Effects of antihyperalgesic drugs on experimentally induced hyperalgesia in man

Pain 76 (1998) 317–325

Effects of antihyperalgesic drugs on experimentally induced hyperalgesia in man A. Bickel a ,*, S. Dorfs a, M. Schmelz a, C. Forster a, W. Uhl b, H.O. Handwerker a a

Department of Physiology I, University of Erlangen/Nu¨rnberg, 91054 Erlangen, Germany b Merck KGaA, 64271 Darmstadt, Germany

Received 23 September 1997; received in revised form 25 February 1998; accepted 5 March 1998

Abstract In a double-blind, cross-over study, ibuprofen (600 mg), a peripherally-acting selective k-opioid receptor agonist (7.5 mg), or placebo were given orally in experiments on healthy volunteers 1 h before assessment of pain thresholds to radiant heat and of pain ratings to controlled mechanical impact stimuli. Mechanical and thermal hyperalgesia had been induced 24 h before by irradiating skin patches on the ventral side of the upper leg. UVB irradiation induced mechanical and thermal hyperalgesia at radiation dosages of three times the minimal erythema dose. UVA irradiation resulted in an immediate erythema and a delayed tanning of the skin, however, no hyperalgesia was observed. For comparison another model of mechanical hyperalgesia was applied in the same experiments which has been previously proven sensitive to non-steroidal anti-inflammatory drugs (NSAIDs). In this model hyperalgesia was assessed, which develops during repetitive pinching of skin folds (pinch model). Ibuprofen significantly diminished heat and mechanical hyperalgesia induced by UVB, but had no effect on pain responses obtained from untreated skin. It also had an antihyperalgesic effect in the pinch stimulus paradigm. In contrast, the k-agonist showed no antihyperalgesic efficacy in the chosen models. It is concluded that the UVB model, as the pinch model, is suitable for establishing antihyperalgesic effects of NSAIDs, but probably not of k-receptor agonists, in healthy human volunteers. Compared to the pinch stimulus model, the UVB model offers additional advantages: (a) drugs may be tested after induction of the skin trauma by UV and this situation is more similar to the clinical use of antihyperalgesic drugs. (b) Since mechanical and thermal hyperalgesia is induced by UVB, drug effects can be tested upon both forms of hyperalgesia.  1998 International Association for the Study of Pain. Published by Elsevier Science B.V. Keywords: Non-steroidal anti-inflammatory drug; k-Receptor-agonist; Pain; Hyperalgesia

1. Introduction In a series of studies we have shown that antihyperalgesic actions of non-steroidal analgesic drugs (NSAIDs) can be demonstrated in experimental algesimetric studies on healthy humans. In these studies an experimental model has been established in which an interdigital web is repetitively pinched, resulting in increasingly painful sensations. In double-blind, cross-over studies it has been proven

* Corresponding author. Institut fu¨r Physiologie und experimentelle Pathophysiologie, Universita¨tsstrasse 17, 91054 Erlangen, Germany. Tel.: +49 9131 852692; fax: +49 9131 852497; e-mail: [email protected]

that under control conditions hyperalgesia develops within 10–15 min after application of a suitable first pinch stimulus, leading to increased pain ratings during the second and the following stimuli. This hyperalgesia can be prevented by oral intake of acetyl-salicylic acid, dipyrone and ibuprofen before starting the experiment (Forster et al., 1988, 1992; Handwerker et al., 1990; Kilo et al., 1995). The tested NSAIDs did not reduce the painfulness of a stimulus applied to unimpaired skin. This led to the conclusion that NSAIDs act by reducing the sensitization of nociceptors in the skin due to inflammatory processes and perhaps also by reducing the augmentation of spinal transmission of nociceptive signals. The antihyperalgesic action of ibuprofen in this model has recently been confirmed by another group (Petersen et al., 1997).

0304-3959/98/$19.00  1998 International Association for the Study of Pain. Published by Elsevier Science B.V. PII S0304-3959 (98 )0 0062-1

318

A. Bickel et al. / Pain 76 (1998) 317–325

In our previous studies the antihyperalgesic medication was effective to prevent sensitization of nociceptors. However, little is known about the reduction of an already existing hyperalgesia by antihyperalgesic drugs in such experimental models. Only in one study freezing of skin patches was used for inducing hyperalgesia which was subsequently reduced by ibuprofen (Kilo et al., 1995). In the present study ultraviolet (UV) irradiation of small skin patches was employed, a more pertinent and highly controlled inflammatory model. Two different qualities of UV irradiation, UVA with a wavelength of 290–320 nm and UVB (320–400 nm) were applied to different spots. An erythema and hyperalgesia were induced by controlled UVB irradiation and 24 h later the sensitivity of the treated skin patch was tested with mechanical and thermal stimuli. Apart from the tanning effect no lasting skin changes were induced. The antihyperalgesic drug was given 1 h before testing and hence applied long after inflammation and hyperalgesia had developed. As a positive control, we employed the established repetitive pinching of an interdigital web, a model in which ibuprofen is known to prevent the development of hyperalgesia. A second purpose of this study was to compare the antihyperalgesic effects of ibuprofen and a peripherally acting opioid in inflammatory hyperalgesia models. The involvement of peripheral opioid receptors in anti-nociception has recently been demonstrated (Przewlocki et al., 1992). This effect was limited to inflamed tissue (Barber et al., 1994), although opioid receptors have been demonstrated on peripheral terminals of sensory nerves not only in inflamed, but also in normal tissue (Antonijevic et al., 1995). In the present study, a potent k-opiate agonist was employed. It penetrates the blood-brain barrier only in limited amounts and therefore elicits its anti-nociceptive activity mainly due to binding at peripheral k-receptors. Sideeffects like sedation and diuresis, known from centrally acting k-opiate agonists, are less pronounced. The substance has been shown to be anti-nociceptive in animal models of inflammatory pain, e.g. the formalin-test and the tail pressure test (Barber et al., 1994). In this study the k-agonist was tested for the first time in experimental inflammatory painmodels in humans.

2. Methods 2.1. Study-design The study was performed in a randomized, double-blind, cross-over design. Each subject participated in four experimental sessions at intervals of at least 7 days. All sessions took place in the early afternoons to minimize circadian influences on pain ratings. A training session without medication was performed 1 week before the first session including medication, to intro-

duce the subjects to the algesimetric testing and pain-rating procedures. The data obtained in this training session was checked for plausibility, but not included in further analysis. Prior to each of the three following experimental sessions the subjects swallowed 3 capsules and 3 tablets: (a) 3 tablets of 200 mg ibuprofen (Ibu-Vivimed, Dr. Mann) and 3 capsules of placebo, (b) 3 capsules of 2.5 mg of the k-agonist EMD61753 (N-[(1S)-2-[(3S)-3-hydroxypyrrolidin-1-yl]-1phenylethyl]-N-methyl-2,2-diphenylacetmide, hydrochloride; Merck, Darmstadt, Germany) and 3 tablets of placebo, (c) 3 capsules and 3 tablets of placebo. These three medication schedules were applied in a randomized Latin square design. Verum and placebo tablets and capsules could not be distinguished by either the experimental subjects or the experimenter. The medication was taken together with 150 ml water, 4 h after the intake of a standardized meal. Algesimetric tests were performed between 1 and 2 h after medication and blood samples for assessment of ibuprofen or k-agonist plasma-levels were taken after 1, 3, 6 and 24 h. 2.2. Subjects Twenty-one male volunteers (22–32 years) participated in the study after giving their informed consent. Subjects had been informed that they could withdraw from the study at any time and were financially compensated for the time spent. The study was approved by the local ethics committee. Before entering the study, medical history was taken and the subjects underwent medical examination and screening of blood parameters. Exclusion criteria were any kind of clinically relevant diseases, especially concerning the gastrointestinal tract, the dermis, the central nervous system and relevant abnormalities in the blood parameters. All subjects had a body weight (BMI) in the normal range. They were not allowed to take any kind of medication within the last two weeks before the experiments nor during the study time. 2.3. Hyperalgesia models 2.3.1. UV-Erythema model Mode of UV application. One week before the first experimental session the minimal erythema dose (MED) for UVB (wavelength 290–320 nm) irradiation was established employing a calibrated UV source (Saalmann multitester SBB LT 400, Saalmann Medizintechnik, Herford, Germany). For that purpose, five circular spots, 15 mm in diameter, at the ventral side of the upper leg were irradiated with increasing intensities of UVB light (0.02–0.06 J/cm2). For each session different spots were used, since the sensitivity of the skin to UV may have changed due to preceding exposures.

A. Bickel et al. / Pain 76 (1998) 317–325

For the algesimetric tests UVB was applied in two doses 24 h before the experimental session at dosages of 1× and 3× the individual MED. In the following we refer to these dosages as UVB-1 and UVB-3. As the study was performed during winter and early spring and all subjects were instructed to avoid UV-exposition at the legs, individual MEDs did not change during the 4 weeks of study time. Immediately after UVB exposure the relevant skin areas showed no alterations. Neither spontaneous ongoing pain nor allodynia was reported from any spot. The erythema developed at about 6 h after irradiation and reached maximum intensity after 12 and 36 h, which fits well with data from literature (Benrath et al., 1995). Afterwards the erythema faded away slowly; the irradiated sites remained tanned for several weeks but in no case were blisters, longer lasting inflammation or permanent marks observed. The skin reaction to UVA was different. In pilot experiments UVA-application led to warming of the skin of about 42°C and further increase of UVA-dose was limited by painfulness of higher doses. To minimize unspecific sideeffects of the heating, UVA-irradiation was limited to fixed doses of 16.8 and 36 J/cm2 with the higher intensity eliciting a hot, but not painful sensation. In the following text, we refer to these two intensities as UVA-1 and UVA-3. After UVA-irradiation (about 500× higher intensity than UVB) the skin showed an immediate erythema reaction, which declined within a few hours. Pain thresholds were not changed immediately after irradiation. After 24 h the skin was tanned at these spots, in no case were a longer-lasting erythema or blisters observed, although a minor thermal lesion was probably already induced with this regimen. The skin stayed tanned at these spots for some weeks, therefore pre-treated spots were easily identified and could be avoided in subsequent sessions. Laser Doppler flow measurements. The grade of increased blood flow on the test spots was assessed with a Laser Doppler device immediately before the algesimetric session. For this purpose a laser probe (Periflux PF2, Perimed, Sweden) was affixed to the skin on the previously irradiated spots together with a control spot and the relative flux (measured in voltage output of the device) was measured. At each spot at least two values were obtained from slightly different probe positions. Heat pain thresholds. Heat stimulation was performed by infrared radiation from a halogen bulb (diameter 1 cm). The skin surface temperature was feedback controlled by a thermocouple attached to the skin (Beck et al., 1974). The thermocouple was fixed to the lamp at a distance of 5 cm and was located exactly centered in the focus of the emitted light (diameter of the focus: 1 cm). To measure heat pain thresholds, the thermocouple was gently attached to the skin of the subject and the skin was warmed to a baseline temperature of 32°C for 5 s. Then temperature was linearly increased at 0.67°C/s by feedback controlled regulation of the emitted light. Subjects were advised to stop the heating by pressing a

319

button as soon as the heat became painful and the threshold temperature, shown on a display, was noted. In no case was the cut-off temperature of 52°C reached. The spots UVA-1, UVA-3, UVB-1, UVB-3 and a control spot on untreated skin between the test spots were tested in a randomized order. Mechanical impact stimulation. Mechanical hyperalgesia was assessed by delivering mechanical impact stimuli. For this purpose a plastic cylinder with a mass of 0.5 g and a surface at its base of 0.2 cm2 was pneumatically driven at controlled speed through a barrel toward the skin, as described elsewhere in detail (Kohllo¨ffel et al., 1991). The UV-treated skin patches and the control spot on untreated skin were tested twice in random order with an impact velocity of 5, 9 and 13 m/s applied in ascending order. Subjects were instructed to give oral sensory ratings after each impact on an open scale, where zero indicated ‘no sensation’ and a rating of 100 a ‘just painful’ stimulus. Other figures were assigned proportionally to sensory magnitudes so that, e.g. doubled ratings indicated doubled painfulness. The subjects had no problems in rating painful and non-painful sensations along one rating scale. The rationale for this kind of rating scale has been discussed elsewhere (Koltzenburg and Handwerker, 1994). The obtained ratings were generally in the range of 5–200. To compensate for inter-individual differences, the grand mean of all ratings during the three experimental sessions was computed for each subject and the ratings were then computed as percentages of this grand mean. Repetitive pinching. The skin web between middle and ring finger was pinched with a probe in the form of a forceps with faces 6 mm in diameter. A force of 12 N was applied at these faces for 2 min and the procedure was repeated twice, after 20 and 40 min. In this experiment the subjects rated the stimulus induced pain at intervals of 10 s on a visual analogue scale (VAS), with the end points ‘just painful’ and ‘worst imaginable pain’ as in previous studies. The 12 ratings given during a squeeze stimulus were averaged and the mean values were statistically analysed (Forster et al., 1988, 1992; Kilo et al., 1995). In this study, in addition, a second evaluation step was performed in which ratings during the 1st and 2nd minute were analysed separately to search for possible effects of the medication on the time course of pain within stimuli. Again, ratings were normalized by subtracting individual values from the grand mean in each subject. 2.4. Experimental protocol All experiments were performed between 1 and 2 h after medication intake. Pinching was applied 60, 80 and 100 min after medication intake. Between the 2nd and 3rd pinching trial, blood flow was measured, the thermal thresholds were assessed and impact induced pain was tested on the UVirradiated spots. Blood samples were taken 1, 3, 6 and 24 h after medication.

320

A. Bickel et al. / Pain 76 (1998) 317–325

All statistical analysis was done using the STATISTICA 5.0 software package (StatSoft, Tulsa, USA). The significance level was set at P , 0.05, if not stated otherwise.

3. Results 3.1. Plasma levels of analgesic compounds Ibuprofen plasma concentrations were determined from blood samples taken 1 and 3 h after drug intake immediately before and after the algesimetric session. The median concentration was 38.53 mg/ml (range 15.85–57.08 mg/ml) after 1 h and 27.08 mg/ml (19.49–37.10 mg/ml) after 3 h. Four out of 21 subjects had higher ibuprofen levels in the 3 h probe than after 1 h, possibly reflecting delayed resorption. Interestingly, the plasma levels obtained in this study after a single dose of 600 mg ibuprofen were in the same range as those obtained in a previous study in which the subjects took 3 doses of 400 mg at 1 h intervals before the experimental session, but less compared with an intake of 3 doses of 800 mg at 1 h intervals (Kilo et al., 1995). The mean plasma concentration for the k-agonist was 86.57 ng/ml (range 23.54–208.94 ng/ml) after 1 h and 66.02 ng/ml (range 27.88–108.94 ng/ml) after 3 h. No significant correlations were found between the plasma levels of the two analgesic compounds and pain ratings in any particular model.

Fig. 1. UV-erythema model. Superficial blood flow at 24 h after UVA and UVB irradiation. Flux was only slightly increased at the UVA-3 and UVB1 spot. UVB irradiation at 3 times the minimal erythema dose (UVB-3) significantly increased blood flow. Ibuprofen significantly inhibited the UVB erythema, whereas the k-agonist was ineffective.

2.5. Statistical procedures 2.5.1. Laser Doppler flux measurements and heat-painthresholds ANOVA-testing was applied to compare the measurements obtained in ibuprofen, k-agonist and placebo sessions. Blood flux values and heat pain thresholds at UVB spots in the three sessions were used as dependent variables while intensity of irradiation (2 levels) and medication were used as independent variables(3 levels). 2.5.2. Mechanical impact stimulation Normalized ratings were analysed by ANOVA with pain ratings as dependent variable and medication (ibuprofen vs. k-agonist vs. placebo), day of session (1st, 2nd, 3rd) and impact velocity (5, 9 and 13 m/s) as independent variables (repeated measurement factor with 3 levels each).

3.2. UV-erythema-model 3.2.1. Laser Doppler flow measurements The relative blood fluxes on the irradiated skin spots were assessed with a Laser Doppler probe before algesimetric tests started. Average flux was increased at the UVA-3, UVB-1 and UVB-3 spots, although only at the UVB-3 spot were the differences to the control spot significant (Fig. 1). At this spot also a clear reddening of the skin was visible and an effect of medication was significant in ANOVA-testing (P = 0.008) (Table 1). Post-hoc tests showed that this effect was due to inhibition of the erythema by ibuprofen which caused significantly lower blood flow as compared to k-agonist (P = 0.015) and placebo (P = 0.046).

2.5.3. Pinch stimulation Normalized VAS pain ratings were analysed again by ANOVA with ratings as dependent, medication (ibuprofen vs. k-agonist vs. placebo) as independent variable and the three stimulus repetitions in each session as repeated measurements factor. Post-hoc Scheffe´ tests were performed on significant factors, where suitable. Table 1

Results of ANOVA testing of different changes in superficial blood flow as measured by a Laser Doppler probe, depending on intensity of UV-irradiation and medication

1. Medication 2. Intensity of UVB-irradiation 1×2

d.f.-effect

MS-effect

d.f.-error

MS-error

F

P-level

2* 1* 2*

15354.0* 606458.9* 6795.9*

40* 20* 40*

2844.010* 2443.140* 1758.568*

5.3987* 248.2293* 3.8644*

0.008402* 0.000000* 0.029213*

*Significant differences were obtained for different intensity (P , 0.0001) and medication (P = 0.008). The post-hoc Scheffe´ test performed on the factor medication revealed, that ratings obtained with ibuprofen pre-treatment differed significantly from placebo (P = 0.046) and the k-agonist session (P = 0.015), while no differences were found between placebo and k-agonist (P = 0.89).

321

A. Bickel et al. / Pain 76 (1998) 317–325

velocity 13 m/s were experienced as close to pain threshold (mean = 73.57). At UVB-3 treated spots the rating to impacts of low velocity was virtually unchanged (mean rating = 13.36), whereas for higher velocities mechanical hyperalgesia was observed (mean = 50.49 and 98.33 for impacts of 9 and 13 m/s, respectively) (Fig. 3). As explained in the Methods section, normalized ratings were used for further analysis. Only UVB-3 treatment caused a mechanical hyperalgesia (UVB 3 vs. control, P , 0.0001, Scheffe´ post-hoc test) which was slightly more prominent with stronger impact stimuli. While the pretreatments with UVA-1 and UVB-1 were completely ineffective, at the UVA-3 skin patch slightly higher ratings were obtained. However, the difference to untreated skin was not significant (P = 0.15 for an impact velocity of 13 m/s). ANOVA analysis (Table 3) revealed a medication effect on the mechanical hyperalgesia at the UVB-3 spot (P = 0.0096) and according to post-hoc testing this anti-hyperalgesic effect was due to ibuprofen (P = 0.003). No differences were found between ratings from the k-agonist session and ratings from the placebo session (P = 0.979).

Fig. 2. UV-erythema model. Heat pain thresholds at 24 h after UVA and UVB irradiation. Only UVB irradiation at 3 times the minimal erythema dose significantly induced heat hyperalgesia. Ibuprofen significantly reduced the UVB-induced heat hyperalgesia, whereas the k-agonist was ineffective.

No significant differences were found between placebo and k-agonist (P = 0.89).

3.3. Repetitive pinch model

3.2.2. Heat pain thresholds The more intense UVB-irradiation significantly lowered the heat pain threshold (mean = 39.05°C ± SEM compared to 45.32°C), whereas UVA or low dose UVB-irradiation did not induce heat hyperalgesia (Fig. 2). Thus, antihyperalgesic effects could only be observed at the UVB-3 spot. Ibuprofen significantly elevated heat pain threshold whereas the k-agonist was ineffective. ANOVA-analysis (Table 2) revealed significant differences for different intensity (P , 0.0001) and medication (P = 0.001). The post-hoc Scheffe´ test performed on the factor medication revealed significant reduction of heat hyperalgesia by ibuprofen as compared to placebo (P = 0.040) and the k-agonist (P = 0.002). No differences were found between placebo and k-agonist (P = 0.52).

The pinch model has previously been used for demonstrating the antihyperalgesic effects of NSAIDs, e.g. ibuprofen (Kilo et al., 1995). In this study it was included as a control procedure. As expected from previous studies, ibuprofen did not affect the painfulness of the first pinch stimulus and hence showed no analgesic effect under control conditions. Although the chosen dose of 600 mg was lower than in previous studies, ibuprofen was effective in reducing the development of hyperalgesia upon stimulus repetition (Fig. 4). The increase of pain ratings during the 2nd and 3rd stimulus was not completely prevented, but significantly reduced compared to the placebo session (P = 0.011). Again, the k-agonist was ineffective and pain ratings did not significantly differ from those in the placebo session (P = 0.798). The ANOVA-analysis (Table 4) disclosed an additional effect of the session, consisting of a general decrease of ratings from session 1 to session 2 (P = 0.011). Such an effect of habituation or ‘learning’ was also observed in previous studies in this model, but had only minor impact on the result, since the greatest change in rating levels occurred

3.2.3. Impact stimuli Increasing velocity and strength of impacts resulted in increased pain-ratings, as previously described (Kohllo¨ffel et al., 1991; Koltzenburg and Handwerker, 1994). In uninjured skin impacts of low velocity (5 m/s) were usually felt as light touch (mean rating = 10.45), impacts of 9 m/s as stronger touch (mean rating = 38.76), whereas impacts of Table 2

Results of ANOVA testing of different changes of heat-pain thresholds, depending on intensity of UV-irradiation and medication

1. Medication 2. Intensity of UVB-irradiation 1×2

d.f.-effect

MS-effect

d.f.-error

MS-error

F

P-level

2* 1* 2

31.2022* 526.2401* 2.5260

40* 20* 40

4.134139* 3.337913* 3.134115

7.5475* 157.6554* 0.8060

0.001656* 0.000000* 0.453771

*Significant differences were obtained for different intensity (P , 0.0001) and medication (P = 0.001). The post-hoc Scheffe´ test performed on the factor medication revealed significant differences between placebo (P = 0.040) and k-agonist session (P = 0.002). No differences were found between placebo and k-agonist (P = 0.52).

322

A. Bickel et al. / Pain 76 (1998) 317–325

Fig. 3. UV-erythema model (UVB-3 spot). Normalized pain ratings at 24 h after UVB irradiation to mechanical impact stimuli of 5, 9 and 13 m/s. Mechanical hyperalgesia was significantly reduced by ibuprofen, whereas the k-agonist was ineffective.

between the training session, which was excluded from the statistical analysis and the 1st experimental session (Forster et al., 1988). In the pinch model, pain ratings did increase not only from 1st to 2nd and 3rd stimulus, but there is also an increase of ratings within the 2-min stimulus periods. This ‘increase within’ was found not to be influenced by NSAIDs in previous studies (Handwerker et al., 1990; Forster et al., 1992). To test a possible differential effect of the k-agonist on this increase of painfulness within stimuli, additional data evaluation was performed. However, the increases of ratings within stimulus periods were influenced neither by ibuprofen nor by the k-agonist (Fig. 5).

4. Discussion In this study we compare antihyperalgesic drug effects on hyperalgesias developing during experimentally controlled local cutaneous inflammation due to UV-irradiation or to repeated tonic pressure stimulation. An important difference between these two models is the time course of development of inflammation. While pinching induces acute effects,

detectable as early as 20 min after the first stimulus, the maximal inflammation after UVB-exposure was found 12–24 h after irradiation (Benrath et al., 1995). To our knowledge, the UV model has been employed for the first time in an experimental therapeutic double-blind study on healthy volunteers, albeit it has already been used for screening anti-inflammatory drugs in animal models (Woodward et al., 1981). In humans, the UV-model has the advantage of being easily applied and well controlled, thus allowing repeated testings. Compared to another model of subacute inflammation introduced by our group, the freeze model (Kilo et al., 1994, 1995), induction of the inflammation with UVB is not painful. While UVB (290–320 nm) is largely absorbed in the epidermis, UVA (320–400 nm) penetrates into the deeper layers of the dermis. Erythema development after UVA irradiation has been discussed controversially. In general, the doses necessary for induction of an UVA-erythema are found to be very high, i.e. up to 1000 times the required UVB-doses (Hrza and Pentland, 1993). However, with higher UVA-intensities more pronounced unspecific thermal effects are induced. This might explain differences found in histological sections of UVA- and UVB-induced erythema (Willis and Cylus, 1977). In our study UVA-irradiation at a dose effective in tanning the skin did not produce a significant mechanical or thermal hyperalgesia. Due to the lack of significant hyperalgesias after 24 h, UVA pretreatment was not appropriate for the assessment of antihyperalgesic effects in the present study. UVB led to a visible erythema as local sign of inflammation after 6–12 h. However, for drug testing the threshold dose producing a visible erythema (UVB-1) apparently was not appropriate, since signs of inflammation and hyperalgesias were marginal. At UVB-1 spots we found no significant changes in painfulness to heat or mechanical impact stimuli. Therefore, no antihyperalgesic drug effects could be found. However, when using a higher UVB-dose of 3 × MED, significant increase in superficial blood flow was observed, the heat pain threshold was significantly lowered and a hyperalgesia to mechanical impacts became evident. On this background hyperalgesic drug effects were tested.

Table 3 Results of ANOVA testing of different effects of velocity and medication in the impact model

1. Day of session 2. Medication 3. Impact velocity 1×2 1×3 2×3 1×2×3

d.f.-effect

MS-effect

d.f.-error

MS-error

F

P-level

2 2 2 4 4 4 8

60.7 2142.2 121554.9 75.5 321.5 283.7 430.8

54 54 108 54 108 108 108

422.6284 422.6284 328.5106 422.6284 328.5106 328.5106 328.5106

0.1436 5.0688 370.0183 0.1785 0.9787 0.8635 1.3115

0.866543 0.009607* 0.000000* 0.948531 0.422357 0.488437 0.245452

*Significant differences were obtained for different velocities (P , 0.0001). The post-hoc Scheffe´ test performed on the factor medication revealed significant differences in ratings between ibuprofen and either placebo (P = 0.037) or k-agonist pre-treatment (P = 0.022), while no differences were found between placebo and k-agonist session (P = 0.979).

323

A. Bickel et al. / Pain 76 (1998) 317–325

Fig. 4. Pinch model. Normalized pain ratings to 3 subsequent pinch stimuli (2 min, 12 N). Pain ratings increase from stimulus to stimulus. Ibuprofen significantly inhibits the development of this mechanical hyperalgesia, whereas the k-agonist was ineffective.

In control skin, no drug effects were observed on heat and mechanical pain thresholds indicating a lack of analgesic action of both medications in uninjured skin. Not surprisingly, at the UVA- and UVB-1-spots in which inflammation was absent or marginal, no drug effect was observed. In contrast, UVB-induced inflammation (blood flow increase) and hyperalgesia was significantly reduced by ibuprofen. The lowering of heat pain threshold and the increases in pain ratings after mechanical impacts were significantly reduced by ibuprofen on this spot, proving an antihyperalgesic action on thermal and mechanical hyperalgesia. For the impact pain this drug effect was dependent on the impact strength. Whereas ratings of strong impacts (9 and 13 m/s) were significantly reduced, sensation of light touch (as caused by impacts of 5 m/s) was not changed significantly. The k-agonist showed no effect in the UVB model. In the pinch model, two forms of hyperalgesia have been observed. Pain ratings in the pinch model increase between stimuli (resulting in higher average ratings during the 2nd and 3rd stimulus), but also within the 2 min of each tonic pinching (resulting in higher ratings during the second half of each stimulus compared to the first half). Only the increase of ratings between stimuli was influenced by

NSAIDs in previous studies (Handwerker et al., 1990; Forster et al., 1992; Petersen et al., 1997). Both phenomena, increase of painfulness ‘between stimuli’ and ‘within stimuli’, may reflect different types of hyperalgesia due to still unknown different peripheral and/or central nervous mechanisms. Since NSAIDs diminish only the hyperalgesia developing between stimuli, this type of hyperalgesia has been hypothetically attributed to cyclo-oxygenase (COX) dependent sensitization of cutaneous nociceptors. A further argument for a mainly peripheral action of NSAIDs was provided by the finding that ibuprofen did not influence the secondary hyperalgesia around a burn injury when 600 mg were applied before the induction of the burn lesion (Petersen et al., 1997). Secondary hyperalgesia has been mainly attributed to central nervous sensitization mechanisms. However, it is unclear if primary hyperalgesia (as tested in the pinch model) is an exclusively peripheral phenomenon and NSAIDs have been proven to suppress COX induction also in the central nervous system (Malmberg and Yaksh, 1992). The increase of painfulness ‘within stimuli’ has been tentatively attributed to central nervous mechanisms, since in the periphery many nociceptors show adaptation rather than increase of discharges during tonic pressure stimulation (Adriaensen et al., 1984; Handwerker et al., 1987). However, recent microneurography data in humans provided evidence for a peripheral mechanism, i.e. the recruitment of silent nociceptors during prolonged stimulation (Schmelz et al., 1997). Regardless of the mechanism, it was of interest in the context of this study, whether a peripherally acting k-agonist influences this pain phenomenon which apparently is insensitive to the action of ibuprofen and other NSAIDs (Forster et al., 1988, 1992; Handwerker et al., 1990). In this study, ibuprofen showed an expected diminution of the hyperalgesia developing between the 1st and the following pinch stimuli. The antihyperalgesic effect was somewhat smaller than in a previous study in which a higher dose of ibuprofen was applied (Kilo et al., 1995). An effect on the increase in pain ratings ‘within stimuli’ was not observed,

Table 4 Results of ANOVA testing in the pinch model

1. Day of session 2. Medication 3. No. of stimulus within the session 1×2 1×3 2×3 1×2×3

d.f.-effect

MS-effect

d.f.-error

MS-error

F

P-level

2 2 2

80284.3 135523.7 451744.7

54 54 108

16379.17 16379.17 5493.43

4.90161 8.27415 82.23357

0.011063* 0.000734* 0.000000*

4 4 4 8

5611.8 2248.7 14208.1 5916.6

54 108 108 108

16379.17 5493.43 5493.43 5493.43

0.34262 0.40935 2.58637 1.07704

0.847984 0.801572 0.040938* 0.384668

*An unexpected session effect was found, which was due to significant different ratings between session 1 and 2 (post-hoc Scheffe´ test: P = 0.0114). The medication effect was due to significant differences between ibuprofen and placebo (P = 0.0109 in the post-hoc Scheffe´-test) or the k-agonist (P = 0.0016). The k-agonist and placebo did not differ significantly (P = 0.7976).

324

A. Bickel et al. / Pain 76 (1998) 317–325

References

Fig. 5. Mean pain ratings in the first and second minute of 3 subsequent pinch stimuli are shown. Ibuprofen significantly inhibits the increase between the subsequent stimuli as shown before. However, the increase during the stimuli did not differ significantly between ibuprofen and placebo (ANOVA, P = 0.64).

also in parallel with previous studies. The k-agonist was ineffective in this model. The lack of pharmacodynamic effects of the peripherally acting k-agonist was unexpected in particular in the UV model, since antihyperalgesic effects of this agent have been proven in animal studies (Barber et al., 1994). One of the following reasons may explain this lack of efficacy: (1) species differences; (2) inefficient low dose. It is possible that the concentration of the k-agonist in the relevant compartment, the inflamed skin, was not sufficient. There is not always a good correlation between plasma levels and analgesic effects of peripherally acting compounds; (3) according to the available animal data k-agonists should act antihyperalgesic in acute inflammations. However, it has been shown that peripheral opioid receptors are upregulated only some days after induction of inflammation (Hassan et al., 1992; Stein et al., 1995). Therefore the kagonist might show its potency better in more chronic inflammations. In this case, acute models of inflammatory pain may be inappropriate for testing the analgesic potency of k-receptor-agonists. For the testing of antihyperalgesic effects of NSAIDs, however, UVB-irradiation might provide a suitable model which can easily be applied in healthy human volunteers. Due to the well controlled character and the small size of the experimentally induced inflammation this model is ethically justified. Since the antihyperalgesic medication is applied in the pinch model before and in the UVB model after induction of the inflammatory processes, the two models can be used in combination to distinguish between ‘therapeutic’ and ‘prophylactic’ effects of a medication. This might be valuable for future phase-II studies.

Acknowledgements This study was supported by Merck KGaA, Darmstadt and by the Deutsche Forschungsgemeinschaft, SFB 353.

Adriaensen, H., Gybels, J., Handwerker, H.O. and Van Hees, J., Nociceptor discharges and sensations due to prolonged noxious mechanical stimulation – a paradox, Hum. Neurobiol., 3 (1984) 53–58. Antonijevic, I., Mousa, S.A., Schafer, M. and Stein, C., Perineurial defect and peripheral opioid analgesia in inflammation, J. Neurosci., 15 (1995) 165–172. Barber, A., Bartoszyk, G.D., Bender, H.M., Gottschlich, R., Greiner, H.E., Harting, J., Mauler, F., Minck, K.O., Murray, R.D., Simon, M. and Seyfried, C.A.A., Pharmacological prophile of the novel, peripherallyselective k-opioid agonist, EMD 61753, Br. J. Pharmacol., 113 (1994) 1317–1327. Beck, P.W., Handwerker, H.O. and Zimmermann, M., Nervous outflow from the cat’s foot during noxious radiant heat stimulation, Brain Res., 67 (1974) 373–386. Benrath, J., Eschenfelder, C., Zimmermann, M. and Gillardon, F., Calcitonin gene-related peptide, substance P and nitric oxide are involved in cutaneous inflammation following ultraviolet irradiation, Eur. J. Pharmacol., 293 (1995) 87–96. Forster, C., Anton, F., Reeh, P.W., Weber, E. and Handwerker, H.O., Measurement of the analgesic effects of aspirin with a new experimental algesimetric procedure, Pain, 32 (1988) 215–222. Forster, C., Magerl, W., Beck, A., Geisslinger, G., Gall, T., Brune, K. and Handwerker, H.O., Differential effects of dipyrone, ibuprofen and paracetamol on experimentally induced pain in man, Agents Actions, 35 (1992) 112–121. Handwerker, H.O., Anton, F. and Reeh, P.W., Discharge patterns of afferent cutaneous nerve fibers from the rat’s tail during prolonged noxious mechanical stimulation, Exp. Brain Res., 65 (1987) 493–504. Handwerker, H.O., Beck, A., Forster, C., Gall, T. and Magerl, W., Analgesic effects of dipyrone as compared to placebo. In: K. Brune (Ed.), New Pharmacological and Epidemiological Data in Analgesics Research, Birkha¨user Verlag, Basel, 1990, pp. 19–28. Hassan, A.H., Pzewlocki, R., Herz, A. and Stein, C., Dynorphin, a preferential ligand for kappa-opioid receptors, is present in nerve fibers and immune cells within inflamed tissue of the rat, Neurosci. Lett., 140 (1992) 85–88. Hrza, L.L. and Pentland, A.P., Mechanisms of UV-induced inflammation, J. Invest. Dermatol., 100 (1993) 41S–53S. Kilo, S., Forster, C., Geisslinger, G., Brune, K. and Handwerker, H.O., Inflammatory models of cutaneous hyperalgesia are sensitive to effects of ibuprofen in man, Pain, 62 (1995) 187–193. Kilo, S., Schmelz, M., Koltzenburg, M. and Handwerker, H.O., Different patterns of hyperalgesia induced by experimental inflammation in human skin, Brain, 117 (1994) 385–396. Kohllo¨ffel, L.U.E., Koltzenburg, M. and Handwerker, H.O., A novel technique for the evaluation of mechanical pain and hyperalgesia, Pain, 46 (1991) 81–87. Koltzenburg, M. and Handwerker, H.O., Differential ability of human cutaneous nociceptors to signal mechanical pain and to produce vasodilatation, J. Neurosci., 14 (1994) 1756–1765. Malmberg, A. and Yaksh, T.L., Antinociceptive actions of spinal nonsteroidal anti-inflammatory agents on the formalin test in the rat, J. Pharmacol. Exp. Ther., 263 (1992) 136–146. Petersen, K.L., Brennum, J. and Dahl, J.B., Experimental evaluation of the analgesic effect of ibuprofen on primary and secondary hyperalgesia, Pain, 70 (1997) 167–174. Przewlocki, R., Hassan, A.H.S., Lason, W., Epplen, C., Herz, A., Stein, C. and Hassan, A.H., Gene-expression and localization of opioid-peptides in immune cells of inflamed tissue – functional-role in antinociception, Neuroscience, 48 (1992) 491–500. Schmelz, M., Schmidt, R., Bickel, A., Handwerker, H.O. and Torebjo¨rk, H.E., Differential sensitivity of mechanosensitive and -insensitive Cfibers in human skin to tonic pressure and capsaicin injection, Soc. Neurosci. Abstr., 23 (1997) 1004.

A. Bickel et al. / Pain 76 (1998) 317–325 Stein, C., Schafer, M. and Hassan, A.H., Peripheral opioid receptors, Ann. Med., 27 (1995) 219–221. Willis, I. and Cylus, L., UVA erythema in skin: is it a sunburn?, J. Invest. Dermatol., 68 (1977) 128–129.

325

Woodward, D.F., Raval, P., Pipkin, M.A. and Owen, D.A.A., Re-evaluation of the effect of non-steroidal antiinflammatory agents on UVinduced cutaneous inflammation, Agents Actions, 11 (1981) 711–717.