The Interaction Between Gabapentin and Amitriptyline in the Rat Formalin Test After Systemic Administration

Page created by Derek Salazar
 
CONTINUE READING
The Interaction Between Gabapentin and Amitriptyline in the
Rat Formalin Test After Systemic Administration
Caroline E. Heughan,                 MSc,   and Jana Sawynok,     PhD
Department of Pharmacology, Dalhousie University, Halifax, Nova Scotia, Canada

       We examined the effects of systemically administered       doses of gabapentin and amitriptyline produced an ad-
       gabapentin on flinching and biting/licking behaviors       ditive effect against biting/licking behaviors, as re-
       produced by 2.5% formalin in the rat, compared these       vealed by isobolographic analysis. These increments
       with those of amitriptyline, and determined the effects    had no effect on flinching behaviors except at the ED25
       of combinations of gabapentin with amitriptyline.          ⫹ ED25 doses, at which flinching was increased, again
       Gabapentin produced a dose-related inhibition of           revealing additivity between the two drugs. Flinching
       Phase 2, but not Phase 1, flinching and biting/licking     behaviors in rats do not reflect the analgesic properties
       behaviors. In contrast, amitriptyline produced an in-      of systemically administered amitriptyline observed in
       crease in Phase 2 flinching behaviors while simulta-       humans and may not be useful for predicting an effect
       neously decreasing biting/licking behaviors. Fifty per-    of combinations of drugs with amitriptyline. Biting/
       cent effective dose (ED50) values against biting/licking   licking behaviors do reflect analgesic properties for
       behaviors were 22.9 ⫾ 1.3 mg/kg and 8.5 ⫾ 1.3 mg/kg        both drugs and may be more useful in this regard.
       for gabapentin and amitriptyline, respectively. Combi-
       nations of increasing fractional increments of ED50                                (Anesth Analg 2002;94:975–80)

A
       variety of animal test models have been devel-             produces synergistic effects, minimal doses could be
        oped to examine various aspects of chronic                used, and this would minimize adverse effects. How-
        pain. Chronic pain is not simply continuing or            ever, if two drugs simply exhibit additive effects, but
persistent acute pain; it involves adaptive changes in            their side effect profiles are different and noninteract-
pain signaling in primary afferent neurons, in the                ing, this could still serve to augment pain relief and
spinal cord, and at supraspinal sites, and it can exhibit         minimize side effects.
distinct neurochemical features (1– 4). The pharmaco-                Amitriptyline (an antidepressant) and gabapentin
logical treatment of chronic pain is not as effective as          (an anticonvulsant) represent different classes of ad-
the treatment of acute pain, because these underlying             juvants used for the treatment of chronic pain. Ami-
changes alter the sensitivity of the system to tradi-             triptyline is widely used, and its efficacy has been
tional analgesics (nonsteroidal antiinflammatory                  established with metaanalysis and systematic ap-
drugs and opioids), and chronic pain therapy has                  proaches (9,10). Gabapentin is used to treat neuro-
come to rely on a group of drugs classified as adju-              pathic pain in particular (11). Both drugs exhibit a
vants (e.g., antidepressants, anticonvulsants, local an-          comparable efficacy in diabetic neuropathy when
esthetics, and ␣-adrenergics) (5,6). These drugs have             compared directly in a single study (12) and when
other primary pharmacological indications, but they               number-needed-to-treat scores are generated (6). Al-
can modify aspects of the changes that occur with                 though these drugs produce a comparable incidence
chronic pain and can lead to pain relief. Adjuvants,              of side effects, the pattern of these effects differs some-
however, generally do not produce complete pain re-               what (12).
lief, and their actions can be limited by adverse effects.           The formalin test is a model of continuing pain
One potential strategy for minimizing side effects is to          involving continued activation of sensory afferents,
use combinations of drugs (7,8). If a drug combination            inflammation, and central sensitization (13,14). Acti-
                                                                  vation of excitatory amino acid receptors in the spinal
  This work was supported by the Canadian Institutes of Health    cord contributes to central sensitization and chronic
Research.                                                         pain, and the formalin test has proved useful for re-
  Accepted for publication December 7, 2001.                      vealing the pharmacology of this system (14,15). Both
  Address correspondence and reprint requests to Jana Sawynok,
Department of Pharmacology, Dalhousie University, Halifax, Nova   amitriptyline (16) and gabapentin (17,18) may modify
Scotia B3H 4H7, Canada. Address e-mail to sawydalu@is.dal.ca.     aspects of excitatory amino acid function in the spinal

©2002 by the International Anesthesia Research Society
0003-2999/02                                                                                      Anesth Analg 2002;94:975–80   975
976   PAIN MEDICINE HEUGHAN AND SAWYNOK                                                              ANESTH ANALG
      GABAPENTIN AND AMITRIPTYLINE                                                                     2002;94:975–80

cord to produce analgesia. We recently reported the           The dose of drug required to produce 50% of the
analgesic properties of amitriptyline in the rat forma-    maximum possible decrease in behavior (ED50) was
lin test after systemic administration (19). In this       obtained for each drug by using linear regression.
study, we determined the analgesic properties of           Subsequently, dose-response information for the con-
gabapentin after systemic administration in the rat        current administration of both drugs was obtained by
formalin model and compared these with those pro-          using constant dose ratios of the respective ED50 value
duced by amitriptyline. We also examined the effects       for each drug (i.e., ED50 ⫹ ED50, ED37.5 ⫹ ED37.5, ED25
of combinations of gabapentin with amitriptyline to        ⫹ ED25, and ED12.5 ⫹ ED12.5). The ED50 value for the
determine whether these two adjuvants could pro-           drug combination was calculated by linear regression
duce additive or synergistic effects. Gabapentin has       and isobolographic analysis, as previously described
been examined in combination with morphine (20),           (27,28). The theoretical and actual ED50 values for the
ibuprofen (21), clonidine (22), and a non-N-methyl-d-      drug combination were compared by using Student’s
aspartate excitatory amino acid receptor antagonist        t-tests.
(23) in various pain models, whereas amitriptyline has        Gabapentin was supplied by Parke Davis (Ann Ar-
been administered in combination with opioids,             bor, MI). Amitriptyline hydrochloride and formalin
neostigmine, and clonidine (24,25). There have been        were purchased from the Sigma Chemical Co. (St.
no reports on the effects of the combination of gaba-      Louis, MO).
pentin with amitriptyline in any model, but the effects
of this combination in the formalin model are of par-
ticular interest because of its potential relevance to
chronic pain. Results of this study have been pre-         Results
sented in abstract form (26).                              The IP injection of gabapentin 10 –50 mg/kg before
                                                           2.5% formalin produced a reduction in both flinching
                                                           and biting/licking behaviors, particularly in the later
Methods                                                    part of the second phase of activity (Figs. 1 and 2).
Experiments were conducted with male Sprague-              Phase 1 behaviors were unaltered. A larger dose
Dawley rats (125–150 g; Charles River, Quebec, Can-        (70 mg/kg) produced sedation (somnolence and flac-
ada). Rats were housed in pairs with a 12:12-h light/      cidity) (data not shown). Amitriptyline augments
dark cycle at 22°C, with free access to food and water.    flinching but decreases biting/licking behaviors (Fig.
All experiments were approved by the University            2), and these actions were expressed uniformly
Committee on Laboratory Animals, and experiments           throughout the second phase of activity. Regression
were conducted in accordance with International As-        analysis revealed an ED50 of 22.9 ⫾ 1.3 mg/kg for
sociation for the Study of Pain and the Canadian           gabapentin and 8.5 ⫾ 1.3 mg/kg for amitriptyline for
Council on Animal Care standards.                          Phase 2 biting/licking responses. Combinations of
  Formalin 2.5% was administered in a volume of 50         fixed increments of ED50 values for the two drugs had
␮L by subcutaneous injection into the dorsal surface of    no significant effect on flinching behaviors except at
the hindpaw at Time 0. Gabapentin and drug combi-          the ED25 ⫹ ED25 doses, at which an augmentation in
nations were administered systemically as a 20-min         flinching was observed (Fig. 3A). However, these
intraperitoneal (IP) pretreatment in a volume of 5 mL/     combinations exhibited an additive effect against
kg. Each behavioral testing session was preceded by a      biting/licking behaviors (Fig. 3B). This was verified by
20-min acclimatization to the 28 ⫻ 28 ⫻ 28-cm Plexi-       isobolographic analysis, with no significant difference
glas observation chamber. After injection of formalin,     between theoretical and actual ED50 values for the
rats were returned to the chamber. Flinching behav-        drug combination (Fig. 4).
iors (lifting, shaking, or rippling of the haunch, with
episodes scoring as a single event) were recorded as a
cumulative number of events, and biting/licking be-
haviors were recorded as time spent exhibiting that        Discussion
behavior, for 60 min after formalin injection (Phase 1     This study demonstrates that systemic administration
⫽ 0 –12 min; Phase 2 ⫽ 16 – 60 min). Two rats in           of gabapentin in rats inhibits both flinching and
adjacent chambers were observed at a time, with ob-        biting/licking behaviors produced by formalin. Other
servations occurring in alternate 2-min bins. Values       studies have previously noted the efficacy of gabap-
were not corrected for time, so scores represent ap-       entin in the formalin test when given both systemi-
proximately half of the total behaviors. Statistics were   cally (21) and spinally (29 –31). The suppression of
performed on the time-course and dose-response data        behaviors was observed in Phase 2, but not in Phase 1,
by using analysis of variance followed by the Student-     as noted in each of those reports, and was prominent
Newman-Keuls test for multiple groups and time             in the later part of Phase 2, as also noted previously
courses.                                                   (21). The observation that spinal gabapentin is less
ANESTH ANALG                                                                       PAIN MEDICINE  HEUGHAN AND SAWYNOK                 977
2002;94:975–80                                                                              GABAPENTIN AND AMITRIPTYLINE

Figure 1. Time course of suppression of 2.5% formalin-evoked (A)
flinching and (B) biting/licking behaviors produced by systemic
(intraperitoneal) pretreatment with gabapentin (GP). Values depict
mean ⫾ sem for six per group; ns indicates P ⬎ 0.05 compared with
the Saline-Pretreated group.
                                                                     Figure 2. Comparison of the effects of systemically-administered
                                                                     gabapentin (GP) and amitriptyline (AMI) in the 2.5% formalin
potent when administered after the formalin and pref-                (FOR) test against Phase 2 (A) flinching and (B) biting/licking
                                                                     behaviors. Data for amitriptyline redrawn (19) with permission.
erentially suppresses Fos-like immunoreactivity in-                  Lines were fit by linear regression analysis of individual groups. *P
duced by larger concentrations of formalin suggests                  ⬍ 0.05 compared with the Saline-Pretreated group; n ⫽ 6 per group.
that gabapentin acts on elements of central sensitiza-
tion that are recruited during the second phase of
activity (31). Gabapentin also produces analgesic
                                                                     could contribute to analgesia. Even though gabapentin
properties in other inflammatory pain tests and nerve
injury models that are considered to be relevant mod-                interacts with specific heterodimers of ␥-aminobutyric
els of chronic pain in humans (11). Although the spe-                acid type B receptors (35), a selective antagonist for
cific mechanisms by which gabapentin produces an-                    this receptor did not block antihyperalgesic effects of
algesia are not resolved, a spinal component of action               gabapentin (36).
is clearly involved. Thus, gabapentin is not only active               In contrast to gabapentin, systemic administration
in behavioral models after spinal application, but it                of amitriptyline augments flinching behaviors while
also inhibits the formalin-evoked expression of Fos-                 simultaneously inhibiting biting/licking behaviors;
like immunoreactivity in the spinal cord (31) and af-                the mechanisms underlying this duality of action are
fects excitatory transmission mediated by glutamate in               considered elsewhere (19). Multiple mechanisms, in-
the superficial spinal cord (17,18). Gabapentin binds to             cluding inhibition of biogenic amine reuptake, inter-
the ␣2␦ subunit of voltage-dependent Ca2⫹ channels                   actions with opioid and adenosine systems, and inter-
(32) and inhibits Ca2⫹ influx and neurotransmitter                   actions with N-methyl-d-aspartate receptors and ion
release at central sites (33,34), and these actions also             channels, are thought to contribute to antinociception
978    PAIN MEDICINE HEUGHAN AND SAWYNOK                                                                                 ANESTH ANALG
       GABAPENTIN AND AMITRIPTYLINE                                                                                        2002;94:975–80

                                                                       Figure 4. Isobologram for the gabapentin (GP) and amitriptyline
                                                                       (AMI) combination against biting/licking behaviors in Phase 2 of
                                                                       the 2.5% formalin test in rats. Points represent 50% effective dose
                                                                       (ED50) ⫾ sem. There was no significant difference between the
                                                                       theoretical and actual ED50 values for the drug combination, indi-
                                                                       cating an additive drug interaction.

                                                                       licking behaviors were reduced in a dose-dependent
                                                                       manner with fixed-dose increments of ED50 values,
                                                                       and isobolographic analysis indicated a strictly addi-
                                                                       tive effect. The drug combination also produced an
                                                                       additive effect on flinching behaviors, even though
                                                                       effects of the two individual drugs on this behavior
                                                                       differed qualitatively. Thus, at the ED25 ⫹ ED25 dose,
                                                                       there was an increase in flinching responses reflecting
                                                                       an earlier expression of the facilitating effect of ami-
                                                                       triptyline, whereas at larger doses, flinches were no
                                                                       longer augmented, indicating that some degree of
                                                                       suppression of enhanced behaviors was now occur-
Figure 3. Effects of gabapentin and amitriptyline at fixed-dose com-   ring because of the activity of gabapentin.
binations of 50% effective dose (ED50) (1 ⫹ 1), and fractional doses
thereof, on Phase 2 (A) flinching and (B) biting/licking behaviors
                                                                          Given that gabapentin and amitriptyline both pro-
produced by 2.5% formalin (FOR). Data are presented as mean ⫹          duce pain relief in humans (9 –11), biting/licking be-
sem for six per group. *P ⬍ 0.05 compared with the Saline-             haviors seem to be a more appropriate end point to
Pretreated Formalin group.                                             consider in the formalin model for amitriptyline com-
                                                                       binations. Whether the additive analgesia observed
resulting from systemic administration of amitripty-                   with this behavior also manifests clinically as an ad-
line (37,38).                                                          ditive pain relief remains to be determined in a con-
   Although the details of the mechanisms by which                     trolled and systematic manner. Before such a study, it
gabapentin and amitriptyline produce analgesia are                     would be prudent to determine (a) the effects of gaba-
not precisely defined, they represent two distinct                     pentin and amitriptyline combinations in other pre-
classes of drugs and seem to interact with different                   clinical models of chronic pain (there are more chronic
endogenous pain regulatory systems. Thus, it was                       models of inflammation and nerve injury in which
initially anticipated that these two drugs might recruit               both gabapentin and amitriptyline exhibit analgesic
parallel or sequential mechanisms in pain transmis-                    activity), (b) whether there is an interaction with ad-
sion pathways and exhibit a synergistic analgesia.                     verse effects (although sedation was not quantitated
However, when gabapentin and amitriptyline were                        or examined systematically in this study, no sedation
combined in fixed fractional proportions, an additive                  was noted in the combination groups), and (c) the
effect was obtained with both behaviors. Biting/                       effects of chronic administration of these two drugs (it
ANESTH ANALG                                                                        PAIN MEDICINE  HEUGHAN AND SAWYNOK                 979
2002;94:975–80                                                                               GABAPENTIN AND AMITRIPTYLINE

is not entirely clear whether drug effects that require               13. Tjølsen A, Berge OG, Hunskaar S, et al. The formalin test: an
                                                                          evaluation of the method. Pain 1992;51:5–17.
chronic administration paradigms represent a phar-
                                                                      14. Coderre TJ, Katz J, Vaccarino AL, Melzack R. Contribution of
macokinetic or a pharmacodynamic mechanism).                              central neuroplasticity to pathological pain: review of clinical
   The differential effects of gabapentin and amitrip-                    and experimental evidence. Pain 1993;52:259 – 85.
tyline on flinching behaviors in the formalin test raise              15. Dickenson AH, Chapman V, Green GM. The pharmacology of
the question of the usefulness of this outcome score for                  excitatory and inhibitory amino acid-mediated events in the
                                                                          transmission and modulation of pain in the spinal cord. Gen
predicting outcomes in clinical paradigms. The flinch                     Pharmacol 1997;28:633– 8.
response is, to some extent, a spinally-mediated re-                  16. Eisenach JC, Gebhart GF. Intrathecal amitriptyline acts as an
sponse, because it can be elicited in chronically-                        N-methyl-D-aspartate receptor antagonist in the presence of
spinalized animals (39). However, some supraspinal                        inflammatory hyperalgesia in rats. Anesthesiology 1995;83:
                                                                          1046 –54.
integration, modulation, or both are involved as well,                17. Shimoyama M, Shimoyama N, Hori Y. Gabapentin affects glu-
because acute spinalization can reduce flinch re-                         tamatergic excitatory neurotransmission in the rat dorsal horn.
sponses (40,41). The fact that a drug that is widely                      Pain 2000;85:405–14.
used to treat chronic pain (amitriptyline) actually en-               18. Gu Y, Huang LYM. Gabapentin actions on N-methyl-D-
hances flinching behaviors indicates that the sole use                    aspartate receptor channels are protein kinase C-dependent.
                                                                          Pain 2001;93:85–92.
of this particular indicator as a measure of pain in the              19. Sawynok J, Reid AR. Antinociception by tricyclic antidepres-
formalin test may not be appropriate for determining                      sants in the rat formalin test: differential effects on different
analgesic activity. A MEDLINE search of the literature                    behaviours following systemic and spinal administration. Pain
for 1997–1998 revealed that 43% of studies that used                      2001;93:51–9.
                                                                      20. Shimoyama M, Shimoyama N, Inturrisi CE, Elliott KJ. Gaba-
the formalin test used flinching as a single measure,                     pentin enhances the antinociceptive effects of spinal morphine
and it was argued that multiple behavioral measures                       in the rat tail-flick test. Pain 1997;72:375– 82.
should be used with the formalin test (42). The results               21. Yoon MH, Yaksh TL. Evaluation of interaction between gabap-
of this study support that conclusion.                                    entin and ibuprofen on the formalin test in rats. Anesthesiology
                                                                          1999;91:1006 –13.
                                                                      22. Cheng JK, Pan HL, Eisenach JC. Antiallodynic effect of intra-
We thank Parke-Davis for their generous supply of gabapentin.             thecal gabapentin and its interaction with clonidine in a rat
                                                                          model of postoperative pain. Anesthesiology 2000;92:1126 –31.
                                                                      23. Chen SR, Eisenach JC, McCaslin PP, Pan HL. Synergistic effect
                                                                          between intrathecal non-NMDA antagonist and gabapentin on
                                                                          allodynia induced by spinal nerve ligation in rats. Anesthesiol-
                                                                          ogy 2000;92:500 – 6.
References                                                            24. Botney M, Fields HL. Amitriptyline potentiates morphine anal-
 1. Woolf CJ, Doubell TP. The pathophysiology of chronic pain:            gesia by a direct action on the central nervous system. Ann
    increased sensitivity to low threshold A ␤-fibre inputs. Curr         Neurol 1983;13:160 – 4.
    Opin Neurobiol 1994;4:525–34.                                     25. Eisenach JC, Gebhart GF. Intrathecal amitriptyline: antinocicep-
 2. Devor M. The pathophysiology of damaged peripheral nerves.            tive interactions with intravenous morphine and intrathecal
    In: Wall PD, Melzack R, eds. Textbook of pain. Edinburgh:             clonidine, neostigmine, and carbamylcholine in rats. Anesthesi-
    Churchill Livingstone, 1994:79 –100.                                  ology 1995;83:1036 – 45.
 3. Woolf CJ, Mannion RJ. Neuropathic pain: aetiology, symptoms,      26. Heughan CE, Reid AR, Sawynok J. The interaction of systemic
    mechanisms, and management. Lancet 1999;353:1959 – 64.                amitriptyline and gabapentin in the rat formalin test [abstract].
 4. Basbaum AI. Distinct neurochemical features of acute and per-         Soc Neurosci Abstr 2000;26:1216.
    sistent pain. Proc Natl Acad Sci U S A 1994;96:7739 – 43.         27. Tallarida RJ, Porreca F, Cowan A. Statistical analysis of drug-
 5. MacFarlane BV, Wright A, O’Callachan J, Beson HAE. Chronic            drug and site-site interactions with isobolograms. Life Sci 1989;
    neuropathic pain and its control by drugs. Pharmacol Ther             45:947– 61.
    1997;75:1–19.                                                     28. Tallarida RJ, Stone DJ, Raffa BB. Efficient designs for studying
 6. Sindrup SH, Jensen TS. Efficacy of pharmacological treatments         synergistic drug combinations. Life Sci 1997;61:417–25.
    of neuropathic pain: an update and effect related to mechanism    29. Shimoyama M, Shimoyama M, Davis AM, et al. Spinal gaba-
    of drug action. Pain 1999;83:389 – 400.                               pentin is antinociceptive in the rat formalin test. Neurosci Lett
 7. Solomon RE, Gebhart GF. Synergistic antinociceptive interac-          1997;222:65–7.
    tions among drugs administered to the spinal cord. Anesth         30. Yoon MH, Yaksh TL. The effect of intrathecal gabapentin on
    Analg 1994;78:1164 –72.                                               behavior and hemodynamics on the formalin test in the rat.
 8. Yaksh TL, Malmberg AB. Interactions of spinal modulatory
                                                                          Anesth Analg 1999;89:434 –9.
    receptor systems. In: Fields HL, Liebskind JC, eds. Progress in
                                                                      31. Kaneko M, Mestre C, Sanchez EH, Hammond DL. Intrathecally
    pain research and management. Vol 1. Seattle: IASP Press, 1994:
                                                                          administered gabapentin inhibits formalin-evoked nociception
    151–71.
 9. Onghena P, Van Houdenhove BJ. Antidepressant-induced an-              and the expression of fos-like immunoreactivity in the spinal
    algesia in chronic non-malignant pain: a meta-analysis of 39          cord of the rat. J Pharmacol Exp Ther 2000;292:743–51.
    placebo-controlled studies. Pain 1992;49:205–19.                  32. Gee NS, Brown JP, Dissanayake VU, et al. The novel anticon-
10. McQuay HJ, Tramer M, Nye BA, et al. A systematic review of            vulsant drug, gabapentin (Neurontin), binds to the ␣2␦ subunit
    antidepressants in neuropathic pain. Pain 1996;68:217–27.             of a calcium channel. J Biol Chem 1996;271:5768 –76.
11. Mao J, Chen LL. Gabapentin in pain management. Anesth             33. Meder WP, Dooley DJ. Modulation of K⫹-induced synaptoso-
    Analg 2000;91:680 –7.                                                 mal calcium influx by gabapentin. Brain Res 2000;875:157–9.
12. Morello CM, Leckband SG, Stoner CP, et al. Randomized             34. Fink K, Meder W, Dooley DJ, Göthert M. Inhibition of neuronal
    double-blind study comparing the efficacy of gabapentin with          Ca2⫹ influx by gabapentin and subsequent reduction of neuro-
    amitriptyline on diabetic peripheral neuropathy pain. Arch In-        transmitter release from rat neocortical slices. Br J Pharmacol
    tern Med 1999;159:1931–7.                                             2000;130:900 – 6.
980    PAIN MEDICINE HEUGHAN AND SAWYNOK                                                                               ANESTH ANALG
       GABAPENTIN AND AMITRIPTYLINE                                                                                      2002;94:975–80

35. Ng GYK, Bertrand S, Sullivan R, et al. ␥-Aminobutyric acid type   39. Coderre TJ, Yashphal K, Henry JL. Specific contribution of
    B receptors with specific heterodimer composition and postsyn-        lumbar spinal mechanisms to persistent nociceptive responses
    aptic actions in hippocampal neurons are targets of anticonvul-       in the formalin test. Neuroreport 1994;5:1337– 40.
    sant gabapentin action. Mol Pharmacol 2001;59:144 –52.            40. Wheeler-Aceto H, Cowan A. Standardization of the rat paw
36. Patel S, Naeem S, Kesingland A, et al. The effects of GABAB           formalin test for the evaluation of analgesics. Psychopharma-
    agonists and gabapentin on mechanical hyperalgesia in models          cology 1991;104:35– 44.
    of neuropathic and inflammatory pain in the rat. Pain 2001;90:    41. Wheeler-Aceto H, Cowan A. Neurogenic and tissue-mediated
    217–26.                                                               components of formalin-induced edema: evidence for supraspi-
37. Sindrup SH. Antidepressants as analgesics. In: Yaksh TL, Maze         nal regulation. Agents Actions 1991;34:264 –9.
    M, Lynch C, et al., eds. Anesthesia: biologic foundations.
                                                                      42. Abbott FV, Ocvirk R, Najafee R, Franklin KBJ. Improving the
    Philadelphia: Lippincott-Raven, 1997:987–97.
                                                                          efficiency of the formalin test. Pain 1999;83:561–9.
38. Sawynok J, Esser MJ, Reid AR. Antidepressants as analgesics: an
    overview of central and peripheral mechanisms of action. J Psy-
    chiatry Neurosci 2001;26:21– 8.
You can also read