Free
Meeting Abstracts  |   April 2008
Cutaneous Amitriptyline in Human Volunteers: Differential Effects on the Components of Sensory Information
Author Affiliations & Notes
  • Christian Dualé, M.D., Ph.D.
    *
  • Julie Daveau, M.B.
  • Jean-Michel Cardot, Pharm.D., Ph.D.
  • Anne Boyer-Grand, Pharm.D.
    §
  • Pierre Schoeffler, M.D.
  • Claude Dubray, M.D., Ph.D.
    #
  • * Hospital Practitioner, Centre Hospitalier Universitaire (CHU) Clermont-Ferrand, Service d’Anesthésie-Réanimation, Hôpital Gabriel-Montpied; CHU Clermont-Ferrand, Centre de Pharmacologie Clinique; Inserm, CIC 501. † Anesthesiology Resident, CHU Clermont-Ferrand, Service d’Anesthésie-Réanimation, Hôpital Gabriel-Montpied. ‡ Professor of Biopharmacy, Univ Clermont1, Fac Pharmacie, Biopharmacie. § Lecturer, CHU Clermont-Ferrand, Pharmacie Centrale, Hôpital Gabriel-Montpied; Univ Clermont1, Fac Pharmacie. ∥ Professor of Anesthesiology and Intensive Care, CHU Clermont-Ferrand, Service d’Anesthésie-Réanimation, Hôpital Gabriel-Montpied; Univ Clermont1, Fac Médecine. # Professor of Clinical Pharmacology, CHU Clermont-Ferrand, Centre de Pharmacologie Clinique; Univ Clermont1, Fac Médecine; Inserm, CIC 501.
Article Information
Meeting Abstracts   |   April 2008
Cutaneous Amitriptyline in Human Volunteers: Differential Effects on the Components of Sensory Information
Anesthesiology 4 2008, Vol.108, 714-721. doi:10.1097/ALN.0b013e3181672632
Anesthesiology 4 2008, Vol.108, 714-721. doi:10.1097/ALN.0b013e3181672632
TRICYCLIC antidepressants (TCAs) are effective in relieving neuropathic pain, the best evidence being available for amitriptyline.1,2 In animal neuropathic models, TCAs reduce thermal hyperalgia,3 mechanical allodynia,4 and ectopic discharges.5 Therefore, TCAs are commonly used for neuropathic pain, where opiates are classically considered as ineffective at the common doses.6 However, only 30% of neuropathic patients experiencing pain have a reduction to half of the pretreatment pain level, and the fairly high incidence of side effects could explain the poor compliance for this treatment.2,7 
The cellular/synaptic action of TCAs is considered to be largely due to inhibition of reuptake of norepinephrine, serotonin,8–10 and adenosin.11,12 Other mechanisms have been also suggested, such as blockade of N  -methyl-d-aspartate receptors,13,14 calcium,15 and sodium voltage-dependent channels.15–20 Animal studies have provided diverse information about the site of action for TCAs on neuropathic pain. Supraspinal effects are supported,21 but part of the supraspinal analgesia may be linked to the relief of associated depressive symptoms.22 Spinal effects are also suggested because intrathecal amitriptyline selectively reduces thermal hyperalgesia by spinal ligation,3 and the analgesic effects of clomipramine are reduced by intrathecal injection of opiate receptor antagonists.21 Recently, action of TCAs on peripheral nerves has been suggested because (1) locally injected amitriptyline has antinociceptive adenosine-dependent effects on inflammatory and neuropathic animal pain models3,23,24; (2) long-lasting local anesthetic effects have been shown after perineural and transcutaneous amitriptyline in rats and humans25–29; and (3) systemic amitriptyline reduces specifically ectopic discharges, which are linked to hyperactivity of sodium channels in the fiber, in the same way as lidocaine.5 
To understand the mechanisms of action of TCAs on sensory fibers, the effects of transcutaneous amitriptyline on the different components of cutaneous sensitivity were studied according to a protocol already validated in human volunteers.29 These data are preliminary to the potential development of topical TCAs in the treatment of neuropathic pain. As secondary endpoints, the systemic absorption of the drug after this type of application was evaluated, as well as the possible residual effects on the skin.
Materials and Methods
The study was conducted in accordance with the Declaration of Helsinki and was approved by the regional Research Ethics Committee (Comité Consultatif pour la Protection des Personnes se prêtant à la Recherche Biomédicale d’Auvergne, Clermont-Ferrand, France). Healthy male patients, aged 18–40, were recruited for the study. This population was chosen to reduce interindividual variability, which could be induced by old age or menstrual cycle. The exclusion criteria were weight out of the 80–120% limits of the Lorentz Ideal Body Weight [height in cm − 100 − (height in cm − 150)/4]; inability to undergo the psychophysical trials; exclusion period from another clinical trial; acute skin disease; excessively hairy back skin; drug, antidepressive, or excessive alcohol consumption; allergy; and contraindications for amitriptyline, lidocaine, and prilocaine. Written information was given to the patient, and a signed consent was collected. All of the patients had a previous training session for psychophysical training by electronic von Frey.
Amitriptyline hydrochloride was prepared in a vehicle composed of water/isopropanol/glycerin (45/45/10) solution (adjusted to pH 8.5 with sodium hydroxide) by the investigational drug service of the university hospital. Four solutions of amitriptyline were prepared: 0 (vehicle), 25, 50, and 100 mm. The vehicle used was based on that used in similar studies involving lidocaine or amitriptyline.29,30 The concentrations were chosen in a range thought to have anesthetic effects,29 with the aim of a total dose (see next paragraph) not exceeding the therapeutic range. The prepared solutions were stored at 4°C in a dark, locked refrigerator. The use of the solutions for the study did not exceed 1 week, and stability was tested over this period (concentration of amitriptyline by ultraviolet spectrophotometry method). The two other treatments were 0.9% NaCl (saline) and lidocaine–prilocaine cream (EMLA; Astra-Zeneca, Rueil-Malmaison, France); these were stored in the same conditions.
Patients were randomly assigned to one order, in which one of the six treatments (saline, vehicle, 25 mm amitriptyline [AMI25], 50 mm amitriptyline [AMI50], 100 mm amitriptyline [AMI100], and EMLA) was allocated to one skin area (numbered 1–6). Because each area was to receive 1.5 ml of the solution (or 2 ml EMLA, which is the recommended amount for 18 cm2), the total doses of active drugs for one session were 82.4 mg amitriptyline, 50 mg lidocaine, and 50 mg prilocaine. Randomization was done by a research assistant who was not involved in the observations, and the allocation was kept in a sealed envelope.
Every day of study, two patients were admitted to the research unit at 8:00 am and were examined for noninvasive hemodynamic parameters (blood pressure and heart rate lying at rest and then in the orthostatic position). A venous cannula was inserted in the arm, and a first blood sample was drawn for amitriptyline assay. A plastic pattern perforated with windows displaying the future areas of treatment was set medially on the back of the patient lying in the prone position, its superior end at the level of the spine of scapula. The borders of the six application areas (numbered 1–6 from left to right and top to bottom) were drawn with a dermographic pencil through the windows. After application of cutaneous alcoholic antiseptic, each area (6 × 3 cm) was bordered by a stuck framework made of precut hydrocolloidal sticking plaster (Algoplaque; URGO, Chenôve, France) to avoid the diffusion of the products outside the area. At T0, the six treatments were applied. In one of the areas, determined by randomization, 1.5 ml EMLA was applied by syringe. On each of the other five marked areas, a 6 × 3-cm piece of sterile gauze was placed, and the gauze was saturated via  syringe with 1.5 ml saline, vehicle, and amitriptyline solutions. Overlapping occlusive dressings (Tegaderm, 9 × 10 cm; 3M Santé, Cergy-Pontoise, France) were fixed to the skin. After 1 h, the dressings and gauze were removed (T1), and each area was lightly wiped with absorbent tissue. Because skin redness has been reported with amitriptyline,29 redness was scored (0–10) after visual comparison to a handmade toner (with Microsoft Office PowerPoint 2003; Redmond, WA). To mask the possible redness and to ensure double-blind testing, each area was then painted with eosin diluted in normal saline, and the border of the area was drawn again with dermographic pencil. The opening of the envelope for randomization, the application of the treatments, the assessment of redness, the removal of dressings, and the eosin application were performed by the same investigator, who did not participate in the rest of the study.
The subjects were allowed to move freely in the clinical investigation center between observations and had a standard lunch at midday. At the times T1, T1 + 2 h, T1 + 4 h, T1 + 6 h, T1 + 8 h, and T1 + 24 h, blood samples for amitriptyline dosage were drawn, hemodynamic data and general tolerance were noted, and psychophysical studies were performed. Sensory testing was performed in the following order: touch, mechanical nociception, cold, warm, and heat. For each test, the areas were examined in the same order (1–6). Each kind of monitoring was performed by a dedicated examiner. For the mechanical tests, the subject sat on a stool, and the examiner sat behind. For thermal testing, the subject lay on a bed in a prone position. After the T1 + 8 h examination and checking of discharge criteria, the venous cannula was withdrawn and the subject was allowed to go home. Subjects returned to the research unit at T1 + 24 h. The last blood sample was drawn by direct venous puncture.
The tactile threshold for mechanical static (punctate) stimuli was assessed using calibrated (0.008–300 g/mm2) von Frey filaments (Bioseb, Chaville, France). Care was taken to avoid stroking the skin with the hair and to apply only a pressure stimulus. The sensations evoked by touching hair were not taken into account. Filaments were applied to the designated point on the skin for approximately 1 s. von Frey hair applications were separated by at least 5 s to reduce the likelihood of anticipatory responses. The von Frey filaments were applied in ascending order of stiffness. Tactile threshold was defined as the smallest force (g/mm2) necessary to bend a von Frey hair, which was just perceived as touch for three consecutive times. If tactile pain threshold exceeded hair number 6.65 (300 g/mm2), the sensitivity was censored at that number. Because the scale of the threshold values naturally displays an exponential pattern, a natural logarithmic transformation was made on the measured values for further analysis.
The mechanical pain threshold was assessed using an electronic algometer (electronic von Frey; Bioseb). The strain gauge was connected to a plastic sterile cone (Eppendorf, Hamburg, Germany), the tip of which was applied perpendicularly to the studied skin area. The punctate pressure was gradually increased with a constant slope under visual control of the pressure value up to the detection of mechanical pain threshold. Threshold was defined as the lowest pressure that produced a sensation of pain. Each threshold value was averaged from three separate consecutive measurements at different points in the test area. A cutoff value was set at 500 g/mm2, which was the considered threshold value if no pain appeared at this cutoff.
The Somedic Thermotest apparatus (Somedic AS, Stockholm, Sweden) was used to deliver quantified and reproducible heat impulses via  a thermode to the skin area. The thermode, which was initially set in the middle of the area, was maintained with a loose elastic band surrounding the chest. The thermode temperature could be varied between 10° and 52°C and cools or warms depending on the direction of current applied. Before each test, the resting temperature was set to 30°C (temperature of the neutral skin). The temperature of the thermode was then decreased or increased until the volunteer felt the predetermined sensation (cold, warm, or painful heat). At this point, the decrease or increase in temperature was terminated by the volunteer pressing the handheld button. The temperature attained was recorded by the apparatus, and the thermode temperature returned rapidly to 30°C. This procedure was repeated on four further occasions, the threshold being the mean of the last three responses calculated by the apparatus. The entire process was then repeated using the next area to be explored. For assessment of the cold, warm, and painful heat sensation thresholds, the rate of decrease or increase in temperature was set at 0.4, 0.2, and 1°C/s, respectively.
The blood samples were analyzed by the Laboratory of Toxicology of the Faculty of Medicine of Clermont-Ferrand. Amitriptyline and nortriptyline plasma levels were determined using a specific high-performance liquid chromatographic method. The limit of quantitation was 2 ng/ml for each analyte.
To search for possible residual effects of amitriptyline, the patients came back to the research unit 8 (±1) and 21 (±2) days after the main session for skin examination. The same pattern was used for the identification of the areas. Only the tactile threshold for mechanical static stimuli was assessed using calibrated von Frey filaments.
Sample size calculation was undertaken on the expectation of a local anesthetic effect of the active drugs. This was quantified as an increase in the mechanical pain threshold (+80 g/mm2) on the electronic von Frey. The mean ± SD for a similar population examined in our research unit is 107 ± 5 g/mm2. With initial α= 5% and β= 10%, the sample size was calculated to be 14, considering that each subject was his own control. It was increased to 16 to allow for randomization after a crossover design.
Statistical Analysis
All of the data collected were numeric and were expressed as median and interquartile range. However, the values for redness (scoring on a scale from 0 to 10) and tactile threshold (logeof the measured value in grams) were considered as ordinal data. The normality of the distribution for the other data were checked by the Shapiro-Wilks test. For each parameter, the effects of period, subject, area, and treatment were analyzed by an analysis of variance (or a Friedman test for ordinal data). If a significant difference appeared between treatments (P  < 0.05), a post hoc  analysis was made by a Tukey test (or multiple pair comparison for ordinal data). The statistical analysis was performed with Statistical Analysis Software 9.1 (SAS Institute Inc., Cary, NC), Microsoft Office Excel 2003, and XLStat (Addinsoft, Paris, France). Figures were generated using Microsoft Office Excel 2003, Microsoft Office Paint 2003, and Microsoft Office PowerPoint 2003.
Results
Of the 16 subjects who underwent the study, one of them was excluded for analysis (except for safety parameters) because he showed major difficulties in concentration during the main session. This was confirmed by further analysis of the thermal threshold data, which showed significant intraindividual discrepancies. For the remaining 15 male subjects, the demographic data were as follows (mean ± SD): age, 25.4 ± 4.5 yr; weight, 73.0 ± 5.4 kg; height, 178.6 ± 6.5 cm. The baseline hemodynamic data at rest were as follows (mean ± SD): heart rate, 67.2 ± 13.4 beats/min; systolic arterial pressure, 133.7 ± 10.2 mmHg; diastolic arterial pressure, 64.0 ± 8.5 mmHg. After rapid orthostatism, there was no significant change in the hemodynamic parameters, except for the diastolic arterial pressure, which increased to 75.1 ± 10.0 mmHg (P  < 0.00001).
Throughout the study, the hemodynamic parameters remained stable (P  > 0.05, analysis of variance for repeated measures). One subject, just before the observation time T1 + 6 h, had malaise with sudation, weakness, paleness, and mild hypotension (99/59 mmHg) with a normal heart rate. Hypoglycemia was diagnosed (2.37 mm), and the malaise disappeared 5 min after oral sugar administration. One other subject, 5 min after sitting on the stool at the T1 + 6 h observation, had malaise with paleness, hypotension (80/40 mmHg), and bradycardia (38 min−1). Glycemia was normal (4.84 mm), and vasovagal malaise was diagnosed. This disappeared after 5 min in a lying-down position. In both cases, the electrocardiogram and the amitriptyline blood level were checked in emergency and showed no abnormality, the levels of amitriptyline being below therapeutic levels. As minor events, dry mouth sensation was noted in 8 of the 96 observations (0.8%) of the main session.
The results for skin color are shown in figure 1. Before they were painted with eosin, the skin areas treated with amitriptyline displayed significant redness (P  < 0.0001, Friedman test). The comparison of the confidence intervals for the redness scores confirmed that redness was higher with AMI50and AMI100(compared with saline, vehicle, and AMI25), whereas it was lower with EMLA.
Fig. 1. Scattergrams showing the level of skin redness compared with a dedicated scale, at the end of the application period of the treatments. A significant difference was found (  P  < 0.0001, Friedman test) between treatments.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). A dose effect is suggested (  dotted line  = trend based on logistic regression), although not statistically significant. 
Fig. 1. Scattergrams showing the level of skin redness compared with a dedicated scale, at the end of the application period of the treatments. A significant difference was found (  P  < 0.0001, Friedman test) between treatments.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). A dose effect is suggested (  dotted line  = trend based on logistic regression), although not statistically significant. 
Fig. 1. Scattergrams showing the level of skin redness compared with a dedicated scale, at the end of the application period of the treatments. A significant difference was found (  P  < 0.0001, Friedman test) between treatments.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). A dose effect is suggested (  dotted line  = trend based on logistic regression), although not statistically significant. 
×
Part of the results for tactile sensation is shown in figure 2. A time effect was noted (P  < 0.0001), with a linear tendency of lowering the thresholds from T1 (median = 1 g/mm2) to T1 + 24 h (median = 0.4 g/mm2), in accord with habituation. This time effect was independent of the treatment and was noted with the same tendencies with the inactive treatment saline (P  < 0.0001). At T1, there was a significant effect of treatment (P  < 0.0001), the post hoc  analysis showing higher thresholds with EMLA compared with all of the other treatments. At T1 + 2 h, the effect of treatment was still significant (P  < 0.016), the post hoc  analysis showing higher thresholds with AMI50, AMI100, and EMLA compared with the other treatments. At T1 + 4 h, the significance level was still reached (P  = 0.017), but the post hoc  analysis did not allow us to individualize groups. No effect of the treatment was noted at the later observation times.
Fig. 2. Scattergrams showing the tactile (von Frey hairs;  A  ) and nociceptive (electronic von Frey;  B  ) cutaneous mechanical thresholds after application (T1) of the treatments, in time.  Horizontal line  = median value;  cross  = mean value. The data observed after T1 + 4 h do not appear, because no significant difference between the groups (Friedman test for touch and analysis of variance for mechanical nociception) was noted after this time.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). The letters  a  ,  b  , and  c  signal the groups identified by the  post hoc  analysis (multiple pair comparison or Tukey test). 
Fig. 2. Scattergrams showing the tactile (von Frey hairs;  A  ) and nociceptive (electronic von Frey;  B  ) cutaneous mechanical thresholds after application (T1) of the treatments, in time.  Horizontal line  = median value;  cross  = mean value. The data observed after T1 + 4 h do not appear, because no significant difference between the groups (Friedman test for touch and analysis of variance for mechanical nociception) was noted after this time.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). The letters  a  ,  b  , and  c  signal the groups identified by the  post hoc  analysis (multiple pair comparison or Tukey test). 
Fig. 2. Scattergrams showing the tactile (von Frey hairs;  A  ) and nociceptive (electronic von Frey;  B  ) cutaneous mechanical thresholds after application (T1) of the treatments, in time.  Horizontal line  = median value;  cross  = mean value. The data observed after T1 + 4 h do not appear, because no significant difference between the groups (Friedman test for touch and analysis of variance for mechanical nociception) was noted after this time.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). The letters  a  ,  b  , and  c  signal the groups identified by the  post hoc  analysis (multiple pair comparison or Tukey test). 
×
Part of the results for mechanical nociceptive sensation is shown on figure 2. A time effect was noted (P  < 0.0001), with a linear tendency of decreasing the thresholds from T1 (mean = 357 g/mm2) to T1 + 8 h (mean = 217 g/mm2) and a mild increase at T1 + 24 h (mean = 239 g/mm2), in accord with habituation. This time effect was partly dependent on treatment, because the same tendency was observed with the saline treatment, without reaching significance. At T1, there was a significant effect of the treatment (P  < 0.0001), the post hoc  analysis showing higher thresholds with EMLA compared with all of the other treatments. At T1 + 2 h, the effect of treatment was still significant (P  < 0.002), the post hoc  analysis showing higher thresholds with AMI50, AMI100, and EMLA compared with the other treatments. No effect of treatment was noted at later observation times.
Part of the results for cold sensation is shown in figure 3 and 4. No time effect (independent of treatment) was noted. At T1, there was a significant effect of the treatment (P  = 0.01), the post hoc  analysis showing lower thresholds (i.e.  , lower sensitivity) with AMI25, AMI50, AMI100, and EMLA (mean values = 23.5°, 21.8°, 22.6°, and 20.3°C, respectively) compared with saline and vehicle (mean values = 27.5° and 27.1°C, respectively). The significance level was not reached at later observation times because the threshold increased to normal values in most subjects. However, the thresholds with amitriptyline remained very low at T1 + 2 h and T1 + 4 h in some subjects (fig. 4), and this effect was weaker with EMLA (fig. 3).
Fig. 3. Scattergrams showing the three kinds of cutaneous thermal thresholds (Thermotest; Somedic AS, Stockholm, Sweden) after the end of application (T1) of the treatments, in time (  A  : cold;  B  : warm;  C  : heat).  Horizontal line  = median value;  cross  = mean value. The data observed after T1 + 4 h do not appear, because no significant difference between the groups (analysis of variance) was noted after this time.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). The letters  a  and  b  signal the groups identified by the  post hoc  analysis (Tukey test). 
Fig. 3. Scattergrams showing the three kinds of cutaneous thermal thresholds (Thermotest; Somedic AS, Stockholm, Sweden) after the end of application (T1) of the treatments, in time (  A  : cold;  B  : warm;  C  : heat).  Horizontal line  = median value;  cross  = mean value. The data observed after T1 + 4 h do not appear, because no significant difference between the groups (analysis of variance) was noted after this time.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). The letters  a  and  b  signal the groups identified by the  post hoc  analysis (Tukey test). 
Fig. 3. Scattergrams showing the three kinds of cutaneous thermal thresholds (Thermotest; Somedic AS, Stockholm, Sweden) after the end of application (T1) of the treatments, in time (  A  : cold;  B  : warm;  C  : heat).  Horizontal line  = median value;  cross  = mean value. The data observed after T1 + 4 h do not appear, because no significant difference between the groups (analysis of variance) was noted after this time.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). The letters  a  and  b  signal the groups identified by the  post hoc  analysis (Tukey test). 
×
Fig. 4. Time course of the two kinds of cutaneous thermal thresholds (Thermotest; Somedic AS, Stockholm, Sweden) affected by amitriptyline application (  A  : cold;  B  : heat). Each  line  represents one subject. These diagrams show a tendency to return to normal values at T1 + 8 h, as well as a strong interindividual variability of the sensitivity to amitriptyline, regarding cold sensation thresholds. 
Fig. 4. Time course of the two kinds of cutaneous thermal thresholds (Thermotest; Somedic AS, Stockholm, Sweden) affected by amitriptyline application (  A  : cold;  B  : heat). Each  line  represents one subject. These diagrams show a tendency to return to normal values at T1 + 8 h, as well as a strong interindividual variability of the sensitivity to amitriptyline, regarding cold sensation thresholds. 
Fig. 4. Time course of the two kinds of cutaneous thermal thresholds (Thermotest; Somedic AS, Stockholm, Sweden) affected by amitriptyline application (  A  : cold;  B  : heat). Each  line  represents one subject. These diagrams show a tendency to return to normal values at T1 + 8 h, as well as a strong interindividual variability of the sensitivity to amitriptyline, regarding cold sensation thresholds. 
×
Part of the results for warm sensation is shown on figure 3. No time effect was noted (P  = 0.215). At T1, the analysis of variance did not show any significant treatment effect (P  = 0.113), but there was a tendency for higher thresholds with EMLA (mean = 36.8°C) compared with all of the other treatments (mean value for saline = 34.8°C). No effect of the treatment was noted at the later observation times.
Part of the results for heat sensation is shown in figures 3 and 4. No independent time effect was noted. At T1, there was a significant effect of the treatment (P  = 0.004), the post hoc  analysis showing lower thresholds (i.e.  , higher sensitivity, hyperalgesia) with AMI25(mean = 40.7°C), AMI50(40.1°C), and AMI100(40.1°C) compared with EMLA (42.8°C), saline (43.4°C), and vehicle (43.8°C). The same effect was observed at T1 + 2 h (P  = 0.012). The significance level was not reached at later observation times, as the threshold decreased to normal values in most subjects, but there was still a tendency to difference with amitriptyline at T1 + 4 h (fig. 3). However, the thresholds with amitriptyline remained low at T1 + 2 h and T1 + 4 h in some subjects (fig. 4).
At T1 + 24 h, T1 + 1 week, and T1 + 3 weeks, the tolerance remained excellent. No significant redness or other cutaneous abnormalities were noted at the treated areas, except a mild eczematous reaction to the Algoplaque (only at T1 + 24 h). The tactile thresholds were identical in all areas (P  = 0.27, 0.39, and 0.064 at T1 + 24 h, T1 + 1, week and T1 + 3 weeks, respectively).
At all times and for every subject, blood sample analysis did not show plasma levels of amitriptyline and nortriptyline over the detection threshold of 2 ng/ml. This indicated that no relevant systemic absorption of amitriptyline occurred during the study.
Discussion
It is believed that part of the analgesic action of TCAs on neuropathic pain is the result of a peripheral effect. If this is the case, it may be hypothesized that TCA drugs may be effective when administered in the peripheral field of the involved nerve. This is suggested by animal studies in which locally injected amitriptyline had antinociceptive effects on neuropathic pain models (dorsal root ligation).3,24 However, until now, the attempts of relieving patients’ neuropathic pain with cutaneous antidepressants have been inconclusive because amitriptyline was not superior to the placebo,31,32 whereas some analgesic effects were reported with doxepin.33 
The long-lasting local anesthetic properties of TCAs have been described in preclinical studies25–28 and confirmed in human volunteers.29 In this last study, amitriptyline hydrochloride was prepared as a 45/45/10 water/isopropanol/glycerin solution (adjusted to pH 8.5 with sodium hydroxide). The vehicle used was based on one used in a similar study aimed to test lidocaine.30 Each subject (randomized and blinded) received 0.3 ml of four solutions—0, 10, 50, and 100 mm amitriptyline in vehicle—on the ventral aspect of the upper arm with an occlusive dressing. As the mean pain rate decreased from 8 out of 10 to 4–5 out of 10 at the puncture of the treated area, the authors concluded that amitriptyline induced local anesthesia (i.e.  , suppression of sensations). At the effective concentrations (greater than 50 mm), redness of the skin was often observed. The development of amitriptyline as a local anesthetic could be compromised by reports of axonal damage after perineural administration,34,35 but the project of transdermal administration remains active because no signs of toxicity have been noted in the study on human volunteers.29 For this reason, it was decided to investigate further, with a first-step study of the mechanisms underlying the observed local anesthesia. A psychophysical approach with measurement of thresholds was chosen because it allows a dissociated analysis of the different components of skin sensitivity and seems to be less subjective than assessment of pain on simple stimulus. Because we were expecting a local anesthetic effect, we chose to use EMLA as a positive control.
The results of the current study were somewhat unexpected because the local anesthetic effects of amitriptyline were fairly weak. This was especially true on mechanical tests, in which the effects of EMLA were obviously stronger, and a significant difference between amitriptyline and control was only noted at T1 + 2 h (fig. 2). Furthermore, a strong dissociative effect on the different thermal thresholds was noted, with no effect on warm, impairment of cold, and enhancement of heat sensitivity (fig. 3). The effects observed after amitriptyline application lasted longer than for EMLA, but they remained shorter than those reported by Gerner et al.  ,29 who noted residual anesthesia at the fifth hour. The discrepancies between the two studies might be explained by different assessment methods or by different study populations (the current study considered only young males). There was also a strong interindividual variability of the sensitivity to amitriptyline regarding cold sensation thresholds. This might be due to differing capacities of the skin to absorb molecules, although amitriptyline per se  also has unpredictable analgesic effects when administered systemically in neuropathic patients.2,7 
The local anesthetic effect of amitriptyline is supported by the ability of the compound to block voltage-dependent sodium channels.15–20 However, such properties cannot explain the dissociation observed in the current study. It must be noted that amitriptyline, like other TCAs, has many possible sites of action on the neurone, such as (1) blockade of histaminergic, adrenergic, serotoninergic, muscarinic,36 nicotinic,15 and glutamatergic13,14 receptors; (2) norepinephrine,10,37 serotonin,10,37 and adenosine11,23,24 uptake inhibition; and (3) blockade of voltage-dependent calcium15,38,39 and potassium40–42 channels. Some of these properties may explain our results. Furthermore, the interactions between amitriptyline and thermal receptors (TRPV1–3and TREK for heat, TRPM8and TRPA1for cold) have not been explored yet.43–46 The dissociation between the effects on warm and heat sensations can also be explained by the type of fibers, more than by specific receptors. Indeed, it is known from animal studies that (1) temperatures below 29°C activate dedicated Aδ fibers (“A-cold”), (2) temperatures between 35° and 42°C (nonnoxious range) activate principally low-threshold Aδ fibers and dedicated C fibers (“C-warm”), and (3) temperatures over 41°C (noxious heat) activate multimodal (“C-mechano-heat”) fibers.43,44,47,48 
The current study is only the first step to investigate a concept and cannot be extrapolated at the moment to neuropathic pain patients. The ability of topical amitriptyline to provide pain relief in these patients still needs to be explored, although inconclusive results have been reported with this molecule diluted in cream.31 The different explanations for this failure may be of a physiologic nature, because the site of action may not be the field of the nerve, but the lesioned area. There may also be methodologic reasons, because the patients included in the previously cited study had heterogeneous histories of painful disease, and some of the patients included had already been treated with TCAs orally, without success. Central sensitization may have been involved in the cases of long-lasting neuropathic pain, explaining the poor effect of a peripheral treatment. Other types of peripheral neuropathic pain (such as touch or cold evoked) can be very disabling, causing distress and suffering for individuals, and these are rarely considered in therapeutic trials.49 These could be preferential indications for testing the analgesic properties of topical amitriptyline. Finally, there may be technical reasons for the failure of the study, such as an insufficient concentration (<20 mg/ml) or a bad penetration of the drug through the human skin. Because the skin barrier is made of epidermal cells all linked together by tight junctions, the transdermal penetration is optimal only for lipophilic, nonionized, and unbound molecules. Amitriptyline (molecular weight of hydrochloride = 313.9; pKa = 9.42, nonionized in alkaline medium; logP = 4.64–5.04) has all of these properties, but its penetration is also conditioned by the quality of the vehicle (which must be fat or alcoholic).
Because the transdermal formulation of amitriptyline that we used in the current study seems not to be applicable in the clinical context, creating an acceptable formulation is a necessary step for a future development. The concentration of amitriptyline must be sufficient for good bioavailability at the site of action (i.e.  , the nerve fiber) without risk of massive systemic absorption. All of these challenges must be raised before starting new clinical trials involving neuropathic patients experiencing pain.
The authors thank Jean Chopineau, Pharm.D., Ph.D. (Professor, Centre Hospitalier Universitaire (CHU) Clermont-Ferrand, Pharmacie Centrale, Hôpital Gabriel-Montpied, Clermont-Ferrand, France), François Coudoré, M.D., Pharm.D., Ph.D. (Professor, CHU Clermont-Ferrand, Laboratoire de Toxicologie), Damien Richard, Pharm.D. (Registrar, CHU Clermont-Ferrand, Laboratoire de Toxicologie), Denise Faulks, D.D.S. (Hospital Practitioner, CHU Clermont-Ferrand, Odontologie et Centre de Soins Dentaires), Alain Eschalier, M.D., Ph.D. (Professor, CHU Clermont-Ferrand, Laboratoire de Toxicologie; Inserm, UMR 766, Clermont-Ferrand, France), Sylvia Boulliau, B.S. (Nurse, CHU Clermont-Ferrand, Centre de Pharmacologie Clinique), Sylvie Eschalier, B.S. (Nurse, CHU Clermont-Ferrand, Centre de Pharmacologie Clinique), Marielle Estrade, Ph.D. (Research Assistant, CHU Clermont-Ferrand, Centre de Pharmacologie Clinique), Monique Bardin (Research Assistant, Univ Clermont1, Fac Médecine, Clermont-Ferrand, France), Geneviève Lafontaine, B.Pharm. (Research Assistant, CHU Clermont-Ferrand, Centre de Pharmacologie Clinique), and Nadia Mellier (Research Assistant, CHU Clermont-Ferrand, Centre de Pharmacologie Clinique).
References
Dubinsky RM, Kabbani H, El Chami Z, Boutwell C, Ali H: Practice parameter: Treatment of postherpetic neuralgia: An evidence-based report of the Quality Standards Subcommittee of the American Academy of Neurology. Neurology 2004; 63:959–65Dubinsky, RM Kabbani, H El Chami, Z Boutwell, C Ali, H
Saarto T, Wiffen PJ: Antidepressants for neuropathic pain. Cochrane Database Syst Rev 2005; CD005454Saarto, T Wiffen, PJ
Esser MJ, Sawynok J: Acute amitriptyline in a rat model of neuropathic pain: Differential symptom and route effects. Pain 1999; 80:643–53Esser, MJ Sawynok, J
Ardid D, Alloui A, Brousse G, Jourdan D, Picard P, Dubray C, Eschalier A: Potentiation of the antinociceptive effect of clomipramine by a 5-ht(1A) antagonist in neuropathic pain in rats. Br J Pharmacol 2001; 132:1118–26Ardid, D Alloui, A Brousse, G Jourdan, D Picard, P Dubray, C Eschalier, A
Abdi S, Lee DH, Chung JM: The anti-allodynic effects of amitriptyline, gabapentin, and lidocaine in a rat model of neuropathic pain. Anesth Analg 1998; 87:1360–6Abdi, S Lee, DH Chung, JM
Dickenson AH, Suzuki R: Opioids in neuropathic pain: Clues from animal studies. Eur J Pain 2005; 9:113–6Dickenson, AH Suzuki, R
McQuay HJ, Tramer M, Nye BA, Carroll D, Wiffen PJ, Moore RA: A systematic review of antidepressants in neuropathic pain. Pain 1996; 68:217–27McQuay, HJ Tramer, M Nye, BA Carroll, D Wiffen, PJ Moore, RA
Schreiber S, Backer MM, Pick CG: The antinociceptive effect of venlafaxine in mice is mediated through opioid and adrenergic mechanisms. Neurosci Lett 1999; 273:85–8Schreiber, S Backer, MM Pick, CG
Eschalier A, Montastruc JL, Devoize JL, Rigal F, Gaillard-Plaza G, Pechadre JC: Influence of naloxone and methysergide on the analgesic effect of clomipramine in rats. Eur J Pharmacol 1981; 74:1–7Eschalier, A Montastruc, JL Devoize, JL Rigal, F Gaillard-Plaza, G Pechadre, JC
Sanchez C, Hyttel J: Comparison of the effects of antidepressants and their metabolites on reuptake of biogenic amines and on receptor binding. Cell Mol Neurobiol 1999; 19:467–89Sanchez, C Hyttel, J
Phillis JW, Wu PH: The effect of various centrally active drugs on adenosine uptake by the central nervous system. Comp Biochem Physiol C 1982; 72:179–87Phillis, JW Wu, PH
Sierralta F, Pinardi G, Mendez M, Miranda HF: Interaction of opioids with antidepressant-induced antinociception. Psychopharmacology (Berl) 1995; 122:374–8Sierralta, F Pinardi, G Mendez, M Miranda, HF
Reynolds IJ, Miller RJ: Tricyclic antidepressants block N-methyl-D-aspartate receptors: Similarities to the action of zinc. Br J Pharmacol 1988; 95:95–102Reynolds, IJ Miller, RJ
Tohda M, Urushihara H, Nomura Y: Inhibitory effects of antidepressants on NMDA-induced currents in Xenopus  oocytes injected with rat brain RNA. Neurochem Int 1995; 26:53–8Tohda, M Urushihara, H Nomura, Y
Park TJ, Shin SY, Suh BC, Suh EK, Lee IS, Kim YS, Kim KT: Differential inhibition of catecholamine secretion by amitriptyline through blockage of nicotinic receptors, sodium channels, and calcium channels in bovine adrenal chromaffin cells. Synapse 1998; 29:248–56Park, TJ Shin, SY Suh, BC Suh, EK Lee, IS Kim, YS Kim, KT
Pancrazio JJ, Kamatchi GL, Roscoe AK, Lynch C III: Inhibition of neuronal Na+ channels by antidepressant drugs. J Pharmacol Exp Ther 1998; 284:208–14Pancrazio, JJ Kamatchi, GL Roscoe, AK Lynch, C
Nau C, Seaver M, Wang SY, Wang GK: Block of human heart hH1 sodium channels by amitriptyline. J Pharmacol Exp Ther 2000; 292:1015–23Nau, C Seaver, M Wang, SY Wang, GK
Barber MJ, Starmer CF, Grant AO: Blockade of cardiac sodium channels by amitriptyline and diphenylhydantoin: Evidence for two use-dependent binding sites. Circ Res 1991; 69:677–96Barber, MJ Starmer, CF Grant, AO
Wang GK, Russell C, Wang SY: State-dependent block of voltage-gated Na+ channels by amitriptyline via  the local anesthetic receptor and its implication for neuropathic pain. Pain 2004; 110:166–74Wang, GK Russell, C Wang, SY
Brau ME, Dreimann M, Olschewski A, Vogel W, Hempelmann G: Effect of drugs used for neuropathic pain management on tetrodotoxin-resistant Na(+) currents in rat sensory neurons. Anesthesiology 2001; 94:137–44Brau, ME Dreimann, M Olschewski, A Vogel, W Hempelmann, G
Marchand F, Ardid D, Chapuy E, Alloui A, Jourdan D, Eschalier A: Evidence for an involvement of supraspinal delta- and spinal mu-opioid receptors in the antihyperalgesic effect of chronically administered clomipramine in mononeuropathic rats. J Pharmacol Exp Ther 2003; 307:268–74Marchand, F Ardid, D Chapuy, E Alloui, A Jourdan, D Eschalier, A
Richeimer SH, Bajwa ZH, Kahraman SS, Ransil BJ, Warfield CA: Utilization patterns of tricyclic antidepressants in a multidisciplinary pain clinic: A survey. Clin J Pain 1997; 13:324–9Richeimer, SH Bajwa, ZH Kahraman, SS Ransil, BJ Warfield, CA
Sawynok J, Reid AR, Esser MJ: Peripheral antinociceptive action of amitriptyline in the rat formalin test: Involvement of adenosine. Pain 1999; 80:45–55Sawynok, J Reid, AR Esser, MJ
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; 399:131–9Esser, MJ Sawynok, J
Gerner P, Mujtaba M, Sinnott CJ, Wang GK: Amitriptyline versus  bupivacaine in rat sciatic nerve blockade. Anesthesiology 2001; 94:661–7Gerner, P Mujtaba, M Sinnott, CJ Wang, GK
Khan MA, Gerner P, Kuo WG: Amitriptyline for prolonged cutaneous analgesia in the rat. Anesthesiology 2002; 96:109–16Khan, MA Gerner, P Kuo, WG
Sudoh Y, Cahoon EE, Gerner P, Wang GK: Tricyclic antidepressants as long-acting local anesthetics. Pain 2003; 103:49–55Sudoh, Y Cahoon, EE Gerner, P Wang, GK
Haderer A, Gerner P, Kao G, Srinivasa V, Wang GK: Cutaneous analgesia after transdermal application of amitriptyline versus  lidocaine in rats. Anesth Analg 2003; 96:1707–10Haderer, A Gerner, P Kao, G Srinivasa, V Wang, GK
Gerner P, Kao G, Srinivasa V, Narang S, Wang GK: Topical amitriptyline in healthy volunteers. Reg Anesth Pain Med 2003; 28:289–93Gerner, P Kao, G Srinivasa, V Narang, S Wang, GK
Kissin I, McDanal J, Xavier AV: Topical lidocaine for relief of superficial pain in postherpetic neuralgia. Neurology 1989; 39:1132–3Kissin, I McDanal, J Xavier, AV
Lynch ME, Clark AJ, Sawynok J, Sullivan MJ: Topical 2% amitriptyline and 1% ketamine in neuropathic pain syndromes: A randomized, double-blind, placebo-controlled trial. Anesthesiology 2005; 103:140–6Lynch, ME Clark, AJ Sawynok, J Sullivan, MJ
Lynch ME, Clark AJ, Sawynok J: A pilot study examining topical amitriptyline, ketamine, and a combination of both in the treatment of neuropathic pain. Clin J Pain 2003; 19:323–8Lynch, ME Clark, AJ Sawynok, J
McCleane G: Topical application of doxepin hydrochloride, capsaicin and a combination of both produces analgesia in chronic human neuropathic pain: A randomized, double-blind, placebo-controlled study. Br J Clin Pharmacol 2000; 49:574–9McCleane, G
Estebe JP, Myers RR: Amitriptyline neurotoxicity: Dose-related pathology after topical application to rat sciatic nerve. Anesthesiology 2004; 100:1519–25Estebe, JP Myers, RR
Barnet CS, Louis DN, Kohane DS: Tissue injury from tricyclic antidepressants used as local anesthetics. Anesth Analg 2005; 101:1838–43Barnet, CS Louis, DN Kohane, DS
Hall H, Ogren SO: Effects of antidepressant drugs on different receptors in the brain. Eur J Pharmacol 1981; 70:393–407Hall, H Ogren, SO
Richelson E, Pfenning M: Blockade by antidepressants and related compounds of biogenic amine uptake into rat brain synaptosomes: Most antidepressants selectively block norepinephrine uptake. Eur J Pharmacol 1984; 104:277–86Richelson, E Pfenning, M
Choi JJ, Huang GJ, Shafik E, Wu WH, McArdle JJ: Imipramine’s selective suppression of an L-type calcium channel in neurons of murine dorsal root ganglia involves G proteins. J Pharmacol Exp Ther 1992; 263:49–53Choi, JJ Huang, GJ Shafik, E Wu, WH McArdle, JJ
Kamatchi GL, Ticku MK: Tricyclic antidepressants inhibit Ca(2+)-activated K(+)-efflux in cultured spinal cord neurons. Brain Res 1991; 545:59–65Kamatchi, GL Ticku, MK
Wooltorton JR, Mathie A: Potent block of potassium currents in rat isolated sympathetic neurones by the uncharged form of amitriptyline and related tricyclic compounds. Br J Pharmacol 1995; 116:2191–200Wooltorton, JR Mathie, A
Wooltorton JR, Mathie A: Block of potassium currents in rat isolated sympathetic neurones by tricyclic antidepressants and structurally related compounds. Br J Pharmacol 1993; 110:1126–32Wooltorton, JR Mathie, A
Galeotti N, Ghelardini C, Bartolini A: Involvement of potassium channels in amitriptyline and clomipramine analgesia. Neuropharmacology 2001; 40:75–84Galeotti, N Ghelardini, C Bartolini, A
Green BG: Temperature perception and nociception. J Neurobiol 2004; 61:13–29Green, BG
Tominaga M, Caterina MJ: Thermosensation and pain. J Neurobiol 2004; 61:3–12Tominaga, M Caterina, MJ
Minke B: TRP channels and Ca2+ signaling. Cell Calcium 2006; 40:261–75Minke, B
Alloui A, Zimmermann K, Mamet J, Duprat F, Noel J, Chemin J, Guy N, Blondeau N, Voilley N, Rubat-Coudert C, Borsotto M, Romey G, Heurteaux C, Reeh P, Eschalier A, Lazdunski M: TREK-1, a K+ channel involved in polymodal pain perception. EMBO J 2006; 25:2368–76Alloui, A Zimmermann, K Mamet, J Duprat, F Noel, J Chemin, J Guy, N Blondeau, N Voilley, N Rubat-Coudert, C Borsotto, M Romey, G Heurteaux, C Reeh, P Eschalier, A Lazdunski, M
Valeriani M, Tinazzi M, Le Pera D, Restuccia D, De Armas L, Maiese T, Tonali P, Arendt-Nielsen L: Inhibitory effect of capsaicin evoked trigeminal pain on warmth sensation and warmth evoked potentials. Exp Brain Res 2005; 160:29–37Valeriani, M Tinazzi, M Le Pera, D Restuccia, D De Armas, L Maiese, T Tonali, P Arendt-Nielsen, L
Andrew D, Craig AD: Spinothalamic lamina I neurones selectively responsive to cutaneous warming in cats. J Physiol 2001; 537:489–95Andrew, D Craig, AD
Eisenberg E, McNicol ED, Carr DB: Efficacy of mu-opioid agonists in the treatment of evoked neuropathic pain: Systematic review of randomized controlled trials. Eur J Pain 2006; 10:667–76Eisenberg, E McNicol, ED Carr, DB
Fig. 1. Scattergrams showing the level of skin redness compared with a dedicated scale, at the end of the application period of the treatments. A significant difference was found (  P  < 0.0001, Friedman test) between treatments.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). A dose effect is suggested (  dotted line  = trend based on logistic regression), although not statistically significant. 
Fig. 1. Scattergrams showing the level of skin redness compared with a dedicated scale, at the end of the application period of the treatments. A significant difference was found (  P  < 0.0001, Friedman test) between treatments.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). A dose effect is suggested (  dotted line  = trend based on logistic regression), although not statistically significant. 
Fig. 1. Scattergrams showing the level of skin redness compared with a dedicated scale, at the end of the application period of the treatments. A significant difference was found (  P  < 0.0001, Friedman test) between treatments.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). A dose effect is suggested (  dotted line  = trend based on logistic regression), although not statistically significant. 
×
Fig. 2. Scattergrams showing the tactile (von Frey hairs;  A  ) and nociceptive (electronic von Frey;  B  ) cutaneous mechanical thresholds after application (T1) of the treatments, in time.  Horizontal line  = median value;  cross  = mean value. The data observed after T1 + 4 h do not appear, because no significant difference between the groups (Friedman test for touch and analysis of variance for mechanical nociception) was noted after this time.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). The letters  a  ,  b  , and  c  signal the groups identified by the  post hoc  analysis (multiple pair comparison or Tukey test). 
Fig. 2. Scattergrams showing the tactile (von Frey hairs;  A  ) and nociceptive (electronic von Frey;  B  ) cutaneous mechanical thresholds after application (T1) of the treatments, in time.  Horizontal line  = median value;  cross  = mean value. The data observed after T1 + 4 h do not appear, because no significant difference between the groups (Friedman test for touch and analysis of variance for mechanical nociception) was noted after this time.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). The letters  a  ,  b  , and  c  signal the groups identified by the  post hoc  analysis (multiple pair comparison or Tukey test). 
Fig. 2. Scattergrams showing the tactile (von Frey hairs;  A  ) and nociceptive (electronic von Frey;  B  ) cutaneous mechanical thresholds after application (T1) of the treatments, in time.  Horizontal line  = median value;  cross  = mean value. The data observed after T1 + 4 h do not appear, because no significant difference between the groups (Friedman test for touch and analysis of variance for mechanical nociception) was noted after this time.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). The letters  a  ,  b  , and  c  signal the groups identified by the  post hoc  analysis (multiple pair comparison or Tukey test). 
×
Fig. 3. Scattergrams showing the three kinds of cutaneous thermal thresholds (Thermotest; Somedic AS, Stockholm, Sweden) after the end of application (T1) of the treatments, in time (  A  : cold;  B  : warm;  C  : heat).  Horizontal line  = median value;  cross  = mean value. The data observed after T1 + 4 h do not appear, because no significant difference between the groups (analysis of variance) was noted after this time.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). The letters  a  and  b  signal the groups identified by the  post hoc  analysis (Tukey test). 
Fig. 3. Scattergrams showing the three kinds of cutaneous thermal thresholds (Thermotest; Somedic AS, Stockholm, Sweden) after the end of application (T1) of the treatments, in time (  A  : cold;  B  : warm;  C  : heat).  Horizontal line  = median value;  cross  = mean value. The data observed after T1 + 4 h do not appear, because no significant difference between the groups (analysis of variance) was noted after this time.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). The letters  a  and  b  signal the groups identified by the  post hoc  analysis (Tukey test). 
Fig. 3. Scattergrams showing the three kinds of cutaneous thermal thresholds (Thermotest; Somedic AS, Stockholm, Sweden) after the end of application (T1) of the treatments, in time (  A  : cold;  B  : warm;  C  : heat).  Horizontal line  = median value;  cross  = mean value. The data observed after T1 + 4 h do not appear, because no significant difference between the groups (analysis of variance) was noted after this time.  White circles  = saline;  white squares  = vehicle;  medium gray squares  = 25 mm amitriptyline (AMI);  dark gray squares  = 50 mm amitriptyline;  black squares  = 100 mm amitriptyline;  black triangles  = EMLA (Astra-Zeneca, Rueil-Malmaison, France). The letters  a  and  b  signal the groups identified by the  post hoc  analysis (Tukey test). 
×
Fig. 4. Time course of the two kinds of cutaneous thermal thresholds (Thermotest; Somedic AS, Stockholm, Sweden) affected by amitriptyline application (  A  : cold;  B  : heat). Each  line  represents one subject. These diagrams show a tendency to return to normal values at T1 + 8 h, as well as a strong interindividual variability of the sensitivity to amitriptyline, regarding cold sensation thresholds. 
Fig. 4. Time course of the two kinds of cutaneous thermal thresholds (Thermotest; Somedic AS, Stockholm, Sweden) affected by amitriptyline application (  A  : cold;  B  : heat). Each  line  represents one subject. These diagrams show a tendency to return to normal values at T1 + 8 h, as well as a strong interindividual variability of the sensitivity to amitriptyline, regarding cold sensation thresholds. 
Fig. 4. Time course of the two kinds of cutaneous thermal thresholds (Thermotest; Somedic AS, Stockholm, Sweden) affected by amitriptyline application (  A  : cold;  B  : heat). Each  line  represents one subject. These diagrams show a tendency to return to normal values at T1 + 8 h, as well as a strong interindividual variability of the sensitivity to amitriptyline, regarding cold sensation thresholds. 
×