Lipid emulsion to treat overdose of local anaesthetic: the gift of the glob


  • John Picard,

  • Tim Meek

Fat is bad for you. It may make our food appetising, but is widely consumed in gross surfeit. In excess, fat harms our patients and impedes any anaesthetic manoeuvre. But fat may also be clinically useful. An intravenous bolus of lipid may save lives in one otherwise lethal situation, namely refractory cardiovascular collapse caused by an overwhelming overdose of local anaesthetic.

So, how can lipid help? One lipid emulsion – Intralipid®; (Fresenius Kabi, Runcorn, UK) – is a time-honoured cocktail: an emulsion of soya oil, glycerol and egg phospholipids. It is a component of parenteral nutrition; many anaesthetists inject it every working day as part of their propofol. But for all its familiarity, it has a little known characteristic: in animals it is an effective antidote to the cardiovascular collapse caused by overdose of bupivacaine.

This has been shown by Guy Weinberg and his colleagues at the University of Illinois in Chicago, who gave rats varying doses of bupivacaine [1]. When the rats became asystolic, they received cardiopulmonary resuscitation followed by intravenous doses of either saline or lipid. Comparison of the dose–survival curves for the two groups is impressive. The toxic dose was increased by approximately 50% in the group that received lipid. The curves do not even overlap: a universally fatal dose in the saline group was universally survived by all rats that received lipid.

Weinberg's group then moved up the food chain to dogs [2]. Cardiovascular collapse was induced with intravenous bupivacaine and the dogs were resuscitated with standardised cardiopulmonary resuscitation for 10 min (to mimic the inevitable initial confusion surrounding such an unexpected collapse). Then either saline or lipid was given. All six dogs in the saline group died, whereas all six in the lipid group regained normal heart rhythm (all but one within 5 min of treatment).

So what is happening? At a molecular level, the precise mechanism of the collapse is unclear. Local anaesthetics bind and inhibit a wide variety of both voltage-gated and ligand-gated channels, and other proteins besides. Indeed, because local anaesthetics have hydrophilic and hydrophobic moieties, they are peculiarly able to bind proteins whether they are dissolved in cytosol or in the lipid bilayer of cell or organelle membrane.

One protein that may be particularly important is carnitine acylcarnitine translocase. Normally lost in the mists of preclinical biochemistry, this carrier resumes anaesthetic significance because the heart has a predilection for fatty acids, ketone bodies and lactate. The translocase contributes to the passage of the acyl CoA constituents of the longer fatty acids across the mitochondrial membranes to the site of their oxidation. If it is inhibited by local anaesthetics, then fatty acids will not be fully oxidised, and the heart's adenosine triphosphate supply will dry up. (This may explain why the collapse is so dreadfully refractory to conventional treatment).

Just how lipid reverses local anaesthetic toxicity is also not known. It may act simply as a circulating lipid sink, drawing local anaesthetic out of the plasma. Alternatively, the fat rush may overwhelm the inhibition of the translocase by mass action. This would increase the myocardium's energy supply and so its susceptibility to resuscitation.

Based on these findings, Weinberg has recommended a dose regimen for clinical use (see Box) [3]. But, how safe is it to apply these findings to humans? Sceptics will point out four weaknesses in Weinberg's evidence. First, his group have experimented with racemic bupivacaine. Would lipid work as well if ropivacaine or levo-bupivacaine were responsible for the collapse? We don't know; perhaps the corresponding experiments will one day be done. Meanwhile what we do know suggests that lipid may be an effective antidote to overdoses of the single isomer preparations, too. After all, both bupivacaine enantiomers are similarly hydrophobic, and ropivacaine also is known to inhibit carnitine acylcarnitine translocase – albeit less potently than racemic bupivacaine [4].

Second, there is some evidence that intravenous boluses of lipid can adversely affect heart and lungs, causing pulmonary vasoconstriction in particular [5]. Is this a reason now to eschew lipid as an antidote for otherwise overwhelming intoxication by local anaesthetic? We believe not: the animal evidence is that lipid does very much more good than harm.

Third, Weinberg's resuscitation did not include epinephrine, which would surely be used in humans according to current protocols. Would the advantages of lipid persist if used in conjunction with epinephrine? We don't know; the experimental difficulty is that small animals intoxicated by local anaesthetics are relatively easily resuscitated with epinephrine, whereas humans are not. To demonstrate that any adjunct to epinephrine has statistically significant benefits, the experimenter can either sacrifice large numbers of animals, or omit epinephrine from resuscitation – the more humane, economic and realistic option.

Fourth, if small animals react differently to local anaesthetic overdose, is it appropriate to base human clinical practice on these animal studies alone? The stark truth is that we may have no choice. There are currently no case reports of lipid emulsion use in human resuscitation. The very unpredictability, rarity and severity of overwhelming intoxication with local anaesthetic make controlled and ethical human experimentation impossible. Ethical human experimentation would only allow us to gauge the effects of lipid in volunteers who are relatively mildly intoxicated with local anaesthetic; we have no way of knowing how this would compare with its use in overwhelming cardiovascular collapse.

If such scepticism is unfounded, some caution may be appropriate. The regimen suggested by Weinberg may be refined. Further experiments may better define the optimal dosage. It may even be that in this peculiar breed of cardiac arrest, the large boluses of epinephrine prescribed by protocol do more harm than good. And, of course, the long-term effects of a large bolus of lipid emulsion have not been studied in humans – though they could hardly be worse than a failed conventional resuscitation. In short, these are reasons eagerly to await refinements, but not to deny lipid's possible benefit to patients in the meantime.

Do any other likely treatments already exist? Bretylium was long quoted as a candidate, but it is no longer available in the UK. The single published study of amiodarone's use in this setting is inconclusive [6]. Propofol is not a suitable alternative because to give a sufficient dose of its carrier lipid, an overdose of propofol would be necessary. A Korean group has shown that combined boluses of glucose, insulin and potassium work in a dog model [7]. But the dose of insulin they propose is large −2−1. So, while their results are impressive, applying their protocol to clinical practice must be fraught with risk. Few anaesthetists would lightly give 140 iu of insulin stat intravenously to anyone, let alone a collapsed patient with an ischaemic brain.

If the evidence in favour of lipid is so persuasive, why has word not spread? Perhaps there are two answers. First, although some of Weinberg's pertinent papers have appeared in Anesthesiology, accounts of lipid as an antidote have been published in Regional Anesthesia and Pain Medicine, a journal which – lamentably – is not widely read in the general anaesthetic community.

A second answer is more complex. Occasionally a treatment such as this emerges which appears to benefit patients in extremis, and therefore is not susceptible to standard prospective randomised controlled trials. There is currently no clear mechanism for publicising such therapies. Specific issues of treatment are clearly beyond the remit of the Royal College of Anaesthetists. The Association of Anaesthetists does not have the means to review all possible novel treatments, and nor do other learned societies. The National Institute for Health and Clinical Excellence might seem the appropriate body, but its ambitions far outstrip its funding (the authors' own approaches have over 6 months elicited no more than standardised acknowledgement). If the usefulness of lipid is established, it may be adopted in the guidelines of the European Resuscitation Council, but that is a long way off.

The dilemma of the Gordian knot lies in the evidence: high grade human evidence will never be available, yet, without recommendation from some authoritative institution or alternative publicity, little low grade evidence will be published. and without at least this, institutions will be reluctant to promulgate guidelines.

The knot may perhaps only be cut by individual clinicians in league with journal editors. Clinicians should be ready to test such treatments, and publish their experience; editors should welcome such case reports. Both may seem anathema now that high-grade scientific evidence has become such a gold standard, but otherwise patients will be denied what may be useful treatment.

Interestingly, dantrolene followed a similar path. First it seemed beneficial in a porcine model; second, it was given to all suspected malignant hyperthermia patients in a multicentre collaborative study [8] and then, third, became orthodox treatment recommended by national institutions. It never was subjected to a randomised prospective controlled trial. The use of lipid as an antidote to intoxication by local anaesthetics may be poised between first and second steps – between use in experimental animals and use in humans in extremis.

So we believe that lipid emulsion, along with Weinberg's protocol, should be immediately available wherever patients receive large boluses of local anaesthetic (in practice, labour wards and theatre suites), and that patients in cardiac arrest attributable to overdose with local anaesthetic should be resuscitated following current advanced life support guidelines, with the addition of a lipid emulsion.

Introducing lipid emulsion is cheap and easy. Intralipid®; 20% costs under £20 (€30) per 500 ml bag and has a long shelf life. Hospital pharmacies will often be able to help with timely rotation of bags.

In short, we exhort all readers to read the original papers, and then get lipid into their theatre suites and labour wards now. Fat may be bad for you, but it may just once be very good for your patient – and whether it is or not, tell us all.

Weinberg's dose regimen for use in humans [3]

In cardiac arrest secondary to local anaesthetic toxicity which is unresponsive to standard therapy, intravenous administration of a lipid such as Intralipid® 20% is recommended in the following regimen:

  • 1give 1−1 over 1 min;
  • 2repeat twice more at 3–5 min intervals;
  • 3then (or sooner if stability is restored), convert to an infusion at a rate of 0.25−1.min−1, continuing until haemodynamic stability is restored;
  • 4increasing the dose beyond 8−1 is unlikely to be useful;
  • 5in practice, in resuscitating an adult weighing 70 kg:
  •  • take a 500-ml bag of Intralipid® 20% and a 50-ml syringe;
  •  • draw up 50 ml and give it stat intravenously, and then draw up and give another 20 ml;
  •  • do exactly the same thing up to twice more as you give epinephrine – if necessary or appropriate;
  •  • then attach the Intralipid bag to a giving set and run it intravenously over the next 15 min.