Secondary hyperalgesia after skin injury was described by Lewis in 1936 . The phenomenon still receives considerable attention, as the mechanisms of secondary hyperalgesia involve changes in neurotransmission in the spinal cord dorsal horn (central sensitization), possibly of importance for the understanding of the pathophysiological mechanisms of pain of long duration [2,3]. These are conditions in which the mechanisms are not fully understood, and in which current treatment programs are often insufficient. Experimental secondary hyperalgesia could be used to study pharmacological substances, possibly reflecting effects in clinical pain of long duration .
Several experimental models have been developed to study the mechanisms of sensory dysfunction. Human models, such as intradermal injection of capsaicin , induction of a superficial skin burn using heat , or topical application of a chemical agent such as mustard oil (MuO)  have been used. These models also allow the study of pharmacological interventions .
The endogenous compound adenosine (ADO) and its analogs induce a variety of both excitatory and inhibitory effects in the nociceptive system, mediated via specific cell membrane-bound receptors . When administered systemically to humans in large doses, ADO activates peripheral nociceptors and induces pain . However, clinical studies using a systemic small-dose infusion of ADO [10,11] and animal studies primarily in rodents (systemic as well as intrathecal administration of adenosine analogs) [8,12] have shown that antinociceptive effects attributed ADO receptor stimulation. These results refer to clinical and experimental models of acute nociceptive pain. Less data are available concerning the effects of ADO and its analogs in pain of long duration, in which the mechanisms include changes in neuronal excitability, such as central sensitization. Presumed analgesic effects after systemic and intrathecal administration of ADO receptor agonists have been demonstrated in animal models of chronic pain induced through both peripheral nerve and spinal cord injury [13,14]. In a human experimental study using a model with topical MuO application, we have shown an ADO-induced attenuation of the area of secondary hyperalgesia, possibly reflecting a modulatory effect on central sensitization . A placebo-controlled clinical study also reported reduced pain after systemic ADO administration in patients suffering neuropathic pain . The primary aim of the present study was to evaluate the effects of systemic ADO administration on tactile and thermal sensory function in two experimental models of cutaneous hyperalgesia, possibly reflecting effects on central sensitization.
The study was approved by the institutional human investigation committee, and the subjects were included after written, informed consent. Ten healthy volunteers (five male and five female), aged 19-31 yr, were included for each model (the same subject was never used for both models). A randomized, placebo-controlled, cross-over design was used. No caffeinated beverages were allowed for 12 h before experiments, and no subject was receiving any medication. The subjects participating in the heat burn injury model were familiar with the test procedure from earlier participation in a study using the same protocol. For the chemical burn injury model, an initial single-blinded placebo procedure was performed for training purposes. All other procedures were double-blinded, and all tests in a single individual were performed by the same investigator. The interval between experimental sessions (ADO or placebo) was 1-2 wk. In a few cases, sensory disturbances of the skin (slight numbness) were still present after this time. In these cases, the corresponding contralateral skin area was used. ADO 60 [micro sign]g [center dot] kg-1 [center dot] min-1 (Item Development AB, Stocksund, Sweden) and saline (placebo) was infused IV via a cubital vein. Each sensory modality was tested separately in both the primary and secondary hyperalgesic areas, when present. The areas stained by ink (chemical burn) or the area still showing erythema at the end of the observation period (thermal burn) was defined as the primary area. These were excluded when the area of secondary hyperalgesia was calculated. All limits of sensory function were marked directly on the skin using a soft pen. After the experiment, the marks were transferred to plastic film, and the areas were measured planimetrically.
Spontaneous pain was rated using a visual analog scale (VAS) graded 0-100, where 0 = no pain and 100 = maximal pain. The area of tactile allodynia was mapped by stroking a soft brush radially from the periphery toward the burn injury until the subject reported pain or increased pain. This was performed from eight directions encircling the injured skin. The area of secondary hyperalgesia to mechanical stimuli was mapped with calibrated von Frey filaments in a similar pattern until the patient reported pain. When the skin of the arm was tested (MuO model), a filament calibrated at 20 g was used, whereas a 46-g filament was used on the leg (thermal model). The threshold to tactile pain was assessed using the method of limits. The mean of four pairs of the von Frey filament producing the lowest pressure perceived as pain and the filament producing the highest pressure not perceived as pain was used. The pain (VAS scoring) evoked by suprathreshold stimuli was assessed by using a 20-g (MuO model) or 46-g (thermal injury model) von Frey filament. The mean rating from three applications was used. Thermal sensory function was assessed using a computerized system (Thermotest[registered sign]; Somedic AB, Horby, Sweden) with a Peltier thermode (stimulation area 12.5 cm) . Thermal test procedures started at 32[degree sign]C, and the skin was allowed to adapt until this was perceived as neutral. For thermal threshold tests, the Peltier thermode changed 1[degree sign]C/s. The subject pressed a button to indicate the threshold, at which time the thermode automatically returned to 32[degree sign]C. The thermal sensory tests were the threshold for (a) cold and (b) warmth detection (mean of five measurements); (c) threshold to heat pain (mean of three measurements); and (d) pain rating (VAS) for suprathreshold thermal stimulation (single 30-s pulse of 45[degree sign]C, thermal injury model). This temperature was perceived as nonpainful in the baseline test.
In the chemical burn injury model, ADO and saline were infused IV for 60 min on separate occasions, starting 15 min before MuO application. The skin of the volar side of the dominant forearm 10 cm distal to the cubital fossa was used. A 2 x 3 cm compress was soaked with 400 [micro sign]L of MuO stained with ink (Markblack, ACO, Stockholm, Sweden). The compress was fixed to the skin for 4 min [11,17]. Sensory tests were performed after 10 min of infusion (baseline before MuO application) and 30 and 60 min after the MuO application. Tests included spontaneous pain rating and tests of tactile sensory function as described above. Thus, sensory tests after MuO application were performed both during and after ADO/placebo infusions. The subjects were repeatedly asked to report adverse effects during the infusion. If adverse effects were bothersome, the infusion rate was reduced by 10%.
In the thermal burn injury model, ADO and saline were infused IV for 120 min on separate occasions, starting 15 min before the induction of burn injury. The skin area of the lateral side of the lower leg (preferably left side) 10 cm distal to the knee joint was used. Burn injury was induced by using the computerized system (Thermotest[registered sign]) used for thermal sensory testing [5,16]. The Peltier thermode was applied firmly to the skin of the test area. After a 15-s increase in temperature from 32[degree sign]C to 47[degree sign]C, the temperature was kept constant for 7 min, after which the thermode returned to 32[degree sign]C, and the thermode was removed from the skin. Spontaneous pain was assessed, and the described tests of both mechanical and thermal sensory function were performed before and 0, 60, and 120 min after the induction of burn injury. This includes sensory testing both during and 15 min after the infusion. The subjects were informed about adverse reactions from ADO infusion, such as chest oppression or warmth sensation, and were instructed to report if they were bothered by such effects. If so, the infusion rate was reduced by 10%.
Data are presented as mean value +/- SD, except for tactile pain thresholds, for which median value (25-75) quartiles are shown. In the MuO model, the mean results of the two placebo treatments were compared with those for active treatment. A Wilcoxon signed rank test for matched pairs was used for statistical analyses. A P value <0.05 was considered significant.
An erythema of the skin developed in the injured area in all subjects. In two patients, minor blisters occurred after the observation period. A considerable interindividual variation was found in thresholds to sensory perception and pain, as well as in VAS ratings and area of secondary hyperalgesia, as indicated by the often large SD in the statistical analyses (Table 1 and Table 2). However, the intraindividual variation was small (Table 1). The variation in baseline assessments of tactile pain threshold was 16 (3-43) g. This was no more than the difference between two adjacent von Frey filaments in any individual.
The application of MuO produced spontaneous and stimulus-evoked pain throughout the 60-min observation period. The hypersensitivity to tactile stimuli in both the primary and secondary areas included pain induced by light stroking of the skin with a soft brush in this model. The mapped areas of tactile allodynia (pain evoked by soft brush) and hyperalgesia (pain evoked by a 20-g von Frey filament), as well as the results from quantitative sensory tests in the MuO model, are shown in Table 2. The area of secondary hyperalgesia after MuO application was 25% +/- 12% (P < 0.05) and 39% +/- 13% (P < 0.05) smaller when ADO was administered compared with placebo at 30 and 60 min, respectively (Figure 1A).
The baseline measurements in the thermal skin injury model were stable between sessions (Table 1). Spontaneous pain disappeared quickly when the heated thermode was removed from the skin. No dynamic tactile allodynia, i.e., pain induced by the soft brush, developed. The area of secondary hyperalgesia was reduced by ADO infusion compared with placebo treatment (Figure 1B). The reduction was 26% +/- 65%, 58% +/- 20% (P < 0.05), and 51% +/- 34% (P < 0.05) after thermal burn induction 0, 60 and 120 min respectively. Tactile pain thresholds (baseline values in Table 1) in the area of primary hyperalgesia were reduced to a lowest value of 11 (6-61) g (placebo) and 16 (6-27) g (ADO). In the area of secondary hyperalgesia, the corresponding values were 12 (6-54) g (placebo) and 16 (5-41) g (ADO). These reductions were significant (P < 0.05) compared with baseline only for placebo treatment. The perception threshold for cold and warmth (both primary and secondary hyperalgesic area) and the heat pain threshold in the area of secondary hyperalgesia remained unchanged over both treatments with a mean value never more than 1.5[degree sign]C from baseline (data not shown). In the area of primary hyperalgesia, the heat pain threshold was reduced compared with baseline (Table 1), to a minimum of 40.0 +/- 1.9[degree sign]C (P < 0.05) for placebo and to 40.9 +/- 2.9[degree sign]C (P < 0.05) for ADO treatment. No differences in VAS ratings for spontaneous or evoked pain after tactile or thermal stimuli were found between treatments. The pattern was similar to the results in the MuO model, and data are not shown in detail.
The primary finding in this study is that a systemic small-dose ADO infusion reduces the area of cutaneous secondary hyperalgesia after a thermally or chemically induced superficial cutaneous burn injury. ADO infusion did not affect other sensory variables, i.e., tactile and thermal thresholds to perception and pain. The rated intensity of suprathreshold stimuli was also not affected. The study confirms previous results using the MuO model  and adds new information by the use of an additional model and further sensory testing.
In this study, we used established methods to induce cutaneous secondary hyperalgesia [5-7]. Sensory changes occurring after burn injury and the effects of a pharmacological intervention were assessed using validated psychophysical methods . Sensory testing data were stable over time, with only minor intra-individual variations (Table 1). This allowed repeated sensory evaluations to test the effects of a pharmacological intervention on sensory function in the current placebo-controlled study.
Although the main results from both models and the conclusions that can be drawn from them are consistent, there are also differences in the results that should be addressed. First, the tactile hyperalgesia was more pronounced in the MuO model, including dynamic tactile allodynia, i.e., pain induced by the normally nonpainful stroke of a soft brush. This suggests that changes in sensory function induced by MuO application, but not by the heated thermode, lead to that activity in low threshold mechanoreceptors signaling through A beta primary afferent fibers that are involved in the generation of pain . Differences may also be due to different sites of the body (arm versus leg) used for skin injury. The intensity of stimulation may also not have been similar, leading to different degrees of secondary neuronal changes. It is also possible, as suggested by LaMotte et al. , that certain stimuli, such as algogenic chemicals, could be more active than heat in the induction of central sensitization using stimulus of similar intensity. Second, we noted different durations of sensory disturbances in line with earlier results . The longer duration of hyperalgesia in the thermal injury model allowed additional sensory tests. Third, different von Frey filaments (20 and 46 g) were used for mapping the areas and for suprathreshold stimulation within the hyperalgesic area. These filaments and the 45[degree sign]C pulse were chosen close to the normal pain threshold in the respective skin area (Table 1) to enhance the contrast to sensory function in normal skin. Fourth, the IV ADO infusion in most cases induces mild but characteristic adverse effects. Consequently, there was a problem with "unblinding" of the treatment and possible investigator bias in the MuO model. Therefore, for the thermal injury protocol, we did not ask about adverse effects, but the subjects were instructed to report bothersome adverse effects spontaneously. This still gave the subject the option to have the infusion rate lowered but reduced the risk of unblinding. Only one subject reported discomfort, which quickly disappeared when the infusion rate was reduced by 10%. The blinding was therefore preserved. Investigator bias is not likely to have had a major influence on the results, as the area of secondary hyperalgesia was similarly reduced in both models.
We found no difference between ADO and placebo treatments in perception and pain thresholds or VAS ratings to suprathreshold stimuli in the primary and secondary areas of hyperalgesia. Only the size of the area of uninjured skin in which sensory changes occurred was reduced by ADO. All other tested thermal and tactile variables, including sensory changes within the remaining area of secondary hyperalgesia, were unaffected. The size of the area of cutaneous hyperalgesia has been suggested to be a marker for the degree of changes in dorsal horn neuronal excitability, central sensitization . If all other tested sensory variables are unaffected, the ADO effect is unlikely to have been caused by sedation or unspecific effects on neural excitability.
Mechanisms of experimental cutaneous hyperalgesia and their relevance for different clinical pain conditions have been extensively studied and discussed . The secondary hyperalgesia is believed to be caused by changes in excitability of second-order neurons in the dorsal horn depending on continuing activity in sensitized primary afferent mechano-heat sensitive C-fibers in the injured area . Using a combination of electrophysiological studies in anesthetized monkeys and psychophysical tests in awake humans, a correlation between the intensity of pain perception and increased excitability of dorsal horn wide dynamic range neurons has been demonstrated after intradermal capsaicin injection . The capsaicin model is similar to those used in this study, and the secondary hyperalgesia is probably mediated by similar mechanisms .
Data from animal models of acute pain indicate antinociceptive effects of ADO mediated via adenosine A1 receptors, linked to the inhibition of adenylyl cyclase activity, in the spinal cord dorsal horn . Adenosine inhibits synaptic transmission in the substantia gelatinosa via hyperpolarization of the postsynaptic membrane . This mechanism may explain a preferential effect on hyperexcitable spinal cord neurons involved in the development of cutaneous secondary hyperalgesia. The observed selective effect on a sensory disturbance attributed to changes in second-order neurons in the spinal cord dorsal horn indicates that the observed effect involves a site of action on the spinal cord level. Peripheral or supraspinal sites of action with projections to the spinal cord level are also possible. The reduced area of secondary hyperalgesia was still present after the treatment, when no remaining direct effect of circulating ADO could be expected. This finding of postinfusion effects on sensory function is in line with earlier clinical results [10,15], which also suggests ADO receptor involvement in the complex and slowly activated mechanisms of central sensitization.
We conclude that a systemic small-dose infusion of ADO specifically reduces the area of experimental cutaneous secondary hyperalgesia without affecting other injury-induced changes in thermal and tactile perception. The results indicate an ADO receptor-mediated modulatory effect on the mechanisms of central sensitization, with possible relevance to pain of long duration, including neuropathic pain.
We thank Ringvor Hagglof and Anette Ebberyd for excellent technical assistance.
1. Lewis T. Experiments relating to cutaneous hyperalgesia and its spread through somatic nerves. Clin Sci 1936;2:273-421.
2. Treede RD, Meyer RA, Raja SN, Campbell JN. Peripheral and central mechanisms of cutaneous hyperalgesia. Prog Neurobiol 1992;38:397-421.
3. Torebjork HE, Lundberg LE, LaMotte RH. Central changes in processing of mechanoreceptive input in capsaicin induced secondary hyperalgesia in humans. J Physiol 1992;448:765-80.
4. Gracely RH. Studies of pain in normal man. In: Wall PD, Melzack R, eds. Textbook of pain. London: Churchill Livingstone, 1994:315-36.
5. Dahl JB, Brennum J, Arendt-Nielsen L, et al. The effect of pre-versus postinjury infiltration with lidocaine on thermal and mechanical hyperalgesia after heat injury to the skin. Pain 1993;53:43-51.
6. Koltzenburg M, Lundberg LER, Torebjork HE. Dynamic and static components of mechanical hyperalgesia in human hairy skin. Pain 1992;51:207-19.
7. Petersen KL. Experimental cutaneous hyperalgesia in humans. IASP Newsletter 1997;November:4-8.
8. Sawynok J, Sweeney MI. The role of purines in nociception. Neuroscience 1989;32:557-69.
9. Sylven C. Mechanisms of pain in angina pectoris: a critical review of the adenosine hypothesis. Cardiovasc Drugs Ther 1993;7:745-59.
10. Segerdahl M, Ekblom A, Sandelin K, et al. Preoperative adenosine infusion reduces the requirements for isoflurane and post-operative analgesics. Anesth Analg 1995;80:1145-9.
11. Segerdahl M, Ekblom A, Sjolund KF, et al. Systemic adenosine attenuates touch evoked allodynia induced by mustard oil in humans. Neuroreport 1995;6:753-6.
12. Karlsten R, Gordh T, Hartvig P, Post C. Effects of intrathecal injection of the adenosine receptor agonist R-phenylisopropyladenosine and N-ethylcarboxamide-adenosine on nociception and motor function in the rat. Anesth Analg 1990;71:60-4.
13. Sjolund KF, Sollevi A, Segerdahl M, et al. Intrathecal and systemic R-phenylisopropyladenosine reduces scratching behaviour in a rat mononeuropathy model. Neuroreport 1996;7:1856-60.
14. Sjolund KF, von Heijne M, Hao JX, et al. Intrathecal administration of the adenosine A1
receptor agonist R-phenylisopropyl adenosine reduces presumed pain behaviour in a rat model of central pain. Neurosci Lett 1998;243:89-92.
15. Belfrage M, Sollevi A, Segerdahl M, et al. Systemic adenosine alleviates spontaneous and stimulus evoked pain in patients with peripheral neuropathic pain. Anesth Analg 1995;81:713-7.
16. Lindblom U. Analysis of abnormal touch, pain, and temperature sensation in patients. In: Boivie J, Hansson P, Lindblom U, eds. Touch, temperature, and pain in health and disease: mechanisms and assessments-progress in pain research and management. Seattle, WA: IASP Press, 1994:63-84.
17. Koltzenburg M, Torebjork E, Wahren LK. Nociceptor modulated central sensitization causes mechanical hyperalgesia in acute chemogenic and chronic neuropathic pain. Brain 1994;117:579-91.
18. LaMotte RH, Shain CN, Simone DA, Tsai EP. Neurogenic hyperalgesia: psychophysical studies of underlying mechanisms. J Neurophysiol 1991;66:190-211.
19. Simone DA, Sorkin LS, Oh U, et al. Neurogenic hyperalgesia: central neural correlates in responses of spinothalamic tract neurons. J Neurophysiol 1991;66:228-46.
20. Salter MW, DeKoninck Y, Henry JL. Physiological roles for adenosine and ATP in synaptic transmission in the spinal dorsal horn. Prog Neurobiol 1993;41:125-56.
21. Li J, Perl ER. Adenosine inhibition of synaptic transmission in the substantia gelatinosa. J Physiol 1994;72:1611-21.