The July 2004 issue of Anesthesiology contained a report by Estebe and Myers1demonstrating neurotoxic effects of amitriptyline when injected in high doses immediately adjacent to the rat sciatic nerve. Local anesthetic properties have been demonstrated with high doses of amitriptyline,2,3and local anesthetics also have been shown to produce neurotoxicity after similar methods of administration.4The demonstration of neurotoxic effects following local administration is an important reminder of the need for careful assessment of novel routes of drug administration.

Unfortunately, there is a significant error in the calculation of the total dose administered in the Estebe and Myers1study. Thus, they inject 0.2 ml of 25 mg/ml or 79.6 mm amitriptyline as their highest dose and compute this to correspond to a dose of 16 nmol. However, this dose actually corresponds to a dose of 16 μmol; therefore their dose computation is in error by a factor of 1000 (table 1, section A).1The doses administered are comparable to those administered in demonstrating local anesthetic properties of amitriptyline (table 1, sections B1, B2)2,3but are significantly higher than those administered in producing antinociception (table 1, section C).5,6Much of their discussion regarding amitriptyline systematically perpetuates the thousand-fold error in calculations, and this results in some very misleading considerations.

Table 1. Doses of Amitriptyline Administered Locally Immediately Adjacent to the Sciatic Nerve (A,B1) or Locally into Tissue (B2,C) to Determine Local Anesthetic Properties or Antinociceptive Actions 

Table 1. Doses of Amitriptyline Administered Locally Immediately Adjacent to the Sciatic Nerve (A,B1) or Locally into Tissue (B2,C) to Determine Local Anesthetic Properties or Antinociceptive Actions 
Table 1. Doses of Amitriptyline Administered Locally Immediately Adjacent to the Sciatic Nerve (A,B1) or Locally into Tissue (B2,C) to Determine Local Anesthetic Properties or Antinociceptive Actions 

After peripheral administration, by injection locally into the dorsal or plantar surface of the hindpaw, amitriptyline produces antinociception in rat models of ongoing (formalin test) or neuropathic pain (spinal nerve ligation model).5,6Efficacy of locally administered amitriptyline has also been demonstrated in a rat model of diabetic neuropathy.7In each of those studies, doses of 100 nmol amitriptyline produced antinociception against spontaneous (formalin) and evoked (thermal, mechanical) behaviors. This dose, however, had no effect on thermal latencies in the uninjured state8and did not produce tissue edema.9Increasing the dose of amitriptyline to 300 and 1000 nmol leads to increases in thermal thresholds in the uninjured state, which may reflect involvement of local anesthetic properties.9These higher doses also cause tissue edema that, at the 1000 nmol dose, persists to some degree at 24 h, but mechanisms involved in edema are unclear.9 

The Estebe and Myers1study evaluates doses of 2,000–16,000 nmol amitriptyline injected immediately adjacent to the sciatic nerve in rats. When amitriptyline up to 100 nmol is administered locally into the hindpaw of rats in models of antinociception, it is administered at a much lower total concentration and into tissue that is not necessarily in the immediate vicinity of the nerve. There is no data to suggest that neurotoxic or overt local anesthetic properties are involved in antinociception at doses up to 100 nmol. Indeed, the actions of amitriptyline at these doses are substantially blocked by methylxanthine adenosine receptor antagonists,5–7and it is very unlikely that such antagonists would block neurotoxic or local anesthetic actions. This issue is raised in the context of implicating a receptor-operated mechanism in the action of amitriptyline rather than in the context of implicating any single particular mechanism in its action. Thus, amitriptyline is a complex drug with a range of pharmacological actions (given acutely it blocks noradrenaline, 5-HT, and adenosine uptake, inhibits α-adrenergic, histamine H1, 5-HT2, and N -methyl-d-aspartate receptors, and blocks Na+, Ca2+, and even K+channels), and many of these effects could contribute to peripheral antinociception. Clearly, any form of topical application of amitriptyline to the skin in humans will need to proceed with due caution regarding the potential for local toxicity.

Dalhousie University, Halifax, Nova Scotia, Canada.

Estebe JP, Myers RR: Amitriptyline neurotoxicity: Dose-related pathology after topical application to rat sciatic nerve. Anesthesiology 2004; 100:1519–25
Gerner P, Mujtaba M, Sinnott CJ, Wang GK: Amitriptyline versus bupivacaine in rat sciatic nerve blockade. Anesthesiology 2001; 94:661–7
Khan MA, Gerner P, Wang GK: Amitriptyline for prolonged cutaneous analgesia in the rat. Anesthesiology 2002; 96:109–16
Myers RR, Kalichman MW, Reisner LS, Powell HC: Neurotoxicity of local anesthetics: Perineurial permeability, edema, and nerve fiber injury. Anesthesiology 1986; 64:29–35
Sawynok J, Reid AR, Esser MJ: Peripheral antinociceptive action of amitriptyline in the rat formalin test: Involvement of adenosine. Pain 1999; 80:45–55
Esser MJ, Sawynok J: Caffeine blockade of the thermal antihyperalgesic effect of acute amitriptyline in a rat model of neuropathic pain. Eur J Pharmacol 2000; 399:131–9
Ulugol A, Karadag HC, Tamer M, Firat Z, Aslantas A, Dokmeci I: Involvement of adenosine in the anti-allodynic effect of amitriptyline in streptozotocin-induced diabetic rats. Neurosci Lett 2002; 328:129–32
Esser MJ, Sawynok J: Acute amitriptyline in a rat model of neuropathic pain: differential symptom and route effects. Pain 1999; 80:643–53
Oatway M, Reid A, Sawynok J: Peripheral antihyperalgesic and analgesic actions of ketamine and amitriptyline in a model of mild thermal injury in the rat. Anesth Analg 2003; 97:168–73