- Email: [email protected]
Qualitative and quantitative characterization of the thermal grill
Pain 116 (2005) 26–32 www.elsevier.com/locate/pain
Qualitative and quantitative characterization of the thermal grill Albert Y. Leung*, Mark S. Wallace, Gery Schulteis, Tony L. Yaksh Department of Anesthesiology, University of California, San Diego School of Medicine, 9500 Gilman Drive #0924, La Jolla, CA 92093-0924, USA Received 7 October 2004; received in revised form 4 March 2005; accepted 14 March 2005
Abstract Concurrent applications to the skin of spatially adjacent bands of innocuous warm and cool stimuli would elicit a peculiar sensation, known as the ‘thermal grill illusion’. To validate the thermal grill as a research tool, this two-phase study qualitatively characterizes this peculiar sensation and further quantitatively establishes the temperature matching of the most intense/noxious thermal grill stimulations at two different time points. The temperature combinations (8C) tested were: 18/18, 42/42, 18/42, 20/20, 40/40, 20/40, 22/22, 38/38, 22/38, 24/24, 36/36 and 24/36. None of the subjects reported pain with single temperature combinations. However, hot associated with pain and burning sensations were reported in all mixed temperature combinations tested. The VAS scores for pain were significantly elevated for 20/40 and 18/42 in comparison to 22/38 and 24/36 (P!0.007). At the 3-second time point, the matching temperatures (GSD) of 20/40 and 18/42 were 45.7G1.8 (range 44-48) and 46.6G1.5 (range 44-48) 8C, respectively, whereas the matching temperatures for the single temperature combinations were similar to the set temperatures. Importantly, at the 10-second time point, none of the combinations were significantly greater than the highest of the pair of stimuli. The time course variation in the perception of the combined stimuli suggests an adaptation occurred in central processing. q 2005 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. Keywords: Thermal grill illusion; Central pain; Quantitative sensory testing; Neuropathic pain; Pain model
1. Introduction In 1896, Thunberg first described that concurrent applications to the skin of spatially adjacent bands of innocuous warm and cool stimuli would elicit a noxious sensation, currently described as a ‘cold burning pain sensation’ (Thunberg, 1896). Over the past few decades, several descriptions and corresponding theories have been hypothesized to explain this so called ‘thermal grill illusion.’ Altrutz proposed that the sensation represented a fusion resulting from the simultaneous activation of sensory channels for warm and cold (Boring, 1942). Other studies suggested that the grill provided a ‘synthetic heat’ or ‘hot’ sensation. In addition, Green’s work hinted the impact of tactile sensation on cooling induced nociception (Green, 1977; Green and Pope, 2003). However, the unspecified duration of exposure to the simultaneous warm and cold * Corresponding author. Tel.: C1 858 657 7030; fax: C1 858 657 7035. E-mail address: [email protected] (A.Y. Leung).
generated by various kind of devices and wide range of temperatures (sometimes over the range of hot or cold pain thresholds) make the results of these studies hard to interpret (Burnett and Dallenbach, 1927; Jenkins, 1938; Sullivan, 1930). More recently, Craig and Bushnell, however, citing evidence from neurological recordings from two classes of ascending spinothalamic tract neurons that are sensitive to innocuous or noxious stimulation and result from functional imaging study, proposed that the thermal grill illusion resulted from the integration of ascending pain and sensory channels (Craig and Bushnell, 1994; Craig et al., 1996). Given the potential role of central integrative processes, Heavner and colleagues suggested that the thermal grill stimulation paradigm might be useful in studying conditions where changes in central processing have occurred as, for example, following peripheral nerve injury (Heavner et al., 1997). Recently, in a case report, Morin and colleagues suggested that the absence of thermal grill-evoked pain is consistent with the hypothesis that in some cases of central pain the loss of the thermosensory pathway results in
0304-3959/$20.00 q 2005 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.pain.2005.03.026
A.Y. Leung et al. / Pain 116 (2005) 26–32
the disruption of the normal cold inhibition of burning pain (Morin et al., 2002). Despite the potential utility as a probe for defining changes in central processing, no studies have been done to systemically characterize the stimuli–response characteristics of the thermal grill stimuli at different time points. To address the need for characterization, we utilized quantitative peripheral neurosensory testing to systematically characterize the response of subjects to different stimulus parameters. The study was divided into two phases with the following objectives: Phase (I). Establish the innocuous temperature combinations that quantitatively would generate the maximal noxious stimulation and to qualitatively characterize the sensation of the thermal grill. Phase (II). Establish a temperature matching of the initial sensation generated by the warm and cold temperature combination(s) that will generate the most intense thermal grill stimulation at two different time points.
2. Methodology 2.1. Subjects Thirteen subjects, 6 males and 7 females, with the age range of 18–50 were recruited for the study. Exclusion criteria were: 1. Pregnancy 2. History of cardiac, hepatic, neurological and endocrinological disorders 3. Evident psychological disorders including diagnosed depression 4. Pending litigation 5. Lack of ability to understand the experimental protocol and to adequately communicate in English 6. History of opiate abuse. The Institutional Human Subjects Committee of the University of California, San Diego, approved all components of this study. 2.2. Study paradigm 2.2.1. Phase (I). Testing of different temperature combinations of thermal grill stimuli The volunteers were first asked to place the dominant hand in an unheated thermal glove, which is made of the insulating material used in oven gloves, for one minute to ensure constant baseline temperature and then place the palm of the dominant hand with all five fingers slightly on the thermal grill for 10 s. The subjects were then asked to characterize the initial sensation as warm, cold, neutral, or hot at the palm of the hand. If the subject reported pain of any kind, then he or she was asked to rate the intensity of pain on a Visual Analog Scale (VAS). Subjects were also asked if the pain is associated with any burning component. Each temperature setting was tested at least three times with 1 min between each stimulation epoch while the subjects placed their dominant hands in
27
the thermal glove. Any sensation that was recorded two or more times out of three trials was considered significant. The following temperature combinations (8C) were tested in randomized fashion: 18/18, 42/42, 18/42, 20/20, 40/40, 20/40, 22/22, 38/38, 22/38, 24/24, 36/36 and 24/36. 2.2.2. Phase II In this phase of the study, baseline thermal thresholds of the palms including cold, warm, cold pain and hot pain were tested bilaterally in healthy subjects with a Thermal Sensory Analyzer (Medoc Advanced Medical Systems, Minneapolis). Subjects with similar bilateral upper extremity thermal thresholds were asked to participate in the thermal matching test at two different time points. 2.2.2.1. Quantitative thermal matching at 3 s. The temperature combination(s) that generated the most intense noxious stimulation were tested for thermal matching at 3-s intervals to assess whether the thermal intensity of the stimulation varied with time. The following temperature combinations were tested randomly: 20/20, 40/40, 20/40, 18/18, 42/42 and 18/42 8C. First, the subject was asked to place both hands in the thermal gloves for at least 1 min. The subject then placed his or her dominant hand on the thermal grill for 3 s, after which the subject immediately placed the nondominant hand on the thermal analyzer with a pre-set temperature for 3 s. At the end of the 3 s, the subject was asked whether the temperature at the thermal analyzer matched the temperature felt on the thermal grill. Each pre-set thermal analyzer temperature was tested three times with both hands in the thermal gloves with at least 1 min breaks between testing. The test was repeated at different pre-set temperatures until a match (at least two out of three) occurred. The range of pre-set temperatures that were used randomly to match the thermal grill was from 10 to 48 8C at onedegree increments. 2.2.2.2. Quantitative thermal matching at 10 s. The temperature combination(s) that generated the most intense noxious stimulation were also tested for thermal matching at the 10-s intervals to assess whether the thermal intensity of the stimulation varied with time. The following temperature combinations were tested randomly: 20/20, 40/40, 20/40, 18/18, 42/42 and 18/42 8C. The subject was first asked to place both hands in the thermal gloves for 1 min. Subsequent to that, the subject placed the dominant hand on the thermal grill for 10 s and then immediately placed the nondominant hand on the thermal analyzer with a pre-set temperature for 10 s. The subjects were then asked whether the temperature at the thermal analyzer matched that felt on the thermal grill. Each pre-set thermal analyzer temperature was tested three times with both hands in the thermal gloves for at least 1 min in between the testing. The test was repeated at different pre-set temperatures until the match occurred. Two or more matchings out of three testings were considered a match. The range of pre-set temperatures that were used randomly to match the thermal grill was from 10 to 48 8C at one-degree increments. Thermal grill. The thermal grill stimulator consisted of two systems of 0.75 cm-flattened copper tubing cut into approximately 10 cm in length and placed 1 cm apart. Each system consisted of five tubes. The systems were then mounted such that every other tubing was connected to a common inflow and outflow. The array was then mounted in a non-metallic frame of 15!20 cm (See Fig. 1). The inflow and outflow of the two separate systems were
28
A.Y. Leung et al. / Pain 116 (2005) 26–32
2.3. Data analysis Friedman analysis (a non-parametric one-way ANOVA test) was used to select the temperature combination(s) that generated the statistically significant VAS scores. In addition, temperature combination comparisons that generated any noxious sensation were ranked in order by using the Wilcoxin Rank Sign Test. Data is expressed as the meanGSD.
3. Results 3.1. Phase I study
Fig. 1. Thermal grill picture.
connected via a series of four-way valves to two separate heating and cooling water baths. The exact temperatures at the copper tubings were monitored with two separate temperature probes (Mallinckrodt Medical, Inc., St Louis, MO), which were attached to the corresponding copper tubings. By appropriately adjusting the valves and settings of the temperatures in the water baths, the following running water temperature combinations (8C) were tested in a randomized fashion: 18/18, 42/42, 18/42, 20/20, 40/40, 20/40, 22/22, 38/38, 22/38, 24/24, 36/36, and 24/36. The temperature range used in this study (from 18 to 42 8C) was not known to cause any human skin damage. 2.2.2.3. Thermal sensory testing. The sensations that were tested include cold, warm, cold pain and hot pain. These sensations were measured by using a Thermal Sensory Analyzer (Medoc Advance Medical Systems, Minneapolis). This device consisted of a thermode measuring 46!29 mm. The temperature of the thermode could either rise or fall (at a rate of 1.2 8C/s for cold and warm sensation, and 3 8C/s for cold and warm pain), depending on sensations that were being tested. The subject signaled the onset of feeling the tested sensation by pressing a switch, which will in turn reverse the temperature change and return the temperature of the thermode to 32 8C baseline. The computer would then record the temperature of the thermode at which the switch was pressed. The average value of testing result would be automatically calculated by the computer and displayed on the screen. The probe was placed at the palm of the subject’s hand during testing. 2.2.2.4. Visual analog scale. The Visual Analog Scale (VAS) is a horizontal linear scale with the length of 100 mm. At one end of the scale is marked ‘No Pain’ and at the other end of it is marked ‘Worst Pain Imaginable.’ If the subject reports pain in testing, he or she was asked to mark the intensity of the painful sensation by drawing a perpendicular line across the linear scale. The length from the ‘no pain’ end to the subject’s marking was then measured and the average of the three measurements was used for analysis.
3.1.1. Visual analog scale After obtaining IRB approval and subjects’ informed consent, a total of 13 healthy subjects were enrolled in the Phase I part of the study. None of the subjects reported pain with single temperature combinations ranging from 18 to 42 8C. Hot sensation with pain is reported in all innocuous warm and cold temperature combinations tested only. All reported hot pain sensation has an associated burning sensation. Pain with cold sensation was not reported. The mean VAS scores of the pain with hot and burning sensation for the 18/42, 20/40, 22/38 and 24/36 (8C) temperature combinations were 15.04G19.48, 11.87G16.17, 2.89G 6.21 and 2.6G8.7, respectively. Of the different warm and cold temperature combination settings, the VAS scores of pain at 20/40 and 18/42 (8C) combinations were significantly elevated in comparison to the mixed 22/38 and 24/36 (8C) temperature combinations with the Friedman Test (P!0.007). With the Chi-square value of 11.61 and P! 0.0089, the mean ranks in order for the VAS scores for the mixed temperature stimuli under the subsequent Wilcoxin Rank Sign Test were shown in Table 1. There is no statistically significant difference of the VAS scores for either between 20/40 and 18/42 or between 22/38 and 24/26 (8C) temperature combinations. No comparison was made to the single temperature combinations as none of the subjects perceived pain with the reported sensation of cold, neutral, warm and hot during the stimuli. 3.1.2. Qualitative characterization of the thermal grill None of the subjects reported pain with any of the single temperature combinations tested. Hot sensation associated with burning pain is present as the predominant sensation in all innocuous warm and cold temperature combinations Table 1 Noxious sensation ranking with Wilcoxin Rank Sign Test (Chi-square value of 11.61 and P!0.0089) Temperature combinations (8C)
Mean rank order
18/42 20/40 22/38 24/36
3.35 2.81 2.00 1.85
A.Y. Leung et al. / Pain 116 (2005) 26–32
29
Table 2 Thermal grill temperature matching at 3 s of stimulation Subjects
18/18 (8C)
42/42 (8C)
18/42 (8C)
20/20 (8C)
40/40 (8C)
20/40 (8C)
1 2 3 4 5 6 7 Mean SD
18 17 18 18 19 18 18 18 0.6
41 42 40 42 42 42 42 41.6 0.8
46 48 46 48 46 48 44 46.6 1.5
20 21 19 20 19 22 21 20.2 1.1
39 39 41 40 40 40 41 40 0.8
46 46 44 48 44 48 44 45.7 1.8
40/40 and 42/42 combinations were 18G0.6 (range 17–19), 20.2G1.1 (range 19–21), 40G0.8 (range 39–41) and 41.6G0.8 (range 41–43) 8C, respectively. In addition, all seven subjects reported a ‘surprise feeling’ or ‘stunning’ sensation when they placed their hands on the thermal grill (Table 2).
Fig. 2. (a) Thermal sensations with single temperature combinations (NZ 13). (b) Thermal sensations with mixed temperature combinations (NZ13).
(Fig. 2). At both 18/42 and 20/40 (8C) temperature combinations, hot burning pain was felt in about 70% of the volunteers as the initial sensation. Most subjects felt a ‘surprise’ or ‘stunning’ sensation with the initial placement of the hand on the thermal grill. However, only 20% of the subjects at both 22/38 and 24/36 (8C) experienced a hot burning pain sensation. The percentage breakdowns for the predominant sensations that were felt by the subjects at the single and mixed temperature combination tested are shown in Fig. 2a and b, respectively. 3.2. Phase II study 3.2.1. Thermal grill and QST temperature matching at 3 s A total of seven right-handed dominant healthy subjects who participated in the Phase I part of the study were screened for the Phase II part of the study. Two subjects were excluded from the study because of the unmatched right and left thermal thresholds. Of the remaining seven subjects who completed the study, the matching temperatures (GSD) of 20/40 and 18/42 were 45.7G1.8 (range 44–48) and 46.6G1.5 (range 44–48) 8C, respectively. The matching temperatures (GSD) for 18/18, 20/20,
3.2.2. Thermal grill and QST temperature matching at 10 s The same seven subjects who completed the Phase II study were enrolled. In this phase of the study, the combinations of thermal grill temperatures that were tested in random order included the following: 18/18, 42/42, 18/42, 20/20, 40/40, and 20/40. Of the seven subjects who have similar upper extremities thermal thresholds, the matching temperature (GSD) for the thermal temperatures combinations (8C) of 18/18, 42/42, 18/42, 20/20, 40/40 and 20/40 (8C) were 18.0G0.6 (range 17–19), 41.7G0.9 (range 40–43), 36.7G10.6 (range 18–43, two no match), 20.1G1.1 (range 19–21), 40.1G0.8 (range 39–41) and 37.3G9.0 (19–42, one no match), respectively. One subject was unable to match any temperature at both 18/42 and 20/40 (8C) temperature combinations and another subject was unable to match at 18/42 (8C) combinations. Thus, at 10 s, none of the combinations were significantly greater than the highest of the pair of stimuli (Table 3).
Table 3 Thermal grill temperature matching at 10 s of stimulation NM-No thermal matching Subjects
18/18 (8C)
42/42 (8C)
18/42 (8C)
20/20 (8C)
40/40 (8C)
20/40 (8C)
1 2 3 4 5 6 7 Mean SD
19 18 18 18 17 18 18 18 0.6
41 42 42 42 43 40 42 41.7 0.9
43 39 18 NM 41 NM 42.5 36.7 10.6
20 21 19 19.5 19 22 20.5 20.1 1.1
39 39 41 40 41 40 40.5 40.1 0.8
41 40 19 NM 41 42 41 37.3 9.0
30
A.Y. Leung et al. / Pain 116 (2005) 26–32
4. Discussion In the present study, the exposure of normal human subjects to the thermal grill stimulation paradigm resulted predominantly in a transient hot burning sensation with a statistically significant mild elevation of VAS scores of pain at the 20/40 and 18/42 8C temperature combinations. This observed mildly painful sensation was independently validated by an increase in the matched temperature when parallel exposures were made to fixed thermal stimuli for 3 s. The effect at 3 s may occur in part by an additional component of adaptation that is suggested by the initial ‘surprise’ or ‘stunning’ feeling that most subjects reported acutely after applying the palm to the grill. Importantly, after 10 s, the initial noticed sensation was noted to dissipate and the matched temperature was no greater than the highest stimulus intensity of the pair. 4.1. Qualitative characterization of the thermal grill In the present study, temperatures up to 42 8C were not reported to be noxious or hot but rather to be warm by all the subjects in the Phase I of the study. This is in parallel to the previous QST work, where the hot pain threshold was in the range of greater than 45 8C (Ando et al., 2000; Leung et al., 2001). In the present work, pairing of temperatures in the range of 20/40 and 18/42 8C was reported to elicit a report of a mild hot burning pain sensation. Contrary to Craig and Bushnell’s report that the thermal grill illusion presented with a burning sensation resembling cold pain, the current study suggests that the predominant sensation in the thermal grill is hot and burning as initially suggested by Altrutz and others (Boring, 1942; Heavner et al., 1997). This sensation appears to be significant at higher temperature differences at 20/40 and 18/42 (8C) than at lower temperature differences such as 22/38 and 24/36 (8C). 4.2. Thermal grill temperature matching The close matching of the single temperature combination in the thermal grill with the QST provides the validation that the thermal grill system as described here provides an appropriate quantifiable thermal stimulus, which matches a validated thermal stimulator system (Verdugo and Ochoa, 1992). The specific issue under consideration in this series of studies is the characteristic of the sensory experience reported when combinations of nonnoxious warm and cold stimuli were presented to the palm. As noted, the concurrent delivery of 20/40 and 18/42 (8C) at 3-s intervals resulted in an aversive experience. This assertion is based on three observations. First, these combinations, but not either temperature stimuli alone, resulted in a significant incidence of assigning the burning pain descriptor to the stimulus condition. Secondly, the ranking of the VAS scores appears to be proportional to the percentage of subjects who reported pain with the different
combined temperature stimuli. Third, these temperature combinations were matched with the QST studies at test temperatures of approximately 46 8C. Systemic QST studies from several laboratories have independently reported that this temperature range falls within the known range of hot pain threshold obtained in the healthy subjects (Ando et al., 2000; Wallace et al., 2002). 4.3. Dynamic aspects of thermal grill stimulation The application of acute thermal stimuli of up to 42 8C typically results in a stable sensation of warmth. Concurrent application of the cool (18–20 8C) and warm (40–42 8C) temperatures to the palm resulted in three additional components. The first, as discussed above, was the appearance of a noxious component at 3 s. Secondly, the observer noted that the noxious character of the sensation diminished by 10 s. The loss of the noxious nature of the effect noted over the 10-s intervals is in contrast to the stable nature of the sensation produced by the single thermal stimulus. This time course variation has not been reported by other investigators. We discount the likelihood of a change in the effective stimulus due to the heat sink capacity of the hand because examination of the grill temperature via the temperature monitoring probes during the stimulus exposure reveals no shifts in temperature. This observation is consistent with the high temperature capacity of the perfusion systems. We suggested that the loss of the acute burning sensation reflects upon an adaptation. As the apparent adaptation was not observed with the single stimulus conditions, we suggest that this alteration reflects upon a central change in processing, which underlies the aversive aspect of the combined stimulus presentation. 4.4. Underlying mechanism of the thermal grill stimulus-response Broadly speaking, high frequency and low-threshold mechano-stimulation is transmitted by the myelinated A-fibers. Cool and well-localized pain is carried by the less myelinated A-d fiber, whereas warm, hot and cold pain sensations are carried by the unmyelinated C-fiber (Verdugo and Ochoa, 1992; Yarnitsky and Ochoa, 1991). Previous electrical stimulation and nerve block studies suggested that cold-specific A-d afferent activity can attenuate C-fiber mediated cold pain sensation (Fruhstorfer, 1984; Ochoa and Yarnitsky, 1994; Wahren et al., 1989). In the thermal grill stimulation paradigm, the presence of warm stimuli in close spatial proximity to the cold stimuli may diminish the spatial summation of A-d fiber input, which has a spinal modulatory role of the C-fiber mediated cold pain threshold. The initial observed hot burning sensation at 3 s is in line with the results of previous nerve compression study in which blockade of the A-d mediated cold sensation leads to elevation of the cold pain threshold and the change of cold pain sensation to a hot burning rather than cold pain alone
A.Y. Leung et al. / Pain 116 (2005) 26–32
(Wahren et al., 1989). Accordingly, the presence of the continuous C-fiber input from the warm sensation could contribute to the spinal encoding of the sensation of hot burning rather than cold burning pain. Finally, as the input from the slow-conducting C-fiber becomes more dominant as observed by the subjects’ report of feeling both warm and cold sensation as stimulus intensities approach the nonnoxious magnitudes, the warm and cold stimuli interaction of the thermal grill sensation becomes less dominant. This reported transient noxious finding from the current study in comparison to a former study by Fruhstorfer, who suggested that the thermal grill sensation as ‘non-noxious heat’, may be due to the fact that the former study: (1) utilized an ischemic nerve block, which could potentially impact the C-fiber; (2) exposed the subjects to continuous testing conditions; and (3) had a significant skin–thermode surface temperature variation (Fruhstorfer et al., 2003). Nevertheless, the thermal matching of the current study did demonstrate that the thermal grill sensation is hardly a noxious one at 10 s. In addition, the findings of this current study are also complimentary to previous work performed by Craig and Bushnell, who measured the electrical activities of spinal cord neurons that projected to the brain in the spinothalamic tract (STT) of anesthetized cats (Craig and Bushnell, 1994). Their group examined all three major types of cat lamina I STT neurons: (1) nociceptive-specific (NS) cells that are responsive only to noxious heat and pinch and receive input from heat nociceptors; (2) thermoreceptive-specific (COLD) cells that are responsive to cooling and receive input from specific cold receptors; and (3) multimodal cells that are responsive to noxious heat, pinch, and cold (HPC) and receive input from cold-sensitive C polymodal nociceptors. Their result showed that NS cells were unaffected by the Warm, Cool, and Grill stimuli. In addition, both COLD and HPC lamina I STT cells responded briskly to the Cool stimuli. However, the response of COLD cells and HPC cells to the Grill stimulus differed significantly. The response of COLD cells to Grill were strongly reduced from their response to Cool but the response of HPC cells to Grill remained nearly the same as their response to Cool, suggesting that the presence of the interlaced warm bars to a cool stimulus not only reduces the activity in the coldspecific channel but also shifts the relative pattern of activity in favor of the HPC channel. Although work by Craig and Bushnell provided physiological evidence of the thermal grill illusion, the potential impact of general anesthesia on the primary afferent neurons and STT cells of the cat has not been discussed. It is known that anesthesia can have significant effect on the neuronal activities of subcortical structure which includes sensory and motor neurons and dorsal horn cells (Antognini and Carstens, 1998; Antognini and Schwartz, 1993; Antognini et al., 1999). In Craig’s study, the electrophysiological activities were measured over the duration of 100 s. Although the duration at which the animals were exposed to the Grill stimulation is not well
31
specified, it appeared that the most robust response from the dorsal horns (Cold and HPC) with spike measurements started at around 30 s and gradually declined over the next 60 s. This phenomenon suggests a change of both the Cold and HPC cell activities at the dorsal horn with prolonged exposure of the stimuli and supports the dynamic nature of the sensory integration at the spinal cord level. Given the above concerns, the current study does provide additional quantitative temperature comparison of the Grill stimuli and identifies the transient nature of the thermal grill sensation.
5. Conclusion The current study provides qualitative and quantitative evidence that the combination of two non-noxious temperature combinations in the thermal grill can generate a noxious sensation, which resembles a hot burning pain sensation generated by a temperature of around 46 8C. However, this initial noxious sensation appears to last for only a short duration. These newly identified dynamic quantitative and qualitative characteristics of the thermal grill needs to be considered if one chooses to use the device as a research tool for studying the anomalous events after nerve or tissue injury.
Acknowledgements Special thanks to Dr Yaksh and Michael Rathbun for constructing the thermal grill and Linda Sutherland for her editorial assistance.
References Ando K, Wallace MS, Braun J, Schulteis G. Effect of oral mexiletine on capsaicin-induced allodynia and hyperalgesia: a double-blind, placebocontrolled, crossover study. Reg Anesth Pain Med 2000;25:468–74. Antognini JF, Carstens E. Macroscopic sites of anesthetic action: brain versus spinal cord. Toxicol Lett 1998;100–101:51–8. Antognini JF, Schwartz K. Exaggerated anesthetic requirements in the preferentially anesthetized brain. Anesthesiology 1993;79:1244–9. Antognini JF, Carstens E, Buzin V. Isoflurane depresses motoneuron excitability by a direct spinal action: an F-wave study. Anesth Analg 1999;88:681–5. Boring E. Sensation and perception in history of experimental psychology. In: Geldard F, editor. The human senses. New Yok: Appleton-CenturyCrofts; 1942. p. 504–21. Burnett NC, Dallenbach KM. The experience of heat. Am J Psychol 1927; 38:418–31. Craig AD, Bushnell MC. The thermal grill illusion: unmasking the burn of cold pain. Science 1994;265:252–5. Craig AD, Reiman EM, Evans A, Bushnell MC. Functional imaging of an illusion of pain. Nature 1996;384:258–60. Fruhstorfer H. Thermal sensibility changes during ischemic nerve block. Pain 1984;20:355–61. Fruhstorfer H, Harju EL, Lindblom UF. The significance of A-delta and C fibres for the perception of synthetic heat. Eur J Pain 2003;7:63–71.
32
A.Y. Leung et al. / Pain 116 (2005) 26–32
Green BG. Localization of thermal sensation: an illusion and synthetic heat. Percept Psychophys 1977;22:331–7. Green BG, Pope JV. Innocuous cooling can produce nociceptive sensations that are inhibited during dynamic mechanical contact. Exp Brain Res 2003;148:290–9. Heavner JE, Calvillo O, Racz GB. Thermal grill illusion and complex regional pain syndrome type I (reflex sympathetic dystrophy). Reg Anesth 1997;22:257–9. Jenkins WL. Studies in thermal sensitivity: 5. The reactions of untrained subjects to simultaneous warm and cold stimulation. J Exp Psychol 1938;22:451–61. Leung A, Wallace MS, Ridgeway B, Yaksh T. Concentration–effect relationship of intravenous alfentanil and ketamine on peripheral neurosensory thresholds, allodynia and hyperalgesia of neuropathic pain. Pain 2001;91:177–87. Morin C, Bushnell MC, Luskin MB, Craig AD. Disruption of thermal perception in a multiple sclerosis patient with central pain. Clin J Pain 2002;18:191–5.
Ochoa JL, Yarnitsky D. The triple cold syndrome. Cold hyperalgesia, cold hypoaesthesia and cold skin in peripheral nerve disease. Brain 1994; 117:185–97. Sullivan AH. The experimental fusion of warmth and cold in heat. J Exp Psychol 1930;13:208–16. Thunberg T. Uppsala Lakforen. Forh 1896;2:489. Verdugo R, Ochoa JL. Quantitative somatosensory thermotest. A key method for functional evaluation of small calibre afferent channels. Brain 1992;115:893–913. Wahren LK, Torebjork E, Jorum E. Central suppression of cold-induced C fibre pain by myelinated fibre input. Pain 1989;38:313–9. Wallace MS, Ridgeway 3rd B, Leung A, Schulteis G, Yaksh TL. Concentration–effect relationships for intravenous alfentanil and ketamine infusions in human volunteers: effects on acute thresholds and capsaicin-evoked hyperpathia. J Clin Pharmacol 2002;42:70–80. Yarnitsky D, Ochoa JL. Warm and cold specific somatosensory systems. Psychophysical thresholds, reaction times and peripheral conduction velocities. Brain 1991;114:1819–26.
Qualitative and quantitative characterization of the thermal grill Albert Y. Leung*, Mark S. Wallace, Gery Schulteis, Tony L. Yaksh Department of Anesthesiology, University of California, San Diego School of Medicine, 9500 Gilman Drive #0924, La Jolla, CA 92093-0924, USA Received 7 October 2004; received in revised form 4 March 2005; accepted 14 March 2005
Abstract Concurrent applications to the skin of spatially adjacent bands of innocuous warm and cool stimuli would elicit a peculiar sensation, known as the ‘thermal grill illusion’. To validate the thermal grill as a research tool, this two-phase study qualitatively characterizes this peculiar sensation and further quantitatively establishes the temperature matching of the most intense/noxious thermal grill stimulations at two different time points. The temperature combinations (8C) tested were: 18/18, 42/42, 18/42, 20/20, 40/40, 20/40, 22/22, 38/38, 22/38, 24/24, 36/36 and 24/36. None of the subjects reported pain with single temperature combinations. However, hot associated with pain and burning sensations were reported in all mixed temperature combinations tested. The VAS scores for pain were significantly elevated for 20/40 and 18/42 in comparison to 22/38 and 24/36 (P!0.007). At the 3-second time point, the matching temperatures (GSD) of 20/40 and 18/42 were 45.7G1.8 (range 44-48) and 46.6G1.5 (range 44-48) 8C, respectively, whereas the matching temperatures for the single temperature combinations were similar to the set temperatures. Importantly, at the 10-second time point, none of the combinations were significantly greater than the highest of the pair of stimuli. The time course variation in the perception of the combined stimuli suggests an adaptation occurred in central processing. q 2005 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. Keywords: Thermal grill illusion; Central pain; Quantitative sensory testing; Neuropathic pain; Pain model
1. Introduction In 1896, Thunberg first described that concurrent applications to the skin of spatially adjacent bands of innocuous warm and cool stimuli would elicit a noxious sensation, currently described as a ‘cold burning pain sensation’ (Thunberg, 1896). Over the past few decades, several descriptions and corresponding theories have been hypothesized to explain this so called ‘thermal grill illusion.’ Altrutz proposed that the sensation represented a fusion resulting from the simultaneous activation of sensory channels for warm and cold (Boring, 1942). Other studies suggested that the grill provided a ‘synthetic heat’ or ‘hot’ sensation. In addition, Green’s work hinted the impact of tactile sensation on cooling induced nociception (Green, 1977; Green and Pope, 2003). However, the unspecified duration of exposure to the simultaneous warm and cold * Corresponding author. Tel.: C1 858 657 7030; fax: C1 858 657 7035. E-mail address: [email protected] (A.Y. Leung).
generated by various kind of devices and wide range of temperatures (sometimes over the range of hot or cold pain thresholds) make the results of these studies hard to interpret (Burnett and Dallenbach, 1927; Jenkins, 1938; Sullivan, 1930). More recently, Craig and Bushnell, however, citing evidence from neurological recordings from two classes of ascending spinothalamic tract neurons that are sensitive to innocuous or noxious stimulation and result from functional imaging study, proposed that the thermal grill illusion resulted from the integration of ascending pain and sensory channels (Craig and Bushnell, 1994; Craig et al., 1996). Given the potential role of central integrative processes, Heavner and colleagues suggested that the thermal grill stimulation paradigm might be useful in studying conditions where changes in central processing have occurred as, for example, following peripheral nerve injury (Heavner et al., 1997). Recently, in a case report, Morin and colleagues suggested that the absence of thermal grill-evoked pain is consistent with the hypothesis that in some cases of central pain the loss of the thermosensory pathway results in
0304-3959/$20.00 q 2005 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.pain.2005.03.026
A.Y. Leung et al. / Pain 116 (2005) 26–32
the disruption of the normal cold inhibition of burning pain (Morin et al., 2002). Despite the potential utility as a probe for defining changes in central processing, no studies have been done to systemically characterize the stimuli–response characteristics of the thermal grill stimuli at different time points. To address the need for characterization, we utilized quantitative peripheral neurosensory testing to systematically characterize the response of subjects to different stimulus parameters. The study was divided into two phases with the following objectives: Phase (I). Establish the innocuous temperature combinations that quantitatively would generate the maximal noxious stimulation and to qualitatively characterize the sensation of the thermal grill. Phase (II). Establish a temperature matching of the initial sensation generated by the warm and cold temperature combination(s) that will generate the most intense thermal grill stimulation at two different time points.
2. Methodology 2.1. Subjects Thirteen subjects, 6 males and 7 females, with the age range of 18–50 were recruited for the study. Exclusion criteria were: 1. Pregnancy 2. History of cardiac, hepatic, neurological and endocrinological disorders 3. Evident psychological disorders including diagnosed depression 4. Pending litigation 5. Lack of ability to understand the experimental protocol and to adequately communicate in English 6. History of opiate abuse. The Institutional Human Subjects Committee of the University of California, San Diego, approved all components of this study. 2.2. Study paradigm 2.2.1. Phase (I). Testing of different temperature combinations of thermal grill stimuli The volunteers were first asked to place the dominant hand in an unheated thermal glove, which is made of the insulating material used in oven gloves, for one minute to ensure constant baseline temperature and then place the palm of the dominant hand with all five fingers slightly on the thermal grill for 10 s. The subjects were then asked to characterize the initial sensation as warm, cold, neutral, or hot at the palm of the hand. If the subject reported pain of any kind, then he or she was asked to rate the intensity of pain on a Visual Analog Scale (VAS). Subjects were also asked if the pain is associated with any burning component. Each temperature setting was tested at least three times with 1 min between each stimulation epoch while the subjects placed their dominant hands in
27
the thermal glove. Any sensation that was recorded two or more times out of three trials was considered significant. The following temperature combinations (8C) were tested in randomized fashion: 18/18, 42/42, 18/42, 20/20, 40/40, 20/40, 22/22, 38/38, 22/38, 24/24, 36/36 and 24/36. 2.2.2. Phase II In this phase of the study, baseline thermal thresholds of the palms including cold, warm, cold pain and hot pain were tested bilaterally in healthy subjects with a Thermal Sensory Analyzer (Medoc Advanced Medical Systems, Minneapolis). Subjects with similar bilateral upper extremity thermal thresholds were asked to participate in the thermal matching test at two different time points. 2.2.2.1. Quantitative thermal matching at 3 s. The temperature combination(s) that generated the most intense noxious stimulation were tested for thermal matching at 3-s intervals to assess whether the thermal intensity of the stimulation varied with time. The following temperature combinations were tested randomly: 20/20, 40/40, 20/40, 18/18, 42/42 and 18/42 8C. First, the subject was asked to place both hands in the thermal gloves for at least 1 min. The subject then placed his or her dominant hand on the thermal grill for 3 s, after which the subject immediately placed the nondominant hand on the thermal analyzer with a pre-set temperature for 3 s. At the end of the 3 s, the subject was asked whether the temperature at the thermal analyzer matched the temperature felt on the thermal grill. Each pre-set thermal analyzer temperature was tested three times with both hands in the thermal gloves with at least 1 min breaks between testing. The test was repeated at different pre-set temperatures until a match (at least two out of three) occurred. The range of pre-set temperatures that were used randomly to match the thermal grill was from 10 to 48 8C at onedegree increments. 2.2.2.2. Quantitative thermal matching at 10 s. The temperature combination(s) that generated the most intense noxious stimulation were also tested for thermal matching at the 10-s intervals to assess whether the thermal intensity of the stimulation varied with time. The following temperature combinations were tested randomly: 20/20, 40/40, 20/40, 18/18, 42/42 and 18/42 8C. The subject was first asked to place both hands in the thermal gloves for 1 min. Subsequent to that, the subject placed the dominant hand on the thermal grill for 10 s and then immediately placed the nondominant hand on the thermal analyzer with a pre-set temperature for 10 s. The subjects were then asked whether the temperature at the thermal analyzer matched that felt on the thermal grill. Each pre-set thermal analyzer temperature was tested three times with both hands in the thermal gloves for at least 1 min in between the testing. The test was repeated at different pre-set temperatures until the match occurred. Two or more matchings out of three testings were considered a match. The range of pre-set temperatures that were used randomly to match the thermal grill was from 10 to 48 8C at one-degree increments. Thermal grill. The thermal grill stimulator consisted of two systems of 0.75 cm-flattened copper tubing cut into approximately 10 cm in length and placed 1 cm apart. Each system consisted of five tubes. The systems were then mounted such that every other tubing was connected to a common inflow and outflow. The array was then mounted in a non-metallic frame of 15!20 cm (See Fig. 1). The inflow and outflow of the two separate systems were
28
A.Y. Leung et al. / Pain 116 (2005) 26–32
2.3. Data analysis Friedman analysis (a non-parametric one-way ANOVA test) was used to select the temperature combination(s) that generated the statistically significant VAS scores. In addition, temperature combination comparisons that generated any noxious sensation were ranked in order by using the Wilcoxin Rank Sign Test. Data is expressed as the meanGSD.
3. Results 3.1. Phase I study
Fig. 1. Thermal grill picture.
connected via a series of four-way valves to two separate heating and cooling water baths. The exact temperatures at the copper tubings were monitored with two separate temperature probes (Mallinckrodt Medical, Inc., St Louis, MO), which were attached to the corresponding copper tubings. By appropriately adjusting the valves and settings of the temperatures in the water baths, the following running water temperature combinations (8C) were tested in a randomized fashion: 18/18, 42/42, 18/42, 20/20, 40/40, 20/40, 22/22, 38/38, 22/38, 24/24, 36/36, and 24/36. The temperature range used in this study (from 18 to 42 8C) was not known to cause any human skin damage. 2.2.2.3. Thermal sensory testing. The sensations that were tested include cold, warm, cold pain and hot pain. These sensations were measured by using a Thermal Sensory Analyzer (Medoc Advance Medical Systems, Minneapolis). This device consisted of a thermode measuring 46!29 mm. The temperature of the thermode could either rise or fall (at a rate of 1.2 8C/s for cold and warm sensation, and 3 8C/s for cold and warm pain), depending on sensations that were being tested. The subject signaled the onset of feeling the tested sensation by pressing a switch, which will in turn reverse the temperature change and return the temperature of the thermode to 32 8C baseline. The computer would then record the temperature of the thermode at which the switch was pressed. The average value of testing result would be automatically calculated by the computer and displayed on the screen. The probe was placed at the palm of the subject’s hand during testing. 2.2.2.4. Visual analog scale. The Visual Analog Scale (VAS) is a horizontal linear scale with the length of 100 mm. At one end of the scale is marked ‘No Pain’ and at the other end of it is marked ‘Worst Pain Imaginable.’ If the subject reports pain in testing, he or she was asked to mark the intensity of the painful sensation by drawing a perpendicular line across the linear scale. The length from the ‘no pain’ end to the subject’s marking was then measured and the average of the three measurements was used for analysis.
3.1.1. Visual analog scale After obtaining IRB approval and subjects’ informed consent, a total of 13 healthy subjects were enrolled in the Phase I part of the study. None of the subjects reported pain with single temperature combinations ranging from 18 to 42 8C. Hot sensation with pain is reported in all innocuous warm and cold temperature combinations tested only. All reported hot pain sensation has an associated burning sensation. Pain with cold sensation was not reported. The mean VAS scores of the pain with hot and burning sensation for the 18/42, 20/40, 22/38 and 24/36 (8C) temperature combinations were 15.04G19.48, 11.87G16.17, 2.89G 6.21 and 2.6G8.7, respectively. Of the different warm and cold temperature combination settings, the VAS scores of pain at 20/40 and 18/42 (8C) combinations were significantly elevated in comparison to the mixed 22/38 and 24/36 (8C) temperature combinations with the Friedman Test (P!0.007). With the Chi-square value of 11.61 and P! 0.0089, the mean ranks in order for the VAS scores for the mixed temperature stimuli under the subsequent Wilcoxin Rank Sign Test were shown in Table 1. There is no statistically significant difference of the VAS scores for either between 20/40 and 18/42 or between 22/38 and 24/26 (8C) temperature combinations. No comparison was made to the single temperature combinations as none of the subjects perceived pain with the reported sensation of cold, neutral, warm and hot during the stimuli. 3.1.2. Qualitative characterization of the thermal grill None of the subjects reported pain with any of the single temperature combinations tested. Hot sensation associated with burning pain is present as the predominant sensation in all innocuous warm and cold temperature combinations Table 1 Noxious sensation ranking with Wilcoxin Rank Sign Test (Chi-square value of 11.61 and P!0.0089) Temperature combinations (8C)
Mean rank order
18/42 20/40 22/38 24/36
3.35 2.81 2.00 1.85
A.Y. Leung et al. / Pain 116 (2005) 26–32
29
Table 2 Thermal grill temperature matching at 3 s of stimulation Subjects
18/18 (8C)
42/42 (8C)
18/42 (8C)
20/20 (8C)
40/40 (8C)
20/40 (8C)
1 2 3 4 5 6 7 Mean SD
18 17 18 18 19 18 18 18 0.6
41 42 40 42 42 42 42 41.6 0.8
46 48 46 48 46 48 44 46.6 1.5
20 21 19 20 19 22 21 20.2 1.1
39 39 41 40 40 40 41 40 0.8
46 46 44 48 44 48 44 45.7 1.8
40/40 and 42/42 combinations were 18G0.6 (range 17–19), 20.2G1.1 (range 19–21), 40G0.8 (range 39–41) and 41.6G0.8 (range 41–43) 8C, respectively. In addition, all seven subjects reported a ‘surprise feeling’ or ‘stunning’ sensation when they placed their hands on the thermal grill (Table 2).
Fig. 2. (a) Thermal sensations with single temperature combinations (NZ 13). (b) Thermal sensations with mixed temperature combinations (NZ13).
(Fig. 2). At both 18/42 and 20/40 (8C) temperature combinations, hot burning pain was felt in about 70% of the volunteers as the initial sensation. Most subjects felt a ‘surprise’ or ‘stunning’ sensation with the initial placement of the hand on the thermal grill. However, only 20% of the subjects at both 22/38 and 24/36 (8C) experienced a hot burning pain sensation. The percentage breakdowns for the predominant sensations that were felt by the subjects at the single and mixed temperature combination tested are shown in Fig. 2a and b, respectively. 3.2. Phase II study 3.2.1. Thermal grill and QST temperature matching at 3 s A total of seven right-handed dominant healthy subjects who participated in the Phase I part of the study were screened for the Phase II part of the study. Two subjects were excluded from the study because of the unmatched right and left thermal thresholds. Of the remaining seven subjects who completed the study, the matching temperatures (GSD) of 20/40 and 18/42 were 45.7G1.8 (range 44–48) and 46.6G1.5 (range 44–48) 8C, respectively. The matching temperatures (GSD) for 18/18, 20/20,
3.2.2. Thermal grill and QST temperature matching at 10 s The same seven subjects who completed the Phase II study were enrolled. In this phase of the study, the combinations of thermal grill temperatures that were tested in random order included the following: 18/18, 42/42, 18/42, 20/20, 40/40, and 20/40. Of the seven subjects who have similar upper extremities thermal thresholds, the matching temperature (GSD) for the thermal temperatures combinations (8C) of 18/18, 42/42, 18/42, 20/20, 40/40 and 20/40 (8C) were 18.0G0.6 (range 17–19), 41.7G0.9 (range 40–43), 36.7G10.6 (range 18–43, two no match), 20.1G1.1 (range 19–21), 40.1G0.8 (range 39–41) and 37.3G9.0 (19–42, one no match), respectively. One subject was unable to match any temperature at both 18/42 and 20/40 (8C) temperature combinations and another subject was unable to match at 18/42 (8C) combinations. Thus, at 10 s, none of the combinations were significantly greater than the highest of the pair of stimuli (Table 3).
Table 3 Thermal grill temperature matching at 10 s of stimulation NM-No thermal matching Subjects
18/18 (8C)
42/42 (8C)
18/42 (8C)
20/20 (8C)
40/40 (8C)
20/40 (8C)
1 2 3 4 5 6 7 Mean SD
19 18 18 18 17 18 18 18 0.6
41 42 42 42 43 40 42 41.7 0.9
43 39 18 NM 41 NM 42.5 36.7 10.6
20 21 19 19.5 19 22 20.5 20.1 1.1
39 39 41 40 41 40 40.5 40.1 0.8
41 40 19 NM 41 42 41 37.3 9.0
30
A.Y. Leung et al. / Pain 116 (2005) 26–32
4. Discussion In the present study, the exposure of normal human subjects to the thermal grill stimulation paradigm resulted predominantly in a transient hot burning sensation with a statistically significant mild elevation of VAS scores of pain at the 20/40 and 18/42 8C temperature combinations. This observed mildly painful sensation was independently validated by an increase in the matched temperature when parallel exposures were made to fixed thermal stimuli for 3 s. The effect at 3 s may occur in part by an additional component of adaptation that is suggested by the initial ‘surprise’ or ‘stunning’ feeling that most subjects reported acutely after applying the palm to the grill. Importantly, after 10 s, the initial noticed sensation was noted to dissipate and the matched temperature was no greater than the highest stimulus intensity of the pair. 4.1. Qualitative characterization of the thermal grill In the present study, temperatures up to 42 8C were not reported to be noxious or hot but rather to be warm by all the subjects in the Phase I of the study. This is in parallel to the previous QST work, where the hot pain threshold was in the range of greater than 45 8C (Ando et al., 2000; Leung et al., 2001). In the present work, pairing of temperatures in the range of 20/40 and 18/42 8C was reported to elicit a report of a mild hot burning pain sensation. Contrary to Craig and Bushnell’s report that the thermal grill illusion presented with a burning sensation resembling cold pain, the current study suggests that the predominant sensation in the thermal grill is hot and burning as initially suggested by Altrutz and others (Boring, 1942; Heavner et al., 1997). This sensation appears to be significant at higher temperature differences at 20/40 and 18/42 (8C) than at lower temperature differences such as 22/38 and 24/36 (8C). 4.2. Thermal grill temperature matching The close matching of the single temperature combination in the thermal grill with the QST provides the validation that the thermal grill system as described here provides an appropriate quantifiable thermal stimulus, which matches a validated thermal stimulator system (Verdugo and Ochoa, 1992). The specific issue under consideration in this series of studies is the characteristic of the sensory experience reported when combinations of nonnoxious warm and cold stimuli were presented to the palm. As noted, the concurrent delivery of 20/40 and 18/42 (8C) at 3-s intervals resulted in an aversive experience. This assertion is based on three observations. First, these combinations, but not either temperature stimuli alone, resulted in a significant incidence of assigning the burning pain descriptor to the stimulus condition. Secondly, the ranking of the VAS scores appears to be proportional to the percentage of subjects who reported pain with the different
combined temperature stimuli. Third, these temperature combinations were matched with the QST studies at test temperatures of approximately 46 8C. Systemic QST studies from several laboratories have independently reported that this temperature range falls within the known range of hot pain threshold obtained in the healthy subjects (Ando et al., 2000; Wallace et al., 2002). 4.3. Dynamic aspects of thermal grill stimulation The application of acute thermal stimuli of up to 42 8C typically results in a stable sensation of warmth. Concurrent application of the cool (18–20 8C) and warm (40–42 8C) temperatures to the palm resulted in three additional components. The first, as discussed above, was the appearance of a noxious component at 3 s. Secondly, the observer noted that the noxious character of the sensation diminished by 10 s. The loss of the noxious nature of the effect noted over the 10-s intervals is in contrast to the stable nature of the sensation produced by the single thermal stimulus. This time course variation has not been reported by other investigators. We discount the likelihood of a change in the effective stimulus due to the heat sink capacity of the hand because examination of the grill temperature via the temperature monitoring probes during the stimulus exposure reveals no shifts in temperature. This observation is consistent with the high temperature capacity of the perfusion systems. We suggested that the loss of the acute burning sensation reflects upon an adaptation. As the apparent adaptation was not observed with the single stimulus conditions, we suggest that this alteration reflects upon a central change in processing, which underlies the aversive aspect of the combined stimulus presentation. 4.4. Underlying mechanism of the thermal grill stimulus-response Broadly speaking, high frequency and low-threshold mechano-stimulation is transmitted by the myelinated A-fibers. Cool and well-localized pain is carried by the less myelinated A-d fiber, whereas warm, hot and cold pain sensations are carried by the unmyelinated C-fiber (Verdugo and Ochoa, 1992; Yarnitsky and Ochoa, 1991). Previous electrical stimulation and nerve block studies suggested that cold-specific A-d afferent activity can attenuate C-fiber mediated cold pain sensation (Fruhstorfer, 1984; Ochoa and Yarnitsky, 1994; Wahren et al., 1989). In the thermal grill stimulation paradigm, the presence of warm stimuli in close spatial proximity to the cold stimuli may diminish the spatial summation of A-d fiber input, which has a spinal modulatory role of the C-fiber mediated cold pain threshold. The initial observed hot burning sensation at 3 s is in line with the results of previous nerve compression study in which blockade of the A-d mediated cold sensation leads to elevation of the cold pain threshold and the change of cold pain sensation to a hot burning rather than cold pain alone
A.Y. Leung et al. / Pain 116 (2005) 26–32
(Wahren et al., 1989). Accordingly, the presence of the continuous C-fiber input from the warm sensation could contribute to the spinal encoding of the sensation of hot burning rather than cold burning pain. Finally, as the input from the slow-conducting C-fiber becomes more dominant as observed by the subjects’ report of feeling both warm and cold sensation as stimulus intensities approach the nonnoxious magnitudes, the warm and cold stimuli interaction of the thermal grill sensation becomes less dominant. This reported transient noxious finding from the current study in comparison to a former study by Fruhstorfer, who suggested that the thermal grill sensation as ‘non-noxious heat’, may be due to the fact that the former study: (1) utilized an ischemic nerve block, which could potentially impact the C-fiber; (2) exposed the subjects to continuous testing conditions; and (3) had a significant skin–thermode surface temperature variation (Fruhstorfer et al., 2003). Nevertheless, the thermal matching of the current study did demonstrate that the thermal grill sensation is hardly a noxious one at 10 s. In addition, the findings of this current study are also complimentary to previous work performed by Craig and Bushnell, who measured the electrical activities of spinal cord neurons that projected to the brain in the spinothalamic tract (STT) of anesthetized cats (Craig and Bushnell, 1994). Their group examined all three major types of cat lamina I STT neurons: (1) nociceptive-specific (NS) cells that are responsive only to noxious heat and pinch and receive input from heat nociceptors; (2) thermoreceptive-specific (COLD) cells that are responsive to cooling and receive input from specific cold receptors; and (3) multimodal cells that are responsive to noxious heat, pinch, and cold (HPC) and receive input from cold-sensitive C polymodal nociceptors. Their result showed that NS cells were unaffected by the Warm, Cool, and Grill stimuli. In addition, both COLD and HPC lamina I STT cells responded briskly to the Cool stimuli. However, the response of COLD cells and HPC cells to the Grill stimulus differed significantly. The response of COLD cells to Grill were strongly reduced from their response to Cool but the response of HPC cells to Grill remained nearly the same as their response to Cool, suggesting that the presence of the interlaced warm bars to a cool stimulus not only reduces the activity in the coldspecific channel but also shifts the relative pattern of activity in favor of the HPC channel. Although work by Craig and Bushnell provided physiological evidence of the thermal grill illusion, the potential impact of general anesthesia on the primary afferent neurons and STT cells of the cat has not been discussed. It is known that anesthesia can have significant effect on the neuronal activities of subcortical structure which includes sensory and motor neurons and dorsal horn cells (Antognini and Carstens, 1998; Antognini and Schwartz, 1993; Antognini et al., 1999). In Craig’s study, the electrophysiological activities were measured over the duration of 100 s. Although the duration at which the animals were exposed to the Grill stimulation is not well
31
specified, it appeared that the most robust response from the dorsal horns (Cold and HPC) with spike measurements started at around 30 s and gradually declined over the next 60 s. This phenomenon suggests a change of both the Cold and HPC cell activities at the dorsal horn with prolonged exposure of the stimuli and supports the dynamic nature of the sensory integration at the spinal cord level. Given the above concerns, the current study does provide additional quantitative temperature comparison of the Grill stimuli and identifies the transient nature of the thermal grill sensation.
5. Conclusion The current study provides qualitative and quantitative evidence that the combination of two non-noxious temperature combinations in the thermal grill can generate a noxious sensation, which resembles a hot burning pain sensation generated by a temperature of around 46 8C. However, this initial noxious sensation appears to last for only a short duration. These newly identified dynamic quantitative and qualitative characteristics of the thermal grill needs to be considered if one chooses to use the device as a research tool for studying the anomalous events after nerve or tissue injury.
Acknowledgements Special thanks to Dr Yaksh and Michael Rathbun for constructing the thermal grill and Linda Sutherland for her editorial assistance.
References Ando K, Wallace MS, Braun J, Schulteis G. Effect of oral mexiletine on capsaicin-induced allodynia and hyperalgesia: a double-blind, placebocontrolled, crossover study. Reg Anesth Pain Med 2000;25:468–74. Antognini JF, Carstens E. Macroscopic sites of anesthetic action: brain versus spinal cord. Toxicol Lett 1998;100–101:51–8. Antognini JF, Schwartz K. Exaggerated anesthetic requirements in the preferentially anesthetized brain. Anesthesiology 1993;79:1244–9. Antognini JF, Carstens E, Buzin V. Isoflurane depresses motoneuron excitability by a direct spinal action: an F-wave study. Anesth Analg 1999;88:681–5. Boring E. Sensation and perception in history of experimental psychology. In: Geldard F, editor. The human senses. New Yok: Appleton-CenturyCrofts; 1942. p. 504–21. Burnett NC, Dallenbach KM. The experience of heat. Am J Psychol 1927; 38:418–31. Craig AD, Bushnell MC. The thermal grill illusion: unmasking the burn of cold pain. Science 1994;265:252–5. Craig AD, Reiman EM, Evans A, Bushnell MC. Functional imaging of an illusion of pain. Nature 1996;384:258–60. Fruhstorfer H. Thermal sensibility changes during ischemic nerve block. Pain 1984;20:355–61. Fruhstorfer H, Harju EL, Lindblom UF. The significance of A-delta and C fibres for the perception of synthetic heat. Eur J Pain 2003;7:63–71.
32
A.Y. Leung et al. / Pain 116 (2005) 26–32
Green BG. Localization of thermal sensation: an illusion and synthetic heat. Percept Psychophys 1977;22:331–7. Green BG, Pope JV. Innocuous cooling can produce nociceptive sensations that are inhibited during dynamic mechanical contact. Exp Brain Res 2003;148:290–9. Heavner JE, Calvillo O, Racz GB. Thermal grill illusion and complex regional pain syndrome type I (reflex sympathetic dystrophy). Reg Anesth 1997;22:257–9. Jenkins WL. Studies in thermal sensitivity: 5. The reactions of untrained subjects to simultaneous warm and cold stimulation. J Exp Psychol 1938;22:451–61. Leung A, Wallace MS, Ridgeway B, Yaksh T. Concentration–effect relationship of intravenous alfentanil and ketamine on peripheral neurosensory thresholds, allodynia and hyperalgesia of neuropathic pain. Pain 2001;91:177–87. Morin C, Bushnell MC, Luskin MB, Craig AD. Disruption of thermal perception in a multiple sclerosis patient with central pain. Clin J Pain 2002;18:191–5.
Ochoa JL, Yarnitsky D. The triple cold syndrome. Cold hyperalgesia, cold hypoaesthesia and cold skin in peripheral nerve disease. Brain 1994; 117:185–97. Sullivan AH. The experimental fusion of warmth and cold in heat. J Exp Psychol 1930;13:208–16. Thunberg T. Uppsala Lakforen. Forh 1896;2:489. Verdugo R, Ochoa JL. Quantitative somatosensory thermotest. A key method for functional evaluation of small calibre afferent channels. Brain 1992;115:893–913. Wahren LK, Torebjork E, Jorum E. Central suppression of cold-induced C fibre pain by myelinated fibre input. Pain 1989;38:313–9. Wallace MS, Ridgeway 3rd B, Leung A, Schulteis G, Yaksh TL. Concentration–effect relationships for intravenous alfentanil and ketamine infusions in human volunteers: effects on acute thresholds and capsaicin-evoked hyperpathia. J Clin Pharmacol 2002;42:70–80. Yarnitsky D, Ochoa JL. Warm and cold specific somatosensory systems. Psychophysical thresholds, reaction times and peripheral conduction velocities. Brain 1991;114:1819–26.