In Reply:--Tinker and Baker disagree with our analysis that "neither peak systemic fluoride concentrations nor duration of fluoride increase alone can be applied nonselectively to all anesthetics to explain or predict nephrotoxicity." We believe the data support our statement: Enflurane anesthesia in isoniazid-treated humans resulted in peak plasma fluoride concentrations exceeding 50 micro Meter (as great as 130 micro Meter), but there was no evidence of renal dysfunction. Prolonged isoflurane anesthesia resulted in peak plasma fluoride concentrations exceeding 50 micro Meter for 2-3 days but had no deleterious effect on any measure of renal function. Prolonged isoflurane sedation resulted in peak plasma fluoride concentrations exceeding 50 micro Meter (as great as 93 micro Meter) but no adverse effects on renal function. During prolonged isoflurane sedation, in which fluoride concentrations remained increased for as long as 32 days, there were no significant changes in renal function. Sevoflurane anesthesia resulted in peak plasma fluoride concentrations exceeding 50 micro Meter, but no adverse effects on renal function have been observed to date. [5-8]In contrast, enflurane anesthesia can result in significantly diminished urine concentrating ability at plasma fluoride concentrations less than 50 micro Meter. Thus, the methoxyflurane experience does not appear to apply equally to all anesthetics.
Tinker and Baker attribute to us the notion that serum fluoride is no longer important in nephrotoxicity. We have made no such assertion.
Tinker and Baker claim that we "suggest, without proposing any mechanism, that the small amount of fluoride produced in the kidney is relevant to nephrotoxicity, whereas the large amount of serum fluoride that passes through the kidney for excretion is irrelevant." There is no such statement in our paper, and furthermore, there are no data on which to argue the point. Renal parenchymal fluoride concentrations in vivo resulting from either renal anesthetic metabolism or tubular fluoride reabsorption have never been measured with methoxyflurane or any other volatile agent. Tinker and Baker are correct in that we proposed no mechanisms of nephrotoxicity. We did not propose any mechanisms of nephrotoxicity because we did not study nephrotoxicity--we studied metabolism.
Tinker and Baker reject the potential that a metabolite or metabolic consequence of methoxyflurane biotransformation other than plasma fluoride may contribute to nephrotoxicity because, "after many years of methoxyflurane study, none has been found." However, there has been scant study of methoxyflurane nephrotoxicity in the last two decades. The absence of proof is not the proof of absence. Indeed, in one of only two papers published since 1980 which even remotely address this issue, the use of analytical methodologies not available during the methoxyflurane era led to a reevaluation of methoxyflurane hepatic metabolism. .
Methoxyflurane nephrotoxicity is intimately and unquestionably related to biotransformation. Methoxyflurane is biotransformed to a number of metabolites. Identification of fluoride as the nephrotoxic metabolite was based on associations between serum fluoride concentration and toxicity in humans; on correlations between changes in metabolism, serum fluoride concentrations, and nephrotoxicity in rats; and on the ability of fluoride (at unknown serum concentrations) to cause toxicity in animals. However, data in humans establishing a causal link between increased serum fluoride concentrations and nephrotoxicity of methoxyflurane or any other anesthetic has never been published. The clinical observations about enflurane, isoflurane, and sevoflurane cited above are pertinent. They call into question the appropriateness of applying a fluoride hypothesis developed to explain methoxyflurane nephrotoxicity nonselectively to all anesthetics without supporting data. We seek to understand the basic mechanisms of anesthetic nephrotoxicity in the post-methoxyflurane era.
We presented experimental results that human kidneys can metabolize volatile anesthetics. We presented a hypothesis that intrarenal anesthetic metabolism may contribute to nephrotoxicity. All the associative data for methoxyflurane metabolism and toxicity are congruent with such a hypothesis. Formation of another nephrotoxic metabolite with fluoride, whether hepatically or renally, would be congruent with the associative data for methoxyflurane metabolism and toxicity. This, too, would be a hypothesis.
We carefully presented these new hypotheses as such, not as conclusions. We look forward to the rigorous testing and confirmation or refutation of these, or other hypotheses, toward the goal of elucidating the mechanism(s) of volatile anesthetic nephrotoxicity. We hope the editorial by Brown and the letter from Tinker and Baker will draw attention to this issue and stimulate further investigation.
Evan D. Kharasch, M.D., Ph.D., Douglas Hankins, M.A., Kenneth Thummel, Ph.D., Department of Anesthesiology, RN-10, University of Washington School of Medicine, Seattle, Washington 98195.
(Accepted for publication April 29, 1995.)