Neosaxitoxin as a Local Anesthetic: Preliminary Observations from a First Human Trial

Ten healthy male volunteers aged 18–32 yr were enrolled. The study was approved by the University of Chile Clinic Hospital Ethics Committee, and written informed consent was obtained from all the participants. Volunteers were interviewed about their health history and underwent a physical examination by Clinic Hospital physicians. Inclusion criteria were healthy men with the ability to understand and respond to the tests performed. Exclusion criteria were patients who consumed any oral analgesics at least 10 days before the study, patients with drugs abuse history, and patients showing any sign of a psychiatric disorder during the clinical examination. These clinical trials were performed with anesthesiologists in attendance and in a location with immediate access to the operating room, with facilities for respiratory and hemodynamic support.

The study was performed as a double-blind, placebo-controlled, randomized trial. Each volunteer acted as his own control. The volunteers were injected in the posterior middle skin of right and left calves. Subcutaneous injections were made, evenly covering a marked quadrangular area (40 × 40 mm) in a standardized fan-like fashion. The injections were made at the two opposite angles of the square, injecting a total of 10 ml solution. A 50-μg dose of neosaxitoxin in a saline solution of 0.9% NaCl was used as a treatment, and saline solution was used as the placebo. A computer-generated randomized table was used to choose the leg to be injected with either the placebo or the toxin in each volunteer. Neosaxitoxin was purified from toxic shellfish collected in Southern Chilean fjords. Toxin purity (98%) was determined by high-performance liquid chromatography with online fluorescence detection and online mass spectroscopy analysis.14,29The toxin was diluted in 0.9% saline solution. No other additives were used. The doses were produced in the Laboratorio Bioquímica de Membrana, Facultad de Medicina, Universidad de Chile with the approval of Instituto Nacional de Salud (National Institutes of Health, Santiago, Chile). The areas were infiltrated with 10 ml of the assigned treatment using a 0.8 × 50-mm (21-gauge) needle. Injections and preparations of syringes were performed by different researchers who were not involved in the sensory testing. Pain on injection of the solution, not needle insertion, was evaluated using a 0-10 visual analog scale. This scale was anchored by the descriptions “no pain” (0) and “worst pain imaginable” (10).

Assessment of pain and sensory thresholds was made before the injections (baseline) and 1, 3, 6, 9, 12, 16, 24, and 48 h after the injections. All sensory testing was performed at the same time of the day in an isolated quiet room with a controlled temperature of 23°C. The subjects were resting in a relaxed position with their eyes closed during all assessments.

The touch detection thresholds were determined by mechanical stimuli with a series of monofilaments of different strength. Ten hairs that covered the range from 0.1 to 100 force grams on a logarithmic scale (Touch Test Sensory Evaluators; Stoelting Co., Wood Dale, IL) were used. The touch detection threshold was defined as the least force of mechanical stimulation that produced a sensation of touch or pressure. Eight stimuli covering the infiltrated area were made with each hair, beginning with the lightest of the 10, until at least one half of the stimulations with one hair caused the specified sensation. The eight stimuli were applied with a rate of approximately 1 Hz.

Thermal thresholds were determined by using a computerized contact thermode (TSA II Neurosensory Analyzer; Medoc Ltd, Minneapolis, MN). All thresholds were evaluated using a 30 × 30-mm thermode. All thermal thresholds were determined as the average of three assessments performed at 10-s intervals, from a baseline temperature of 32°C and with a rate of change of 1°C/s. The upper cutoff limit was 50°C, and the lower cutoff was 0°C. Cold and warm detection thresholds were defined as the smallest change from baseline that the volunteer could perceive, and the volunteer pressed a button as soon as the specific sensation was perceived. The heat and cold pain detection threshold was the temperature perceived as painful, and the volunteers were instructed to react to the first sensation of pain.

Localized reactions at the injection site (erythema, discoloration, hematoma, induration, swelling, and blisters) and neosaxitoxin intoxication symptoms such as nausea, headache, ataxia, perioral, and distal limb paresthesia were recorded. Two weeks after the injections, the volunteers returned to the Clinical Hospital where physicians evaluated the injection sites for persistent and delayed reactions and the volunteers were questioned regarding any abnormality. Two months after the injections, the volunteers were contacted by telephone and asked about any trouble or incidents that they thought might be related to the study applications.

Blood and urine samples were taken 1 h and 4 h after the injection to determine the amounts of neosaxitoxin with the high-performance liquid chromatography technique using online fluorescence.14,29 

Normality of data was evaluated by using the Shapiro-Wilk test. Comparisons were made by using Student t  test for parametric data and the Mann–Whitney U test for nonparametric data. These analyses were conducted by using Stata Statistical Software (Stata Corp., College Station, TX). P  values less than 0.05 were considered statistically significant. Neural blockade extent was defined as the period with significantly reduced sensory sensitivity compared with baseline values. Assuming a two-sided type I error protection of 0.05 and a power of 0.90, 10 patients in each of the groups were required to reveal a reduction in mean of perception of 3°C, assuming a SD of 2. Mean ± SD values are reported in figures 1–5.

None of the volunteers presented symptoms of neosaxitoxin intoxication. No toxin could be detected in blood and urine samples, using a cutoff limit of toxin detection by online high-performance liquid chromatography technique of 1 × 10⁻¹²mol. Four volunteers had small hematomas in the infiltration zone (two in the toxin group and two in the placebo group) at the 24-h evaluation; all disappeared after 2 weeks. No other local reactions were presented. None of the volunteers noted any motor disability or discomfort during the follow-up period.

There was no significant difference between the groups in any of the baseline evaluations, giving the values a normal distribution in the five parameters evaluated.

There were no significant differences between the neosaxitoxin and placebo injection pain. The mean pain for the neosaxitoxin injection was 6.5 ± 1.6 and for placebo 5.8 ± 2.1 (P = 0.41). Blinding was considered adequate as no differences were observed.

The five parameters presented a complete block at 1 h from the infiltration, showing P < 0.0001 in all cases. All parameters reverted to the baseline level in a steady way.

All volunteers almost reached the limit of 50°C (mean, 49.1 ± 0.5°C) with no warm sensation (fig. 1). The threshold gradually returned to baseline values 9 h from the infiltration, being significantly elevated for 6 h.

All participants reported complete heat pain blockade 1 h post injection, reaching 50°C with no pain sensation (fig. 2). This parameter presented the fastest return to normal values, as the threshold remained elevated for only 3 h after the injection.

All participants reported a complete cold sensation blockade the first 3 h post injection, exhibiting the threshold down to 0°C, showing no dispersion (fig. 3). Return to the baseline value was measured 12 h after injection.

This parameter took the longest to return to the baseline value (fig. 4). As with the cold sensation, all patients reported a complete block of cold sensation for the first 3 h. The cold pain threshold then slowly returned to its baseline value, achieving a significant 24 h of neural blockade

Data of the applied forces are shown on a logarithmic scale (fig. 5). This parameter presented a significant reduction until 9 h post injection, showing a progressive return to normal values, in accordance with the all the other parameters measured.

This is the first report of neosaxitoxin testing as a local anesthetic in a human randomized trial. Neosaxitoxin is an extremely potent agent, capable of producing, at a microgram level of dosage, clinically sensitive blockage in a completely reversible manner. Typically, local anesthetics like lidocaine and bupivacaine are used in the 25- to 500-μg range, 1,000 to 10,000 times greater than the neosaxitoxin dose tested in the present study.

In this trial, we found a time-dependant reversal of sensory parameters. The heat pain and warm sensation were the earliest to revert, followed by mechanical perception and cold sensation, and lastly the cold pain sensation. In studies of spinal and epidural anesthesia with lidocaine, a sequential temporal recovery of sensations has been noted, with the pinprick being the first to revert, followed by touch and cold.30,31Using transdermal lidocaine applications, it has been reported that pinprick and cold sensations are more strongly affected than the warm sensation measured with the receptor site self-reporting of perceived intensity test.32According to nerve fibers involved in perception, C fibers (small unmyelinated) convey afferent sensory information, including that of warm temperatures and a diffuse pain response, whereas A delta fibers (small myelinated) relay the sharp pain response and cold sensation.33C fibers are the most resistant to local anesthetics, as shown in the rat sciatic nerve model, and have been shown to be less sensitive to lidocaine than A delta fibers.34C fibers are also less sensitive to tetrodotoxin blocking than A delta fibers in the frog sciatic nerve preparation and in the rat dorsal roots.35,36 

A possible physiologic explanation for this phenomenon could be the fact that C fibers have a higher presence of Na^V1.8 and Na^V1.9 (tetrodotoxin-resistant sodium channels) than myelinated fibers.3,37–39In Ranvier nodes, there is also a predominance of Na^V1.6, a tetrodotoxin-sensitive sodium channel.40These data explain the time sequence of sensation reversal shown in this study, in which the warm sensation was the first to revert because of the higher paralytic shellfish poisoning toxin resistance of C fibers.

The anesthetic effect of outer sodium channel-blocking compounds such as tetrodotoxin and saxitoxin has long been known and only been previously tested in animal models5,13,22,41–43. Neosaxitoxin’s anesthetic effect has been studied only in an animal in vivo  preparation, and in this experimental model, neosaxitoxin was the most potent anesthetic when compared with saxitoxin, tetrodotoxin, and dc-saxitoxin.5 

The amide-type local anesthetics like lidocaine, ropivacaine, and bupivacaine show highly toxic effects, mainly cardiovascular (dysrhythmias and cardiac depression), central nervous system toxicity (seizures and coma), and direct nerve cell cytotoxicity.44,45These undesirable clinical effects have stimulated the search for safer local anesthetics to reduce the risk of cardiac and central nervous system toxicity produced by accidental intravenous injection.46–48Neosaxitoxin should have less cardiotoxic effect than amide anesthetic agents because of its lower affinity for cardiac sodium channel receptors. Relative to axons, cardiac Purkinje fibers have extremely low affinity to tetrodotoxin:49the cardiac sodium channels bind tetrodotoxin with 200-fold lower affinity mainly by one amino acid change in domain I, when compared with the binding measured in brain and skeletal channel membrane preparations.1In a model of rats with intrathecally placed catheters, tetrodotoxin also achieved an adequate anesthetic effect without causing local neurotoxicity. This result was opposite that of bupivacaine, which produced moderate to severe nerve root injury when evaluated for histologic damage.22 

The systemic toxicity of the outer pore blockers (tetrodotoxin and saxitoxin) has been described extensively.13,43Unlike that with conventional local anesthetics, death from these toxins results primarily from respiratory failure due to diaphragmatic paralysis. The latter occurs by direct action on the diaphragm and phrenic nerves, rather than a central depressing effect.13Respiratory support is the mainstay of treatment. On average, 24-72 h are usually required for a complete neurologic recovery.50Moreover, severely paralytic shellfish poisoning-intoxicated patients have recovered without neurologic impairment or other clinical damages.51If inadvertent intravenous injection were to occur in the presence of a vigilant anesthesiologist, most episodes could be managed successfully without sequelae by proper attention to ventilation.43Is important to keep in mind that international official regulations allow consumption of shellfish with a safe limit of 80 μg saxitoxin, equivalent to 100 g shellfish meat.‡‡

In conclusion, this is the first report of neosaxitoxin use as local anesthetic in a human trial. Neosaxitoxin showed effective local anesthetic properties, blocking the five sensitive parameters measured in this trial. Ongoing studies using highest doses of neosaxitoxin will show its limits and maximal effects as a local anesthetic compound. Our data will open a whole new line of research in acute and chronic pain management with this phycotoxin.

The authors acknowledge the support of The Henry Mayer Center, Hospital Clínico Universidad de Chile, and thank David Sisco, B.S., Santiago, Chile, for revision of the manuscript.