away for re-alignment after trauma, and have had to replace the batteries occasionally. Our maintenance costs are £3372.65 for 25 machine years, one-tenth of the £1350 p.a. used in the economic analysis. Assuming a lifespan of 10 yr, a more realistic figure for equipment in today’s NHS, the initial capital cost comes down to £9 500 000, and the annual capital and maintenance cost of having a machine to £835.

Against the saving in machine costs must be set the number of line insertions performed per machine. The model assumed each machine would be used to insert 15 lines per week, based on the local cardiac surgery unit inserting 30 lines per week and presumably needing two machines. I think that this is unrealistic and one machine must be sited at each of the places where central lines are regularly inserted, i.e. each theatre performing major surgery and each intensive care area. Our four machines are used for an average of seven lines each per week. We found it very difficult to ensure that a SiteRite was immediately available for every line insertion when the machines had to be moved between theatres, and to the ICU.

Training costs are the area that I have the most disagreement with. NICE estimated that training costs alone would amount to £14 000 000 in England and Wales, ‘not including any capital cost requirement for training laboratories or dummies’. They suggest that consultant anaesthetists should perform 10 insertions under the supervision of a consultant radiologist, giving a training cost of £1090 per consultant anaesthetist. Trainees were charged £3357, the equivalent of one consultant radiologist doing nothing else for 1 week but train them. Of necessity, there was no one to train me, as I was the first person to get a suitable machine in Scotland, but it was very easy to use. In my experience, one half-hour session provided during the course of normal work is more than sufficient to get my consultant colleagues started on the use of the SiteRite. Most of them pick it up in 5 min. In terms of training junior staff, ultrasound actually saves time as you can quickly show the trainees the anatomy and ensure that they get into the right place very quickly. We have neither training laboratories nor dummies. I note that the only clinical authors of the assessment report were radiologists. The report’s own figures,3 show that anaesthetists insert the majority of central lines. They should certainly be the ones to train other anaesthetists. As they are already training them, and it takes less time to train with ultrasound than without, the training costs per central line insertion of up to £17 should be removed from the model, if not replaced by a negative amount.

The economic model uses arterial puncture as its only complication, with a landmark rate of 12%, an ultrasound rate of 3%, and a cost of £40. Only two references were eventually used to produce the figures for the model. Troianus and colleagues7 reported on 160 patients about to undergo cardiac surgery. Seven of 83 (8.4%) ‘Landmark’ patients suffered carotid puncture, compared with one of 77 (1.4%) in the ultrasound group. The authors said that they recorded the experience of the operators, but they omitted to report it. Sulek and colleagues8 described the results of a four-armed trial using landmark or ultrasound to approach the right or the left internal jugular vein. All operators had experience of over 60 cannulations, so they could not be classed as novices. The patients were undergoing elective abdominal, vascular, or cardiothoracic surgery, and were anaesthetized and mechanically ventilated. The carotid artery was punctured in 10 of 60 (17%) ‘Landmark’ patients, but only 3/60 (5%) using ultrasound. The high rate of carotid puncture used in the model has drawn criticism from colleagues in cardiac surgery. Webber SJ, Doidge CL and Morgan-Hughes NJ, in a poster presentation at the Association of Cardiothoracic Anaesthetists Meeting on June 21, 2002,9 studied 349 patients whose internal jugular veins were cannulated using the landmark method before cardiac surgery. Of the patients 3.4% suffered a carotid puncture; exactly the same rate as is obtained by summing the patient numbers and the complications in the papers quoted by Latto, in what must be the bible of central venous catheterization.10

Complications other than carotid puncture were not considered in the economic model. The School of Health and Related Research say in the discussion,3 ‘Inclusion of other complications would be likely to favour ultrasound further. Death for example is uncommon but the possible outcome of insertion complication in high risk patients, for example ventilated patients or patients undergoing cardiac pacing procedure.’ The source of the School’s mortality rate, the NCEPOD study on Interventional Vascular Radiology,11 may not be the best place to look for serious and fatal complications of central venous cannulation. Litigation work and personal communication within the UK has informed me of about four cases of cerebrovascular accident following carotid artery puncture during attempts at central venous cannulation; a complication I have not found reported in the medical literature. I also know of one young patient who lost an arm when central venous cannulation went wrong. The cost to the nation of the two cases that have already been settled is over £500 000. None of these cases have been formally reported in the medical literature.

Applying more realistic values to the model, the additional cost of using ultrasound comes out as £3, compared with the model’s £6.64, the lower capital and maintenance costs more than making up for the reduced number of procedures per week. At these prices, the use of ultrasound becomes economic even with complication rates as low as those reported by Webber and colleagues to the Association of Cardiothoracic Anaesthetists.9

The Assessment Report’s recommendations3 only relate to internal jugular venous cannulation in adults, because they have no randomized controlled trials demonstrating benefit in any of the other sites. They do consider the legal and ethical implications about whether ultrasound guidance should be used in children and infants, and for subclavian and femoral vein access. I personally have no doubt that once ultrasound machines become available, sensible practitioners trained in their use will use them at all sites, and for all ages. The NICE guidelines1 12 have gone through several metamorphoses; the first time I saw them they were marked ‘Strictly Confidential’. Now they are available to all on the Internet at

http://www.nice.org.uk/article.asp?a=36753

The latest version abandons the cautious ethical approach of the authors of the assessment report, and recommends the use of ‘2-D imaging ultrasound guidance as an adjunct for placing CVCs in the majority of clinical scenarios’. Very soon, complications of central venous catheterization where ultrasound has not been used will be very difficult to defend in court.

The Appraisal Committee addressed the problem of deskilling practitioners in the landmark method, and stressed that its theory should still be taught.3 I was taught the theory of using chloroform with a Schimmelbusch mask, ether with an Ogston’s inhaler, and ethyl chloride with a Guy-Ross bag. I have used chloroform, ether, trichloroethylene, halothane, and enflurane in Boyle’s bottles. I have no regrets whatsoever about the passing of these techniques, and much prefer to use modern agents from accurate, temperature-controlled vaporizers. The situation with ultrasound guided central venous cannulation is little different. I encourage all those concerned with inserting central venous catheters to take the necessary steps to get hold of appropriate numbers of portable ultrasound machines, not only to offer their patients ‘best practice’, but to save their trust’s money in the long run.

References

1 National Institute for Clinical Excellence. Guidance on the use of ultrasound locating devices for placing central venous catheters. Technology Appraisal Guidance No. 49, September 2002

2 Scott DHT. ‘In the country of the blind, the one-eyed man is king.’ Erasmus (1466–1536). Br J Anaesth 1999; 82: 820–1[Free Full Text]

3 Calvert N, Hind D, McWilliams RG, Thomas SM, Beverley C. The effectiveness and cost effectiveness of ultrasound locating devices for central venous access. National Institute for Clinical Excellence, http://www.nice.org.uk/Docref.asp?d=3019, 2002

4 Denys BG, Uretsky BF. Anatomical variations of internal jugular vein location: impact on central venous access. Crit Care Med 1991; 19: 1516–19[ISI][Medline]

5 Reeves AR, Seshadri R, Trerotola SO. Recent trends in central venous catheter placement: a comparison of interventional radiology with other specialties. J Vasc Interv Radiol 2001; 12: 1211–14[Abstract/Free Full Text]

6 Armstrong PJ, Cullen M, Scott DHT. The ‘SiteRite’ ultrasound machine—an aid to internal jugular vein cannulation. Anaesthesia 1993; 48: 319–23[ISI][Medline]

7 Troianos CA, Jobes DR, and Ellison N. Ultrasound-guided cannulation of the internal jugular vein. A prospective, randomized study. Anesth Analg 1991; 72: 823–6[ISI][Medline]

8 Sulek CA, Blas ML, Lobato EB. A randomized study of left versus right internal jugular vein cannulation in adults. J Clin Anesth 2000; 12: 142–5[CrossRef][ISI][Medline]

9 Webber SJ, Doidge CL, Morgan-Hughes NJ. Internal jugular line insertion in cardiac surgical patients using the landmark technique: is there a NICE alternative? Association of Cardiothoracic Anaesthesists, http://www.acta.org.uk/absCam02.htm, 2002

10 Latto IP. The internal jugular vein. In: Latto IP, Ng WS, Jones PL and Jenkins BJ, eds. Percutaneous Central Venous and Arterial Catheterisation. London: W B Saunders, 2000; 180–2

11 Interventional vascular radiology and interventional neurovascular radiology. National Confidential Enquiry into Perioperative Deaths, http://www.ncepod.org.uk/20003.htm, 2000

12 Barnett D, Adam J, Akehurst RL, et al. Appraisal consultation document: ultrasound locating devices for placing central venous catheters. National Institute for Clinical Excellence, http://www.nice.org.uk/article.asp?a=30194, 2002