Lipid Emulsion and Recovery from Local Anesthetic–Induced “Cardiac Arrest”

To the Editor The findings of Zausig et al.1 describing the effects of lipid emulsion infusion on recovery from local anesthetic–induced cardiac toxicity support the general observation that lipid infusion reverses many effects of bupivacaine toxicity, and we agree that discrepancies between their results and those of Weinberg et al.2 in another Langendorff model probably reflect differences in protocol. However, we believe that their key finding that the benefit of lipid emulsion was limited to bupivacaine toxicity and is not applicable to treating ropivacaine or mepivacaine overdose is model specific and directly contradicts published clinical experience.3,4 It is important to point out several factors that limit translating findings in their isolated heart model to the clinical setting. First, the authors correctly state that their model of local anesthetic–induced asystole is distinct from real-life cardiac arrest in that normoxic coronary perfusion is maintained throughout the experiment. Mazoit et al.5 and Weinberg et al.2 have previously shown that myocardial bupivacaine washout occurs very rapidly in the perfused, isolated heart. This implies that buffer flow itself is a major cause of reversing toxicity in a Langendorff model with continued normal flow despite asystole. Therefore, adding lipid emulsion in this in vitro model can only exert a very minor effect by comparison—particularly for less-lipophilic drugs. Hence, the model used by Zausig et al. is biased in its design to yield precisely the results they report. Second, we question the dose and regimen for administering lipid. The authors do not report the final buffer triglyceride concentration and therefore we do not know if sufficient lipid reached the coronary circulation. Because lipid may operate by mass effect, an inadequate concentration of lipid could explain the observed failure to reverse mepivacaine and ropivacaine toxicity more rapidly than controls. Did they give enough lipid? A dose-response study across a range of lipid infusions would have answered this question. The authors did not use an initial lipid bolus, a key element of lipid therapy. Furthermore, they did not offer a pharmacokinetic rationale for infusing the emulsion at 0.25 mL/kg/min, a rate recommended for patients, not the isolated heart. Moreover, the authors chose Lipofundin (B Braun, Melsungen, Germany), a formulation not reported in any published examples of successful lipid resuscitation to date. Lipofundin 20% comprises a 1:1 ratio of medium- and long-chain triglycerides, and Mazoit et al.6 previously showed that a lipid formulation containing medium-chain triglycerides had less than half the local anesthetic binding capacity of a formulation containing 100% long-chain fatty acids. Would a different formulation have yielded different results? Third, the authors implicitly refute alternatives to the lipid sink as contributing mechanisms when they state, “The crucial lipid sink effect only seems to be relevant in bupivacaine-induced cardiac toxicity, but not in mepivacaine- or ropivacaine-induced cardiac toxicity.” This seems to preclude the possibility that lipid infusion could reverse toxicity caused by ropivacaine, mepivacaine, or other less-lipophilic local anesthetics. However, there is ample evidence of exactly this phenomenon.3,4,7,8 We assert that it is equally likely that mechanisms other than simple partitioning, such as oxidative metabolic and positive inotropic effects, might contribute to and explain more completely the phenomenon of lipid resuscitation, particularly when involving the less-lipophilic local anesthetics. Finally, we believe the term “cardiac arrest” in the article title is highly misleading, because the study only presented data from an isolated heart model of local anesthetic toxicity. Although asystole in the isolated heart might technically be “cardiac arrest,” we're concerned that the casual reader might conclude that the authors' findings apply to resuscitating a patient. Cardiac toxicity in a Langendorff preparation is entirely distinct from cardiac arrest in vivo, and many important differences preclude the direct translation of data derived from an isolated organ to the clinical setting. Accumulating animal studies9,10 and clinical experience3,4,7,8 clearly demonstrate the effectiveness of lipid resuscitation with a variety of local anesthetics and lipid emulsions.4,8 Perhaps we will ultimately find that particular lipid emulsions are better suited for specific local anesthetics, and the work of Zausig et al. is a start in this direction. Lisa Warren, MD Department of Anesthesia, Critical Care and Pain Medicine Massachusetts General Hospital Boston, Massachusetts [email protected] Guy Weinberg, MD Department of Anesthesiology University of Illinois at Chicago Jesse Brown VA Medical Center Chicago, Illinois