To the Editor:—
The article by Taniguchi et al. 1was informative and provocative. However, we take issue with their contention that the data published by Gissen et al. 2have little, if any, relevance to the clinical situation.
Only one of us (D. H. L.) is a coauthor of that earlier study, but we both agree with all of the criticisms of Taniguchi et al. of our methodology. However, the intent of Gissen et al. was to do exactly what Taniguchi et al. criticized us for doing—overwhelming neurons with potentially injurious agents. We did that to simulate a catastrophic situation that would occur clinically, namely the accidental intrathecal injection of an amount of low-pH chloroprocaine that contained sodium bisulfite intended for epidural delivery.
Taniguchi et al. made six criticisms of the study of Gissen et al. It is true that (1) “experiments were conducted on isolated segments of nerve that lack a cell body, a blood supply, and normal physiologic defenses” and (2) “the model is, by nature, unstable, and conduction will deteriorate and fail without intervention within a few days.” Nonetheless, it is also true that it was experiments with a similar in vitro preparation (the isolated frog sartorius-sciatic nerve) that provided the initial data that resulted in today’s clinical use of muscle relaxants in anesthesia practice. Both therapeutic and toxicologic events that occur in vivo can often be simulated by drug exposure of isolated tissues in vitro . For this reason alone, the authors should not be so quick to condemn the data of Gissen et al.
Nevertheless, even the four remaining criticisms that Taniguchi et al. argue do not detract from the value of the study of Gissen et al. :
“Conduction failure (as used by Gissen) is an imperfect endpoint.” Inasmuch as most of the clinical sequelae from exposing nerve tissue to high concentrations of intrathecal local anesthetics, inter alia , are neurologic deficits, they most likely arise from conduction failure, so this seems a logical physiologic endpoint to measure.
“It is difficult to know relevant concentrations in an in vitro system devoid of normal physiologic processes.” This is true, but does not preclude an investigation. The concentration of putatively toxic substances in the system used by Gissen et al. was assumed to be equal to the concentration-injected intrathecally that likely caused the cauda equina syndrome. Although it is true that the isolated vagus nerve in a physiologic solution is not in the same environment as, for example, nerve roots passing through the cerebrospinal fluid near blood vessels and other drug adsorbing tissues, we do not know the extent to which the “normal physiologic” processes in vivo are compromised by the toxic actions of the deleterious substances.
“Nerves are exposed to a bath containing undiluted bisulfite …an exposure that is not likely to occur in vivo , in which …cerebrospinal fluid buffers are present.”
“The composition of the in vitro bath remains relatively constant over time because it lacks redistribution or any appreciable uptake.”
Both statements 5 and 6 rest on an assumed mixing pattern of the injected solution over time. When a large volume of concentrated drug intended for the epidural space is accidentally injected intrathecally, over a short time, a relatively high drug (or adjuvant) concentration may be present around the spinal roots for minutes. Little hydrodynamic mixing occurs after the initial bolus injection,3,4and the diffusion of substances that controls their dilution occurs on the same time scale as their penetration into nerve tissue, the likely site for toxic actions. Even if these conditions exist for only several minutes, that is potentially long enough to cause cauda equina syndrome. In fact, another study has shown that as little as 3 min exposure to 5% lidocaine can cause irreversible nerve conduction failure.5Furthermore, that is exactly what Taniguchi’s coauthor believed happened when excessive amounts of 5% lidocaine caused cauda equina syndrome (conduction failure) during continuous spinal anesthesia.6,7
Just as Taniguchi et al. criticizes the methodology of Gissen et al. for not being “clinical or physiologic,” a recent editorial8raised similar concerns about the methodology of Taniguchi et al. Just how clinical and physiologic is the continuous 2-h intrathecal infusion that Taniguchi et al. used? What clinical (and physiologic) scenario does that represent? It represents a nonideal way of studying a rare complication8and in that sense it is similar to the approach of Gissen et al.
In fact, this particular situation is one example of the larger strategic question: How do we account for the causes of the occasional adverse clinical events that are clearly not the fault of the physician’s technique but occur frequently enough to suggest causative linkages with drugs, adjuvants, or devices? There can be no prospective approaches, and retrospective studies all suffer from the pitfall of low numbers of events and heterogeneous patient populations with differing demographics and anatomical variabilities that are invisible to physical examination. Carefully conducted animal studies with models closest to the clinical circumstance provide the opportunity to simulate the clinical sequelae and thereby validate the approach, but these experimental effects are often difficult to explain mechanistically. In vivo recordings of neuronal activity for the long times over which toxic effects may develop is extraordinarily expensive and unlikely to find funding from the National Institutes of Health, the pharmaceutical industry, or professional societies. The next best approach to reach a mechanistic explanation is to study simpler systems, such as the isolated nerve in vitro . We suggest to those who use behavioral phenomenology to study the toxicity of intrathecal agents that their knowledge would be well advanced by electrophysiologic investigations of the nerves exposed to drug, either in vivo or in vitro , and we would be glad to advise them on how to conduct such studies.
Although not scientific, it is nevertheless noteworthy that after the publication of the article of Gissen et al. and the removal of sodium bisulfite by manufacturers from the epidural chloroprocaine formulation, there have been no reports of chloroprocaine-induced cauda equina syndrome. There are probably several reasons for this observation, such as the now routine slow and incremental epidural injections. Therefore, although this is not proof of the safety of chloroprocaine per se , it is inconsistent with the conclusion of Taniguchi et al. that “clinical deficits associated with unintentional intrathecal injections of chloroprocaine likely resulted from a direct effect of the anesthetic, not the preservative.” Before making this claim, should Taniguchi first simulate in rats the same clinical conditions9–11(i.e. , the intrathecal injection of massive amounts of chloroprocaine, bisulfite, or both) that prompted Gissen et al. to do their study?
* Boston University Medical School, Boston Medical Center, Boston, Massachusetts. firstname.lastname@example.org