Sensitivity of repeated interdigital web pinching to detect antinociceptive effects of ibuprofen

Abstract

Aims

The aim of the current study was to assess the viability of the interdigital web pinch model as a test for analgesic activity in volunteer-based early phase drug development.

Methods

Pain thresholds and sensitization to a series of four sessions of interdigital web pinching (12 Newtons force) were measured in 26 male volunteers before and 1 and 3 h after oral dosing with ibuprofen (800 mg) or placebo to ibuprofen. Within each time point, the pain thresholds were measured by calculating the average visual analogue scores (VAS) for the first session of pinching (VAS-1). Sensitization to pinching was assessed by calculating the average changes in these scores for the three subsequent sessions of pinching (VAS-2). Moreover, the difference between the VAS score after the first session of pinching and that obtained at the end of the fourth session of pinching was calculated as a secondary endpoint (VAS-3).

Results

Treatment with ibuprofen had no significant effect on VAS-1 at either 1 or 3 h after dosing. However, the mean values of VAS-2 and VAS-3, were significantly reduced (P < 0.05) following treatment with ibuprofen.

Conclusions

This model has been able to detect an antinociceptive effect with ibuprofen. However, large numbers of subjects were required in order to demonstrate this effect and this feature would restrict the model's utility in early phase clinical trials where small numbers of subjects are normally employed.

Introduction

The use of experimental models in normal subjects to demonstrate analgesic efficacy of nonsteroidal anti-inflammatory drugs (NSAIDs) has provided, at best, equivocal results. Perhaps this is due to the inability of some of the models to faithfully mimic the true pain condition [1]. Indeed, whereas the sensory dimension of pain is predominant in most experimental models, affective and evaluative dimensions assume the predominant role in clinical pain [2]. A further confounding factor could be that the analgesic properties of NSAIDs do not solely relate to their anti-inflammatory actions and may reflect actions at the spinal level [3, 4] or plasticity changes within the ascending pathways induced by increased nociceptor input to the central nervous system [5].

Repeated tonic pinching of interdigital webs is an experimental technique that has been used to assess pain and antinociceptive properties of NSAIDs in human volunteer trials [6–9] and the technique appears to have particular merit as the pain modalities appear to change with repeated stimulation. Indeed a degree of sensitization to the pain response has previously been noted [6], presumably and initially due to peripheral sensitization [10–12] and, with continuation of stimulus, central sensitization could be evoked [12]. Hence, repeated pinching is likely to evoke a variety of pain modalities that could serve as targets for pharmacological intervention. This experimental model may therefore provide a similar variety of pain components to those modalities evoked in clinical states of pain. However, there are indications that the sensitivity of this model would preclude its usefulness in early phase clinical pharmacology studies. For example, previous studies by other workers have shown that relatively large numbers of subjects are needed (up to 22 subjects were required to identify analgesic activity in this model [7]) and previous studies of our own have indicated that up to 28 subjects would be required to show a reduction in pain sensitization [13]. Moreover, the single doses of analgesic agents used in some of these previous studies are higher that would normally be used for treatment of clinical pain, i.e. 800 mg ibuprofen [7], vs single doses of between 200 and 400 mg that are recommended to treat mild to moderate pain.

The key objective of the present study therefore was to evaluate, under rigorous clinical trial conditions, the potential for employing the interdigital web pinch model to assess antinociceptive activity of agents in early clinical development. A preliminary communication of these data has been presented by the author at the IUPHAR meeting in Munich, July 1998.

Methods

Volunteers

Twenty-eight male volunteers participated in the study. All subjects were judged to be healthy by medical examination and laboratory tests. Their mean age was 31.8 years (range 21–44 years), their mean weight was 74.5 kg (range 59–102 kg), and their mean height was 179.3 cm (range 163–190 cm). The study was carried out in accordance with the Declaration of Helsinki and written informed consent was obtained from all the participants. The study was approved by the local ethics committee (Bayerische Landersärztkammer [Bavarian Chamber of Physicians, Munich]).

Drug supply and dosing

Ibuprofen was supplied as ibuprofen BP, 400 mg from Norton Pharmaceuticals. Placebo tablets were provided by Boots Pharmaceuticals.

Oral doses of ibuprofen 800 mg and matching placebo were given according to a double-blind, randomized two-period crossover design. Each dose composed of two identical appearing tablets. All tablets of active drug contained 400 mg ibuprofen. The tablets were taken with 150 ml water at 08.00 h on each of two trial treatment days. The volunteers received both trial treatments in random order, and at least 6 days elapsed between prior doses. The volunteers were fasted from midnight until 4 h after dosing. No other medication was allowed within 4 weeks subsequent to the start up to the end of the study period.

Interdigital web pinching and analgesimetric testing

The mechanical stimulator used in this study was similar to that described previously [14]. At each time point, during the trial, a different interdigital web was used. Before dosing, the webs between the 2nd and 3rd digit of the nondominant hand were used. At 1 h after dosing the same interdigital webs of the dominant hand were used and at 3 h after dosing the webs between the 3rd and 4th digit of the nondominant hand were used.

On each occasion the interdigital webs were pinched with a constant force of approximately 12 N for 2 min during each session of pinching. If required, the pinching could be released immediately. Before entry into the trial, subjects were completely familiarized with the test procedures by undergoing one complete set of assessments. However, subjects were not entered or eliminated on the basis of this familiarization process.

During each session of pinching the subject rated his current pain intensity every 10 s after being prompted by an acoustic signal. This was accomplished by movement of a lever which, in turn, controlled a luminous VAS. Immediately after setting the score, the lever was returned to the ‘no pain’ position. On recruitment, the subject was instructed how to use this device and the scale represented ‘no pain’ (left) and ‘intolerable pain’ (right limit). The total scale was equal to 5 V. All data reported were expressed as a percentage of the total scale. Before pinching, the subject set his baseline perception score following three demands at 20 s each. All pain scores set on the visual analogue scale (VAS) were then captured automatically and transferred to a personal computer that steered the test and collected and stored all VAS data.

On each trial day, before dosing (between 2 and 0.40 h prior to receipt of dose) and at 1 and 3 h after dosing, each analgesimetric test consisted of four periods, each of 2 min duration, where stimuli at 12 N were delivered to the interdigital web as shown in Figure 1.

Figure 1.

Schema of pinching sequence during the pain assessments. Each solid bar represents the 2 min of interdigital web pinchingat a force of 12 N, conducted on four occasions at each time point. The open bars represent periods between pinching when the12 N force was released. The time scale is shown as elapsed time during the pain assessment sessions.

After each 10 s of stimulation, during each 2 min period of pinching, the subject scored his pain perception on the VAS. The VAS scores and the VAS responses (i.e. changes from prepinching baseline) after each 10 s of pinching were recorded.

Variables assessed and endpoints

For each 2 min period of pinching the arithmetic mean VAS-responses during the pinching sessions were calculated. The following variables were assessed: average VAS responses for the first 2 min of pinching at each time point were calculated (VAS-1); average changes in these responses for the three subsequent sessions of pinching at each time point (VAS-2) and the difference between the VAS responses after the first session of pinching and that obtained at the end of the fourth session of pinching at each time point (VAS-3).

The primary endpoints were VAS-1 and VAS-2. The secondary endpoint was VAS-3.

Statistical considerations

Model

The endpoints of the test at each time point subsequent to dosing were fitted to the following model (detailed for VAS-2):

$$\begin{array}{c}\text{VAS}-2 \left(1.00 \text{h}\right)=\text{VAS}-2 \left(\text{predose}\right)+ \text{sequence}=\\ \text{subject} \left(\text{sequence}\right)+ \text{period}+\text{ibuprofen}\end{array}$$

The assumptions underlying this parametric approach (normality and constancy of variance) were evaluated. It was necessary to adopt a nonparametric approach to validate the main results. Treatment specific carry-over of drug effect was examined by testing sequence effect using the subject (sequence) mean square as the error term.

Effects of ibuprofen

The mean square error of the model described above was used as the variance estimate to calculate the 95% confidence interval (95% CI) of the true difference between treatment means (ibuprofen least squares mean minus placebo least squares mean) for the tests at 1 and 3 h after dosing.

Sample size estimates

The sample size required to detect with 95% confidence and 80% power, an effect (δ) on sensitization in order that either no sensitization occurs (i.e. VAS-2 = 0) or that the same effect as presently observed, was calculated as follows:

$$\text{N}=(2.{\sigma }^2{\left(t_{\alpha }+t_{\beta }\right)}^2/{\delta }^2$$

where t_α = t value for 2-sided α: 0.05, and t_β = t value for 1-sided β: 0.20, each time for the residual degrees of freedom and σ² equals the mean square error. A previous study employing the interdigital web pinching technique [13] had demonstrated that a sample size of 28 would be required to demonstrate a decrease in any sensitization produced by repeated interdigital pinching.

Results

Twenty-six volunteers completed the trial successfully with only minor adverse events. Two volunteers were excluded from the evaluation of test criteria as they found the test procedures to produce unbearable pain. Their final test period was aborted. An analysis of covariance showed that there was no carryover between treatments.

Analgesimetric assessment

Pre-dose assessments

The sessions of predose assessments of interdigital web pinching revealed that in both placebo and ibuprofen groups, there was evidence of sensitization to repeated pinching (Figure 2).

Figure 2.

Visual analogue scale (VAS) pain assessments in healthy male subjects following four sessions of interdigital web pinching prior to receipt of either placebo or ibuprofen 800 mg. Each session of pinching was as described in Figure 1. There was at least 1 week between pinching sessions. The data from predose assessments prior to subjects receiving an oral dose of either placebo (♦) or ibuprofen 800 mg (▪) are mean values of VAS responses expressed as a percentage ± s.d. n = 26 in both cases.

Drug assessments

Treatment with ibuprofen had no significant effect on either the individual VAS responses, during each 2 min period of pinching, or VAS-1 at either 1 or 3 h after dosing (Figures 3a, 3b and 4a, respectively). The ls adjusted mean treatment differences (95% confidence intervals) were −3.90 (−8.68, 0.88) and −1.64 (−7.23, 3.96) at 1 and 3 h, respectively. However, mean values for VAS-2 and VAS-3 were significantly reduced (P < 0.05) following treatment with ibuprofen (Figure 4b,c, respectively). The ls adjusted mean treatment differences (95% confidence intervals) were −3.99 (−8.38, 0.40) and −5.66 (−9.85, −1.46) for VAS-2 at 1 and 3 h, respectively, and −4.50 (−10.35, 1.34) and −7.96 (−13.66, −2.26) for VAS-3 at 1 and 3 h, respectively.

Figure 3.

VAS pain assessments in healthy male subjects during four periods of interdigital web pinching, each of 2 min duration (a) 1 h and (b) 3 h after receiving an oral dose of either placebo (♦) or ibuprofen 800 mg (▪). Each value is a mean value of VAS responses expressed as a percentage ± s.d. n = 26 in both cases.

Figure 4.

VAS pain assessments in healthy male subjects during 4 periods of interdigital web pinching, each of 2 min duration, 1 and 3 h after receiving an oral dose of either placebo (□) or ibuprofen 800 mg (▪). (a) VAS-1, (b) VAS-2 and (c) VAS-3 variates, expressed as mean percentage values ± s.d. n = 26 in all cases. * indicates a significant difference (P < 0.05) compared with placebo.

Sample size estimates

Using the VAS-2 variate, with 95% confidence and 80% power, it has been calculated that 58 and 28 subjects would be required to obtain the observed (treatment) effect at 1 and 3 h, respectively. If no sensitization of the pinching response was observed, then between 62 and 32 subjects would be required to obtain an effect at 1 and 3 h, respectively.

Tolerability of test procedure and adverse events

In general the test procedure was well tolerated. All subjects exhibited erythema at the pinching site and although the test procedure caused discomfort, this was thought to be bearable by all with the exception of three volunteers who demanded interruption to the procedure at various times during the trial. Two of these subjects were excluded from the data analysis. The third subject experienced the discomfort during the pretrial familiarization process but successfully completed the drug treatment periods of the trial.

Discussion

The present study has demonstrated that ibuprofen, albeit at a relatively high dose, was able to produce an antinociceptive effect in the interdigital web pinching model. These data are in agreement with previously published work where a similar effect has been demonstrated with ibuprofen, dipyrone [7] and aspirin [6]. Interestingly, the current study demonstrated that at best, ibuprofen could reduce the pain response by ≈ 10%, whereas previous studies have found that reductions in the pain rating were between ≈ 10 and 15% depending on the strength of the stimulus, i.e. 8 or 10 N, respectively [7]. It is likely that as 12 N force was used as a stimulus in the current study this may account for this difference between our data and those obtained by this previous study [7]. However, in preliminary experiments with this model, we found that use of a 9 N force did not appear to produce sufficient ‘sensitization’ to the pinching that might have allowed a potential antinociceptive effect to be detected (unpublished observations).

The dose of ibuprofen that was employed in the current study was higher than that normally prescribed for treatment of mild to moderate pain (i.e. as single doses of 200–400 mg). In the current study it was decided to use the dose of 800 mg in order that a comparison could be made to the previous study [7]. More recently, it has been shown that 400 mg ibuprofen given on a thrice daily basis, produced a significant reduction in the hyperalgesia evoked by repeated pinching of the interdigital webs [15]. However, it ought to be noted that in this latter study, plasma concentrations of R-and S-enantiomers of ibuprofen were comparable to those observed follow- ing a single dose of 800 mg ibuprofen (i.e. 24.7 and 15.1 µg ml⁻¹, R and S-enantiomers, respectively, following 3 × 400 mg ibuprofen [15] vs 16.5 and 12.8 µg ml⁻¹, R-and S-enantiomers, respectively, following 1 × 800 mg ibuprofen [7]). It would be expected therefore to see a similar degree of antinociceptive action with both these dosing regimens and suggests that a single 400 mg dose of ibuprofen would not be effective.

The success of this model appears to depend on the ability of the interdigital webs becoming more sensitive to the pressure placed upon them by repeated pinching. It is likely that the pinching stimulus would initially stimulate c-fibre afferents leading to an increased excitability of spinal neurones due to, for example, release of tachykinins and activation of the NMDA receptor complex. This would, in turn, lead to a reduction in the mechanical receptive field-threshold resulting in activation of Aβ-fibres. Moreover, repeated pinching would lead to release of inflammatory mediators, both at the site of pinching, and spinally. It is known that cyclooxygenase activity removes the descending inhibitory influence of noradrenrenergic neurones terminating in the spinal cord [16] which, in turn, would lead to an increased nociceptive input. The localized inflammation of the interdigital web, indicates that this phenomenon did take place. Sensitivity of this model to pharmacological intervention would, therefore be dependent upon the degree of inflammation and peripheral sensitization that had occurred. It follows that there is a better chance that a cyclooxygenase inhibitor would be more beneficial than a receptor antagonist in the current study as the former agents would be expected to reduce the release of neuro-active mediators from inflammatory cells [12], which, in turn, would lead to a reduction the total pain per se. Moreover, the effectiveness of a cyclooxygenase inhibitor would be expected to be more pronounced when inflammation had been established, i.e. after repeated interdigital web pinching. This was the case in the current study and suggests that this model would be relatively insensitive to analgesic agents without anti-inflammatory activity. The question unanswered by the current study is whether or not the ibuprofen-antinociceptive effect was due to a peripheral or central effect or both.

Although the interdigital web pinch model has detected an antinociceptive effect with ibuprofen, the major disadvantages of this type of model lie in the large numbers of subjects that are required in order to demonstrate this effect and the sensitivity of the model restricting it to agents with anti inflammatory actions. These features reduce the utility of the model in clinical pharmacology trials where small numbers of subjects are normally employed and agents with a diversity of action are evaluated.